Acknowledgments. Financial support for the project was also provided by the Everglades Agricultural Area- Environmental Protection District.

Transcription

1 EVERGLADES AGRICULTURAL AREA BMPs FOR REDUCING PARTICULATE PHOSPHORUS TRANSPORT FINAL REPORT SUBMITTED TO THE FLORIDA DEPARTMENT OF ENVIRONMENTAL PROTECTION By Samira H. Daroub, Timothy A. Lang. Orlando A. Diaz, Ming Chen, and James D. Stuck Everglades Research and Education Center Institute of Food and Agricultural Sciences University Of Florida June 2005

2 Acknowledgments This project, and the preparation of this report, was funded in part by a Section 319 h Nonpoint Source Management Program grant from the U.S. Environmental Protection Agency through a contract with the Stormwater/Nonpoint Source Management Section of the Florida Department of Environmental Protection. The total cost of the project was $3,000,000 of which $750,000 or 25% was provided by the U.S. Environmental Protection Agency. Financial support for the project was also provided by the Everglades Agricultural Area- Environmental Protection District. Page 2

3 TABLE OF CONTENTS Acknowledgments... 2 Executive Summary... 6 Quality Assurance / Quality Control Overall Objectives Farm Descriptions CHAPTER 1 Farm Verification of BMP Efficacy List of Figures Chapter List of Tables Chapter Introduction Objective Materials and Methods Results and Discussion Conclusions References CHAPTER 2 Specific Conductance in the Everglades Agricultural Area List of Figures Chapter List of Tables Chapter Introduction Objectives Materials and Methods Specific Conductance Monitoring Program in the EAA Statistical Data Analysis Historical Specific Conductance Information on South Florida Results I. General Characteristics of Specific Conductance in the EAA II. Potential Sources of Specific Conductance in the EAA III. Impact of Phosphorus BMPs on Specific Conductance Discussion Page 3

4 Conclusions References CHAPTER 3 On-Farm Particulate Phosphorus Measurement and Control List of Figures Chapter List of Tables Chapter Introduction Biological Contribution Mechanism Particle Erosion and Transport Primary Processes and Illustrative Examples Material and Methods Intensive-Study Program Elements Event Data Analysis Farm Sediment Surveys Concentrated Discharge Sampling Results Event Summary Statistics Particulate Phosphorus Load Distribution Analysis Farm Management Practices that May Impact Particulate P export Analysis of Major Event Contributing to Top 50% Particulate P Loads. 186 Event Analysis for UF9200A Event Analysis for UF9206B Event Analysis for UF9209A Concentrated Suspended Solids (Bulk Samples) Farm Sediment Surveys Discussion and Summary Farms Summary Phosphorus Content Dominant Events, Velocity, and Response Times Key Processes Demonstrated at Each Farm Conclusions and General Recommendations References Page 4

5 CHAPTER 4 Floating Aquatic Weeds and EAA Farm P Loads List of Figures Chapter List of Tables Chapter Introduction Aerial Survey of Emergent Aquatic Weeds BMP Demonstration Farm Objectives Materials and Methods Aerial Survey of Emergent Aquatic Weeds BMP Demonstration Farm Materials and Methods Results Aerial Survey of Emergent Aquatic Weeds BMP Demonstration Farm Results Conclusions References CHAPTER 5 Effects of BMPs on Crops and Soil List of Figures Chapter List of Tables Chapter Introduction Materials and Methods Results and Discussion Phosphorus Budgets Water Tables Soil Fertility Conclusions References Project Conclusions Educational and Extension Activities List of Publications Page 5

6 Executive Summary Agricultural Best Management Practices (BMPs) have reduced phosphorus (P) loading from the Everglades Agricultural Area (EAA) by over 50% since 1995 as reported by the South Florida Water Management District (SFWMD) data and evidenced by the University of Florida/Institute of Food and Agricultural Sciences (UF/IFAS) research. A definition of a BMP pertinent to the EAA was defined as: An alternative management practice that is technically feasible, economically viable, socially acceptable, and scientifically sound, that when implemented, will lead to reduced P concentrations and loads leaving farms in the EAA, while not threatening the viability of the agricultural industry. Initial studies by UF/IFAS that started in 1986 lead to the development of BMPs for reducing P concentrations and loads in the EAA. It was determined that fertilizer practices and a combination of improved drainage uniformity and a reduction in drainage pumping could yield significant reductions in P concentrations and loads for all crops. Using the results of these studies, and best professional judgment, expected reductions in P loading were attached to each BMP. It was hypothesized that P load reductions ranging from 20 to 60% could be realized for farms in the EAA and for the basin as a whole. The BMPs suggested by the UF/IFAS research, and others proposed by industry and the SFWMD, were selected by the SFWMD for inclusion in the table of BMP options used in compliance of Rule 40E-63. Mandatory BMP implementation in the EAA started in January In 1992, the UF/IFAS started a wide-scale implementation and BMP efficacy verification project aimed at quantifying the load reductions that could be achieved. Ten farms located through out the EAA were selected as being representative of soils, geographic location, crop rotations, and water management philosophies. Best management practice packages were developed for each farm and implemented. Monitoring of farm drainage volumes and total P concentrations began in Water Year The number of farms included in the project was reduced to seven in 2000 and to three in January Results of the BMP implementation and efficacy in the EAA study are reported in Chapter 1. Specific conductance monitoring was started on two farms in 1997 and eight farms in 1998 with a main goal of determining if specific conductance in EAA farm canals is man induced Page 6

7 and if it can be abated by additional BMPs. Related results on specific conductance in EAA canals are reported in Chapter 2. Particulate P source, transport and control demonstration started in 2000 on three farms with the objective of determining conditions and farm operations that resulted in high loads of particulate P. Another objective was to recommend management practices to reduce particulate P and therefore total P loads. Due to the importance of aquatic weeds in the particulate P loads, aerial surveys and surveys of aquatic weed coverage in two farm canals were also conducted over a two year period. In addition, demonstration plots were established at the EREC farm to demonstrate optimized BMP practices to growers. These optimized BMPs include aquatic weed control as well as velocity control in the canals. Results of particulate P source, transport and control demonstration are presented in Chapter 3 and results of Aquatic weed and BMP demonstration are presented in Chapter 4. The sugarcane lysimeter study was designed to demonstrate the effects of higher than traditional water table levels (that will occur naturally under BMP implementation) on 3 popular sugarcane cultivars as well as the effect of delivering nutrient-rich drainage waters (P-fertigation) to sugarcane. The vegetable/rice lysimeter study was designed to demonstrate short- and long-term soil fertility and crop nutrient uptake trends for different vegetable/rice/flooded fallow crop rotations. Drainage waters from the vegetable/rice lysimeters served as the P-fertigation source into select sugarcane lysimeters. The lysimeter study is presented in chapter 5. Numerous BMP training seminars and workshops were conducted. A listing of these workshops is presented at the end of the report. Research and extension publications are also listed at the end of the report. BMP Implementation and Verification All indicators of BMP efficacy have shown that consistent and sustained reductions in total P concentrations and loads have occurred due to the implementation of BMPs in the EAA. Basin-level numbers presented annually by the SFWMD reinforce the effectiveness of the BMP program, showing a sustained 50% reduction in total P loading from the EAA. In WY 2004, the TP load reduction from the EAA was 64% compared to the pre-bmp baseline Page 7

8 period. The three year average load reduction is 55%. Phosphorus concentrations are also reduced. In WY 2004, P concentrations from the EAA averaged 69 ppb compared to the pre-bmp base period P concentration of 173 ppb. This major and sustainable reduction is credited to the BMP program. Adjusted unit area loads on project farms averaged 0.73 lbs total P/acre after BMP implementation compared to 1.30 lbs total P/acre prior to WY95. This represents a project average reduction in adjusted unit area loads of approximately 44% which is close to the EAA basin load reduction. The UF/IFAS project baseline WY94 total P concentration was mg/l. Phosphorus concentrations in WY02 from project sites was 0.10 mg/l which was very similar to the basin level P concentration of mg/l. On the project sites, however, average concentrations are variable with years and reach 0.2 mg/l in certain years. In addition to the above, a simpler method of normalizing data was used. Volumes of drainage and P UALs were indexed to the rainfall amounts received during the year. This yielded figures for how much drainage occurred for each inch of rainfall as well as how much P left the farms on a per inch of rain basis. The Volume: Rainfall (V:R) ratio for the pre-bmp period was 0.45 inches of drainage for each inch of rain received. After BMP implementation, the ratio fell 13% to an average of 0.39 inches per inch of rain. This is indicative of the reduction in farm drainage that occurred due to BMP implementation. The P UAL to rainfall ratio dropped from lbs of P discharged per acre per inch of rain to , a reduction of approximately 6.5%. The Everglades Forever Act of 1994 mandated a research and monitoring program on the evaluation of water quality standards in the EAA (Chapter 40E-63). The goal of this research was to evaluate the constituents that have been previously identified as elements of water quality concern that will likely not be significantly improved by the Stormwater Treatment Areas and current Best Management Practices being widely implemented throughout the EAA; and to identify strategies needed to address such parameters (40E (2)). These parameters were identified by the Florida Department of Environmental Protection (FDEP) as specific conductance, particulate P, and the pesticides Atrazine and Ametryn. The Everglades Agricultural Area-Environmental Protection District (EAA-EPD) and the SFWMD are responsible for the monitoring of Atrazine and Ametryn from the EAA basin. The Page 8

9 UF/IFAS implemented a research project to investigate the specific conductance and particulate P issues in the EAA. Specific Conductance in EAA Farm Canals: The objectives of the specific conductance monitoring and research program as stated by Chapter 40E-63, Part III: "the farm-scale research shall be expanded to include monitoring for specific conductance at all points where total phosphorus is currently being monitored. The expanded research program shall include the development, testing, and implementation of BMPs to address reduction of specific conductance. Specific conductance was monitored at ten EAA farms (12 pump structures). All data were collected using Hydrolab DataSonde (series 3, 4, and 4a) multi-parameter water quality data loggers. In order to identify the specific ions and ion ratios that comprise specific conductance, weekly grab samples were taken in 2001 and 2002 from eight farms (10 pump structures) and analyzed for ionic composition. Summary statistics showed that mean specific conductance above ms/cm occurred at only two out of the ten farms monitored. The farms with conductance above ms/cm were UF9206A&B and UF9208A. Higher concentrations of sodium (Na + ) and chloride (Cl - ) were also observed at these two farms. Of the two farms, UF9208A, also showed high levels of sulfate (SO 2-4 ). Determination of ion compositions in grab samples at the ten pump structures indicated that the major anions are bicarbonate (HCO - 3 ), Cl - and SO 2-4 and the major cations are Na + and calcium (Ca 2+ ) in farm canal water of the EAA farms. Potential sources of specific conductance were evaluated. These included geological influences, drainage pumping, irrigation water and fertilizer application. Comparing average specific conductance data points of the study sites to historical Cl - concentration maps of shallow groundwater revealed that the current elevated farm conductance readings of UF9208A coincided with historically high Cl - concentrations in ft ground water wells. UF9206A&B also is located in an area that has wells of high Cl - concentration. The Na/Cl ratio in the farm canals ranged from 0.57 to The Na/Cl ratio in seawater is It has been reported that connate seawater underlies the area and exchanges with the surface Page 9

10 water where canals are cut into the limestone. Shallow ground water hydrology and quality has a major impact on specific conductance in the EAA. The effect of drainage pumping on specific conductance was variable and site specific. There was a low correlation between drainage pumping and conductance when all the sites were combined. Irrigation had a low negative correlation with specific conductance. Statistical analysis of the daily average specific conductance at three intensively monitored farms indicated that drainage pumping increased specific conductance at UF9200A and UF9209A, but not at UF9206A&B. Irrigation decreased specific conductance at all three farms, UF9200A, UF9206A&B and UF9209A. Drainage event analysis on the two elevated specific conductance farms (UF9206A&B and UF9208A) also demonstrated the variable effect of pumping. For example, out of six selected drainage events on UF9206A, three were observed to have increased conductance with volume pumped. Specific conductance had no relationship with drainage pumping to rainfall ratio. One farm that had the lowest drainage pumping to rainfall ratio, showed the highest specific conductance. This strengthened the conclusion that farm conductance is strongly influenced by underlying ground water composition. Irrigation had a weak negative correlation with specific conductance. On the three intensely monitored farms, UF9200A, UF9206A&B, and UF9209A, irrigation had the effect of decreasing specific conductance. The irrigation water utilized by the farms with the highest specific conductance (UF9206A&B and UF9208A) was also characterized by higher specific conductance. Farm UF9208A received irrigation water via a secondary canal that connects to the Hillsboro canal. Farm UF9206A&B received irrigation water from a secondary canal that connects to the Ocean canal. The Ocean canal may source its water from either the West Palm Beach Canal to the east, or the Hillsboro Canal to the west. Both the Ocean and the Hillsboro Canals have historically had relatively high specific conductance compared to other major district canals in the EAA. Previous research in the EAA indicated that potassium chloride (KCl) fertilizer application contributed less than 3% to the total dissolved solids (TDS) concentrations in canal waters. It is also reported that a sugarcane crop at harvest takes up more P and K from the soil than that applied by fertilizers. Our results show KCl fertilizer application in one of the high conductance farms with mixed cropping systems contributed less than 6.5% of the TDS in Page 10

11 drainage water. This was calculated assuming that all the KCl fertilizer ended up in the drainage water which is highly unlikely as crops take up K + and Cl - in large quantities. To assess the impact of current P load reduction BMPs on specific conductance, nonparametric Mann-Kendall trend analyses and Sen s slope analysis of specific conductance at different pump structures in the EAA were conducted. Both of these analyses indicated that downward trends were statistically significant at structures UF9202A, UF9205A and UF9207B during the study period. One farm UF9208A showed an upward trend using the Mann-Kendall trend analysis, however there was no significant trend using the Sen s slope analysis. In conclusion, specific conductance in the EAA canals is strongly influenced by the composition of the shallow ground water, historically reported to be high in Na + and Cl - due to connate seawater entrapment and the mixing of surface and ground water. The effect of drainage pumping was variable and site specific. Irrigation, in general, decreased specific conductance. Canal specific conductance is governed mainly by the quality and the hydrology of the underlying shallow ground water, which is farm specific. Fertilizers contributed a very small percentage to the total dissolved solids in the drainage water therefore had no substantial contribution to specific conductance in the EAA. Current P load reduction BMPs have reduced specific conductance in some locations in the EAA. It is the conclusion of this study that no further BMPs can be identified by additional research that would provide abatement of specific conductance in the discharge in the EAA. The issue of specific conductance in the EAA is a geological one, and shallow ground water is the major factor controlling the level of specific conductance in the EAA farm canals. Particulate P Measurement and Control in the EAA The objective of the particulate P research as stated by Chapter 40E-63, Part III: In recognition that substantial particulate matter such as sediments are being discharged from farms, given that published University of Florida Institute of Food and Agricultural Sciences data has demonstrated that particulate phosphorus constitutes a significant portion of total phosphorus, the farm-scale research shall be expanded to include the development, testing, and implementation of BMPs for reducing discharge of particulate phosphorus (i.e. sedimentation basins). Page 11

12 Phosphorus transport in runoff can occur in soluble and particulate forms. Particulate P consists of all solid phase forms including P sorbed by sediment particles and organic material transported during runoff. Particulate P accounts for 20% to 70% of the total P load exported from EAA farms and is frequently the cause of spikes in farms total P loads. The conclusion of our earlier studies suggests that a significant fraction of the particulate P in the EAA originates from in-stream biological growth rather than from field soil erosion. Recently deposited biological sediment material such as settled plankton, filamentous algae, and macrophyte detritus is the fraction that contributes the most to particulate P export. Exported solids may also be contributed directly from loosely bound material detached by turbulent shear forces of floating aquatic vegetation. Other contributions to particulate P loads come from submerged aquatic vegetation and planktonic growth. One of the primary goals of this study was to identify conditions that cause increased particulate P load rates, and analyze those conditions to determine operating procedures that may be optimized to reduce particulate P export, and therefore overall P export at the farm level. Load rate is the product of flow and concentration over a given unit time period. The particulate P demonstration study was conducted on three farms in the EAA: a sugarcane farm in the northern EAA (UF9200A), a mixed-crop operation in the eastern EAA (UF9206A&B), and a sugarcane farm in the western EAA (UF9209A). Each pump station was fully instrumented, and data was continually recorded for key parameters such as rainfall, pump flow rates, and inlet and outlet water levels. All pump stations are equipped with ISCO 3700 portable automatic samplers that collect water samples every 15 or 30 minutes and composite them into one- or two-hour discrete samples for analysis. All collected samples are analyzed for total suspended solids (TSS), total P, and total dissolved P (TDP). Particulate P is calculated as the difference between total P and TDP. The complexity and the diversity of the systems included are considerable. The approach that has been adopted here was to conduct various forms of cluster analysis to attempt to identify primary parameters that have had the most impact on particulate P transport at the study farms. Event analysis was conducted on the fraction of the total P load contributed by particulate P for each pump station over the four-year study period. The annual contributions from the particulate P loads to the total P loads have decreased in two of the three farms in Page 12

13 Particulate P at UF9200A decreased from an average of 50% over the last three years ( ) to 28% in The particulate P load contributions of UF9206A increased from 26% in year 2000 to 36% in years 2001 and 2002, and decreased to 27% in Particulate P load contributions from farm UF9206B decreased from 40% in 2000 to an average contribution of 36% during the last three years. At UF9209A the contribution from particulate P to total P load was almost constant, around 67% in 2001 and In 2003, UF9209A pumped its canals lower and longer than previous years, causing more sediments to be dislodged from the bottom of the canal and transported out of the farm, resulting in a particulate P contribution of 80% to the total P load. Load Distribution Analysis of the cumulative hydraulic and particulate P loads generated for each farm and year, showed that 50% of the annual particulate P loads was contributed by less than 25% of the hydraulic load. Process Distribution Analysis was conducted to determine the most probable mechanism for particulate P transport in the sub-events that contribute most to the annual loads, i.e. those in the top 50% of the load distribution. The objective of this analysis is to identify conditions that give rise to the increased particulate P transport events. The most distinctive pattern observed from this analysis is the number of farm-years that were dominated by few events. Data over the four-year study shows that six of the 15 farm-years sampled had a single event that contributed 30% or more to the top 50% particulate P load. Three farm-years had two events that contributed a total of 30% or more. Three farm-years had three events that contributed a total of 30% or more. Only two of the 15 farm-years had their load rates distributed such that it took more than three events to contribute a total of 30% or more to the top 50%. Periodically, large volume ( liter) composite samples were taken at each of the study farms. These samples were concentrated by sedimentation in the field. The sedimented solids were collected and further concentrated in the lab, after which they were analyzed for the same physical and chemical properties as the farm sediments, including bulk density, solids content, particle specific gravity, organic matter content, and P content. Selected samples were analyzed for particle size distribution and settling velocity distribution. The analysis from the concentrated suspended solids (bulk samples) showed that the exported suspended solids volume is relatively small when present in its settled state. However, the contribution to the total annual P load of the farm could be significant. Page 13

14 Thus the importance of the suspended solids on the overall water quality of the farm must be considered when solids removal and control plans are being evaluated. Farm Sediment Surveys were conducted with the objective of determining the P storage in the main canal sediments, to evaluate farm sediments properties, and to monitor changes in sediment character and inventory over time. Quarterly inventories were conducted of canal sediment volume, mass, and P content at each farm. A number of transect locations were set up at regular intervals upstream of the pump station at each farm. Canal sediment surface elevation and depth was determined at each location. Core samples of the sediment were taken at each transect, sectioned and analyzed for key physical and chemical parameters, including bulk density, solids content, particle specific gravity, organic matter content, and P content. The surveys reported cover the 22-month period from November 2000 through August 2002 for UF9200A, UF9206B, and UF9209A. It appears that there was a trend toward sediment accumulation over the study period at UF9206B and UF9209A, while at UF9200A sediment depth remained relatively constant. The P content (on a dry weight basis) typically decreases as depth increases, but the bulk density of the sediment increases as depth increases. To understand the transport of particulate P in farm canals in the EAA it is necessary to identify and state the primary processes of movement. Major processes identified affecting particulate P movement were first flush, cumulative high velocity, restart flush, particulate phosphorus spike, and pump cycling. The first flush includes biological material accumulated during the quiescent period between pumping events. This highly mobile material causes an increase in the concentration of suspended solids during the first hours of pump events. Cumulative high velocity produces a steadily increasing discharge concentration of suspended solids as the water farther upstream has a longer time to accumulate eroded suspended solids as it moves downstream to the discharge pump station. Restart flush is similar to first flush. When pumping is terminated, suspended solids in the canal system settle out in place. If there has been a significant concentration of suspended solids in the downstream reaches of the canal system at shutdown, there will be a high initial concentration in the discharge when the pump is restarted. Particulate P spikes occur occasionally. A particulate P spike is defined when the particulate P concentration for a particular sample is more than twice that of either the preceding or Page 14

15 succeeding samples. The spike is assumed to originate from a random release of particulate material from upstream sources, such as a collection of floating macrophytes or a removal of a flow obstruction. Pump Cycling differs from pump restart in that the pump cycles through on-off oscillations over relatively short time periods, e.g. 30 minutes to two hours. This condition occurs when a farm pump is on automatic on-off control that is tied to canal level. The diversity of the farms has allowed a number of observations to be made regarding the importance of various operating parameters affecting particulate P loads. Dominant events started when pumping operations deviated from typical practices, but these deviations were specific to each particular farm. A detailed operational summary of study farms with specific recommendations to each farm is presented in Chapter 3. But following is a short description of recommendations to reduce particulate P loads. Recommendations to Reduce Particulate P Loads: Velocity in Canals Velocity is a key control parameter for reducing particulate P export. Recommended velocities are relative, in that they must be within the operating framework of the configuration of the farm. Velocities should be as low as possible, and velocity excursions should be avoided, regardless of the average or typical velocity of the canal system. Velocities greater than 0.4 m/s (1.3 ft/sec) have been associated with greater transport rates at the study farms. Given the parabolic relationship between velocity and erosion, slow and long periods is preferred than fast and short periods for pumping a given volume of water. Pump Cycling and Reduced Run Times Long-run period cycling of about 8-16 hours, which reduces continuous pumping duration, has been shown to be beneficial in interrupting continued high velocity transport. This was evidenced on farms where the response time of the farm hydraulic system (i.e., the time required from pump start-up to the time when the equivalent of one volume of farm canal water is exported) is greater than the pump cycling period. Short period cycling of one hour or less is detrimental and should be avoided. Level Control Control of canal water levels is critical in avoiding major velocity excursions, and also to stay away from large deviations of the normal farm canal velocities. Lack of level control or major changes in minimum canal levels have resulted in dominant events at the two farms that did not practice strict canal water level control. Canal levels Page 15

16 should be controlled to give minimum canal depths that do not exceed the maximum velocity recommendation. Aquatic Weed Control Weed control programs in the main canals is one of the most productive techniques in reducing the supply of high P content biomass. Physical removal along the entire length of the main canals is expensive to implement and not practical. For that reason, installation of weed-retention booms is recommended to be located at a distance >300 m (984 ft) upstream the main pump station. Spot spraying of weeds closest to the pump station is also recommended. Chemical treatment of major weed infestations will lead to the accumulation of transportable material into the bottom of the canal and is not recommended. Influence of Floating Aquatic Weeds on Everglades Agricultural Area Farm P Loads To achieve additional reductions in the EAA farm P exports through improvements in BMP implementation, the processes of P cycling, especially particulate P production in farm canals require better elucidation. In EAA farm canals the predominant floating or emergent aquatic plant species are water lettuce (Pistia stratiotes), water hyacinth (Eichhornia crassipes), and water pennywort (Hydrocotyle verticillata). These three fast-growing aquatic plants are capable of quickly covering a farm s entire system of canals and ditches, effectively inhibiting drainage of farm fields and increasing canal drainage velocities. To begin the process of garnering the necessary information to enable the determination of the feasibility of controlling the growth and senescence of aquatic weeds, monitoring of the weed growth mass in the main canals and field ditches over the course of two weed cycles was determined. Additionally, the P content of the weeds was determined in order to calculate P removal or P re-introduction potentials. Aerial surveys of the farm main canals, coupled with ground surveys of the field ditches, were undertaken for a two-year period. These surveys yielded the total area of surface waters covered by aquatic weeds. The P mass estimate was determined through aerial surveys of the farm main canals coupled with ground surveys and physical sample collection. Aerial photographs of the main farm canals and field ditches were taken once per month from October through March and twice per month during April through September. The aerial reconnaissance cycle began in July 2000 and ended in June Page 16

17 As additional assessment of possible P load reductions from EAA farms, the influence of these floating aquatic weeds on farm P loads was demonstrated. In theory the elimination of emergent aquatic weeds should provide conditions that optimize P co-precipitation with calcium carbonate from the canal water column, a process that occurs during active photosynthesis by submerged aquatic plants growing in waters saturated with calcium carbonate (DeBusk and Dierberg, 2003). Photosynthesis-induced calcium carbonate precipitate contains P that is of low bio-availability and relatively low transportability. Optimizing P co-precipitation in main farm canals was identified as a means to encourage the sequestering of P in less mobile canal sediments and to allow for eventual recycling of canal sediments back to farm fields. The impacts of controlling emergent aquatic weed and drainage flow velocity were assessed and demonstrated at the Everglades Research and Education Center s BMP Demonstration Farm. The BMP demonstration sugarcane farm consists of two hydraulically isolated sugarcane blocks of 125 and 200 acres each. Each block is equipped with identical drainage pumps and monitoring instrumentation to record rainfall, flow, canal levels and to collect discrete hourly drainage water samples. The BMP farm was established to demonstrate to growers the operational differences between an optimized BMP sugarcane farm and a conventional BMP sugarcane farm. Demonstration farm data, i.e., drainage volume, P species concentrations, total suspended solids, canal levels, flow velocities, and rainfall were analyzed to assess the effectiveness of optimized BMPs. This study compared the effects of velocity control and floating aquatic weed management on particulate, dissolved, and total P farm loads. It was demonstrated that an aquatic weed crop serves as a substantial reservoir for P during a season. A process by which the weeds could be grown as a crop and incorporated into a management practice to mitigate total P loading is not currently foreseeable. It is apparent that controlling the growth of weeds by conscientious and consistent spraying, to minimize huge infestations of macrophytes in the canals, will reduce particulate production and subsequent particulate P loads. Equally apparent from this study and past studies, is that aquatic plants and animals can significantly affect the P cycle, and hence P loading, in EAA farm canals. The physical removal of weeds from open channels should help limit the available P in the water column. However, a system must be devised that minimizes major dislodgment of detritus during weed removal. There are many factors that contribute to a very complicated phosphorus cycle in an EAA farm canal and ditch system. With that Page 17

18 understanding, and beyond the obvious notion that the physical removal of aquatic weeds from the canal and ditch systems will remove a substantial amount of P from the water column, there is no clear indication that a simple and sustainable management practice can be devised for incorporating the intentional growth and removal of aquatic weeds into a total P load reduction BMP. A second goal of this task was to assess and demonstrate the combined effects of drainage flow velocity and floating aquatic weeds on the P loads exported in the drainage waters from sugarcane fields. Results confirm the hypothesis that particulate P source control (removal of floating aquatic weeds) and application of critical velocity limits lead to measurable P load reductions. Drainage water concentrations of total P, total dissolved P, and particulate P for the BMP block were 54, 58 and 51 % lower than Control block concentrations. BMP block unit area loads for total P, total dissolved P, and particulate P were 28, 21, and 32 % lower than corresponding loads from the Control block. The observed P load reduction in the BMP block most likely is a result of decreases in easily transportable particulate P as well as the absence of conditions that allow export of less transportable P sources (canal sediments). Under the study s present arrangement it is difficult to determine what fraction of the P load reduction is due to source control and what fraction is due to critical velocity control. The efficacy of each practice may be determined separately. Blocks could be operated with identical velocity controls and differing levels of aquatic weed control. Contributions to P load from controlled aquatic weed growth vs. uncontrolled weed growth could then be compared. Conversely, both blocks could be operated with identical levels of control of floating aquatic weeds and different drainage velocities to compare the effect of velocity on P load. Effects of BMPs on Crops and Soil A large-scale lysimeter demonstration project was started in December 1997 to understand the efficacy of various proposed BMP strategies and potential impacts of BMP implementation on long-term soil fertility and crop production trends. The lysimeter site included 25 lysimeters: 11 smaller units dedicated to vegetable (crisphead lettuce) and rice cropping systems and 14 larger units planted to sugarcane. The sugarcane lysimeter assessment was designed to demonstrate the effects of higher than traditional water table (WT) levels (that occur under BMP implementation) on 3 popular sugarcane cultivars as well Page 18

19 as the effects of delivering nutrient-rich drainage waters (P-fertigation) to sugarcane. The vegetable/rice lysimeter study was designed to demonstrate short- and long-term soil fertility and crop nutrient uptake trends for different vegetable/rice/flooded fallow crop rotations. Drainage waters from the vegetable/rice lysimeters served as the P-fertigation source into select sugarcane lysimeters. Averaged across all lysimeter treatments for the entire duration of the 36-month study period the vegetable/rice P export (122 lbs P/ac) was almost eleven times greater than the P amount exported by sugarcane (11 lbs P/ac). This large difference arose from the great difference in fertilizer P input into the two systems. The vegetable/rice treatment received approximately 6.6 times greater fertilizer P input than the sugarcane treatments. Water extractable P soil test values in the sugarcane lysimeters did not change appreciably over the course of the project and the treatment that included vegetable irrigation waters did not show any marked increase in soil test P. Water extractable soil test P levels for the vegetable/rice treatments increased while fertilizer P was being applied at high rates, but once water extractable soil P levels reached a plateau (~40 lbs P/ac), resultant fertilizer P additions were subsequently reduced, soil P levels decreased to approximately 30 lbs P/ac. Depth of water table had no effect on P load exported from sugarcane grown under two water table regimes, 18 to 24 inch and 14 to 20 inch water tables. The sugarcane treatment that received vegetable drainage water exported 2.7 times more P in drainage waters than sugarcane that did not receive vegetable drainage water. From this field lysimeter assessment it appears that routing vegetable and rice drainage waters through sugarcane fields is an effective practice to lower vegetable drainage water P loads, but is somewhat limited by the timing and intensity of the specific rainfall event and the stage of growth of the sugarcane receiving the drainage water. Project Conclusions Results of over 10 year BMP research on EAA farms have shown the efficacy of implemented BMPs. Each farm in the EAA implements a suite of BMPs totaling 25 points. Success is observed on individual farms, and on the EAA basin as a whole. The Everglades Forever Act (EFA) mandates a 25% load reduction of P from the Page 19

20 EAA basin. The average reduction is measured and calculated by the SFWMD after adjusting to rainfall. The total P load reduction in WY 2004 was 64% and the last three-year average total P reduction is 55%. Total P concentrations from the EAA have also decreased. The actual WY 2004 total P concentration from the EAA with BMPs implemented is 69 ppb. Prior to BMP implementation, the average EAA total P concentration was 173 ppb. Specific conductance was not an issue in the majority of the EAA farm canals monitored. Out of the ten farms that were monitored, only two had average specific conductance higher than ms/cm. Specific conductance in the EAA canals is strongly influenced by the composition of the shallow ground water, historically reported to be high in Na + and Cl - due to connate seawater entrapment and the mixing of surface and ground water. The effect of drainage pumping was variable and site specific. Canal specific conductance is governed mainly by the quality and the hydrology of the underlying shallow ground water, which is farm specific. Fertilizers contributed a very small percentage to the total dissolved solids in the drainage water therefore had no substantial contribution to specific conductance in the EAA. Current P load reduction BMPs have reduced specific conductance in some locations in the EAA. It is the conclusion of this study that no further BMPs can be identified by additional research that would provide abatement of specific conductance in the discharge in the EAA. The issue of specific conductance in the EAA is a geological one, and shallow ground water is the major factor controlling the level of specific conductance in the EAA farm canals. Results from the particulate P research, that included monitoring of three farm operations as well as aerial survey of two farms, indicate that there are certain operating procedures when implemented could lead to reductions in the transport of particulate P and therefore overall P export. This should help substantially with reducing P spikes that are occasionally observed in farms concentrations and loads. Aquatic weeds in EAA farm canals are a major source of particulate P loads. Increased particulate P loads may occur from transport of moderate amounts of high P content readily transportable biological material close to the pump station. This light material can be transported at moderate flow rates, for example at pump start- Page 20

21 up after long inter-event time periods. Increased particulate P load rates may also occur from transport of large amounts of lower P content sediment material over a short period of time. This type of increased particulate P load rate could occur during high pumping rate events, that causes canal level to drop close to the bottom, increasing flow velocity, and resulting in the dislodging and transport of base sediment material in the canal. Recommendations to reduce particulate P loads and occasional spikes include reducing velocity in canals, maintaining a minimum canal level, and controlling the growth of macrophytes in farm main canals. Reducing velocity in the canals will reduce the loads of particulate P transported off the farm especially the lower P content sediment material. Reducing velocity in canals can be achieved by pumping at a slower rate for a longer period of time and practicing canal level control. In the EAA, pumping rates may be easily doubled or tripled by running multiple pumps or switching from small to large capacity pumps. Velocities may also change rapidly when canals are drawn down to low levels. Floating aquatic weeds contribute to the readily transportable biological materials close to the pump station, but the best approach for aquatic weed management may not be straight forward. Weed booms are recommended to keep the floating aquatic weeds away from the pump station. Spot chemical treatment is also recommended to control aquatic weeds and prevent major infestation. Chemical treatment of major aquatic weed infestations, however, will lead to the death and accumulation of highly transportable sediments. It was demonstrated that an aquatic weed crop serves as a substantial reservoir for P during a season. It is apparent that controlling the growth of weeds by physical removal and a conscientious and consistent spraying will reduce particulate production and subsequent particulate P loads. Spraying heavily infested canals with aquatic weeds will lead to the accumulation of highly transportable biological materials and is not recommended. A process by which the weeds could be grown as a crop and incorporated into a management practice to mitigate total P loading is not currently foreseeable. Data from the demonstration farm at the Everglades Research and Education Center confirm the hypothesis that particulate P source control (removal of floating aquatic weeds) and application of critical velocity limits Page 21

22 lead to measurable P load reductions. The observed P load reduction in the BMP block most likely is a result of decreases in easily transportable particulate P as well as the absence of conditions that allow export of less transportable P sources (canal sediments). Results from the lysimeter study show that the vegetable/rice P export was approximately 11 times greater than the P amount exported by sugarcane. This difference arose from the great difference in fertilizer P input into the two systems. The vegetable/rice treatment received approximately 6.6 times greater fertilizer P input than the sugarcane treatments. Water extractable P soil test values in the sugarcane lysimeters did not change appreciably and the treatment that included vegetable irrigation waters did not show any marked increase in soil test P. Water extractable soil test P levels for the vegetable/rice treatments increased while fertilizer P was being applied at high rates but decreased as fertilizer P additions decreased. Depth of water table had no effect on P load exported from sugarcane grown under two water table regimes, 18 to 24 inch and 14 to 20 inch water tables. The sugarcane treatment that received vegetable drainage water exported 2.7 times more P in drainage waters than sugarcane that did not receive vegetable drainage water. From this field lysimeter assessment it would appear that routing vegetable and rice drainage waters through sugarcane fields is an effective practice, but is limited by the timing and intensity of the specific rainfall event and the stage of growth of the sugarcane receiving the drainage water. Page 22

23 Introduction The Everglades Forever Act (EFA) requires that the Everglades Agricultural Area (EAA) basin reduces total phosphorus (P) loads by 25 percent compared to the pre-best Management Practice (BMP) base period. The BMP program was initiated in the EAA in 1995 with a suite of BMPs required to be implemented by each grower in the area. Since the program initiation, the EAA s annual percentage load reduction has averaged greater than 50 percent as reported by the South Florida Water Management District (SFWMD). This success is also evidenced by the University of Florida/Institute of Food and Agricultural Sciences (UF/IFAS) research. A definition of BMP pertinent to the EAA was defined as: An alternative management practice that is technically feasible, economically viable, socially acceptable, and scientifically sound, that when implemented, will lead to reduced P concentrations and loads leaving farms in the EAA, while not threatening the viability of the agricultural industry. The UF/IFAS agricultural BMP research and education program to reduce P concentrations and loads in the EAA began in Earlier studies conducted at UF/IFAS had focused mainly on determining the effects of fertilizer rates and application methods and water management practices on drainage water P loads and concentrations. Results of this initial work showed that sugarcane fertilizer related practices did not lead directly to short-term P concentration and load increases since application amounts are low. Vegetable crop fertilization practices, however, were found to be capable of causing short-term P spikes given certain hydrologic and cultivation conditions. Banding fertilizer for vegetable crops at reduced rates, and using proper fertilizer handling and application methods for all crops, could lessen the occurrences of fertilizer related short-term increases in P concentrations and loads. The study also showed that water management practices greatly affected both P concentrations and loads. Obviously, pumping less water off farms yielded large and immediate reductions in P loads. However, drainage rates and uniformity could also greatly affect P loads and concentrations positively or negatively. It was, therefore, determined that a combination of improved drainage uniformity and a reduction in drainage pumping could yield significant reductions in P concentrations and loads for all crops. Page 23

24 In addition to the above, it was determined that sugarcane land yielded P concentrations and loads that were equal to, or lower than, fallow land that was allowed to flood and drain. Also, in spite of the effects of the higher fertilization rates, and the more demanding water management requirements of vegetable crops, vegetables could be grown in the EAA without greatly affecting P concentrations and loads. To accomplish this, the growers were advised to block farms into hydraulically isolated units and to rotate crops accordingly. By doing so, vegetable field drainage water would not directly enter the farm drainage stream, but rather would be diverted to sugarcane blocks for retention or detention. Additionally, it was determined that growing rice in rotation with vegetables, which requires no additional P fertilizer, yielded a major export of P from the EAA in grain and held nutrients in plant matter, which could then be reincorporated into the soil. It was cautioned, however, that drain down of the rice floodwater be accomplished in a manner such that it did not directly enter the drainage stream. Allowing the floodwater to subside naturally leaves residual nutrients in the field for use by the succeeding crops. Additional drainage required to lower water table levels should be discharged to sugarcane field blocks, keeping the nutrient enriched water from directly entering the farm drainage stream. This practice was extrapolated to pertain to all floodwaters, regardless of whether they originated from rain, or purposeful flooding of fallow fields. The initial studies lead to the development of BMPs for reducing P concentrations and loads in the EAA. Using the results of the study, and best professional judgment, expected reductions in P loading were attached to each BMP. It was hypothesized that P load reductions ranging from 20 to 60% could be realized for farms in the EAA and for the basin as a whole. The BMPs suggested by the UF/IFAS research and others proposed by industry and the SFWMD, were selected by the SFWMD for inclusion in the table of BMP options used in compliance of Rule 40E-63 (Table 1). Mandatory BMP implementation started in January In 1992, the project moved to wide-scale implementation and BMP efficacy verification aimed at quantifying the load reductions that could be achieved. Ten farms around the EAA were selected as being representative of soils, geographic location, crop rotations, and water management philosophies. Best management practice packages were developed for each farm and implemented. Monitoring of farm drainage volumes and total P Page 24

25 Table 1. Best Management Practices summary and BMP equivalent points (Everglades Consolidate Report, SFWMD 2003.) BMP PTS DESCRIPTION WATER MANAGEMENT PRACTICES ½ Inch Water Detention 1 Inch Water Detention 5 10 MINIMIZES THE VOLUME OF OFF-SITE DISCHARGES Delay pumping based on rain gage measurements. Detention (in farm canals and soil profile) measured on a per event basis rainfall vs. runoff. Improved Infrastructure 5 Water table management plan; controlling levels in canals and field ditches using internal water control structures, fallow fields, aquatic cover crop fields, prolonged crop flood; effective irrigation and discharge plans. Other NUTRIENT CONTROL PRACTICES Fertilizer Application Control tbd 2 ½ Properly constructed and maintained storage system; greater detention with water management plan having target water table levels and structure operating procedures; monitored water table. MINIMIZES THE MOVEMENT OF NUTRIENTS OFF-SITE * Limited Applicability Uniform and controlled boundary fertilizer application (e.g. banding at the root zone; pneumatic controlled-edge application such as AIRMAX); calibrated application equipment; setbacks from canals. Fertilizer Spill Prevention 2 ½ Formal spill prevention protocols (handling, transfer, education). Soil Testing 5 Avoid excess application by determining P requirements of soil. Plant Tissue Analysis 2 ½ Avoid excess application by determining P requirements of plant. Split P Application* 5 Applying P proportionately at various times during the growing season. Total application not exceeding recommendation. Slow Release P Fertilizer* 5 Applying specially treated fertilizer that breaks down slowly thus releasing P to the plant over time. PARTICULATE MATTER AND SEDIMENT CONTROLS MINIMIZES THE MOVEMENT OF PARTICULATE MATTER AND SEDIMENTS OFF-SITE (Each consistently implemented across the entire basin acreage.) Any 2 Any 4 Any 6 2 ½ 5 10 leveling fields cover crops ditch bank berm raised culvert bottoms sediment sumps in canals stabilized ditch banks sediment sumps in field ditches aquatic plant management canal/ditch cleaning program debris barriers at outfall slow drainage velocity near pumps sediment sump upstream of drainage structure PASTURE MANAGEMENT PLAN FOR ON-FARM OPERATION AND MANAGEMENT PRACTICES Pasture Management 5 reduce cattle waste nutrients in discharges by "hot spot" management, i.e. plans for placement of drinking water, feed and supplements, cowpens and shade. low cattle density OTHER BMPs OTHER PRACTICES PROPOSED Urban Xeriscape 5 Use of plants that require less water and fertilizer. Det. Pond Littoral Zone 5 Vegetative filtering area for on-site stormwater runoff. Other BMP Proposed tbd BMP proposed by permittee and accepted by SFWMD. Page 25

26 concentrations began in WY1993. During the first few years of this part of the study, farm and basin data showed that BMP packages could reduce P loads by 50% on a recurring basis. The number of farms included in the project was reduced to seven and eventually to three in January All indicators of BMP efficacy have shown that consistent and sustained reductions in total P concentrations and loads have occurred due to the implementation of BMPs in the EAA. Basin-level numbers presented annually by the SFWMD reinforce the effectiveness of the BMP program, showing a sustained 50% reduction in total P loading from the EAA over the course of this project. However, during any given year, there appears to be between 3 and 10 events, generally associated with heavier rainfalls and particular antecedent conditions that show substantial total P concentration and load spikes. Although these events occur relatively infrequently during the year, their inordinately high total P loads account for a substantial percentage of the total annual load. These spikes may be attributed to elevated particulate P concentrations, elevated dissolved P concentrations, larger drainage volumes, or a combination of the three. Assessing methods to minimize these spikes would help to further reduce farm total P loading. Since most growers have already taken many of the steps to reduce total P loads from the 0.5 to 1.0 inch rainfall events, focusing attention on these generally larger rainfall events seems to be appropriate. There appears to exist a real potential for reducing the magnitude of the total P concentration factor in the total P load equation using BMPs that focus on the management of farm canal P sources and sinks relative to particulate matter. This particulate matter has been shown to be responsible for about 60% of the total P leaving a farm during drainage. Furthermore, the particulate matter appears to be sourced primarily from floating aquatic vegetation. Hence, it appears that the EAA farm canal systems act as temporary P sinks, but that their usefulness as P sinks is negated at the onset of the next drainage event or by the onset of the aquatic plants senescence phase. Demonstrating the ability to beneficially use aquatic plant growth and the canal P-cycling phenomenon at the farm level, could lead to even greater reductions in farm P concentrations. This reduction should then be reflected at the basin level. The Biological Contribution Mechanism states that the majority of the P associated with particulate matter in the EAA is sourced from in-stream biological growth rather than soil erosion. The Biological Contribution Mechanism includes sediment erosion Page 26

27 as a source of exported particulate P. Sediments that contribute significantly to the P export is, for the most part, recently deposited biological material such as settled plankton, filamentous algae, and macrophyte detritus. An agricultural BMP is an alternative practice that constantly undergoes modifications and refinements such that it: a) improves performance of the primary function of crop production; b) reduces detrimental side effects of the performance of the primary function; c) optimizes the input versus output relationships with respect to labor, time, direct expenses, and natural resources; and d) continues to improve upon its ability to perform the first three functions. In this particular case, the BMPs are aimed at optimizing the use of water and nutrients (both naturally occurring and applied), while reducing specific conductance and P loading to off-farm canals and maintaining production levels and adequate profitability to ensure the sustainability of agriculture in the EAA. The Everglades Forever Act of 1994 mandated a research and monitoring program on the evaluation of water quality standards in the EAA (Chapter 40E-63). The goal of this program was to evaluate the constituents that have been previously identified as elements of water quality concern that will likely not be significantly improved by the Stormwater Treatment Areas and current Best Management Practices being widely implemented throughout the EAA; and to identify strategies needed to address such parameters (40E (2)). These parameters were identified by the Florida Department of Environmental Protection (FDEP) as specific conductance, particulate phosphorus, and the pesticides Atrazine and Ametryn. The Everglades Agricultural Area-Environmental Protection District (EAA-EPD) and the South Florida Water Management District (SFWMD) are responsible for the monitoring of Atrazine and Ametryn. The UF/IFAS undertook a research and monitoring project sponsored by the Everglades Agricultural Area Environmental Protection Distirct (EAA-EPD) and the FDEP to answer the questions regarding particulate P and specific conductance in the EAA. Page 27

28 Quality Assurance / Quality Control The Everglades Research and Education Center laboratory is certified by the State of Florida Department of Health (FDOH), Bureau of Laboratories in the categories of nonpotable water general chemistry (conductivity and phosphorus, total) with successful ongoing compliance with the NELAC Standards and FAC Rule 64E-1 regulations (FDOH, 2005) since Specific methods and analytes certified are cited on the Laboratory Scope of Accreditation for our laboratory and are on file at the Bureau of Laboratories, P. O. Box 210, Jacksonville, Florida Based on the general requirements, proficiency test samples from Environmental Resource Associate are analyzed and results reported to the FDOH twice a year. In addition, our laboratory is involved in the Everglades Round Robin Inter-laboratory comparison program initiated by the Florida Department of Environmental Protection since 1995 for the purpose of assessing the comparability of P data from laboratories engaged in the analysis of samples from the Everglades. All field and laboratory activities and analyses in this project strictly followed relevant Standard Operating Procedures and Quality Assurance / Quality Control criteria documented in the quality manual for the laboratory of the Everglades Research and Education Center (Chen, 2001). References Chen, M Quality manual for the laboratory of the Everglades Research and Education Center (NELAC Certification #E76463). University of Florida, Belle Glade, FL. FDOH The Department of Health Rule, 64E-1, FAC - Certification of Environmental Testing Laboratories. Page 28

29 Overall Objectives The overall objectives of this project were: 1. To demonstrate the long-term viability of on-farm BMPs, including improvements on the current BMPs that are already implemented. This objective includes assessing and implementing BMPs to mitigate total P concentration and load spikes that occur a few times during the year. 2. To aid in the enhancement of the uniformity of BMP implementation across the EAA through demonstration and educational programs. 3. To identify the P sources, P content, and P cycling characteristics of the main farm canal floating aquatic plants, sediments, and suspended particulate matter and to demonstrate how they can be used to further reduce farm-level P loading. 4. To demonstrate the effects of BMPs designed to control growth and remove floating and suspended aquatic plants and animals on farm level drainage water total P concentrations and loads. 5. To demonstrate farm water management systems that could lead to greater levels of particulate matter retention. 6. To demonstrate the long-term effects of fertilizer, water management, and surface water fertigation BMPs on sugarcane plant and ratoon crop P uptake, production, and soil fertility. This portion of the project shall be done in controlled lysimeters (water and nutrient inputs and outputs). The report is divided into 5 chapters. Each chapter addresses one or more objectives. Chapter 1, Farm verification of BMP efficacy, addresses Objective 1 which asks for the demonstration of long term viability of BMPs in the EAA. A list of BMP training workshops and extension activities to address Objective 2 is provided at the end of the report. A list of publications including extension publications are also listed at the end of the report. Specific conductance in the EAA canals and the effect of current BMPs on specific conductance is Page 29

30 covered in Chapter 2. Chapter 3, On-farm particulate phosphorus measurement and control, addresses objectives 3 and 5 which include identifying the P sources, P content, and P cycling characteristics of the main farm canal floating aquatic plants, sediments, and suspended particulate matter and demonstrating farm water management systems that could lead to greater levels of particulate matter retention. Chapter 4, Influence of floating quatic weeds on EAA farm P loads, addresses objective 4 which it to demonstrate the effects of BMPs designed to control growth and remove floating and suspended aquatic plants and animals on farm level drainage water total P concentrations and loads. Chapter 5, BMP crop and soil effects- lysimeter study, addresses objective 6: To demonstrate the long-term effects of fertilizer, water management, and surface water fertigation BMPs on sugarcane plant and ratoon crop P uptake, production, and soil fertility. This portion of the project was conducted in controlled lysimeters (water and nutrient input and output). A separate section on the description of EAA farms used in this project and BMPs implemented on each farm is included in the report. Page 30

31 Farm Descriptions All or some of the ten farms described below were used for the different demonstration studies in this project. The BMP programs for the ten farms operating under project sponsored BMP technologies were developed cooperatively between project personnel and the grower. Emphasis was placed on designing practical farm management programs that would accommodate the growers capabilities and goals regarding P reductions in off-farm discharge. The BMP programs have been tailored to each site in cooperation with the growers, who continue to experiment and fine-tune their management systems. Annual BMP trainings are provided and customized to individual companies by project personnel to better improve uniformity of implementation. UF9200A Site UF9200A (1280 acres) started BMP operations in January The grower was not comfortable with placing a booster pump in his main farm canal (Figure 0A-BMP) and, upon further inspection, it was determined that the canal might be too shallow to accommodate an adequate installation. Additionally, the grower felt that drainage capabilities had to be maintained at the head of his farm since muck depths are believed to be shallower there. In the interest of reducing his discharge TP concentrations and loads, the grower modified three primary management activities. First, the grower focused on reducing the inordinate amount of irrigation water being let into the farm. In the past, this grower tended to irrigate frequently in his attempt to micro-manage water table levels. This micro-management led to increased drainage pumping and often involved pumps being switched directly from the irrigation to the drainage mode. The grower has also monitored farm drainage more closely, allowing at least one inch of rain to accumulate prior to pumping initiation and, at the end of the drainage event, allowing evapotranspiration (ET) to rid fields of at least one inch of water that would otherwise have been pumped off-farm (by turning the pumps off earlier and letting water tables redistribute). The grower has also altered the stage-controlled settings for pump start and stop on his electric pump. Page 31

32 Ditch level recorder (inoperative) UF9200W3 N Water table recorder (inoperative) UF9200W4 Temporary pump 4,360 ft Main Farm Canal Water table recorder UF9200W8 UF9200W9 Ditch level recorder LEGEND: Field ditch Roadway Canal Pump Riser Culvert Flow direction Rainfall collector Weed Boom, ~ ft upstream Ditch level recorder UF9200W5 Water table recorder UF9200W7 Temporary monitoring station UF9200B * all field ditches are joined to the main farm canal and the perimeter ditch by riser and culvert connections not shown. Water table recorder UF9200W6 5,032 ft Water level recorder UF9200W11 UF9200W10 Water table recorder UF9200W2 Ditch level recorder UF9200W1 634 ft 2,560 ft Farm pump UF9200A Figure 0A-BMP: UF9200A hydraulic system and farm layout. Page 32

33 The second major BMP activity involved the cleaning of the majority of the farm ditches. The cleaning reduced sedimentary material that had accumulated in the ditches, increased farm water storage capacity, and enhanced the ability of the grower to route drainage flows through his ditch and main canal system. An early BMP involved the setting aside of two field blocks (about 80 acres) at the back of the farm to serve as a water retention area. The grower left these field blocks flooded fallow and pumped water into them for storage as well as for increased drainage capabilities at the back of the farm. This system was operational from mid-may through mid-november Since then, the fields have been replanted to sugarcane and the practice has been discontinued. A recent BMP includes the installation (June 1998) of a weed boom across the main farm canal roughly 100 feet from the discharge pump structure. The purpose of this device is to trap floating aquatic plants, sequestering them downstream from the scouring zone of the pump structure. The effective use of weed booms should reduce the component of off-farm P loading attributable to floating plant parts (and attendant biological detritus) and the potentially high-p aquatic ecosystem associated with the submerged root mass. Preliminary investigations with weed booms will aid in the future design of more comprehensive aquatic weed control BMPs. Components of these BMPs may require additional traps to limit the import of aquatic plants during irrigation, the deliberate encouragement of weed growth within farm canals to sequester dissolved P into plant biomass, and an effective and efficient means to harvest this biomass in order to remove accumulated P from the aquatic system. Intensive work was done at the site with respect to reducing particulate P transport. The site was inventoried for aquatic weed coverage, aquatic plant species, and sediment characteristics. It is suspected that lowering pump start up energy imparted to the water, pumping at a slower and continuous rate to reduce scouring, and interevent hydraulic vacuuming of the scour zone can lead to reductions in particulate P export. This would, in turn, greatly reduce the P first-flush effects and event spikes that occur during the course of a year. UF9201A Site UF9201A (1280 acres) started BMP operations in September 1994 and concentrated on improved fertility BMP practices for vegetable production. The two-section vegetable unit Page 33

34 N Ditch level recorder UF9201W7 Water table recorder UF9201W8 Main Farm Canal Farm Pump UF9201A UF9201W6 Water table recorder Ditch level recorder UF9201W5 Off Site Canal LEGEND: Field ditch Roadway Canal Pump Riser Culvert * all field ditches are connected to the main farm canal and the perimeter ditch by risers and culverts. Ditch level recorder UF9201W3 Water table recorder UF9201W4 Water table recorder UF9201W2 Ditch level recorder UF9201W1 Figure 1A-BMP: UF9201A hydraulic system and farm layout. was hydraulically isolated from the rest of the farm (Figure 1A-BMP). Management of irrigation and/or floodwaters was monitored through two pumps discharging into a major Page 34

35 private canal. Detailed records of crop production and fertilizer applications are being kept by the grower. In addition to the fertilizer BMPs, the grower was reducing TP discharge through controlled drainage pumping. The two sections of land were split apart, allowing independent (although temporary) control of water levels in each section. Traditional (baseline) water management practices involved active pumping discharge of summer fallow floodwaters from both sections simultaneously. During the BMP period, the grower routed water from the first section to the second section for temporary storage while cultivation activities progressed in the first section. Water was then drained back into the first section once cultivation activities are completed. In order to prepare the second section nearest the pump station for planting, off-farm drainage was usually necessary to remove excessive water that had not been redistributed by seepage or lost to ET. When possible, floodwaters were allowed to subside naturally through ET and percolation prior to pumping. This practice encouraged retention of particulate matter in the fields, which in turn lead to lower TP concentrations and loads in discharged drainage water. During the 59-day period from August 7 through October , the grower engaged in almost continuous pumping that was unrelated to crop production events within the monitored site. Throughout most of this 59-day period, this production area was actually maintained under flooded fallow conditions. The grower used the monitored internal pump station to pass large volumes of water from one part of the farm to another. This internal transfer of water was reflected as discharge loads in the monitoring database, but in actuality, these discharge loads are completely unrelated to crop production requirements within the monitored site. When questioned, the grower informed project staff that the site was to be used as a water holding and drainage overflow site for the entire farm BMP package. At that time, the value of the site as a vegetable BMP site was lost and all monitoring efforts were ceased. From the original 10 sites, UF9201A, the vegetable monoculture, was dropped because the project focus shifted to one of solely evaluating discharges to Works of the District. This site became an internal water holding area for the much larger farm, essentially making it a component of their total farm water management BMP package. Page 35

36 UF9202A Site UF9202A (320 acres) started BMP operations in April The grower had voiced concerns that he could not drain the back four fields of his farm using his single main farm Farm Pump UF9202A N Water table recorder UF9202W1 624 ft Ditch level recorder UF9202W2 Main Farm Canal 2,618 ft Water table recorder UF9202W7 LEGEND: Field ditch Roadway Canal Pump Riser Culvert Flow direction Rainfall collector * all field ditches are connected to the main farm canal and the perimeter ditch by risers and culverts. Water table recorder UF9202W6 UF9202B Temporary pump and monitoring station 1,824 ft Water table recorder UF9202W4 Ditch level recorder UF9202W3 640 ft Off Site Canal Figure 2A-BMP: UF9202A hydraulic system and farm layout. Page 36

37 pump. To address this concern, the backup pump that was originally at the main farm pump station was simply moved back internal to the farm. The grower plugged his main farm canal (Figure 2A-BMP) and installed this booster pump a distance approximately two-thirds of the way down the length of his main canal. Water quality and discharge monitoring equipment were also installed at this location. The grower can now drain the front of his farm using his main pump while the rear of the farm, where in the past the grower experienced major seepage problems requiring operation of his main farm pump station, can be dewatered using the internal farm pump. In situations where the back of the farm requires drainage but the front of the farm does not, the grower leaves the main pump off and simply uses the internal pump to drain water from the back to the front of the farm. This also enables the grower to keep more water on the front fields of the farm when desired. This water management scheme enables the grower to achieve more uniform drainage and water table levels while inhibiting over-drainage of areas of the farm near the main pump station. Additionally, water tables remain higher and soil oxidation and subsequent P mineralization is reduced. The grower was informed about the role of ET on field drainage and how to recognize when pumps are not responsible for drying out fields. This grower also practices minimum tillage for sugarcane by planting the new crop between old rows of stubble. UF9202A generally tends to be dry since it is located on a knoll. During the 1994 crop season, a neighboring farm grew rice immediately adjacent to UF9202A. The perimeter of UF9202A is highly susceptible to seepage and the resulting inflow of water from neighboring rice fields actually helped the grower by supplying much needed irrigation water. The net effect to the grower's crop was positive because he reported substantial yield increases. It should be noted that due to seepage problems, shallow muck depths, and his position at the end of the private canal that serves him and numerous other small farms draining to the North New River Canal, there is a potential for UF9202A to rapidly encounter farm-wide flooding problems during very wet periods. Pumping activity by neighboring farms during very wet periods tend to keep adjacent off-farm canals at capacity levels. Subsequent seepage into UF9202A can create major drainage problems. Page 37

38 UF9203A Site UF9203A (4608 acres) started BMP operations in January During time periods when rice is grown in rotation with sugarcane, this site can be used to examine the N LEGEND: Main canal Farm canal Field ditch Pump Flow control structure UF9203W1 UF9203W2 North New River Canal U.S. 27 Booster pump To Palm Beach Farm D UF9203W3 C Main Pump Station UF9203A UF9203W4 UF9203W5 UF9203W6 A B-1 B-2 UF9203W7 UF9203W8 A UF9203W9 UF9203W10 UF9203W11 Figure 3A-BMP: UF9203A hydraulic system and farm layout. Page 38

39 effectiveness of using rice crops for drainage water filtration as well as for supplying nutrients to sugarcane. The water management BMPs focus on improved movement of water around the farm in an effort to reduce off-farm drainage pumping. Site UF9203A is also being used to determine the effects of deep drainage canals and sediment control methods on farm P discharge. Each of the back two sections of the farm has the capability of being managed as separate hydraulic units (Figure 3A-BMP). Installations of water control structures at the heads of the main canals serving each section have been completed. Portable pumps located at points B-1 and B-2 (Figure 3A-BMP) are used to move water around the farm and essentially partition the farm into three water management units (the front, central, and back) for drainage purposes. In other words, the back of the farm can be drained independently of, and to, the front of the farm. Removing an inch of water from the back two sections results in less than half-an-inch of water spread over the rest of the farm. Once again, blocking the farm into smaller hydraulic units and conducting water management activities accordingly allows the maintenance of more uniform drainage, the avoidance of over-drainage, higher water tables, and a reduction in soil oxidation and P mineralization. The main farm pump station is no longer solely relied on to drain the back of the farm, which should greatly reduce the amount of pumping required. This system highlights an important BMP management strategy that can be used at many different EAA farms. The concept is quite simple. Once growers install the ability to manage their farms as two or more independent units (depending on the farm size), the frequency and duration of main pump station operations will be reduced. Hydraulically isolating different production units within a farm greatly improves farm-wide drainage uniformity over space and time. This hydraulic system will also be effective for UF9203A should sugarcane or rice ever be rotated into vegetables. UF9204A This site (640 acres) pursued independent BMP implementation strategies and was assigned a final date of December 31, 1994 for the baseline monitoring period and a starting date of January 1, 1995 for the BMP monitoring period, consistent with regulatory deadline requirements for BMP program implementation. The site began the monitoring program as a Page 39

40 Ca al sugarcane monoculture but in April 1995, half (320 acres) of the site was rotated into rice (Figure 4A-BMP). The effect of this major change in cropping system to half of the farm LEGEND: Field ditch Roadway Canal Pump Culvert * all field ditches are connected to the main canal by culverts not shown N Highland-Glades n Farm Pump UF9204A Main Farm Canal Figure 4A-BMP: UF9204A hydraulic system and farm layout. Page 40

41 area, and the projected difficulty with respect to managing rice drainage waters in the absence of a comprehensive hydraulic control strategy, will be assessed. In December 1995, the rice acreage was rotated back into sugarcane. This site has also modified decision-making criteria used to determine pumping operations by adopting strict stage and antecedent rainfall condition criteria for pumping initiation and cutoff. UF9205A This site (320 acres) began the monitoring program as a sugarcane-pasture system (Figure 5A-BMP). In early 1994, a portion of the pastureland was converted to sod and the remaining pasture was rotated into corn in March After the corn harvest, land was fallowed, and then rotated into melon production in May After the melon harvest, land was fallowed in October 1995 and subsequently rotated back into sugarcane in February Melons were planted in several field blocks from May through August 1996, including a field located immediately adjacent to the main farm pump. Following the melon harvest, the farm has remained under a combination of sugarcane production and fallow fields. The effects of growing corn and melons (or any other water management intensive crop) in rotation with sugarcane, without a high level of water control, has been shown. Trends in monitoring data illustrate the importance of blocking farms into hydraulic units that can then be independently controlled when cultivating crops other than sugarcane. Also shown is the importance of maintaining a large enough percentage of farm acreage planted to sugarcane to maintain the flexible treatment of excess water in water-intolerant crop areas. Operations at this site, along with those at UF9206A&B, demonstrate that by blocking the farm into units, traditionally high TP loaders can greatly reduce discharge P loads. Observations at UF9205A during the corn/melon production period confirmed that much greater volumes of water had to be discharged from the entire farm in order to maintain desirable water table levels in the specific field blocks planted to the non-sugarcane crops. This site has always been an enigma for the project. In fact, it was used to demonstrate reverse BMP implementation and recovery after switching back to BMP operations. The site was dropped when it was implemented into one of the STA projects. Page 41

42 N Main Farm Canal LEGEND: Field ditch Roadway Canal Pump Riser Culvert * all field ditches are connected to the main farm canal by riser and culverts but are not joined to the perimeter side ditch Sump Miami Canal Figure 5A-BMP: UF9205A hydraulic system and farm layout. UF9206A&B Site UF9206A&B (1754 acres) started BMP operations in May The grower has installed a sophisticated hydraulic system with culverts, risers, and pumps placed strategically throughout the farm (Figure 6A&B-BMP). The grower has essentially partitioned Page 42

43 the farm into seven hydraulic units. Water can be moved from any production block into any of the seven hydraulic units for temporary storage, irrigation, or drainage. The grower continues to maintain the existing farm drainage capacity at the farm outlets. The hydraulic layout allows for water from low volume or minor rainfall events, or from localized drainage events during cultural practices, to be kept on-farm and subsequently bled off into the atmosphere through ET. During minor rainfall events, vegetable drainage, rice drainage, or flooded fallow field waters can be diverted to adjacent sugarcane production units instead of being pumped directly off-farm. Virtually every field ditch has an operable riser and board structure allowing for maximum water control, uniform drainage, and the utilization of the longest paths for routing water to the main pump stations. These physical changes to the farm should allow the grower to reduce both TP concentrations and loads. Additionally, site UF9206A&B grows flooded rice following vegetables. These rice fields serve as additional water storage capacity for the farm. The rice crop also serves as a nutrient sink, removing excessive TP from fields previously planted to vegetables. Rice drainage waters with high nutrient loads are to be recycled internal to the farm and put to use in fields requiring water and fertilization. Over the next couple of years, the grower has been requested to bring his crop rotation into synchronization with his hydraulic units as much as possible. It was also suggested that the grower arrange cropping patterns that place so called "high P" crops a suitable distance from the off-farm discharge points. The role that ET plays in drying out hard-to-drain fields was discussed with the grower to ensure that he understands that some fields do not drain by pumping. Put simply, operating pumps continuously for days to dry out an area of the farm is counterproductive. If the grower accounted for ET rates during these lengthy pumping events, he would realize that a substantial amount of the water table lowering was actually accomplished through ET. The grower maintains higher water tables in the sugarcane fields while still achieving his drainage needs in the vegetable blocks. Page 43

44 New interceptor ditch S. R. 80 N LEGEND: Main canal Farm canal Field ditch Flow direction Rainfall collector Water control structure Water table recorder UF9206W8 Weed Boom, ~110 upstream Pump and monitoring station A Interior Pump Station B Culverts and Risers C Enlarged Drainage Ditches B B UF9206D Ditch level recorder UF9206W1 Water table recorder UF9206W2 Ditch level recorder UF9206W3 UF9206C Water table recorder UF9206W5 A SOD B C B B Ditch level recorder UF9206W4 UF9206W9 C UF9206W6 Water table recorder UF9206W7 Ditch level recorder UF9206W10 Water table recorders UF9206B UF9206A North Pump Station Obsolete Pump Station Lysimeter Experiment Site South Pump Station Mill Road Canal Figure 6A&B-BMP: UF9206A & B hydraulic system and farm layout. Ambitious field ditch excavations were begun in February This maintenance work continued for a number of months that coincided with a major portion of the wet season. Page 44

45 Excavation work on portions of the farm s perimeter canal network have also taken place during 1996 and Throughout the October 1995 through April 1996 period, field blocks closest to the two main farm pump stations were planted to vegetables (radish, leaf, carrots). Following a summer fallow period, some of these blocks were rotated into rice. The combination of extensive ditch/canal excavation work and planting vegetables and rice immediately adjacent to off-farm discharge sites was expected to impact drainage water quality. Subsequent review of drainage water quality data confirms appreciable P loading increases throughout this time period. In June 1998, floating weed booms were installed across the surface of the two main farm canals served by farm pump structures. Booms were located downstream of the pump scouring zone. In theory, trapping weeds (and the high-p biological community associated with the submerged root fraction) should reduce the overall P stream entering the pumps during discharge events. Some additional design work needs to be implemented in order to address weeds that enter the scouring zone from a feeder ditch located between the weed boom and the pump structure. In 2000, the grower made further changes to his water distribution system. About a mile of the north main farm canal was cleaned and widened. The southern perimeter ditch is also being widened to improve farm hydraulics. These activities, however, greatly increased the site P concentrations and loads during the past water year. In fact, the concentrations and loads from this single farm greatly skewed project composite results. This site is the second farm being used in the intensive particulate P transport studies and demonstrations. Aquatic weed coverage and inventories were conducted on a monthly basis for two years (Chapter 4). Unlike UF9200A, this farm has implemented an aggressive aquatic weed management program. However, the weed management program may be adversely affecting the TP concentrations and loads by increasing particulate P transport. UF9207A&B This site (2500 acres) pursued independent BMP implementation strategies and was assigned a final date of December 31, 1994 for the baseline monitoring period and a starting date of January 1, 1995 for the BMP monitoring period, consistent with regulatory deadline Page 45

46 Hillsboro Canal North Pump Station Chemical injection feeder and flow direction South Pump Station W30 ft x L200 ft x D25 ft pit N LEGEND: Main Canal Farm Canal Field Ditch Figure 7A&B-BMP: UF9207A hydraulic system and farm layout. Page 46

47 requirements for BMP program implementation. The site is engaged in sugarcane-vegetable rotations, generally involving sweet corn (Figure 7A&B-BMP). The BMP initiatives include the operation of a chemical injection program (August 1993 through December 1994) to precipitate P out of the drainage stream, the installation of a 200 by 30 ft (length x width) sediment trap pit in his main farm canal in March 1995, and routing drainage water around the farm boundaries to allow particulates more time to settle out of the drainage stream. UF9208A This site (262 acres) pursued independent BMP implementation strategies and has been assigned a final date of December 31, 1994 for the baseline monitoring period and a starting date of January 1, 1995 for the BMP monitoring period, consistent with regulatory deadline requirements for BMP program implementation. This is a small (262 acre) sugarcane monoculture site (Figure 8A-BMP) that has a history of infrequent pumping events involving minor discharges of water relative to other project sites. It should be noted that UALs and TP loads have been historically extremely low at this site due to minimal pumping. This site changed management as a result of the STA development program. Project monitoring was slated to be discontinued in January However, the new management changed the crop rotation to include corn and it was decided that monitoring should continue through August 2000, in order to demonstrate the effects, if any, of the crop rotation change. Site UF9208A was terminated because it did not appear to add useful information to the data set and drainage pumping was minimal due to active drainage flows through the shallow aquifer to an off-farm canal. Page 47

48 N LEGEND: Field ditch Roadway Canal Pump Screw gate Culvert no junction Conservancy District Canal Figure 8A-BMP: UF9208A hydraulic system and farm layout. UF9209A Site UF9209A (3072 acres) (Figure 9A-BMP) independently began BMP operations in May The grower initially implemented the pump BMP developed by industry consultants, Page 48

49 which set canal stage conditions under which drainage pumping should occur. Since then, the grower has enhanced the pump operation BMP. The program uses established upper and lower main canal stages (measured at the main pump station and at the geographic Pump Station N LEGEND: Main Canal Farm Canal Field Ditch Booster pump Miami Canal UF9209W6 UF9209W5 UF9209W7 UF9209W2 UF9209W1 UF9209W3 UF9209W4 Figure 9A-BMP: UF9209A hydraulic system and farm layout. Page 49

50 farm center) that serve as criteria for pump operations. In early 1995, the project began monitoring field water table levels around the farm in order to supply the grower with additional information useful for the development of a comprehensive water management plan that addresses crop water requirements and farm-wide drainage uniformity. An internal booster pump has been installed approximately two-thirds of the way down the length of the main farm canal to block the farm into two hydraulic units (Figure 9A-BMP). This farm site was added to the intensive particulate P transport studies. Sediment core samples were acquired and analyzed. The site is important in this regard since it has a history of extremely low TP concentrations and loads, of which about 80% can be attributed to particulates. Farm canal bottoms appear to be relatively sediment free although field ditches are not. Further information about the site is needed prior to speculation regarding the source of the particulate P. Crop Rotations Crop mixes at all farms have been tracked by calendar year since 1992 (Ch 1 Table 2). Growers have experimented with adding rice, corn and other vegetables, sod, and fallow to their rotations. Trends appear to be up and down, with some growers increasing corn and/or sod. Rice acreage does not appear to have increased, while fallow acreage has decreased at most sites, except where corn production has risen. Of note are the consistent decrease in sugarcane production at UF9206A&B and UF9208A, and the increase at UF9207A&B. Site UF9203A s experimentation with rice appears to have waned while Sites UF9206A&B and UF9207A&B have increased sod production. Fallowing fields appears to have waned, except where necessary to fit corn into the crop rotation. Monthly crop mixes for the farms are submitted in electronic format in Appendix A. Page 50

51 CHAPTER 1 Farm Verification of BMP Efficacy List of Figures Chapter 1 Ch 1 Figure 1. Location of the ten farm sites in the EAA Ch. 1 Figure 2. Event total P and total dissolved P concentration data for UF9206A (Event 1= November 1996; Event 250= Febraury 2002)...58 Ch. 1 Figure 3. Event total P and total dissolved P concentrations for UF9209A (Event 1 = January 16, 1998; Event 310= Decmeber 30, 2002)...60 Ch. 1 Figure 4. Event total P concentrations for UF9207A (Event 1= August 31, 1992, Event 252= November 5, 2002)...60 Ch. 1 Figure 5. Event total P concentrations for UF9209A (Event 1= July 24, 1992; Event 450 = Decmeber 30, 2002)...61 Ch. 1 Figure 6. Average total P concentration for water years, derived from area-weighted concentration composites of all project sites, except UF9201A, UF, 9205A, and UF 9208A Ch. 1 Figure 7. Average total P concentration for water years, derived from area-weighted concentration composites of all project sites, except UF9201A, UF, 9205A, and UF 9208A Ch. 1 Figure 8. Average total P adjusted unit area loads for water years, derived from areaweighted AUAL composites of all project sites, except UF9201A, UF9205A, and UF9208A Ch. 1 Figure 9. Area-weighted average drainage volume to rainfall ratios for project sites Ch. 1 Figure 10. Area-weighted average P UAL to rainfall ratios for project sites...65 List of Tables Chapter 1 Ch.1 Table 1. Major BMPs implemented at the ten farms in the EAA Ch.1 Table 2. Percentage of total crop acres planted to various crops at the ten project sites for the specified years Page 51

52 Introduction There are a myriad of factors that can affect the magnitude of phosphorus (P) concentrations leaving a particular farm, on both event-wise and long-term basis, in the Everglades Agricultural Area (EAA). For the sake of developing a discussion regarding factors that can affect farm drainage P concentrations and loads, the factors will be separated into four main categories: 1) fertilizer; 2) soil; 3) water; and 4) other. It should be noted that as the management of individual items within the four categories can elevate P concentrations, conversely, some of them might also be managed effectively to reduce P concentrations. Additionally, it should be noted that because of a mixture of certain factors, particular areas of land might yield background levels of P concentrations that cannot be readily altered through the better management of any of the BMPs. Each farm has a characteristic lowest achievable discharge P concentration that cannot be reduced without an extensive time period and major expense. For the purposes of developing the list of factors that have the potential of elevating farm-level P loads, this discussion will focus on UF9200A (sugarcane monoculture) and UF9206A&B (mixed cropping system with significant vegetable acreage). Factors that can affect P concentrations are listed below. Fertilizer: Soils: Type (liquid, granular, slow release, alternative) Amount based on adequate soil test results Timing of application to match crop needs Application method (banding, banding at reduced rates, broadcast) Uniformity of distribution over field Distribution around farm often depending on crop rotation Application timing relative to subsequent rainfall and drainage events Storage site(s) on-farm Disposal method for excess fertilizer in application machinery Cleaning method of application machinery Occurrences of leaks and spills Inadvertent direct application to canals and ditches Depth of soil Mineral content of soils Presence of marl layer P saturation status Decomposition status Prevailing oxidation rates Page 52

53 Historical land use and associated P loading rate of soil Moisture content Water: Pumping rates Pumping durations Pump start-up energy imparted to water Minimum water levels allowed in canals near the main pump station Amount of rainfall to trigger pumping Antecedent soil moisture conditions Ditch and canal configurations Area served by the main pump station Length of run of main canal system Tortuosity of flow paths Average farm water table and seasonable variability Drainage uniformity Seepage rates and P concentrations to and from farm Proximity to flooded areas and area canals Adjacent farms water management practices Rainfall amounts and uniformity across farm Bulk precipitation P concentrations Irrigation amount and frequency Irrigation water P concentrations Time interval between irrigation and drainage events Area canal water levels at times of irrigation and drainage Degree of impermeability of marl and limestone layers Other: Ditch cleaning activities and timing relative to rainfalls Ditch excavation activities Soil erosion tendencies Ditchbank stability Pump station modifications Impacts of area canal dredging on farms Farm and area canal aquatic plant growth and senescence Crop rotation and proximity of certain crops to main pump station Migratory birds There are undoubtedly other factors, which could affect farm-level P concentrations both negatively and positively. Additionally, although the entire list above pertains to UF9206A&B, several items are not relevant to UF9200A. For example, crop rotations and proximity of certain crops will have no real effect on a sugarcane monoculture while it could be one of the predominant factors for a mixed crop rotation farm. Also, the area served by the main pump station may greatly affect UF9200A which is two sections of land served by a single main pump station, but would not be quite so important in the case of UF9206A&B Page 53

54 which hydraulically isolates blocks of the farm and makes good use of portable booster pumps to separately control smaller blocks within the total farm area. Fertilizer application factors would play an almost insignificant role at UF9200A, which is a sugarcane monoculture, simply because of the non-intensive nature of sugarcane fertilization practices. On the other hand, fertilization practices at UF9206A&B could greatly affect drainage water P concentrations, especially if heavily fertilized areas are adjacent to the main pump stations. Furthermore, it has been learned that farm canal aquatic plant growth and senescence leading to P-rich particulate matter production and entrainment contributes greatly to P concentrations and loads at all sites. Objective The objective of this facet of the program was to maintain a continuing database on drainage flows, cropping patterns, and water quality for farm sites chosen that are representative of EAA agriculture. This activity helped to track the effectiveness of the BMP program and provide indicators of potential foci for extension programs. Side benefits are that P concentration and load data were being accumulated for atypical hydrologic, infrastructure improvement, and crop rotation change events. The data also served as a baseline for work on assessing the effects of particulate P transport on P concentrations and loads. Further, the data help to illustrate concentration and load trends at the farm level relative to those at the basin level. Materials and Methods Farm descriptions, cropping history and BMPs implemented are presented in the previous section. All of the farms were monitored for a number of variables including total P, rainfall, pumping volume, canal levels. The monitoring and laboratory analysis portions of this project contributed directly to all other project activities. Site locations, BMP activities, and cropping systems are presented in Figure 1, Table 1, and Table 2, respectively. Page 54

55 UF9200A UF9204A UF9206A&B UF9209A UF9208A UF9201A UF9207A&B UF9205A UF9202A UF9203A Ch 1 Figure 1. Location of the ten farm sites in the EAA. Page 55

56 Ch. 1 Table 1 Major BMPs implemented at the ten project sites. Site EAA Major BMPs Implemented Sub-basin UF9200A S-5A Reduced frequency of irrigation/drainage events. Attenuated water table micro-management. Cleaned ditches and removed sedimentary material from ditchways. Installed weed boom in main farm canal. Used calibrated soil tests and improved fertilizer practices. UF9201A S-6 Routed water internally between field blocks during planting season. Allowed summer fallow flood waters to recede naturally through ET/percolation. Used internal monitored area (1998) for water storage for larger farm water requirements. Used calibrated soil tests and improved fertilizer practices. UF9202A S-7 Installed internal booster pump which improved drainage uniformity and reduced off-farm discharge requirements. Practiced minimum-tillage sugarcane planting practices. Used calibrated soil tests and improved fertilizer practices. UF9203A S-7 Installed control structures which allowed improved farm drainage and greater hydraulic control between internal field blocks. Increased capacity of main farm ditches. Used portable booster pump. Used calibrated soil tests and improved fertilizer practices. UF9204A S-6 Implemented new off-farm pumping protocol (January 1995). Rotated half of farm sugarcane acreage into rice (April 1995) in the absence of concurrent hydraulic BMP implementation. Rotated rice back to sugarcane (December 1995). Used calibrated soil tests and improved fertilizer practices. UF9205A S-8 Rotated almost half of farm sugarcane acreage into corn (March 1994) followed by fallow and then melons (May 1995) in the absence of concurrent hydraulic BMP implementation. Rotated melons back to fallow (October 1995) and then planted sugarcane (February 1996). Used calibrated soil tests and improved fertilizer practices. UF9206A&B S-5A Installed control structures, compartmentalized farm into 7 hydraulically isolated units which allowed internal routing of vegetable and/or rice draindown waters to other areas of farm for storage and/or natural removal through ET/percolation. Installed weed booms in main farm canals. Used calibrated soil tests and improved fertilizer practices. UF9207A&B S-6 Implemented independently designed BMP tests (chemical additive injection and sediment trapping pit). Internally rerouted water around farm. Blocked crops in rotation. Used calibrated soil tests and improved fertilizer practices. UF9208A S-6 Increased on-farm water retention in soil profile and open waterways. Implemented field ditch and farm canal sediment control strategies. Used calibrated soil tests and improved fertilizer practices. UF9209A S-8 Adopted improved industry pump BMP (pulse pumping). Enforced stricter protocols for off-farm discharge. Installed internal booster pump for improved drainage. Used calibrated soil tests and improved fertilizer practices. Page 56

57 Ch. 1 Table 2. Percentage of acres planted by crop at the ten project sites. Page 57

58 Results and Discussion P Concentrations and Loads P concentration and load data are tabulated and plotted for all sites in Appendix A on the accompanying CD. Prior to looking at TP concentration and load reductions effected by BMP implementation, it is of great benefit to qualitatively study the data to see how they are affected by management practices and other farm characteristics. It is also possible to gain an appreciation for the variability in the data and an idea of what factors may be responsible for elevated TP concentrations on EAA farms. Figure 2 displays the event TP and TDP data for one site, UF9206A. One should immediately note that TDP concentrations appear to be approaching and exceeding the average project TP value of mg/l. It is also obvious that TDP concentrations are greatly responsible for the higher TP concentrations seen at the site. P concentration spikes are prevalent and it appears that concentrations rose on either side of Event 150. It is also evident that TDP concentrations were greatly reduced after Event 175. Valid explanations are available for all characteristics. Total and Total Dissolved Phosphorus Concentrations UF9206A Events Concentrations TP and TDP Concentrations, mg/l Event number TP TDP Ch. 1 Figure 2. Event total P and total dissolved P concentration data for UF9206A (Event 1= November 1996; Event 250= Febraury 2002). Page 58

59 The graph represents samples collected from November 1996 (event one) through February 2002 (event 250). This project farm is a mixed cropping system, with only about 30% of the land area in sugarcane. A large amount of vegetables are grown during the course of the year. Rice and sod are also grown extensively. Both of these crops generally yield higher TDP concentrations in drainage water. Rice cultivation can also lead to high levels of particulate P (PP) when drain down water is allowed to directly enter the farm canal system during pumping. Near Event 150, the grower also was in the midst of extensive farm canal and field ditch modifications. This lead directly to increased TP concentrations in drainage water. Finally, there was a change in management of rice drainage waters that occurred after Event 175 that could account for the markedly lower TDP concentrations. Note that, although TDP excursions do occur, they generally stay below 0.10 mg TDP/L. This shows the importance of the role that PP plays in the TP loading equation. In contrast with Figure 2, the TP and TDP characteristics shown in Figure 3 are for a large monoculture sugarcane farm, UF9200A. One can see that TP concentrations are much lower, generally staying below mg TP/L. A few TP excursions do occur, but they are far fewer and lesser in magnitude than those seen at UF9206A. Note also that the TDP concentrations are extremely low, typically staying below 0.05 mg/l. The farm, UF9209A, represented in Figure 3, is a sugarcane monoculture. In addition, the grower pumps relatively small volumes on a more frequent basis as water tables are brought down carefully to avoid over-drainage. Spikes in TP are caused primarily by PP. Particulate-P accounts for nearly 75% of the TP concentration and load at this site. Total P concentrations at this site have remained consistently around 0.08 mg/l throughout the project duration. Page 59

60 Total and Total Dissolved Phosphorus Concentrations UF9209A Events Concentrations TP and TDP Concentrations, mg/l Event Number TP TDP Ch. 1 Figure 3. Event total P and total dissolved P concentrations for UF9209A (Event 1 = January 16, 1998; Event 310= Decmeber 30, 2002) Total-Phosphorus Concentrations UF9207A Events TP Concentrations, mg/l Event Number Ch. 1 Figure 4. Event total P concentrations for UF9207A (Event 1= August 31, 1992, Event 252= November 5, 2002). Page 60

61 The breadth of the differences in TP and TDP concentration profiles that exist in the EAA can be seen in Figures 2 and 3. It is evident that there is a lowest reasonably achievable level in TP concentrations at the farm level. Furthermore, this level varies by farm, probably dependent on geographic location, soil types, land use history, hydrology, and current crop mix. For example, at UF9207A, TP concentrations cluster along a straight line at around 0.12 mg/l (Figure 4). However, at UF9209A, the lowest reasonably achievable TP concentration might be on the order of 0.06 mg/l (Figure 5). Clearly, a difference exists. To further complicate matters, achieving a reduction in TP concentrations at UF9209A from 0.08 to 0.06 mg/l will probably not be cost effective, whereas achieving the same reduction of 0.02 mg/l TP concentration at UF9207A and other farms might be possible by reducing a minimal number of TP spikes. Clearly, the BMP potential to reduce P concentration is not equal among different farms, for reasons that are sometimes beyond the control of the farmer. This lowest reasonably achievable concentration varies between 0.05 and 0.15 mg/l at the project sites. Total-Phosphorus Concentrations UF9209A Events TP Concentrations, mg/l Event Number Ch. 1 Figure 5. Event total P concentrations for UF9209A (Event 1= July 24, 1992; Event 450 = Decmeber 30, 2002). Page 61

62 A common parameter used to compare P loads is Unit Area Load (UAL). This parameter is the normalized P load exported for a given land area, usually reported in lbs P per acre per year. The UALs of farm sites averaged lbs TP/acre post-bmp implementation. Compared with the WY94 baseline rate of 0.921, this represents a negligible 2.5% increase. Since load is dependent upon rainfall and its distribution, an adjusted unit area load (AUAL) parameter has been developed which attempts to compensate for rainfall amount and distribution. The AUALs for study sites averaged 0.73 lbs TP/acre after BMP implementation compared to 1.30 lbs TP/acre prior to WY95. This represents a project farm average reduction in AUAL of approximately 44%. In addition to the above, a simpler method of normalizing data was used. Volumes of drainage and UALs were indexed to the rainfall amounts received during the year. This yielded figures for how much drainage occurred for each inch of rainfall as well as how much P left the farms on a per inch of rain basis. The V:R ratio for the pre-bmp period was 0.45 inches of drainage for each inch of rain received. After BMP implementation, the ratio fell 13% to an average of 0.39 inches per inch of rain. This is indicative of the reduction in farm drainage that occurred due to BMP implementation. The UAL to rainfall ratio dropped from lbs of P discharged per acre per inch of rain to , a reduction of approximately 6.5%. All indicators of BMP efficacy have shown that consistent and sustained reductions in TP concentrations and loads have occurred due to the implementation of BMPs in the EAA. Basin-level numbers presented annually by the SFWMD reinforce the effectiveness of the BMP program, showing a sustained 50% reduction in TP loading from the EAA over the course of this project. Figures 7 through 10 show the area-weighted project average trends for TP concentrations, UALs, AUALs, V:R ratios, and UAL:R ratios, respectively. Clearly, the data are variable, reflecting the variability in hydrologic conditions over the years. Yet, it is clear from every indicator that a downward trend has occurred over the years. Page 62

63 Average TP Concentrations Area-Weighted Project Averages Total Phosphorus Concentrations, mg/l WY94 WY95 WY96 WY97 WY98 WY99 WY00 WY01 WY02 Water Year Ch. 1 Figure 6. Average total P concentration for water years, derived from areaweighted concentration composites of all project sites, except UF9201A, UF, 9205A, and UF 9208A Average TP Unit Area Loads Area-Weighted Project Averages 1.40 TP Unit Area Loads, lb/ac WY94 WY95 WY96 WY97 WY98 WY99 WY00 WY01 WY02 Water Year Fi 5 7 A TP it l d f t d i d f i ht d UAL it f Ch. 1 Figure 7. Average total P concentration for water years, derived from areaweighted concentration composites of all project sites, except UF9201A, UF, 9205A, and UF 9208A. Page 63

64 Average TP Adjusted Unit Area Loads Area-Weighted Project Averages 1.4 TP Adjusted Unit Area Loads, lb/ac WY94 WY95 WY96 WY97 WY98 WY99 WY00 WY01 WY02 Water Year Ch. 1 Figure 8. Average total P adjusted unit area loads for water years, derived from area-weighted AUAL composites of all project sites, except UF9201A, UF9205A, and UF9208A. V:R Ratio, in/in Drainage Volume to Rainfall Ratios Composite of Farm Sites WY94 WY95 WY96 WY97 WY98 WY99 WY00 WY01 WY02 Water Years (May1 through April 30) Ch. 1 Figure 9. Area-weighted average drainage volume to rainfall ratios for project sites. Page 64

65 P UAL:R Ratio (P UAL/in Rain) Phosphorus Load to Rainfall Ratios Composite of Farm Sites WY94 WY95 WY96 WY97 WY98 WY99 WY00 WY01 WY02 Water Years (May 1 through April 30) Figure 5-10: Area-weighted average P UAL to rainfall ratios for project sites. Ch. 1 Figure 10. Area-weighted average P UAL to rainfall ratios for project sites. Conclusions BMPs have proven to be effective and EAA farms have reduced P loads by 50% over shortand long-term historical periods. This major and sustainable reduction is credited to the BMP program. Enhancing the uniformity of implementation could, potentially, reduce loads further. Reducing the few major P export events that occur during a year at any farm could also further reduce P loads. Further efforts towards increasing the reduction at the farm level would require intensive studies and development programs. The majority of the reductions in P loading at the farm level have been due to changes in water management schemes (water detention) that reduce volumes of actively pumped drainage water from farms following rainfall events. The reduction in active drainage volume does not represent a reduction in water supply, since the shallow limestone aquifer beneath the EAA does not inhibit lateral flow. Page 65

66 Crop mixes at all farms have been tracked by calendar year since Growers have experimented with adding rice, corn and other vegetables, sod, and fallow to their rotations. Trends appear to fluctuate with commodity prices with some growers increasing corn, beans, and/or sod. Rice acreage does not appear to have increased, while fallow acreage has decreased at most sites, except where corn production has risen. Fallowing fields appears to have waned, except where necessary to fit corn into the crop rotation. Unit Area Loads (UAL) averaged lbs TP/acre post-bmp implementation. Compared with the WY94 baseline rate of 0.921, this represents a negligible 2.5% increase. Adjusted unit area loads (AUAL) averaged 0.73 lbs TP/acre after BMP implementation compared to 1.30 lbs TP/acre prior to WY95. This represents a project average reduction in AUAL of approximately 44%. In addition to the above method, a truer method of normalizing data was used. Volumes of drainage and UALs were indexed to the rainfall amounts received during the year. This yielded figures for how much drainage occurred for each inch of rainfall as well as how much P left the farms on a per inch of rain basis. The V:R ratio for the pre-bmp period was 0.45 inches of drainage for each inch of rain received. After BMP implementation, the ratio fell 13% to an average of 0.39 inches per inch of rain. This is indicative of the reduction in farm drainage that occurred due to BMP implementation. The UAL to rainfall ratio dropped from lbs of P discharged per acre per inch of rain to , a reduction of approximately 6.5%. All indicators of BMP efficacy have shown that consistent and sustained reductions in P concentrations and loads have occurred due to the implementation of BMPs in the EAA. Basin-level numbers presented annually by the SFWMD reinforce the effectiveness of the BMP program, showing a sustained 50% reduction in TP loading from the EAA over the course of this project. During this part of the study, farm and basin data showed that BMP packages could reduce P loads by 50% on a recurring basis. The major BMP efficacy verification period stopped after WY2002. Three farm pump stations continued to be monitored to satisfy the requirements of Rule 40E-63. However, BMPs have proven to be effective and EAA farms have reduced P loads by 50% over shortand long-term historical periods. This major and sustainable reduction is credited to the BMP program. Enhancing the uniformity of implementation could, potentially, reduce loads further. Reducing the few major P export events that occur during a year at any farm could Page 66

67 also further reduce P loads. It would appear prudent at this time, to focus on understanding the P dynamics within the South Florida Water Management District (SFWMD) canal system, the Stormwater Treatment Areas (STAs) and the Water Conservation Areas (WCAs). There are simply too many factors related to P cycling in the aquatic systems, P transport issues, inherent background P levels, Lake Okeechobee concentrations and loads, and human interaction effects that would occlude or negate the effort towards achieving further farm-level P load reductions. Hence, it can be stated that BMPs in the EAA are capable of sustaining a 50% reduction in agricultural P loads and that further efforts towards increasing the reduction would probably not result in measurable, economically viable alternatives at the farm-level without intensive studies and development programs whose positive results may be masked or made unnecessary by factors beyond the control of the agricultural sector. It can also be stated that the majority of the reductions in P loading at the farm level are due to changes in water management schemes that reduce volumes and P concentrations of water actively pumped off farms following rainfall events. Further, the reduction in pumping does not represent a reduction in water supply since the shallow limestone aquifer beneath the EAA does not inhibit lateral flow and that this lateral flow through the limestone probably provides further reductions in P concentrations. Page 67

68 References Daroub, S.H., T.A. Lang, O.A. Diaz, M. Chen, and J.D. Stuck Annual report Phase XII: Implementation and Verification of BMPs for Reducing P Loading in the EAA and Everglades Agricultural Area BMPs for Reducing Particulate Phosphorus Transport. University of Florida, EREC, Belle Glade, May Submitted to the Everglades Agricultural Area Environmental Protection District and The Florida Department of Environmental Protection. Daroub, S.H., J.D. Stuck, T.A. Lang, and O.A. Diaz, and M. Chen Annual report Phase XI: Implementation and Verification of BMPs for Reducing P Loading in the EAA and Everglades Agricultural Area BMPs for Reducing Particulate Phosphorus Transport. University of Florida, EREC, Belle Glade, April Submitted to the Everglades Agricultural Area Environmental Protection District and The Florida Department of Environmental Protection. Daroub, S. H., J.D. Stuck, R.W. Rice, T.A. Lang, and O.A. Diaz Annual report Phase X: Implementation and Verification of BMPs for Reducing P Loading in the EAA and Everglades Agricultural Area BMPs for Reducing Particulate Phosphorus Transport. University of Florida, EREC, Belle Glade, March Submitted to the Everglades Agricultural Area Environmental Protection District and Florida Department of Environmental Protection. Izuno, F.T., J.D. Stuck, R.W. Rice, T.A. Lang, and O.A. Diaz Annual report Phase IX: Implementation and Verification of BMPs for Reducing P Loading in the EAA and Everglades Agricultural Area BMPs for Reducing Particulate Phosphorus Transport. University of Florida, EREC, Belle Glade, March Submitted to the Everglades Agricultural Area Environmental Protection District and The Florida Department of Environmental Protection. Izuno, F.T., and R.W. Rice Annual Report Phase VIII. Implementation and Verification of BMPs for Reducing P Loading in the EAA and Sustainable Agriculture Coupled with Farm-Level P Reduction BMP Implementation. University of Florida, EREC, Belle Glade, May Submitted to the Everglades Agricultural Area Environmental Protection District. Izuno, F.T., and R.W. Rice Annual Report Phase VII. Implementation and Verification of BMPs for Reducing P Loading in the EAA and Sustainable Agriculture Coupled with Farm-Level P Reduction BMP Implementation. University of Florida, EREC, Belle Glade, June Submitted to the Everglades Agricultural Area Environmental Protection District and The Florida Department of Environmental Protection. Page 68

69 CHAPTER 2 Specific Conductance in the Everglades Agricultural Area List of Figures Chapter 2 Ch. 2 Figure 1. Specific Conductance monitoring sites...75 Ch. 2 Figure 2. Monthly averages of Specific Conductance (SpCond, ms/cm) by site...84 Ch. 2 Figure 3. Mean Specific Conductance from all sites combined (SpCond, ms/cm) by month...87 Ch. 2 Figure 4. Plot for Cl - concentration (mg/l) from Feb. 28, 2001 to Oct. 27, 2002 by site Ch. 2 Figure 5. Study sites with mean Specific Conductance duperimposed upon a map of chloride concentration of shallow wells (20 To 50 Feet Depth) in the EAA (chloride map recreated from Parker, 1955) Ch. 2 Figure 6. Water quality of shallow wells in the EAA and study farm locations (well data from Miller and Lietz, 1976)...95 Ch. 2 Figure 7. Hourly Specific Conductance of drainage event UF9206A plotted against time (Graph A) and cumulative volume pumped per acre (Graph B) Ch. 2 Figure 8. Hourly Specific Conductance of drainage event UF9206B plotted against time (Graph A) and cumulative volume pumped per acre (Graph B) Ch. 2 Figure 9. Hourly Specific Conductance of drainage event UF9208A plotted against time (Graph A) and cumulative volume pumped per acre (Graph B) Ch. 2 Figure 10. Potential contribution of potash fertilizer to total exported dissolved solids from farm UF9206AB (Graph A); potential contribution of Cl- from potash fertilizer to total exported Cl- (Graph B) Ch. 2 Figure 11. Plot of average monthly Specific Conductance (SpCond, ms/cm) from all sites combined by year Ch. 2 Figure 12. Time trend of Specific Conductance at UF9200A Ch. 2 Figure 13. Time trend of Specific Conductance at UF9201A Ch. 2 Figure 14. Time trend of Specific Conductance at UF9202A Ch. 2 Figure 15. Time trend of Specific Conductance at UF9203A Ch. 2 Figure 16. Time trend of Specific Conductance at UF9204A Ch. 2 Figure 17. Time trend of Specific Conductance at UF9205A Ch. 2 Figure 18. Time trend of Specific Conductance at UF9206A Ch. 2 Figure 19. Time trend of Specific Conductance at UF9206B Ch. 2 Figure 20. Time trend of Specific Conductance at UF9207A Page 69

70 Ch. 2 Figure 21. Time trend of Specific Conductance at UF9207B Ch. 2 Figure 22. Time trend Of Specific Conductance at UF9208A Ch. 2 Figure 23. Time trend of Specific Conductance at UF9209A List of Tables Chapter 2 Ch. 2 Table 1. Summary statistics for historical Specific Conductance of Lake Okeechobee, ground water, and canal water in the EAA...80 Ch. 2 Table 2. Summary Statistics for hourly averages of Specific Conductance and other parameters in all sites during the study period...82 Ch. 2 Table 3. Summary of rainfall, water and crop management at the ten farms combined Ch. 2 Table 4. Means of Specifc Conducatnce and other parmeters by site...85 Ch. 2 Table 5. Statistical summary for ion composition and other parameters in grab samples combined for 10 pump structures...88 Ch. 2 Table 6. Mean ion concentrations in grab samples by site...89 Ch. 2 Table 7. Statistical summary of the effects of ambient conditions, pumping, and irrigation on Specific Conductivity at four structures Ch. 2 Table 8. Effects of interactions between site, pumping, rainfall, and irrigation on daily average Specific Conductance at four intensively monitored pump structures Ch. 2 Table 9. Specific Conductance, irrigation source, and pumping to rainfall ratio by site Ch. 2 Table 10. Correlation coefficients between parameters monitored for all sites combined Ch. 2 Table 11. Non-Parametric Mann-Kendall Trend Analysis and Sen s Slope Analysis of Specific Conductance by site Page 70

71 Introduction In April of 1994 the Florida state legislature passed The Everglades Forever Act (EFA) which mandated: 1) the construction of six STAs encompassing 16,188 ha; 2) Everglades water supply and hydroperiod improvement and restoration; 3) an EAA research and monitoring program; 4) evaluation of water quality standards; 5) research and implementation of BMPs in the EAA; and 6) monitoring and control of exotic species (Florida Statute Section , 1994). The research and implementation of BMPs component of the EFA required the SFWMD to conduct research in cooperation with the EAA landowners to identify water quality parameters that are not being significantly improved either by the STAs or the BMPs, and to identify further BMP strategies needed to address these parameters (Florida Statute Section , 1994). In 1997 the SFWMD revised the Everglades Regulatory Program, Chapter 40E-63 in accordance with the EFA, to include a research/monitoring program to address concerns regarding particulate P, specific conductance, and concentrations of the pesticides Ametryn and Atrazine found in surface waters of the EAA. The UF/IFAS project began to monitor particulate P and specific conductance at three farm sites in early 1997; in early 1998 all ten project sites were equipped to monitor particulate P and specific conductance. This report deals with the issue of specific conductance in the EAA. The issue of particulate P will be addressed in a separate report. The monitoring of the pesticides Ametryn and Atrazine is conducted by the EAA-EPD and the SFWMD. The conductivity of a body of water is a function of the quantity of ions contained in it, the equivalent conductivity of each ion, and water temperature. Compensation of this measurement to 25 degrees Centigrade yields specific conductance. Specific conductance is an indirect measure of the total concentration of ionized substances (e.g. Ca2+, magnesium (Mg2+), Na+, potassium (K+), Cl-, HCO3-, SO4 2- and fluoride (F-)) in the water. Although all ions contribute to specific conductance, their valences and mobilities differ, so their actual and relative concentrations affect specific conductance. When the concentration of ions is high, specific conductance is high, and the resistance to electrical passage is low. Page 71

72 Under natural conditions, the specific conductance of a water body is generally based on the geology of the watershed through which the water flows. Water coming in contact with soils and erodible source rock material will dissolve salts, especially when soil drainage is poor. Some rocks and soils release ions very easily when water flows over them. Concentrations generally are greatest in streams draining basins with rocks and soils that contain easily dissolved minerals (Risey and Doyle, 1997). The chemistry of the surface water can be modified by precipitation and evapotranspiration, weathering of geological formation, and chemical changes brought about by biological organisms and chemical equilibria (Flora and Rosendahl, 1981). Naturally occurring geothermal activity can also contribute to high specific conductance level. Streams that run through areas with granite bedrock tend to have lower specific conductance because granite is composed of more inert materials that do not ionize (dissolve into ionic components) when washed into the water. On the other hand, streams that run through areas with clay soils tend to have higher specific conductance because of the presence of materials that ionize when washed into the water (USEPA, 1997). Groundwater inflows can have the same effect depending on the bedrock they flow through. For example, if acidic water flows over rocks containing calcite (CaCO3), such as calcareous shales, Ca2+ and HCO3- will dissolve into the water; therefore, specific conductance will increase (Virginia DEQ, 2003). One important source of salts in EAA waters is from connate sea water entrapped in shallow ground water formation (Parker et al., 1955, Gleason, 1974; Waller and Earle, 1975; CH2MHill, 1978). Saline waters, left by Pleistocene invasions of the area, are present in formations in the Everglades (Parker et al., 1955). Other potential sources of specific conductance ions in the EAA reported in the literature include dissociated ions from fertilizer application, dissolution of limestone, which is abundant in the soils of the area, and solubilization of ions at newly exposed rock faces after blasting or scarping. It has also been reported that the high mineral concentrations in ground waters are related to the occurrence of muck soils and to the low permeability of both the muck and the underlying marls (Parker et al., 1955; Gleason, 1974). The ionic concentration of precipitation is too low to influence the ionic composition of the surface water in the Everglades, but the rainfall dilutes the ionic concentration of the water thereby lowering specific conductance in the Everglades National Park (Flora and Rosendahl, 1981). Page 72

73 Objectives The objectives of this research are as stated by Chapter 40E-63, Part III: the farm-scale research shall be expanded to include monitoring for specific conductance at all points where total phosphorus is currently being monitored. The expanded research program shall include the development, testing, and implementation of BMPs to address reduction of specific conductance. To achieve these objectives required by chapter 40E-63, Part III, the following studies were conducted: 1. Characterization of specific conductance at farms that are representative of the EAA (different cropping systems, geographical locations, farm size and management practices). 2. Identification of potential sources of specific conductance. 3. Determination of the impact of current P load reduction BMPs on specific conductance. 4. Determination if additional BMPs are needed to abate specific conductance in the EAA. Page 73

74 Materials and Methods Specific Conductance Monitoring Program in the EAA A monitoring program was established in January 1997 to measure specific conductance in accordance with Chapter 40E-63. Hydrolab DataSonde (series 3, 4, 4a) multi-parameter water quality data logger was used to measure and record specific conductance in situ. The DataSonde units were also equipped to measure temperature, ph, dissolved oxygen, oxidation-reduction potential, depth, and turbidity. The DataSonde units were calibrated according to instrument specifications and programmed for a six-day run (programmed to record a measurement every hour). The units were transported to the site canals and deployed at a depth of one meter beneath the canal water surface. After the DataSonde s programmed run ended, it was retrieved from the field. A freshly calibrated and programmed unit was then placed in the canal. A DataSonde unit, which completed its six-day deployment, was brought back into the laboratory where the data were downloaded to a computer and stored in electronic format. A post-run assessment for drift of the instrument s sensors was subsequently conducted. The instruments were then cleaned, maintained, and re-calibrated in the laboratory and returned to the field for deployment during the subsequent monitoring cycle. All field and laboratory activities strictly followed relevant Standard Operating Procedures and the National Environmental Laboratory Accreditation Program (NELAP) approved quality manual (Chen, 2001). Quality control criteria regarding sensor drift and biofouling were as follows: ph <± 0.2 at ph=7.0 check; specific conductance < ± 0.1 ms/cm at ms/cm (0.01 M KCl); percent oxygen saturation < ± 7.0% at 100 % air saturation; redox < ± 20 mv of a ph 7.0 and quinhydrone solution; and turbidity < ± 8.0 NTU at 80 NTU check. Ion concentrations in grab samples were determined to identify the major ions that are associated with specific conductance in the EAA. Grab samples were taken from near hydrolab locations. Analyses for Ca 2+, Mg 2+, and K + concentrations were conducted via an atomic absorption spectrometer, and analysis of Na + was conducted with a flame emission spectrophotometer using EPA Method (USEPA, 1983). Analyses for Cl -, SO 2-4, and F - were conducted using an ion chromatography specified by EPA Method A Dionex DX-500 ion chromatograph, equipped with an injection valve, a sample loop, guard column, Page 74

75 and ion separator columns (IONPAC AS-HC 4-mm, AG9, ASRS-ULTRA, C16, CG16, CSRS-ULTRA), was chosen to measure anion concentrations. The cations Ca 2+, Mg 2+, Na +, and K + were analyzed within 6 months holding time. The anions F -, Cl -, and SO 2-4 were analyzed within 28 days of holding time (Chen, 2001). Bicarbonate anion was not analyzed for in our monitoring program since it was used as the eluent in the ion chromatography determinations. However, we did calculate HCO - 3 ion concentrations as the difference between the sum of the cations and anions measured. UF9204A UF9200A UF9209A UF9208A UF9206A&B OCEAN CANAL UF9201A UF9207A&B UF9203A UF9205A UF9202A Ch. 2 Figure 1. Specific Conductance monitoring sites. Page 75

76 Statistical Data Analysis Histogram analyses and goodness-of-fit tests were conducted to check the distribution patterns of specific conductance and ions measured (Gilbert, 1987). Summary statistics were conducted using UNIVARIATE and ANOVA procedures to assess significant differences between different parameters (SAS, 1999). A box and whisker plot was then used to display visual summaries (site by site and aggregates) of: (1) the center of the data (the median = the centerline of the box), (2) the variation or spread (interquartile range = the box height), (3) the skewness (quartile skew = the relative size of box halves) and (4) presence or absence of unusual values ( outliers and extreme values) (Antonopoulos et al., 2001). Correlation analysis was used to relate water quality parameters with management practices, such as monthly irrigation and sugarcane percentage in the cropping system. Monthly averages of specific conductance for ten farms were used for time series trend analysis. To determine if there are trends in the data over the monitoring period, two tests were done. The upward and downward trends over time were evaluated using a nonparametric Mann-Kendall test for zero slope of the linear regression of time-ordered data versus time using the ChemStat 6.0 software (Starpoint Software Inc., Cincinnati, OH). A positive value of Z indicates an upward trend, whereas a negative value of Z indicates a downward trend (Gilbert, 1987). The Sen s slope test was also used to detect yearly trends. Generally, downward trends in specific conductance indicate an improvement in water quality with time; whereas upward trends indicate a general deterioration in water quality with time (Lietz, 1996). The results were examined by plotting the data and by observing the statistically significant trends (SAS Institute, 1999). Page 76

77 Historical Specific Conductance Information on South Florida Historical water quality information on the EAA and surrounding areas provides an important role in evaluating current specific conductance status of farm canals in the EAA. Parker et al. (1955) conducted an exhaustive field study from 1941 to 1943 on surface and ground water in South Florida. The Hillsboro Canal near Deerfield Beach and the North New River Canal near Ft. Lauderdale showed wide fluctuations in specific conductance. For example, the Hillsboro Canal specific conductance ranged from 0.22 to 1.44 ms/cm from July through August. North New River Canal specific conductance ranged from 0.28 to 1.04 ms/cm during the same time period. The Corps of Engineers (1971) reported specific conductance of waters in the Water Conservations Areas (WCA) from 1950 to Hillsboro Canal at S- 6 showed specific conductance ranging from 0.49 to 1.11 ms/cm for its 10 th and 90 th percentiles; with a median value of 0.78 ms/cm. The Diversion Canal at S-143 showed specific conductance ranging from 0.45 to 1.00 ms/cm for its 10 th and 90 th percentiles; with a median value of 0.74 ms/cm. Gleason (1974) measured specific conductance data in marsh water of WCA-2A and canal waters that were actively flowing into WCA-2A. Specific conductance reported ranged from 0.22 to 2.10 ms/cm for the WCA-2A marsh water during June 26 through August 14, The Hillsboro Canal specific conductance ranged from 1.10 to 1.50 ms/cm, whereas North New River Canal specific conductance ranged from 0.93 to 1.20 ms/cm on July 31, 1973 (Gleason, 1974). Several factors were reported to cause the conductance fluctuations: 1) groundwater in the Everglades is highly mineralized; 2) wide variations in canal water composition due to the additions of lake Okeechobee water; 3) the agricultural drainage from the EAA; and 4) active dissolution of the underlying bedrock as water is drained out of bedrock, which is overlain by peat, within the EAA (Gleason, 1974). A water quality study in the EAA sponsored by the Florida Sugar Cane League and the SFWMD demonstrated that higher specific conductance in the pumped drainage water of three intensively monitored EAA farms were attributed mainly to the exchange of soil water with the highly mineralized shallow (8- to 10- ft) ground water, instead of fertilization application (CH2M Hill, 1978). In particular, shallow ground water and the soil solution at the elevated specific conductance sites had greater Na + and Cl - concentrations as well as Page 77

78 specific conductance than those of the other sites (CH2M Hill, 1978). Shallow groundwater specific conductance measurements from 20 shallow wells in Palm Beach County in the vicinity of Lake Okeechobee ranged from 0.92 to 9.08 ms/cm (Parker et al., 1955). Another study indicated that the specific conductance of the shallow aquifer in the EAA varied from 1.10 to 30.0 ms/cm (Scott, 1977). Sodium, K + and Cl - concentrations in this area were the highest in the entire Palm Beach County (Scott, 1977). The saline groundwater was reported to occur in association with overlaying low permeability muck and/or marl deposits (Gleason, 1974). Parker (1955) and Waller and Earle (1975) refer to the ground water immediately south of Lake Okeechobee as being mineralized due to its contact with connate seawater from ancient marine sediments. They reported that during periods of back pumping, this ground water is drawn into the canals, thereby increasing the mineral content of the surface water. A collaborative study by the University of Florida Agricultural Experimental Station and the USDA Soil Conservation Service described an isolated area of fairly permeable rocks underlying about half of Lake Okeechobee and nearby lands to the south and east, perhaps 25 feet thick and encountered at a depth of 12 to 30 feet (Jones, 1948). The water from this area contained total solids concentration of 4,000 to 5,000 ppm, Cl - ion concentrations as high as 1,500 ppm and SO 2-4 ion concentrations up to 500 ppm (Jones, 1948). In a later study conducted jointly by the U.S. Geological Survey and the U.S. Army Corps of Engineers, specific conductance of surface water in the EAA was measured and higher values was found in the northern half of this area than that from the southern half (Waller and Earle, 1975). This historical information indicated that higher specific conductance water in certain areas in the EAA is a natural phenomenon. Summary statistics of specific conductance data obtained from the SFWMD indicate variable and significantly high values of specific conductance in wells of the EAA, with a mean value of 2.45 ms/cm (Table 1). During , mean specific conductance in main canals decreased in the order of: Ocean Canal (1.63 ms/cm); West Palm Beach Canal (1.49 ms/cm); Hillsboro Canal (1.35 ms/cm); North New River Canal (1.09 ms/cm); Miami Canal (1.01 ms/cm). Lake Okeechobee had an average specific conductance of 0.62 ms/cm during the time period of (Table 1). Historically, high and variable concentrations Page 78

79 in the Hillsboro Canal caused the city of Belle Glade difficulty in treating its public water supply (Parker et al., 1955). Geologically, underlying organic soils of the Lake Okeechobee-Everglades depression is the Fort Thompson formation, a series of alternating beds of limestone, shells, sand, and marl of marine, brackish, and fresh water origin (Jones, 1948; Parker et al., 1955). The marine beds represent times when the area was flooded by the sea; the fresh-water beds record times when sea-level was below the present level and fresh-water lakes and marshes occupied the sea; and the brackish-water beds may represent either times of rising or falling sea levels when the water in the area was neither salt nor fresh but was a mixture of the two. The water in most part of the Everglades region comes mostly from precipitation within the basin (~54 in./yr.) and that upon Lake Okeechobee drainage basin (~51 in./yr.). The quantity of flow into the region through sub-surface aquifers is negligible. However, in the EAA, ditches or wells penetrating the underlying rock may release sufficient flow to increase materially the amount of pumping required for drainage. The water yielded by the occasional solution holes and lenses of permeable material in the Fort Thompson formation under the upper Everglades, is usually so highly charged with minerals that it cannot be used for household purposes or irrigation (Jones, 1948). Water from agricultural lands around the Hillsboro Canal is typically high in Na + and Cl - concentrations derived from ground water used for irrigation and from connate seawater (McKenzie, 1995). Page 79

80 Ch. 2 Table 1. Summary statistics for historical Specific Conductance of Lake Okeechobee, ground water, and canal water in the EAA. Water Sources Time No of Specific conductivity, ms/cm Significant Test period Observation Range Median Minimum Maximum 10th pctl 90th Pctl Std Dev Mean Difference Pr>F Wells a < Ocean Canal b West Palm Beach Canal bc Hillsboro Canal bcd North New River Canal cd Miami Canal de Lake Ockeechobee e Note: Raw data from SFWMD DBHYDRO database. Page 80

81 Results I. General Characteristics of Specific Conductance in the EAA Specific Conductance Averages of specific conductance and other parameters over the entire study period are presented in Table 2. Crop and water management characteristics of the farms monitored as well as rainfall are presented in Table 3. The skewness and kurtosis values for turbidity and pump hours are not close to zero, and the mean values of the monthly average of these parameters are above the median. This indicates that data sets for these two parameters are highly positively skewed and a log-transformation of the data is necessary. Geometric means instead of arithmetic means are therefore better representations of the true mean for turbidity and pump hours, while arithmetic mean is used for all other parameters. Monthly averages of specific conductance at the ten farms (12 discharge structures) are presented in Figure 2. Means of specific conductance and other parameters at each pump structure for the entire monitoring period are presented in Table 4. Specific conductance data from two farms (UF9206A&B, and UF9208A) show averages above ms/cm (Table 4). No differences in rainfall, evaporation or irrigation were observed among sites. The percentage of cane in the cropping system was different between sites, varying from less than 30% in UF9206A&B to 99.8% in UF9202A. Arithmetic means of specific conductance at the 12 pump structures decreased in the following order (in ms/cm): UF9208A (1.682)> UF9206A (1.513), UF9206B (1.540) > UF9201A (1.169) UF9204A (1.076), UF9207A (1.091), UF9207B (1.086) UF9202A (0.954) UF9200A (0.888) UF9203A (0.862), UF9209A (0.818) UF9205A (0.738). Historical data have indicated monthly fluctuation in dissolved salts in the Everglades canals (Parker et al., 1955). A box and whisker plot of the monthly specific conductance data Page 81

82 Ch. 2 Table 2. Summary Statistics for hourly averages of Specific Conductance and other parameters in all sites during the study period. Parameter Units No. of Obs. Mean Standard Deviation Standard Error Range Minimum Maximum Median Skewness Kurtosis Pr.>F by Site Pr.>F by Month Sp. Cond. ms/cm < < Temperature 0 C < ph < < TDS g/l < < Diss. Oxy. % Sat < < Diss. Oxy. mg/l < < ORP mv < < Turbidity NTU < Variation as a result of different pump structures Variation as a result of different months. Page 82

83 Ch. 2 Table 3. Summary of rainfall, water and crop management at the ten farms combined. Parameter Units No. of Obs. Mean Standard Deviation Standard Error Range Minimum Maximum Median Skewness Kurtosis Pr.>F by Site Pr.>F by Month Rainfall in < Evaporation in < Irrigation in < Cane % < Pu mp hrs < < Variation as a result of different pump structures. Variation as a result of different months. Page 83

84 S p C o n d A 1A 2A 3A 4A 5A 6A 6B 7A 7B 8A 9A Si t e Ch. 2 Figure 2. Monthly averages of Specific Conductance (SpCond, ms/cm) by site. Page 84

85 Ch. 2 Table 4. Means of Specifc Conducatnce and other parmeters by site. Start Stop Sp. Cond Temp. ph TDS DO DO ORP Turbidity Rainfall Evap. Irrigation Site Date Date (ms/cm ) ( 0 C ) (g/l ) (% Sat. ) (mg/l) (mv) (NTU) (in.) (in.) (in.) UF9200A Nov-96 Dec UF9201A Jan-98 Dec UF9202A Jan-98 Dec UF9203A Jan-98 Dec UF9204A Jan-98 Dec UF9205A Jan-98 Dec UF9206A Nov-96 Dec UF9206B Nov-96 Dec UF9207A Jan-98 Dec UF9207B Jan-98 Dec UF9208A Jan-98 Aug UF9209A Jan-98 Dec Arithmetic mean. Geometric mean. Page 85

86 indicates that specific conductance shows monthly fluctuation (Figure 3). ANOVA of the data (p=0.05) shows the overall monthly trend of specific conductance in canal water of EAA farms decreased as follows (arithmetic mean in ms/cm): August (1.280), September (1.263) and October (1.253) > February (1.200), November (1.185), January (1.180), July (1.175), December (1.171), and March (1.136) > April (1.027) > May (0.934) and June (0.892). The three highest months (August, September, and October) also had the highest rainfall (8.12, 6.52, and 5.77 in., respectively). This is consistent with the observation by Parker et al. (1955) who found that Cl - and Na + concentrations in September samples were considerable higher than concentrations in May. Ion Composition Determination of ion compositions in grab samples at ten pump structures indicated that HCO - 3, Cl - and SO 2-4 are the major anions and Na + and Ca 2+ are the major cations in farm canal water of the EAA farms (Table 5). It is believed that in the current ph range of 7.0 to - 7.5, the principal ion responsible for alkalinity is HCO 3 (Hem, 1985; Gleason, 1974). In this - study, the mean concentrations of HCO 3 varied from mg/l at UF9209A to mg/l at UF9208A (Table 6). The mean concentrations of Cl - varied from 71.6 mg/l at UF9202A to mg/l at UF9208A (Table 6; Figure 4). It was reported that many of the surface waters in southeastern Florida had less than 15 mg/l of Cl -, but ground waters with 100 mg/l, or more, were not uncommon (Parker et al., 1955). Mean concentrations of SO 2-4 ranged from 45.2 mg/l at UF9203A to mg/l at UF9208A. Mean concentrations of Na + ranged from 41.7 mg/l at UF9202A to mg/l at UF9208A. The quantity of Na + in ordinary surface or ground water is reported to be less than 30 mg/l; considerable quantities of Na + would be found in waters contaminated with sea water or in waters with salts enclosed in the older marine deposits (Parker et al., 1955). Mean concentrations of Ca 2+ ranged from less than 51.0 mg/l to 78.8 mg/l (Table 6). Mean concentrations of Mg 2+ ranged from 18.9 mg/l to 46.1 mg/l, and mean concentrations of K + varied from 5.8 mg/l to 11.4 mg/l (Table 6). Page 86

87 S p C o n d Mont h Ch. 2 Figure 3. Mean Specific Conductance from all sites combined (SpCond, ms/cm) by month. Page 87

88 Ch. 2 Table 5. Statistical summary for ion composition and other parameters in grab samples combined for 10 pump structures. Variable N Range Median Minimum Maximum Mean Std Dev K (mg/l) Na (mg/l) Ca (mg/l) Mg (mg/l) Cl (mg/l) SO 4 ( mg/l) F (mg/l) HCO 3 (mg/l) Temperature, o C ph Sp. Cond. (ms/cm) TDS (mg/l) Diss. Oxy. (% Sat.) Diss. Oxy. (mg Sat.) Redox (mv) Turbidity (NTU) Pump volume (galon) Rainfall (in.) Calculated based on the balance of cations and anions. Page 88

89 Ch. 2 Table 6. Mean ion concentrations in grab samples by site. Site K Na Ca Mg Cl SO 4 F HCO mg/l Na/Cl UF9200A 8.7 b 64.0 cd 52.2 c 22.3 d 93.6 cd de 0.61 c ef 0.65 c UF9202A 5.8 d 41.7 e 77.7 ab 26.7 c 71.6 e 70.3 bc 0.57 c c 0.57 d UF9203A 7.3 c 56.7 d 49.1 c 22.2 d 84.2 de 45.2 g 0.54 c def 0.67 bc UF9204A 9.1 b 79.1 c 67.8 b 35.9 b c 76.5 b 0.65 bc b 0.72 ab UF9206A 10.8 a b 51.0 c 25.8 c b 65.8 cd 0.91 a cde 0.68 bc UF9206B 11.4 a b 54.7 c 28.3 c ab 68.6 c 0.82 ab cd 0.68 bc UF9207A 9.7 b 77.5 c 54.9 c 26.6 c c 54.3 f 0.72 abc c 0.72 bc UF9207B 9.6 b 77.6 c 56.4 c 27.6 c c 55.3 ef 0.73 abc c 0.70 bc UF9208A 10.9 a a 78.8 a 46.1 a a a 0.89 a a 0.78 a UF9209A 6.5 cd 49.8 d 51.8 c 18.9 d 82.6 de 49.1 fg 0.61 c f 0.59 d Calculated based on the balance of cations and anions. Page 89

90 Cl -, mg/l Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01 Dec-02 Ch. 2 Figure 4. Plot for Cl - concentration (mg/l) from Feb. 28, 2001 to Oct. 27, 2002 by site. Page 90

91 It is noted that concentrations of Na + and Cl - in farm canal water at UF9208A and UF9206A&B are significantly greater than at other pump structures in the EAA (Table 6). Correlation coefficient analysis shows a strong correlation coefficient (r=0.95) between concentrations of Na + and Cl - 2- in canal water near the EAA farms. The concentration of SO 4 ion is the highest at UF92008A. We also calculated the Na/Cl ratio in these farm canals. The average Na/Cl weight ratio for farm canal water in the EAA ranged between 0.57 and 0.78 (Table 6). The Na/Cl weight ratio for seawater is 0.55 (Stumm and Morgan, 1981). Most of these farm canals are close to this ratio or slightly higher. The Na/Cl weight ratio for Hillsboro Canal water and WCA-2A marsh water varied from 0.76 to 0.85 (Gleason, 1974). The relatively high Na content of canal water in the EAA may partly from the remnants of saline residues that have not been completely flushed out of the ground and partly from cation-exchange processes (Parker et al., 1955). Because the muck and rock of the EAA are much less permeable than the sandstones and limestones, the saline residues have not been entirely flushed out, and the organic colloids are still partly saturated with Na, presumably adsorbed from ancient seawater. When brought in contact with Na + bearing clays, the Ca +2 in the solution is exchanged for Na + in the clays. The water then comes in contact with more lime rock, which dissolves to form more Ca +2 and HCO - 3. Repetition of the process increases HCO - 3 and Na + to high values. Page 91

92 II. Potential Sources of Specific Conductance in the EAA It is clear from Section I (general characteristics of specific conductance in the EAA) that specific conductance was not a problem in the majority of the farms monitored. Eight out of the ten farms monitored showed mean specific conductance values below ms/cm. Only two farms (UF9206A&B, and UF9208A) had mean specific conductance values higher than ms/cm. In this section, the effect of geological influences, water management practices, and fertilizer on specific conductance in the EAA will be investigated. Geological Influences Historically, groundwater in the Everglades is highly mineralized and dissolved solids increase with depth. Saline waters and residues left by Pleistocene invasions of the area by the sea have never been completely flushed out of the formations in much of the Everglades, particularly in areas near Lake Okeechobee (Jones, 1948; Parker et al., 1955). Some wells in the EAA less than 50 ft deep yield water high in specific conductance, SO and Cl - (Parker et al., 1955, Scott, 1997). We superimposed the average specific conductance of the monitored discharge structures on the variation map of the shallow ground water (20 to 50 ft well depths) Cl - levels from Parker et al. (1955) (Figure 5). The current specific conductance data points show that the high elevated conductance at UF9208A are in the area of wells that have a Cl - concentration ranging from mg/ L and those of UF9206A&B are in the area of wells that have a Cl - concentration of mg/l (Figure 5). Chloride concentrations greater than 100 mg/l were considered to be evidence of saltwater mixing with freshwater in the surficial aquifer system and indicate the presence of saltwater interface (Hittle, 1999). Up to 400 mg/l of Cl and 390 mg/l of Na + were reported in shallow ground waters close to UF9206A&B in two separate investigations conducted by Scott (1977) and Miller and Lietz (1976). They also documented specific conductance values of up to 2.35 ms/cm in the ground water samples in the same wells. Specific conductance in three farms in the vicinity of UF9206A&B averaged 1.56, 2.15 and 2.29 ms/cm, respectively, and the shallow groundwater had a mean specific conductance of 1.54 ms/cm (CH2M Hill, 1978). High specific conductance values and ion concentrations were found in shallower wells ( ft.) in an area just south of Lake Okeechobee, between West Palm Beach and Page 92

93 Ch. 2 Figure 5. Study sites with mean Specific Conductance duperimposed upon a map of chloride concentration of shallow wells (20 To 50 Feet Depth) in the EAA (chloride map recreated from Parker, 1955). Page 93

94 Hillsboro canals (Miller and Lietz, 1976)(Figure 6). It is most likely that these two canals (the depth of the canals is between ft.) intersect this highly mineralized area and get the saltwater intrusion by cutting into the saltwater interface. The current specific conductance data correspond well also with documented specific conductance data of the SFWMD (Table 1). This is indicative that the elevated specific conductance in certain canals in the EAA is a result of saline groundwater intrusion from ancient seas. This is especially true for specific conductance in the northern portion of Hillsboro Canal (Figure 1), which agrees with result of McKenzie (1995), who documented that the quality of the water in the southern part of the Boca Raton canal system is affected by the Hillsboro Canal water channeled from conservation and agricultural areas to the west. UF9208A is located near the north end of the Hillsboro Canal. Drainage Pumping Data from three intensively monitored farms (UF9206A&B, UF9200A and UF92009A) were analyzed to determine if there were specific management practices that affect specific conductance; specifically, the effects of drainage pumping and irrigation on specific conductance. Table 7 shows the effect of ambient conditions, irrigation and drainage pumping on specific conductance on these three farms. Drainage pumping had no significant effect on specific conductance at UF9206A&B (Table 7). Specific conductance was higher during drainage pumping than during irrigation or ambient conditions for both UF9200A and UF9209A. The specific conductance values of drainage waters at these two farms, however, were below ms/cm. The pumping effect on specific conductance is site specific as seen by the significant site by pumping interaction (Table 8). To illustrate the effects of drainage pumping on specific conductance at the three pumping structures with mean elevated specific conductance (UF9206A, UF9206B, and UF9208A), graphs of hourly specific conductance observations over time and against cumulative drainage volume were plotted. Six drainage events were plotted for each of the three structures (Appendix 1). Example graphs of drainage events from each structure are presented in Figures 7, 8, and 9. The drainage event chosen for structures UF9206A and UF9206B occurred in March 2001 (Figures 7 and 8). Drainage pumping was initiated after a rainfall of approximately one inch had occurred on the 29 th of March. At both structures, specific conductance prior to drainage initiation was > 1.75 ms/cm. As drainage volume Page 94

95 Ch. 2 Figure 6. Water quality of shallow wells in the EAA and study farm locations (well data from Miller and Lietz, 1976). Page 95

96 Ch. 2 Table 7. Statistical summary of the effects of ambient conditions, pumping, and irrigation on Specific Conductivity at four structures. Water Management No of Statistical Summary (Specific conductivity, ms/cm) Significant Test Practices Observation Range Median Minimum Maximum Std Dev Mean Difference Pr>F Site=UF9200A Ambient b < Pump a Irrigation c Site=UF9206A Ambient a < Pump a Irrigation b Site=UF9206B Ambient a < Pump a Irrigation b Site=UF9209A Ambient b < Pump a Irrigation c All 4 sites (UF9200A, UF9206A&B, UF9209A) Ambient a < Pump a Irrigation b Page 96

97 Ch. 2 Table 8. Effects of interactions between site, pumping, rainfall, and irrigation on daily average Specific Conductance at four intensively monitored pump structures. Variables Degree of Type I hypotheses freedom F Value Pr > F Site <.0001 Irrigation <.0001 Pump Rainfall Site X Irrigation <.0001 Site X Pump <.0001 Site X Rainfall Irrigation X Pump Irrigation X Rainfall Pump X Rainfall Site X Irrigation X Rainfall Site X Irrigation X Pump X Rainfall Page 97

98 Sp Cond (ms/cm) or Rain (in.) UF9206A - Conductivity vs. Time Graph A SpCond Rain Volume Volume (KGal/hr) /29/ :00 03/29/ :00 03/30/ :00 03/30/ :00 03/31/ :00 03/31/ : /01/ :00 Sp Cond (ms/cm) UF9206A - Conductivity vs. Cumulative Volume SpCond Canal level Graph B Cumulative Volume (Kgal/A) Canal level (Ft.) Ch. 2 Figure 7. Hourly Specific Conductance of drainage event UF9206A plotted against time (Graph A) and cumulative volume pumped per acre (Graph B). Page 98

99 Sp Cond (ms/cm) or Rain (in.) /29/ :00 03/30/ :00 UF9206B - Conductivity vs. Time 03/30/ :00 03/31/ :00 03/31/ :00 Graph A SpCond Rain Volume 04/01/ :00 04/01/ : /02/ :00 Volume (KGal/hr) Sp Cond (ms/cm) UF9206B - Conductivity vs. Cumulative Volume SpCond Canal level Graph B Cumulative Volume (Kgal/A) Canal level (Ft.) Ch. 2 Figure 8. Hourly Specific Conductance of drainage event UF9206B plotted against time (Graph A) and cumulative volume pumped per acre (Graph B). Page 99

100 increased (Figure 7 & 8), specific conductance decreased; once pumping ceased specific conductance at UF9206B returned to pre-pumping levels (Figure 8, Graph A), at UF9206A specific conductance did not change (Figure 7, Graph A). Three of six drainage events plotted for UF9206A were observed to have increased conductance with volume pumped; at UF9206B the same trends were observed with three of six events producing trends of increased specific conductance with pumping. This variable effect of pumping on specific conductance explains results reported in Table 7. The inconsistent effect of drainage pumping on specific conductivity within UF9206A&B is the result of the dynamic hydrologic conditions that exist in the area (and within the EAA). Contrary to previous, widely held perceptions that Everglades peat is impermeable and that hydraulic driving forces are too small to cause recharge and discharge, significant exchanges of surface and ground water in Everglades peat can occur at very low pressure differences (Harvey, 2004). The flow of shallow aquifer water into farm canals is dependent on pressure differences. Conditions that create pressure differences and favor groundwater flow into the farm canals are dependent on relative pressures of surface canals and the shallow aquifer, both off-farm and on-farm. As the farm and its adjacent farms are drained or kept flooded and as SFWMD canals are raised and lowered, pressure gradients are created which may or may not favor the movement of shallow groundwater into farm canals. The complex conditions that favor groundwater flow into the farm canals are not thought to be significantly influenced by any management practices on this farm, but are mainly governed by off-farm hydrologic influences, i.e. district conveyance canal levels and adjacent farm water levels. The drainage event used to illustrate the influence of drainage pumping on specific conductance at structure UF9208A occurred in July 2001 (Figure 9). A rain of 2.42 inches fell on July 23, 2001; drainage pumping started on July 23, and ended on July 24, Specific conductance was greater than 2.00 ms/cm prior to the drainage event; during the initial stages of drainage pumping specific conductance decreased slightly then returned to pre-pumping levels. Five of the six events plotted at this structure exhibited a trend towards increased specific conductance with pumping duration and volume. Page 100

101 Sp Cond (ms/cm) or Rain (in.) SpCond Rain Volume UF9208A Conductivity vs. Time Graph A Volume (Kgal/hr) /22/ :00 07/23/ :00 07/23/ :00 07/24/ :00 07/24/ :00 07/25/ : /25/ : UF9208A Conductivity vs. Cumulative Volume Graph B SpConc (ms/cm) Level (Ft.) 0.50 SpCond Canal level Cumulative Volume (Kgal/A) Ch. 2 Figure 9. Hourly Specific Conductance of drainage event UF9208A plotted against time (Graph A) and cumulative volume pumped per acre (Graph B). Page 101

102 This 262-acre sugarcane farm is bordered on the north by the Hillsboro Canal and on the south by a large private canal. As mentioned earlier in this report, water from agricultural lands around the Hillsboro canal has been reported to be high in Na + and Cl - concentrations (McKenzie, 1995). The farm drainage structure discharges into the private canal to the south. During drainage events the Hillsboro canal to the north is normally at a higher stage level relative to the private drainage canal to the south of the farm. The difference in stage level between the two canals directly influences farm canal water level, field water tables, and creates a pressure gradient between the two conveyance canals across the farm. The consistent increase in specific conductance with drainage pumping indicated, that for most drainage events at this farm, conditions exist that favor shallow groundwater intrusion into the farm drainage stream. Irrigation Water At the three intensively monitored farms, UF9200A, UF9206A&B, and UF9209A, the addition of irrigation water decreased specific conductance in farm canals (Table 7). The irrigation water flowing through the structures with the highest specific conductance (UF9206A, UF9206B, and UF9208A) was also characterized by higher specific conductance (Table 1). Site UF9208A received irrigation water via a secondary canal that connects to the Hillsboro canal. Sites UF9206A and UF9206B receive irrigation water from a secondary canal that connects to the Ocean canal. The Ocean canal may source its water from either the West Palm Beach canal to the east, or the Hillsboro Canal to the west. Both the Ocean and the Hillsboro Canals have historically had relatively high specific conductance compared to the other major district conveyance canals in the EAA (Table 1). The effect of irrigation water quality on farm canal mean specific conductance is also presented in Table 9. Sites that received irrigation water directly from low specific conductance district canals (Miami and North New River) had lower mean specific conductance values. Sites that received irrigation water from district canals with relatively higher specific conductance (Ocean and Hillsboro canals) had relatively higher mean specific conductance values. It was also evident that the sites that did not receive irrigation water directly from main conveyance canals, but received irrigation water via secondary or branch canals had relatively higher mean specific conductance values. The quality of Page 102

103 Ch. 2 Table 9. Specific Conductance, irrigation source, and pumping to rainfall ratio by site. Site Non-Drainage Specific Conductance (ms/cm) Irrigation Source Drainage Specific Conductance (ms/cm) Pumping to Rainfall Ratio UF9205A MC UF9209A MC UF9203A NNR UF9200A WP UF9202A interior-nnr UF9204A interior-oc UF9207AB HC UF9201A interior-hc UF9206AB interior-oc UF9208A interior-hc Note: MC=Miami Canal, NNR=North New River, WP=West Palm Beach, OC=Ocean Canal, HC=Hillsborough Canal. Page 103

104 irrigation water that a farm receives appears to have a direct influence on the specific conductance of the subsequent water that leaves the farm during drainage pumping events. The drainage volume to rainfall ratio appeared to have had little or no effect on farm canal specific conductance (Table 9). The farm that had the highest average specific conductance during drainage pumping (UF9208A) also had a low drainage pumping to rainfall ratio (0.24). Site UF9205A had the highest drainage pumping to rainfall ratio (1.04), but had relatively low drainage pumping specific conductance (0.941 ms/cm). Two sites, UF9202A and UF9204A (both farms are 640-acre sugarcane monoculture farms), had the same drainage pumping to rainfall ratios (0.17), similar management practices, yet had different drainage pumping specific conductance values (0.87 and 1.14 ms/cm, respectively). These observations lend support to the conclusion that farm canal specific conductance is governed mainly by the quality and hydrology of the underlying shallow ground water, which is farm specific. Farm canal specific conductance also appears to be influenced to a lesser degree by the quality and quantity of the irrigation water it receives. Fertilizers Based on the current level of fertilizer BMPs in the EAA, K and P applied would be taken up by the crop during the growing season. Coale et al. (1993) evaluated nutrient accumulation and removal by sugarcane grown on Everglades histosols. They concluded that 63 and 64% of total accumulated P and K, respectively, were removed from the field as millable sugarcane. Phosphorus and K removal from the field by crop harvest was higher than the amount applied by fertilizers at that season. For example, an average of 343 kg K/ha was removed by crop harvest, with a fertilizer application of 150 kg K/ha. The predominant fertilizers applied to sugarcane in the EAA are potassium chloride (KCl: potash), and diammonium phosphate (DAP: (NH 4 ) 2 HPO 4 ) or triple super phosphate (TSP: Ca(H 2 PO 4 ) 2.H 2 O). From the fertilizer recommendations of the UF/IFAS Soil Testing Laboratory at the Everglades Research and Education Center, an average application of plant nutrients for a ratoon sugarcane crop in the EAA would be 125 lbs K and 20 lbs P per acre. Utilizing the most commonly applied fertilizers for sugarcane in the EAA, this is approximately equivalent to 240 lbs of KCl (115 lbs Cl - ) and 87 lbs of DAP or 75 lb of TSP per acre. The main fertilizers applied to leafy vegetables in the EAA are ammonium polyphosphate (APP:(NH 4 PO 3 ) n ) and KCl. Following UF/IFAS recommendations, an Page 104

105 average application rate of P and K plant nutrients for a leafy vegetable crop growing in the EAA would be 166 lbs K and 175 lbs P per acre. Utilizing the most commonly applied fertilizers for vegetables in the EAA, this would be approximately equivalent to 320 lbs of KCl (153 lbs Cl - ) and 1080 lbs of APP per acre. A comparison of potential Cl - load additions from fertilizer to the TDS loads produced by farm drainage waters has been reported for vegetable, sugarcane, and cattle farms of the EAA (CH2M Hill, 1978). The comparison showed that even if the entire annual Cl - load from fertilizer application at the highest rate of fertilization site (vegetable farm) was exported in the drainage water, the amount of Cl - derived from fertilizer would account for less than 3% of the annual drainage water TDS loading from the farm. The Cl - load applied to the farm through crop fertilization at the vegetable farm was reported to be 233 lbs/a for the 14-month study period (July 1976 through September 1977). The vegetable farm, located close to UF9206A&B in the current research, had a mean specific conductance value of 2.29 ms/cm during the study period (CH2M Hill, 1978). An analysis of drainage volumes, total dissolved salts, and calculated potash fertilizer applications was conducted for the elevated specific conductance farm UF9206A&B. From 1997 through 2002 the farm planted an average of 45, 30, and 25% of its cultivated acreage to sod, cane, and leaf vegetables, respectively. By applying moderate level potash fertilization rates to each crop and multiplying fertilizer rates by the percentages planted to each crop, a mean fertilizer application rate (353 lb KCl per acre per year) was calculated for potash, the predominant and most commonly applied fertilizer at this farm and throughout the EAA. A graphical comparison between potash fertilizer application and total TDS exported by farm UF9206A&B is presented in Figure 10 (Graph A). Total TDS exported by the farm is normalized to kg per acre per year for ease of comparison with fertilizer rate. Even by assuming the unlikely scenario of total export of all fertilizer KCl via drainage water, the potential average annual contribution of applied KCl fertilizer to total TDS exported in the farm drainage water would only be 6.4%. The potential percent contribution of potash fertilizer to total TDS load varied directly with yearly drainage volume, further emphasizing the minor role of fertilizer application in farm total TDS export (and specific conductance). Chloride is a highly mobile ion in the soils, subsurface strata, and waters of the EAA (CH2M Hill, 1978). Most plant species, however, take up Cl - in relatively high rates. Plant tissues usually contain substantial amounts of Cl - often in the range of 2 to 20 mg Cl/g dry weight Page 105

106 but in the chlorophilic species Cl - may amount to 100 mg Cl/g dry weight (Mengel and Kirkby, 2001). Anderson and Bowen (1990) report that critical level of Cl - in sugarcane is 0.068% and toxic level is 0.5%. Assuming a Cl - concentration in leaves of 0.3% (3 g/kg dry weight) and an average dry matter yield of 15 tons (1 ton = 2000 lbs) per acre, the uptake of Cl - by a sugarcane crop is about 40 Kg (90 lb) per acre. A comparison of total Cl - from calculated potash fertilizer applications to total Cl - exported in drainage waters was conducted for farm UF9206A&B for the calendar years of 2001 and 2002 (Figure 10 Graph B). Yearly Cl - drainage loads were calculated from annual drainage volumes and average Cl - concentrations of weekly water samples collected in 2001 and The comparison shows the potential relative contribution of fertilizer Cl - to the total Cl - exported in drainage water. Even with the implausible assumption that no fertilizer-sourced Cl - is removed by harvested crops, the maximum potential contribution as a percentage of fertilizer-sourced Cl - to exported Cl - load was 14.9 and 24.1 % for 2001 and 2002, respectively. The higher percentage for calendar year 2002 was the result of greatly reduced drainage pumping due to drought. Reductions or changes in fertilizer use are not projected to have a significant impact upon mean specific conductance at this farm. Page 106

107 . Potash and Total TDS (kg acre -1 Yr -1 ) (39.7) 1998 (35.8) Farm UF9206AB Graph A 1999 (38.3) 2000 (33.4) Year (Drainage Volume: Acre-Inch Acre -1 Year -1 ) 2001 (31.5) Potash Fertilizer (KCl) Total Exported TDS 6.4% 5.6% 6.3% 6.5% 6.2% 7.1% 2002 (23.8) Farm UF9206AB Graph B Chloride in Fertilizer Chloride in Drainage Water Chloride (kg acre -1 Yr -1 ) % 24.1% (31.5) Year (Drainage Volume: Acre-Inch Acre -1 Year -1 ) 2002 (23.8) Ch. 2 Figure 10. Potential contribution of potash fertilizer to total exported dissolved solids from farm UF9206AB (Graph A); potential contribution of Cl- from potash fertilizer to total exported Cl- (Graph B). Page 107

108 III. Impact of Phosphorus BMPs on Specific Conductance Current Drainage and fertilizer BMPs Impact on Specific Conductance It was evident from the results presented in section II, that the shallow ground water mineralogy in the EAA plays the major role in determining levels of specific conductance in the EAA. Both drainage pumping and fertilizer application were shown to have either inconsistent or minimal effect on specific conductance. Correlation analysis was conducted to determine relationships between specific conductance and certain management practices. Correlation analysis does not indicate a cause and effect, but merely that a relationship exists. There is a weak correlation between monthly pump hours and specific conductance (r = 0.21), and a weak negative correlation between specific conductance and irrigation (r=-0.25) (Table 10). The correlation analysis showing the weak relationship between specific conductance and pump hours emphasizes the conclusions we had regarding the inconsistent effect of drainage pumping on specific conductance. Farm canal specific conductance is governed mainly by the quality and the hydrology of the underlying shallow ground water, which is farm specific. Therefore the current P load reduction water management BMP of delaying drainage till ½ or 1 of rain has fallen is expected have variable effects on specific conductance. The Fertilizer analysis provided in the pervious section illustrated that Cl from KCl fertilization at the recommended rates in the EAA comprises a small percentage of the total TDS in drainage water. It is concluded that the current soil testing/ fertilizer BMPs currently employed are adequate in maintaining low levels of ions in drainage water coming from fertilizers. Page 108

109 Ch. 2 Table 10. Correlation coefficients between parameters monitored for all sites combined. r value Temp. ph SpCond Diss. Oxy. ORP Turbidity Rainfall Evap. Irrigation Pump Temp. ( o C) 1 ph SpCond. (ms/cm) Diss. Oxy. (% Sat.) ORP (mv) Turbidity (NTU) Rainfall (in.) Evaporation (in.) Irrigation (in.) Pump (hrs) ORP: Oxidation-reduction potential. Page 109

110 Trends of Specific Conductance since 1997 Water quality data possess unique characteristics that may exhibit seasonal variation, which may include a seasonal fluctuant, as well as a yearly trend. This variation may be the result of a diversity of conditions, including specific agricultural land use practices, biological activity, or sources of steam flow or sediment (Lietz, 2000).Trend analysis was conducted to ascertain the impact of current BMPs on specific conductance. A trend in water quality is defined as a monotonic change in a particular constituent with time (Lietz, 2000). Historically, canal waters exhibit widely fluctuating levels of specific conductance, HCO - 3, Ca 2+, nitrate, and orthophosphate rather than strong trends (Gleason, 1974). In the current study, box and Whisker plot of the yearly specific conductance data from 1997 shows a generally decreasing trend (Fig.11). In 1997, only two farms were being monitored and therefore the significance of this decreasing trend using ANOVA analysis is not possible. To determine if this general yearly trend decrease is significant, non-parametric Mann Kendall trend analyses, and Sen s slope analyses were conducted for each site separately (Table 11). The non-parametric Mann-Kendall trend analysis (p values less than 0.05) confirmed statistically significant downward trends in specific conductance at UF9202A, UF9205A and UF9207B (Table 11). Sen s non-parametric estimator of slope analysis indicates reduction rates of specific conductance at UF9202A, UF9205A and UF9207B are at 4.32, 33.5, and 5.00 µs/cm per year, respectively (Table 11). A statistically significant upward trend in specific conductance at UF9208A was also detected using the nonparametric Mann-Kendall trend analysis (p values less than 0.05). However, the upward trend is not significant using the Sen s non-parametric estimator of slope analysis. To visually show the trends, monthly averages of specific conductance over the entire monitoring period were plotted with time. A time series line generated by SAS (SAS, 1999) was also plotted (Fig.12-23). The figures show the downward trend at the previously mentioned three farms. With the exception of UF9208A, all other farms show no change in specific conductance. Although UF9201A showed visually a downward trend (Fig. 13), this trend was not significant due to the variability of the data. In summary, the trend analysis presented in this section demonstrates that a downward yearly trend of specific conductance is obvious in three out of the ten farms monitored. This is an indication that current P load reduction BMPs have had a positive effect on specific conductance on some farms in the EAA. Page 110

111 S p C o n d Year Ch. 2 Figure 11. Plot of average monthly Specific Conductance (SpCond, ms/cm) from all sites combined by year. Page 111

112 Ch. 2 Table 11. Non-Parametric Mann-Kendall Trend Analysis and Sen s Slope Analysis of Specific Conductance by site. Mann-Kendall Test Sen s Slope Analysis Site Z-score 95% confidence Estimator median Q 90% confidence level (µs/cm) level (µs/cm) UF 9200A No trend No trend UF 9201A No trend No trend UF 9202A Downward trend Downward trend UF 9203A 0.05 No trend 0 No trend UF 9204A No trend No trend UF 9205A Downward trend Downward trend UF 9206A No trend No trend UF 9206B No trend No trend UF 9207A No trend No trend UF 9207B Downward trend Downward trend UF 9208A 1.73 Upward trend No trend UF 9209A 1.06 No trend No trend Page 112

113 3.0 Specific Conductance Data: UF9200A Specific Conductance in ms/cm Actual Data Time Trend 0.0 Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01 Dec-02 Time Ch. 2 Figure 12. Time trend of Specific Conductance at UF9200A. Page 113

114 3.0 Specific Conductance Data: UF9201A Specific Conductance in ms/cm Actual Data Time Trend 0.0 Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01 Dec-02 Time Ch. 2 Figure 13. Time trend of Specific Conductance at UF9201A. Page 114

115 3.0 Specific Conductance Data: UF9202A Specific Conductance in ms/cm Actual Data Time Trend 0.0 Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01 Dec-02 Time Ch. 2 Figure 14. Time trend of Specific Conductance at UF9202A. Page 115

116 3.0 Specific Conductance Data: UF9203A Specific Conductance in ms/cm Actual Data Time Trend 0.0 Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01 Dec-02 Time Ch. 2 Figure 15. Time trend of Specific Conductance at UF9203A. Page 116

117 3.0 Specific Conductance Data: UF9204A Specific Conductance in ms/cm Actual Data Time Trend 0.0 Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01 Dec-02 Time Ch. 2 Figure 16. Time trend of Specific Conductance at UF9204A. Page 117

118 Specific Conductance in ms/cm Specific Conductance Data: UF9205A Actual Data Time Trend 0.0 Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01 Dec-02 Time Ch. 2 Figure 17. Time trend of Specific Conductance at UF9205A. Page 118

119 Specific Conductance in ms/cm Specific Conductance Data: UF9206A Actual Data Time Trend 0.0 Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01 Dec-02 Time Ch. 2 Figure 18. Time trend of Specific Conductance at UF9206A. Page 119

120 Specific Conductance in ms/cm Specific Conductance Data: UF9206B Actual Data Time Trend 0.0 Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01 Dec-02 Time Ch. 2 Figure 19. Time trend of Specific Conductance at UF9206B. Page 120

121 Specific Conductance in ms/cm Specific Conductance Data: UF9207A Actual Data Time Trend 0.0 Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01 Dec-02 Time Ch. 2 Figure 20. Time trend of Specific Conductance at UF9207A. Page 121

122 3.0 Specific Conductance Data: UF9207B Specific Conductance in ms/cm Actual Data Time Trend 0.0 Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01 Dec-02 Time Ch. 2 Figure 21. Time trend of Specific Conductance at UF9207B. Page 122

123 3.0 Specific Conductance Data: UF9208A Specific Conductance in ms/cm Actual Data Time Trend 0.0 Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01 Dec-02 Time Ch. 2 Figure 22. Time trend Of Specific Conductance at UF9208A. Page 123

124 3.0 Specific Conductance Data: UF9209A Specific Conductance in ms/cm Actual Data Time Trend 0.0 Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01 Dec-02 Time Ch. 2 Figure 23. Time trend of Specific Conductance at UF9209A. Page 124

125 Discussion Best management practice or BMP means a practice or combination of practices determined by the district, in cooperation with the department, based on research, fieldtesting, and expert review, to be the most effective and practicable, including economic and technological considerations, on-farm means of improving water quality in agricultural discharges to a level that balances water quality improvements and agricultural productivity ( section 2 of the F.A.C.). The questions that this study aimed to address are the following: is the high specific conductance in the EAA man-induced, and can it be abated by additional BMPs? The quality and hydrology of the shallow aquifer are the major factors controlling specific conductance in the EAA. High specific conductance in the EAA is a natural phenomenon. High levels of specific conductance and Cl - concentrations in shallow wells in the EAA have been documented in 1948 and 1955, before the advent of major agricultural operations in the area. High conductance levels in the EAA canals are due to the mixing of surface and ground water, as many of the canals in the EAA are dug into the limestone. Elevated specific conductance is historically attributed to Na + and Cl - ions that are sourced from saline water of the shallow aquifer. The levels of Na + and Cl - concentrations and the Na/Cl ratio found in the EAA canals are a strong indication that the source of these ions is ancient sea water. Because the ions that are contributing to increased specific conductance in the EAA are sourced from ancient saline water entrapped in the Everglades formation, the current management practices employed by EAA farmers to reduce P loads were found to have minor effects on abating specific conductance. Lower specific conductance trends were observed in three out of the ten farms monitored, and with no adverse impacts on the other farms with the exception of one. Drainage pumping and fertilizer application were the two major management practices that we investigated that may lead to high specific conductance. It was shown that although specific conductance is the highest during the high rainfall months of the year, the effect of drainage pumping is inconsistent and site-specific. Specific conductance was not related to the drainage pumping to rainfall ratio. For example, one farm that had the lowest drainage pumping to rainfall ratio had the highest specific conductance among the ten farms. Canal specific conductance is governed mainly by the Page 125

126 quality and the hydrology of the underlying shallow ground water, which is farm specific. Farm canal water appears also to be influenced to a lesser degree by the quality and the quantity of irrigation water it receives. The irrigation water that flows into the high specific conductance farms in this study was also characterized by high specific conductance. One of the potential sources of high specific conductance that was mentioned in the literature is fertilizer application. We investigated this claim and found this to be false. Using data from one of our mixed crops farm that had high specific conductance values, the contribution of KCl fertilization to the total TDS export was negligible. This contribution was less than 6.5% of the total TDS exported in the drainage water assuming that all the KCl fertilizer applied ended up in the drainage water. This of course will never be the case as EAA crops take up large amounts of K + and Cl -. Sugarcane, for example, removes about 40 Kg Cl /acre (90 lb Cl/acre) at harvest. Chloride is an essential element for crops. Keeping the BMP definition in mind, it is the conclusion of this study that no further BMPs can be identified by additional research that would provide abatement of specific conductance for farm discharge waters of the EAA. The currently employed P load reduction BMPs have reduced specific conductance slightly at three out of the ten farms, but we conclude that further BMPs that target specific conductance will not be effective or practical. The issue of specific conductance is a geological problem in the EAA and additional farm management practices will have minimal effect on specific conductance. Conclusions Specific conductance was monitored at ten representative farms in the EAA and ion composition was determined in grab samples from eight of the ten farms. The specific conductance monitoring started in 1997 and continues through 2004 for three farms. The sampling for ion composition was conducted on a weekly basis in 2001 and Historical data on specific conductance in South Florida waters, including the EAA, showed wide variability. Data obtained from the SFWMD showed high specific conductance in EAA wells. Main canals in the EAA showed variable and often high specific conductance levels. The factors reported in the literature that affect conductance in the EAA conveyance canals included groundwater composition, agricultural farm drainage, active dissolution of the underlying bedrock, and the effects of Lake Okeechobee water. Saline waters left by ancient sea invasions are present in the underlying formation of the Everglades. Page 126

127 Specific conductance was not an issue in the majority of the EAA farm canals monitored. Out of the ten farms that were monitored, only two had average specific conductance higher than ms/cm. Higher levels of Na + and Cl - were observed at the two elevated specific 2- conductance farms. Of the two farms, one also contained higher concentrations of SO 4 ions. Ion composition data confirmed that the ions determining specific conductance in the EAA waters are Na +, Ca 2+, HCO - 3, Cl - and SO 2-4. The Na/Cl ratio in most of the EAA canals approximated the ratio of seawater (0.55). This was a strong indication that the source of the salts is from connate seawater entrapped in ground water and not from agrochemicals. Relationships between specific conductance, geographical influences, and management practices were explored. Shallow ground water quality plays a major role in the elevated specific conductance in some of the EAA farm canals. Historical data showed elevated concentration of Cl - and high specific conductance in shallow ground water in the EAA. One farm with an elevated specific conductance level, UF92008A, is located in an area of shallow wells of reported Cl - concentration range of mg /L. This farm is also influenced by the Hillsboro canal, which historically has had high conductance and high levels of Na + and Cl -. The second farm with high specific conductance level, UF9206A&B, is also in an area of shallow wells that have had concentrations of up to 400 mg/l of Cl - reported. The irrigation water flowing into these two farms (UF9206A&B and UF9208A) is characterized by higher specific conductance. The effect of drainage pumping on specific conductance was also investigated. There was a weak correlation between specific conductance and pump hours. This drainage pumping effect was variable and site specific. For example, at UF9206A&B drainage pumping did not affect specific conductance. Drainage event analysis on the two elevated specific conductance farms demonstrated the variable effect of drainage pumping and strengthened the conclusion that specific conductance in farm canals is strongly influenced by the underlying ground water hydrology and quality. In addition, farm drainage volume pumped to rainfall ratio had little or no effect on farm canal specific conductance. So, the current P load reduction BMP of delaying pumping until an accumulated ½ to 1 of rain has fallen may help or have no influence on mitigating increased conductance levels in the EAA farm canals. Irrigation had a weak negative correlation with specific conductance. On the three intensely monitored farms, UF9200A, UF9206A&B, and UF9209A, irrigation had the effect of Page 127

128 decreasing specific conductance. Sites that received irrigation water directly from low specific conductance district canals (Miami and North New River canals) had lower mean specific conductance values. Sites that received irrigation water from district canals with relatively higher specific conductance (Ocean and Hillsboro canals) had relatively higher mean specific conductance values. It was also evident that the sites that did not receive irrigation water directly from main conveyance canals, but received irrigation water via secondary or branch canals had relatively higher mean specific conductance values. The literature suggests that the effect of fertilization is negligible on the specific conductance in EAA farm canals. It has also been reported in the literature that the sugarcane crop takes up more P and K than applied by ferilizers. In our analysis, Cl - from KCl fertilization contributed less than 6.5% of the annual drainage TDS loading from one farm with elevated specific conductance assuming the highly unlikely scenario that all of the KCl fertilization ends up in the drainage water. Crops take up a lot of Cl - and in the case of sugarcane, an average of 90 lb of Cl - per acre is taken up. Yearly trend analysis, conducted on each site over the monitoring period of each site, showed a decreasing trend of specific conductance in three farms and an upward trend in one farm. The full implementation of farm level BMPs in the EAA in 1995 may have contributed to the improvement of water quality by reducing specific conductance in canal water at these three farms (UF9202A, UF9205A and UF9207B). No further BMPs were tested as it was concluded that specific conductance is not a farm management problem, but a geological one. It is the conclusion of this study that no further BMPs can be identified with additional research. Ground water in the EAA is highly mineralized and has high specific conductance values. As many of the canals are dug into the limestone, mixing of surface and ground water occurs and leads to increases of specific conductance. Specific conductance is attributed to certain ions that are sourced from saline water of the shallow aquifer. Drainage pumping could be a factor leading to the high specific conductance in some of the farm canals in the EAA depending on location and the specific conductance of the ground water wells. The currently employed water management BMPs are mitigating the problem as much as can be accomplished given the geological presence of a highly mineralized shallow aquifer. Page 128

129 References Anderson, D. L., and J. E. Bowen Sugarcane Nutrition. Potash and Phosphate Institute, Atlanta, Georgia; Potash and Phosphate Institute of Canada, Saskatoon, Saskatchewan; Foundation of Agronomic Research, Atlanta, Georgia. Antonopoulos, V. Z., D. M. Papamichail, and K. A. Mitsiou Statistical trend analysis of water quality and quantity data for the Strymon River in Greece. Hydrology and Earth System Sciences 5: Chen, M Quality manual for the laboratory of the Everglades Research and Education Center (NELAC Certification #E76463). University of Florida, Belle Glade, FL. CH2M Hill, Water quality studies in the Everglades Agricultural Area of Florida. Engineering report submitted to the Florida Sugar Cane League. CH2M Hill, Gainesville, FL. Coale, F.J., Sanchez, C. A., F. T. Izuno, and A. B. Bottcher Nutrient accumulation and removal by sugarcane grown on Everglades histosols. Agronomy Journal 85: Corps of Engineers Water quality study Everglades National Park. Department of Army, Jacksonville, FL. Flora, M. D. and P. C. Rosendahl Specific conductance and ionic characteristics of the Shark River Slough. National Park Service. Report T-615. State of Florida Florida Statute Section Amending the Marjory Stoneman Douglas Everglades Protection Act. The Everglades Forever Act. Tallahassee, FL. Gilbert, O. R Statistical methods for environmental pollution monitoring. John Wiley & Sons, Inc. NW. Gleason, P.J Chemical quality of water in conservation area 2A and associated Canals. Central and Southern Florida Flood Control District Tech. Pub pp. Harvey, J.W Groundwater-surface water interactions and relation to water quality in the everglades. Hem, J. D., 1985, Study and interpretation of the chemical characteristics of natural water (3d ed.): U.S. Geological Survey Water-Supply Paper Washington, DC. Hittle, C. D Delineation of saltwater intrusion in the surficial aquifer system in eastern Palm Beach, Martin, and St. Lucie Counties, Florida, U.S. Geological Survey, Water-resources Investigations Report Jones, L. A Soils, geology, and water control in the Everglades region. University of Florida Agricultural Experimental Station, Bulletin 442, March, 1948, Gainesville, FL. 168 pp. Lietz, A.C Assessment of water quality in the South Indian River Water Control District, Palm Beach, Florida, U.S. Geological Survey Open-File Report, Washington, DC. Page 129

130 Lietz, A.C Analysis of water-quality trends at two discharge stations - one within Big Cypress National Preserve and one near Biscayne Bay - southern Florida, U.S. Geological Survey WRIR Washington, DC. McKenzie, D.J Water quality of the Boca Raton canal system and effects of the Hillsboro canal inflow, southeastern Florida, U.S. Geological Survey WRIR Tallahassee, FL. Mengel, K., and E. A. Kirkby. Principles of Plant Nutrition. 5 th edition. Kluwer Academic Publishers, the Netherlands. Miller, W. L. and A. C. Lietz Quality of water data, Palm Beach County, Florida U.S. Geological Survey, Open File Report Tallahassee, FL Parker, G.G., G.E. Ferguson, and S.K. Love Water resources of southeastern Florida with special reference to the geology and groundwater of the Miami area. U.S. Geological Survey Water Supply Paper Risey, J. C., and M.C. Doyle Water temperature, specific-conductance, and meteorological data for the Rualatin River basin, Oregon, USGS Open- File Report Portland, OR. SAS Institute, Version 8. SAS Institute, Inc. Cary, NC. Scot, W. B Hydraulic conductivity and water quality of the shallow aquifer, Palm Beach County, Florida. U.S. Geological Survey, Water-Resource Investigations Stumm, W. and J. J. Morgan Aquatic chemistry. An Introduction emphasizing chemical equilibria in natural waters. 2 nd ed. John Wiley & Sons, Inc. New York. USEPA (U.S. Environmental Protection Agency) Volunteer stream monitoring: a methods manual. EPA 841-B USEPA Office of Water. Washington, DC USEPA (U.S. Environmental Protection Agency) Methods for chemical analysis for water and wastes. USEPA, Cincinnati, OH. Virginia DEQ (Department of Environmental Quality), The quality of Virginia non-tidal streams: first year report. Technical Bulletin, WQA/ Virginia DEQ. Richmond, VI. Waller, B.G. and J.E. Earle Chemical and biological quality of water in part of the Everglades. Southeastern Florida Water Resource Investigations 56-75, USGS, Tallahassee, FL. Page 130

131 CHAPTER 3 On-Farm Particulate Phosphorus Measurement and Control List of Figures Chapter 3 Ch. 3 Figure 1. Aquatic weed coverage of an EAA drainage canal Ch. 3 Figure 2. Cleaning program to keep weeds away from the pump station Ch. 3 Figure 3. Mass fraction of typical particulate phosphorus sources Ch. 3 Figure 4. Effect of velocity on erosion rates Ch. 3 Figure 5. Flow, canal level, and Total Suspended Solids profiles for Year 2000 Event 276 at UF 9200A Ch. 3 Figure 6. Velocity and Particulate Phosphorus profiles for Year 2000 Event 276 at UF9200A Ch. 3 Figure 7. Location of study farms in the EAA Ch. 3 Figure 8. A) Annual unit area particulate P loads, and B) annual unit area dissolved P loads from 2000 to Ch. 3 Figure 9. Annual drainage pumping from 2000 to Ch. 3 Figure 10. Annual pumping to rainfall ratios from 2000 to Ch. 3 Figure 11. Equivalent TSS concentrations from 2000 to Ch. 3 Figure 12. Equivalent particulate P concentration from 2000 to Ch. 3 Figure 13. Average P content of TSS from 2000 through Ch. 3 Figure 14. Cumulative hydraulic and particulate P load distributions for UF9200A Ch. 3 Figure 15. Cumulative hydraulic and particulate P load distributions for UF9206A Ch. 3 Figure 16. Cumulative hydraulic and particulate P load distributions for UF9206B Ch. 3 Figure 17. Cumulative hydraulic and particulate P load distributions for UF9209A Ch. 3 Figure 18. Percentage point distribution of contributing processes for top 50% of particulate P load for A) UF9200A, B) UF9209A, C) UF9206A, and D) UF9206B Ch. 3 Figure 19. Percentage point distribution of consolidated processes for top 50% of particulate P load for A) UF9200A, B) UF9209A, C) UF9206A, and D) UF9206B Ch. 3 Figure 20. Hydraulic history for event UF9200A (August 2, 2001) Ch. 3 Figure 21. Profiles for event UF9200A (June 21, 2002) Ch. 3 Figure 22. Profiles for event UF9200A (November 5, 2003) Ch. 3 Figure 23. Profiles for event UF9206B (October 2, 2000) Ch. 3 Figure 24. Profiles for event UF9206B (September 26, 2001) Ch. 3 Figure 25. Profiles for event UF9206B (February 10, 2002) Page 131

132 Ch. 3 Figure 26. Profiles for event UF9206B (August 25, 2002) Ch. 3 Figure 27. Profiles for event UF9206B (March 27, 2003) Ch. 3 Figure 28. Selected pumping profile for UF9209A for year Ch. 3 Figure 29. Profiles for event UF9209A (September 8, 2001) Ch. 3 Figure 30. Selected pumping profile for UF9209A for year Ch. 3 Figure 31. Profiles for event UF9209A (August 29, 2002) Ch. 3 Figure 32. Profiles for event UF9209A (September 30, 2002) Ch. 3 Figure 33. Profiles for event UF9209A (June 19, 2003) Ch. 3 Figure 34. Particle size distributions for bulk samples from UF9200A Ch. 3 Figure 35. Particle size distributions during various time periods Ch. 3 Figure 36 Time sequence of peaks of particle dize distribution curves Ch. 3 Figure 37. Settling velocity distributions Stokes Law vs. settling column Ch. 3 Figure 38. Average sediment depth over time Ch. 3 Figure 39. Sediment P content vs. time and depth at A) UF9200A, B) UF9206B, and C) UF9209A Ch. 3 Figure 40. Wet sediment P concentration vs. depth Ch. 3 Figure 41. Normalized top 12 cm sediment P content variation with time Ch. 3 Figure 42. Sediment P concentration in top 12 cm vs. location in farm canal Ch. 3 Figure 43. Phosphorus content distribution for top 50% loads UF9200A Ch. 3 Figure 44. Correlation of pumping-to-tainfall ratio with P content Ch. 3 Figure 45. Annual average particulate P vs. annual average velocity List of Tables Chapter 3 Ch. 3 Table 1. Hydraulic event statistics for UF9200A Ch. 3 Table 2. Physical-chemical event statistics for UF9200A Ch. 3 Table 3. Hydraulic event statistics for UF9206A Ch. 3 Table 4. Physical-chemical event statistics for UF9206A Ch. 3 Table 5. Hydraulic Event statistics for UF9206B Ch. 3 Table 6. Physical-Chemical event statistics for UF9206B Ch. 3 Table 7. Hydraulic event statistics for UF9209A Ch. 3 Table 8. Physical-chemical event statistics for UF9209A Ch. 3 Table 9. Particulate P contributions to P loads Ch. 3 Table 10. Annual averages of key parameters from study farms Ch. 3 Table 11. Summary of annual rainfall and pumping data Ch. 3 Table 12. Distribution of processes contributing to top 50% of measured loads at study farms Ch. 3 Table 13. Percentage point distributions of top 50% load events Page 132

133 Ch. 3 Table 14. Selected physical and chemical characteristics of concentrated suspended solids samples UF9200A Ch. 3 Table 15. Selected physical and chemical characteristics of concentrated suspended solids samples UF9206B Ch. 3 Table 16. Selected physical and chemical characteristics of concentrated suspended solids samples UF9209A Ch. 3 Table 17. Estimated average sediment P inventories Ch. 3 Table 18. Average velocity and canal depths of study farms Page 133

134 Introduction The Everglades Forever Act of 1994 mandated a research and monitoring program on the evaluation of water quality standards in the Everglade Agricultural Area (EAA) (Chapter 40E-63). The goal of this research was to evaluate the constituents that have been previously identified as elements of water quality concern that will likely not be significantly improved by the Stormwater Treatment Areas and current Best Management Practices (BMPs) being widely implemented throughout the EAA; and to identify strategies needed to address such parameters (40E (2)). These parameters were identified by the Florida Department of Environmental Protection (FDEP) as specific conductance, particulate phosphorus, and the pesticides Atrazine and Ametryn. This report deals with the issue of particulate P in the EAA. The Everglades Agricultural Area-Environmental Protection District (EAA-EPD) and the South Florida Water Management District (SFWMD) are responsible for the monitoring of Atrazine and Ametryn. The objective of this work as stated by Chapter 40E-63, Part III: In recognition that substantial particulate matter such as sediments are being discharged from farms, given that published University of Florida Institute of Food and Agricultural Sciences data has demonstrated that particulate phosphorus constitutes a significant portion of total phosphorus, the farm-scale research shall be expanded to include the development, testing, and implementation of BMPs for reducing discharge of particulate phosphorus (i.e. sedimentation basins). Phosphorus (P) transport in runoff can occur in soluble and particulate forms. Dissolved P is comprised mostly of orthophosphates, which are immediately available for algal uptake (Sharpley et al., 1992). Particulate P consists of all solid phase forms including P sorbed by soil particles and organic material transported during runoff. The primary objective for studying particulate P transport is the potential for developing or modifying management practices to affect significant reductions in P export from EAA farms. In general, particulate P constitutes the major portion of the (75 to 90%) P transported in runoff from conventional tilled land (Sharpley et al., 1987; Sharpley et al., 1993). Studies in the EAA have shown that a great portion of total phosphorus (TP) loading in drainage waters is in the particulate form (Izuno and Bottcher, 1991). Izuno and Rice (1999) reported that particulate P accounted for 20% to 70% of the TP exported from EAA farms, and that particulate P export was frequently the cause of spikes in TP loads. Page 134

135 Biological Contribution Mechanism Stuck et al. (2001) studied farm-scale particulate P transport at an EAA test farm and proposed a supplementary mechanism for particulate P export from EAA farms that possibly explained the difficulty of trapping recently deposited light flocculant organic sediments. They showed that the mass fraction of P in exported suspended solids was frequently much higher than that of farm soil or field litter, and that the chemical characteristics of the exported suspended solids often more closely resembled those of aquatic flora than those of farm soil. They concluded that a significant fraction of particulate P in the EAA originates from in-stream biological growth rather than from soil erosion, and proposed that a major contributing factor to particulate P discharge is the Biological Contribution Mechanism. Sediment that contributes significantly to P export was postulated to be, for the most part, recently deposited biological material such as settled plankton, filamentous algae, and macrophyte detritus. For example, Reddy and DeBusk (1991) showed that detrital production from Eichhornia crassipes (water hyacinths) could be as high as 15 gr/day/m 2 of hyacinth mat. Exported solids may also be contributed directly by floating or suspended plants when loosely bound material is detached by turbulent shear forces. The root structure of aquatic plants such as Pistia stratiotes (water lettuce) can account for a large fraction of the total plant biomass, and may also provide an ideal location for growth of attached microorganisms (epiphytic growth). Engle and Melack (1990) studied mats of mixed aquatic weeds and found epiphyton concentrations as high as 146 gr/m 2 of mat. They also found that up to 70% of the attached epiphyton could be detached by wind-driven movement of the mats, and that the population remaining on the roots regenerated itself completely with in 1-2 weeks. Studies by Stuck (1996) showed that from 29% to 38% of the total mass of Pistia stratiotes could be dislodged by vigorous agitation. In addition to floating aquatic weeds, contributions to particulate P are made by submerged aquatic plants and planktonic growth. The filamentous algae lyngbya is found in water systems throughout Florida. The conditions in the EAA canals of high ph and high temperature are favorable for the growth of lyngbya. Tubea et al. (1981) showed that lyngbya populations could exhibit doubling times of 0.8 to 2.0 days in favorable conditions. Stuck (1996) found that the field ditch surficial sediments in a representative EAA sugarcane farm contained approximately 15% by mass of readily identifiable lyngbya detritus. Page 135

136 The Biological Contribution Mechanism includes sediment erosion as a source of exported particulate P, but it modifies the character of the sediment that contributes particulate P, allowing that sediment to consist of a heterogeneous mixture of organic matter in various stages of decomposition, with various levels of P content and variable transport properties. The mechanism has been supported on the farm scale by the evidence of large differences between the physical and chemical characteristics of farm soils, farm sediments, and exported particulate matter. Specifically, the exported particulate matter frequently had characteristics that were more akin to viable plant matter than to farm soils or most of the farm sediments. Figure 1 shows a scene of total macrophyte coverage in an EAA farm ditch that is representative of macrophyte coverage potential when weed control is not practiced. Aerial reconnaissance studies have shown that the aquatic weed coverage may average as much as 50% of the total drainage conveyance area on a farm that does not practice weed control. On the farm that practiced consistent weed control, the coverage still averaged more than 20% (Daroub et al., 2003). However, many growers in the EAA recognize the problem and implement cleaning programs to prevent weeds to accumulate close to the pump station. Figure 2 shows an example of a typical pump station in the EAA where the grower implements a good cleaning program to prevent the accumulation of aquatic weeds in the vicinity of the pump intake. The current hypothesis is that particulate P is sourced from farm canals rather than from overland flow erosion. Soluble P may be converted to insoluble plant matter, and vice versa, depending on the physiological state of the biota in the canals. Processes inside the canal allow for the immobilization and remobilization of sediments, depending on inter-event times and hydrodynamic conditions. The location of P inventories may change within the canal system as the biological population changes and as collections of biomass change their positions because of flow or wind patterns. The development of the biological contribution mechanism has included the use of the P content of a particulate mass (expressed as a P mass fraction, mg P/kg total dry mass) as an approximate biomarker to estimate the source of the particulate matter. Figure 3 shows a typical P mass fraction range of a number of potential particulate P sources. Soil in the EAA typically has P mass fraction in the range of mg/kg (Fiskell and Nicholson, 1986; Stuck, 1996). Several recent studies have shown that the base sediments in EAA canals typically have P mass fraction in the range of mg/kg (Stuck, 1996; Izuno et al., 1998). Detritus from the floating water-weeds is in the range of mg/kg (Stuck, Page 136

137 1996), while the plants themselves may have a P mass fraction in the range of mg/kg (Stuck, 1996). Previous studies in the EAA have reported that the average floating aquatic plant P content of several farms over a two-year period averaged about 4200 mg/kg. The P mass fraction of planktonic growth may be in the range of mg/kg or higher (Behrendt, 1990; Atkinson, 1991). Figure 3 shows this range graphically. Ch. 3 Figure 1. Aquatic weed coverage of an EAA drainage canal. Page 137

138 Ch. 3 Figure 2. Cleaning program to keep weeds away from the pump station. P Mass Fraction of Various Constituents (Approxim ate-for Illustration) P Mass Fraction (mg/kg) [Soil] [Sediment] [Plant Detritus] [Macrophytes] [Plankton] Ch. 3 Figure 3. Mass fraction of typical particulate phosphorus sources. Page 138

139 It must be stressed that Figure 3 represents typical historical ranges, and that local environmental conditions may give rise to results that may differ from this illustration. The natural result of having heterogeneity of sources is that the resulting collection, for example the surficial layer of farm canal sediment, may contain a diverse collection of particles with various ages, P content, particle size, and specific gravity. This gives rise to selective transport under various conditions, which will be illustrated in the next section. Particle Erosion and Transport The organic sediments of the EAA are similar to cohesive clay sediments in their erosion characteristics (Stuck, 1996). The behavior of cohesive sediments may be illustrated briefly in simplified form as follows. As water flows over material, its energy may cause some of the material to disengage and enter the flowing water. Resistance to this disengagement is referred to as shear strength. An idealized sediment bed will have a shear-strength and a yield-strength. Hydraulic stresses less than the bed shear strength will cause minimal erosion. As shear stress on the bed increases beyond the shear strength, erosion from the surface of the bed will proceed at a rate that is proportional to the excess of the shear stress compared to the shear strength. This is called the Bed Erosion Regime. At some point the shear stress will exceed the yield strength of the bed. At this point the forces on the bed exceed the cohesive forces holding the bed together and the bed starts to break up. As water velocity increases, the bed continues to break up more rapidly. Solids mobilization in this region is much greater than in the Bed Erosion Regime. This phase is called the Bed Transport Regime. In both regimes erosion rate is directly proportional to shear stress. It is extremely important to understand, however, that shear stress is proportional to the square of velocity. In the most simplified form the relationship between erosion rate and velocity in the Bed Erosion Regime is given by the equation ε = 2 K ( v b v c 2 ) (1) where ε = Erosion rate, mass/time/area K = Rate constant and conversion factor v b = Canal velocity Page 139

140 v c = Velocity at which shear stress equals bed shear strength (critical velocity) Figure 4 shows this relationship in qualitative form with arbitrary units. In this idealized case, no erosion takes place until the critical velocity of 0.05 units is reached. At that point, the erosion rate increases as a function of the difference between the square of the canal velocity and the square of the critical velocity. At a velocity of 0.32 units the yield strength of the bed is exceeded, the bed starts to disintegrate, and the regime shifts to Bed Transport. The bed transport region may be described by an equation similar to Equation 1, but with a different critical velocity, and a much higher erosion coefficient. Effect of Velocity on Erosion Rate (Qualitative Illustration w ith relative units) 40 Erosion Rate = Erosion Rate - mass/unit area/time Critical Velocity Erosion Rate = 13 5 Erosion Rate = Velocity - length/time Ch. 3 Figure 4. Effect of velocity on erosion rates. Page 140

141 Several important points may be made from this simplified illustration. First, at relatively low velocities, there is zero to minimal erosion and particle mobilization. Beyond the critical velocity, the particle-mobilization rate increases with the square of the velocity. In our example, the erosion rate at a velocity of 0.15 velocity units is 3 rate units. Doubling the velocity to 0.30 velocity units increases the erosion to 13 rate units, more than a four-fold increase in particle mobilization. An additional increase in velocity to 0.40 units causes the system to enter the Bed Transport Region. Here, where mobilization is even greater, the erosion rate increases to 36 units, almost a three-fold increase arising from a 25% increase in velocity. This example illustrates the significant changes that may occur within relatively narrow flow rates. In the EAA, pumping rates may easily be doubled or tripled by running multiple pumps or switching from small to large capacity pumps. Velocities may also change rapidly when canals are drawn down to low levels. These operating factors can have a significant impact on particle mobilization rates. The amount and physiological condition of the various types of organic matter present in a water conveyance system varies with time and flow. Biological growth incorporates soluble P into particulate matter, represented by the plant biomass. Plant death and decomposition releases some P as soluble P, and releases some plant biomass as mobile particulate matter. In stagnate conditions the particulate matter accumulates in place. In irrigated conditions, this matter may be transported upstream. With drainage conditions, particulates may be transported downstream and ultimately discharged. The response to all these conditions causes a continuing change in the amounts and locations of particulate P. Inter-event time (the time between pumping events) can have an influence on the amount and location of transportable organic material in EAA farm canals. The longer the inter-event time, the more time is available for biological growth and accumulation in the canal system. Stuck et al. (2001) and Izuno et al. (1998) showed that there could be a positive correlation between the length of the inter-event time and the amount of particulate mass available for transport at start up. Page 141

142 Primary Processes and Illustrative Examples This section defines the primary transport processes that occur in the farm canals. These primary processes have been discussed at length elsewhere (Stuck, 1996; Stuck et al., 2001; Daroub et al., 2003). For the purposes of this report, the processes are classed into several categories, and are described as follows: First Flush During the (relatively) quiescent period between pumping events biological material can grow and accumulate in the canals. This fresh material, along with solids that were suspended at the time of shutdown in the preceding event, can be readily suspended under the turbulent conditions that exist at pump start-up. This highly mobile material causes an increase in the concentration of suspended solids during the early periods of pump events. Eventually this highly mobile material flushes out and the process of erosion proceeds on the less mobile particulate matter in the canals. Cumulative High Velocity The normal erosion process at constant velocity produces (in the idealized case) a steadily increasing discharge concentration of suspended solids. The reason for this is that water farther upstream has a longer time to accumulate eroded suspended solids as it moves downstream to the discharge point. If there is a substantial increase in velocity, there will not necessarily be an immediate increase in suspended solids concentration, because of the lag time for the flowing water to accumulate additional suspended solids. There are often circumstances during pumping events when velocity may change significantly, such as when a larger pump is started up or when canal depth becomes shallow, significantly reducing cross sectional area available for flow. If there is a large volume of water in the canal, as is usually the case when a large pump is started up, the effects of this velocity increase are not seen until sometime later, so changes in concentration may be affected by cumulative high velocity. This process also proceeds after first flush, as long as velocity is sufficiently high. First flush mobilizes the material remaining close to the pump station from the previous event as well as the highly mobile new material that was produced in the inter-event period. After first flush, the sustained high velocity will continue to mobilize particulate P, the continued export of which will exhibit the lag just described. Restart Flush When pumping is terminated, suspended solids in the canal system settle out in place. If there had been a significant concentration of suspended solids in the downstream reaches of the canal system at shutdown, there will be an increased initial Page 142

143 concentration in the discharge when the pump is restarted. This is similar to First Flush, except that the time between pump shut down and restart is less than that for First Flush. In fact the break between First Flush and Restart Flush is somewhat arbitrary, in that an event is defined as the start of pumping after more than twenty-four hours of quiescence, so a pump start after twenty-three hours would give rise to a Restart Flush, whereas a pump start after twenty-five hours would give rise to First Flush. Because they are similar, both these phenomena are grouped into the category Start-Up Flush. Particulate Phosphorus Spike This is described by a somewhat arbitrary definition that if the particulate P concentration for a particular sample is more than twice that of either the preceding or succeeding samples, then a spike has occurred. The spike is assumed to originate from a random release of particulate material from upstream sources, such as a collection of floating macrophytes. By this definition, a sustained concentration increase and decrease that covers several consecutive samples would not be classified as a spike. Pump Cycling This category differs from pump restart in that the pump cycles through onoff oscillations over relatively short time periods, e.g. 30 minutes to two hours. This condition occurs when a farm pump is on automatic on-off control that is tied to canal level. One of three of the farms studied employs this control system. The other two rely on manual pump control for start up and shut down. Other Obviously this is the catchall category, but it is relatively narrow. Any sample that exhibits increased particulate P and is not explained by the other categories falls in the Other category. This category can include sustained spikes, which might arise from some upstream disturbance, such as starting a booster pump or unblocking a culvert in a rice field or any other special event. These categories are well illustrated by Year 2000 Event 276 that started at Station UF9200A on October 2, 2000 and lasted for over eight days (Figure 5). The grower typically pumps either with one large or one small electrical pump, both of which are under on-off level control. Prior to this event, he had pumped relatively infrequently, using the small pump for a total of 82 hours and the large pump for less than 2 hours over the preceding 83 days, so there was ample opportunity for biological material to accumulate in the canals. During the event he switched from the large pump to the small one, and back, several times. There were also several instances of pump oscillation under the on-off control mechanism. Page 143

144 The suspended solids, flow, and level profiles for this event are shown in Figure 5. Large pump operation is indicated by flow rates that start at 2.5 m 3 /s, small pump operation is indicated by flow rates that start at 0.5 m 3 /s. The effect of the inter-event buildup is clearly illustrated by the initial suspended solids surge, which rose to more than 3500 mg/l. The figure clearly shows two subsequent waves of suspended solids that are associated with the operation of the large pump, and illustrate the lag effect of prior high velocity. Figure 6 shows the velocity and particulate P profiles of this same event. Here the start-up flush of particulate P is evident at the start of the event. At Decimal Date (DD) there is a particulate P spike. At DD the large pump is started and at DD the effect of this increase in velocity begins to be seen. The particulate P concentration increases steadily until the pump goes into oscillation mode from the level control, at which point concentration starts to decrease because of the decreasing average velocity. Although the concentrations are decreasing, they are still appreciable, because there is still high velocity when the pump is running. The particulate P export that occurred during pump cycling was in fact a continuation of the cumulative high velocity in effect prior to the start of pump cycling because the suspended solids do not have sufficient time to settle and consolidate between pump cycles. At DD , when the large pump is restarted there is an immediate surge of particulate P, which is categorized as re-start flush. When the large pump went into cycle mode on DD , particulate P began to settle in the region just upstream of the pump. This final surge of particulate P is from immediate resuspension of the material that was settled out when the large pump went into cycle mode and then when flow was switched over to the small pump. Comparison of the shape of the TSS curve and the shape of the particulate P curve in the period from DD through DD shows that the particulate P surge led the suspended solids surge. This is an illustration that the light, flocculent, readily transported material, which is high in P content, does not necessarily move in the same way that the bulk of the suspended solids move. This example of categorization is generally representative of what is seen at the study farm UF9200A. The other farms operate under different hydraulic control schemes, and at times present more complicated and extreme conditions, particularly when the canals are drawn down to very low levels. The basic categorization technique, however, applies reasonably well to the major contributing events for all study farms. Page 144

145 400 Event UF9200A-276 Year 2000 Flow, Level, and TSS TSS Flow 2.5 Total Suspended Solids (mg/l) Canal Level Flow (m 3 /s), Level (m) Decimal Date 0.0 Ch. 3 Figure 5. Flow, canal level, and Total Suspended Solids profiles for Year 2000 Event 276 at UF 9200A. Page 145

146 1.0 Event UF9200A-276 Year 2000 Particulate Phosphorus and Velocity Start-Up Flush PP Velocity 0.40 Particulate Phosphorus Concentration (mg/l) Spike Re-Start Flush Cumulative High Velocity Re-Start Flush Velocity (m/s) 0.1 Pump Cycling Decimal Date 0.00 Ch. 3 Figure 6. Velocity and Particulate Phosphorus profiles for Year 2000 Event 276 at UF9200A. Page 146

147 Material and Methods Intensive-Study Program Elements Event Data Analysis The four pump stations on the three study farms are monitored continuously for pumping events. At pump start-up sampling of the pump inlet water is started. Each pump station is equipped with 3700 portable ISCO automatic samplers that collect water samples every 15 minutes. Four consecutive 15-minute samples are composited into one-hour discrete samples for analysis. Sample caddies are collected every 24 hours for the duration of the pumping event. Water samples are placed in a water cooler with ice and immediately transported to the EREC Water Quality Laboratory for analysis. All samples are analyzed for total suspended solids (TSS), total phosphorus (TP), and total dissolved phosphorus (TDP). Particulate P is calculated as the difference between TP and TDP. Water samples for TDP analysis are immediately filtered though a 0.45 µm filter-membrane, samples for TP analysis are not filtered. Analysis for TP and TDP are performed using the mercury oxide digestion method (Method 365.4, EPA 1993). Total suspended solids analysis is done following SOP No 13 from the EREC (EREC-SOP, 2002) and Method (EPA, 1993). The pump stations are instrumented and monitored continuously for, among other variables, rainfall, pump flow rates, and inlet and outlet water levels. The monitored data are downloaded twice daily via telemetry to a central data processing location. Each pump station also has a Hydrolab Datasonde located in close proximity to the water sampling point. The Datasonde monitors a number of variables, including water temperature, ph, conductivity, and turbidity. Data from the Datasondes are downloaded manually on a weekly basis. The illustrations presented in Figures 4-6 of the introduction of this chapter were intended to illustrate the non-linearity of system responses in particulate P transport and the inherent potential for time displacement between cause and effect. The results presented in the particulate P load distribution analysis section confirm this and imply added complexity with year-to-year changes in system responses. The most thorough way to study a non-linear system is with the use of a mathematical system model. A model has been developed for particulate P transport and has been Page 147

148 applied to detailed transport from one EAA farm over a one-year period (Stuck, 1996). This model, however, requires extensive contour mapping of the hydraulic system of a farm, and is dependent on empirically derived parameters that may change from farm to farm, and from year to year. In its present state of development it is not applicable to the analysis of data without extensive calibration to each farm of interest. The approach that has been adopted here is to conduct various forms of cluster analysis to attempt to identify primary parameters that have had the most impact on particulate P transport at the study farms. The analyses performed are: Load Distribution Analysis Divides all events into smaller sub-events, typically onehour increments, and determines the distribution of particulate P loads in comparison to the distribution of hydraulic loads in these sub-events. Process Distribution Analysis Determines the most probable mechanism for particulate P transport in the sub-events that contribute most to annual loads, i.e. those in the top 50% of the load distribution. Event Analysis Evaluates the events that contain the highest number of high contribution sub-events for defining characteristics. Farm System Synthesis Description of the characteristics of each farm that caused the high contribution sub-events. Farm Sediment Surveys This element was intended to generate a reliable database on the character and inventory of sediments in the main canals of each of the study farms, and to determine if seasonal changes could be detected in canal sediment. Quarterly inventories were conducted of canal sediment volume, mass, and P content at each farm. A number of transect locations were set up at regular intervals upstream of the pump station at each farm. Canal sediment surface elevation and depth was determined at each location using a neutrally buoyant disc pad on the end of a calibrated pole. The pad would rest on top of the sediment and give the depth to the top of the sediment. Sediment depth was determined using a calibrated penetrometer probe that was driven through the sediment to the canal bottom. Both depths were referenced to the then-current canal water surface elevation, and then referenced to mean sea level by the then-current pump station staff gage reading. At the same time, core samples of the sediment were taken at each transect. These core samples were then sectioned and analyzed for key physical and chemical parameters, Page 148

149 including bulk density, solids content, particle specific gravity, organic matter content, and P content. Concentrated Discharge Sampling This element was intended to provide an adequate supply of discharged suspended solids to allow the determination of physical properties of the actual suspended solids. Periodically, large volume ( liter) composite samples were taken at each of the study farms. These samples were concentrated by sedimentation in the field. The sedimented solids were collected and further concentrated in the lab, after which they were analyzed for the same physical and chemical properties as the farm sediments, including bulk density, solids content, particle specific gravity, volatile organic matter content, and P content. Selected samples were analyzed for particle size distribution and settling velocity distribution. N Lake Okeechobee West Palm Beach Canal UF9200A UF9209A Bolles Canal EAA UF9206A&B Hillsboro Canal STA-1E STA-1W WCA-1 STA-5 Miami Canal STA-6 North New River Canal STA-3/4 WCA-3 STA-2 WCA-2 Ch. 3 Figure 7. Location of study farms in the EAA. Page 149

150 Results Event Summary Statistics Tables 1-8 detail the event summary statistics for the four pump stations from 2000 to The sampling program did not start until the middle of year 2000, but event hydraulic statistics are included for all of 2000 for the stations at farms UF9200A and UF9206A and B in order to follow annual pumping volumes for each farm. The data sets discussed in this report cover four wet seasons and three and half calendar years for UF9200A and UF9206A and B and three wet seasons and three and quarter calendar years for UF9209A. Pump station at farm UF9209A has followed a pumping scheme that is more regular than the other two farms. Its pumping events tend to be more frequent, of shorter duration, with shorter inter-event times. Thus it has more pumping events than the other two study farms. Data collected during the last four years is considered to be adequate to draw preliminary inferences regarding particulate P transport in the EAA. Unless otherwise noted, subsequent analysis will refer only to those flow events or part of events where samples were successfully collected, with no attempt to estimate the parameters associated with missing samples. The data presented in Tables 1-8 includes equivalent concentrations and P content for each event and each year. The equivalent concentrations are calculated numbers and represent the total sampled mass of the component of interest. e.g. total suspended solids, divided by the total pumping volume (during sampling) of the event or year. They represent a characteristic or mass average concentration of the sampled portions of the event or year. Similarly, the P content is calculated as the total sampled mass of particulate P divided by the total sampled mass of suspended solids, and represent the mass average P content for the sampled portion of each event or year. The characteristic concentrations may be used to estimate the total annual loads, compensating for un-sampled periods. Summary of key elements of the data sets in Tables 1 8 are presented in Tables 9-11 and Figures Table 9 summarizes the annual contributions from the particulate P loads to the TP loads for each farm during the last four years. Values from UF9200A increased from 47% in 2000 to 56% in However in 2003, particulate P contribution to TP loads dropped to 28%. At UF9209A the contribution from particulate P to the TP load was almost constant, 67% in 2001, 68% in But in 2003, UF9209A pumped their canals lower and longer than previous years, resulting in more sediments being scoured from bottom of the canal and Page 150

151 transported out of the farm, ensuing in a particulate P contribution of 80% to the TP load. The particulate P load contributions of UF9206A increased from 26% in year 2000 to 36% in years 2001 and 2002, and decreased to 27% in Particulate load contributions from farm UF9206B decreased from 40% in 2000 to an average contribution of 36% during the last three years. Table 10 summarizes the annual average data for equivalent concentrations, estimated loads, and suspended solids P content for each farm. Page 151

152 Ch. 3 Table 1. Hydraulic event statistics for UF9200A. Start Date Event Number Interevent Time (days) Start Decimal Date Finish Decimal Date Duration (hrs) Volume Pumped (m3) Cumulative Volume (m3) Volume Sampled (m3) 01/05/ A /05/ :20 01/05/ : ,605 17,605 * 01/17/ A /17/ :06 01/17/ :35 6 7,603 25,208 * 02/09/ A /08/ :30 02/09/ : , ,566 * 03/ 20/ A /20/ :10 03/21/ : , ,704 * 03/31/ A /30/ :54 03/30/ : , ,718 * 04/13/ A /13/ :20 04/19/ : , ,047 * 07/03/ A /03/ :05 07/04/ : , ,969 57,923 07/08/ A /08/ :39 07/11/ : , , ,682 09/20/ A /19/ :40 09/22/ : , , ,432 09/25/ A /25/ :05 09/26/ : , ,939 46,957 10/02/ A /02/ :55 10/11/ : ,581 1,408, ,306 Total or Avg. 813,300 92% 07/12/ A /11/ :20 07/12/ : ,256 87,256 82,334 07/15/ A /14/ :00 07/19/ : , , ,271 07/23/ A /23/ :00 07/26/ : ,633 1,127, ,283 08/02/ A /02/ :15 08/06/ : ,974 1,772, ,759 08/08/ A /08/ :00 08/08/ : ,873 1,816,110 ** 09/09/ A /09/ :30 09/11/ : ,324 1,916,434 ** 09/13/ A /13/ :25 09/15/ : ,602 2,083,036 ** 09/27/ A /27/ :40 10/02/ : ,544 2,600, ,837 10/25/ A /25/ :35 10/25/ : ,180 2,639,760 38,825 11/05/ A /05/ :20 11/08/ : ,771 2,753,531 89,456 Total or Avg. 1,762,765 64% 02/11/ A /11/ :10 02/18/ : , , ,547 02/25/ A /25/ :50 02/25/ : , ,989 21,073 03/04/ A /04/ :15 03/04/ : , ,479 43,073 06/22/ A /21/ :45 07/05/ : ,248,944 1,655, ,525 07/08/ A /08/ :20 07/16/ : ,333 2,210, ,122 09/06/ A /06/ :05 09/07/ : ,374 2,288,130 75,913 10/12/ A /12/ :25 10/13/ : ,464 2,332,594 43,401 10/16/ A /16/ :20 10/17/ : ,369 2,371,963 36,452 10/26/ A /25/ :35 10/28/ : ,406 2,493,369 90,502 11/18/ A /17/ :00 11/22/ : ,387 2,651,756 99,087 Total or Avg. 1,831,694 82% 03/17/ A /17/ :00 03/17/ : ,476 42,476 42,476 04/26/ A /26/ :00 05/02/ : , , ,770 05/28/ A /28/ :00 05/30/ : , , ,781 06/20/ A /20/ :00 06/24/ : , , ,503 07/23/ A /23/ :00 07/28/ : , , ,783 07/31/ A /31/ :00 08/02/ : ,440 1,094, ,079 08/05/ A /04/ :00 08/21/ : ,530,276 2,624, ,712 08/26/ A /26/ :00 08/27/ : ,070 2,652,449 ** 08/29/ A /29/ :00 08/30/ : ,819 2,702,268 49,819 09/04/ A /04/ :00 09/09/ : ,008 2,851, ,635 09/29/ A /29/ :00 10/01/ : ,600 2,929,876 78,600 11/05/ A /05/ :00 11/11/ : ,151 3,046, ,151 12/15/ A /14/ :00 12/18/ : ,007 3,066,033 20,007 Total or Avg. 2,533,606 83% * Events prior to installation of sample stations. ** Malfunction of discrete sampling system No samples taken. Page 152

153 Ch. 3 Table 2. Physical-chemical event statistics for UF9200A. Start Date Event Number TSS Load Sampled (kg) TP Load Sampled (kg) TDP Load Sampled (kg) PP Load Sampled (kg) TSS Equiv Conc (ppm) TP Equiv Conc (ppb) TDP Equiv Conc (ppb) PP Equiv Conc (ppb) % PP P Content of TSS (mg/kg) 01/05/ A * * * * * * * * * * 01/17/ A * * * * * * * * * * 02/09/ A * * * * * * * * * * 03/20/ A * * * * * * * * * * 03/31/ A * * * * * * * * * * 04/13/ A * * * * * * * * * * 07/03/ A % /08/ A , % /20/ A % /25/ A , % /02/ A , % 1762 Total or Avg. 45, % /12/ A , % /15/ A , % /23/ A , % /02/ A , % /08/ A ** ** ** ** ** ** ** ** ** ** 09/09/ A ** ** ** ** ** ** ** ** ** ** 09/13/ A ** ** ** ** ** ** ** ** ** ** 09/27/ A , % /25/ A , % /05/ A , % 3356 Total or Avg. 46, % /11/ A , % /25/ A % /04/ A % /22/ A , % /08/ A , % /06/ A , % /12/ A , % /16/ A , % /26/ A , % /18/ A , % 2117 Total or Avg. 81, % /17/ A % /26/ A , % /28/ A % /20/ A , % /23/ A , % /31/ A , % /05/ A , % /26/ A ** ** ** ** ** ** ** ** ** ** 08/29/ A % /04/ A , % /29/ A , % /05/ A , % /15/ A % 4713 Total or Avg. 103, % 1311 * Events prior to installation of sample stations. ** Malfunction of discrete sampling systems No samples taken. Page 153

154 Ch. 3 Table 3. Hydraulic event statistics for UF9206A. Start Date Event Number Interevent Time (days) Start Decimal Date Finish Decimal Date Duration (hrs) Volume Pumped (m3) Cumulative Volume (m3) Volume Sampled (m3) 01/25/ A /25/ :35 01/25/ : ,266 31,266 * 02/09/ A /09/ :55 02/09/ : ,128 54,394 * 03/20/ A /19/ :05 03/22/ : , ,221 * 04/14/ A /13/ :26 04/17/ : , ,305 * 05/09/ A /08/ :40 05/09/ : , ,435 * 05/17/ A /16/ :30 05/17/ : , ,774 * 07/06/ A /05/ :25 07/05/ : , ,464 * 07/08/ A /08/ :45 07/12/ : ,458 1,115, ,712 07/14/ A /13/ :50 07/15/ : ,952 1,276, ,094 07/22/ A /21/ :10 07/23/ : ,857 1,462, ,234 08/03/ A /02/ :25 08/06/ : ,112 1,692, ,795 09/08/ A /08/ :50 09/11/ : ,610 1,771,453 72,952 09/18/ A /18/ :10 09/21/ : ,300 1,903,753 99,201 09/29/ A /28/ :25 10/01/ : ,829 2,009, ,728 10/03/ A /02/ :00 10/10/ : ,066 2,728, ,455 Total or Avg. 1,765,173 81% 03/20/ A /19/ :50 03/21/ : , , ,148 03/30/ A /29/ :00 03/31/ : , ,789 87,641 06/09/ A /08/ :25 06/09/ : , ,728 86,414 06/28/ A /27/ :25 06/29/ : , , ,392 07/10/ A /09/ :25 07/26/ : ,777 1,491, ,294 08/02/ A /01/ :35 08/08/ : ,732 2,082, ,225 09/10/ A /09/ :25 09/10/ : ,426 2,157,941 ** 09/14/ A /14/ :15 09/14/ : ,942 2,180,883 ** 09/27/ A /26/ :35 10/02/ : ,889 2,531, ,870 10/22/ A /22/ :00 10/24/ : ,910 2,623,682 33,308 11/05/ A /04/ :25 11/05/ : ,389 2,650,071 26,287 Total or Avg. 2,338,578 88% 02/11/ A /10/ :39 02/12/ : , ,548 74,531 02/17/ A /16/ :40 02/16/ : , ,485 4,861 02/24/ A /23/ :05 02/24/ : , ,316 46,888 06/16/ A /16/ :10 06/23/ : , , ,761 06/24/ A /24/ :15 06/29/ : , ,796 79,682 06/30/ A /30/ :20 07/02/ : ,925 1,032,721 87,144 07/08/ A /08/ :19 07/12/ : ,958 1,268, ,561 07/15/ A /15/ :30 07/16/ : ,270 1,339,949 44,054 08/22/ A /21/ :40 08/23/ : ,887 1,461, ,784 08/28/ A /27/ :30 09/01/ : ,526 1,713, ,906 10/15/ A /14/ :10 10/16/ : ,784 1,781,146 66,487 10/26/ A /26/ :40 10/28/ : ,119 1,875,265 94,051 11/17/ A /16/ :05 11/17/ : ,227 1,936,492 60,693 11/21/ A /21/ :50 11/21/ : ,472 1,972,964 36,472 Total or Avg. 1,423,874 72% * Events prior to installation of sample stations. ** Malfunction of discrete sampling system No samples taken. Page 154

155 Ch. 3 Table 3. (cont d) Start Date Event Number Interevent Time (days) Start Decimal Date Finish Decimal Date Duration (hrs) Volume Pumped (m3) Cumulative Volume (m3) Volume Sampled (m3) 01/25/ A /25/ :00 01/26/ : ,450 84,450 84,450 03/13/ A /13/ :00 03/19/ : , , ,476 03/27/ A /27/ :00 03/28/ : , ,045 87,119 04/27/ A /27/ :00 04/28/ : , ,661 72,261 04/30/ A /30/ :00 04/30/ : , ,614 15,953 05/18/ A /18/ :00 05/21/ : , , ,243 05/27/ A /27/ :00 05/29/ : , , ,039 06/18/ A /18/ :00 06/19/ : , ,739-08/04/ A /04/ :00 08/06/ : , , ,721 08/11/ A /11/ :00 08/28/ : ,186 1,721, ,686 09/18/ A /18/ :00 09/22/ : ,150 1,795,643 74,150 09/26/ A /26/ :00 10/02/ : ,262 2,162, ,522 11/03/ A /03/ :00 11/03/ : ,015 2,176,920 14,015 11/06/ A /06/ :00 11/07/ : ,077 2,243,997 23,652 11/25/ A /25/ :00 11/25/ :00 3 3,893 2,247,890-12/05/ A /04/ :00 12/04/ :00 1 8,505 2,256,396 8,505 12/14/ A /14/ :00 12/17/ : ,021 2,466, ,056 12/21/ A /21/ :00 12/21/ : ,599 2,493,016 26,599 Total or Avg. 2,340,447 94% Page 155

156 Ch. 3 Table 4. Physical-chemical event statistics for UF9206A. Start Date Event Number TSS Load Sampled (kg) TP Load Sampled (kg) TDP Load Sampled (kg) PP Load Sampled (kg) TSS Equiv Conc (ppm) TP Equiv Conc (ppb) TDP Equiv Conc (ppb) PP Equiv Conc (ppb) 01/25/ A * * * * * * * * 02/09/ A * * * * * * * * 03/20/ A * * * * * * * * 04/14/ A * * * * * * * * 05/09/ A * * * * * * * * 05/17/ A * * * * * * * * 07/06/ A * * * * * * * * % PP P Content of TSS (mg/kg) 07/08/ A , % /14/ A , % /22/ A , % /03/ A , % /08/ A , % /18/ A , % /29/ A , % /03/ A , % 1028 Total or Avg. 148, % 1180 * * * * * * * 03/20/ A , % /30/ A , % /09/ A % /28/ A , % /10/ A , % /02/ A , % /10/ A ** ** ** ** ** ** ** ** ** 09/14/ A ** ** ** ** ** ** ** ** ** 09/27/ A , % /22/ A , % /05/ A % 1821 Total or Avg. 106, % /11/ A , % /17/ A % /24/ A , % /16/ A , % /24/ A , % /30/ A , % /08/ A , % /15/ A , % /22/ A , % /28/ A , % /15/ A , % /26/ A , % /17/ A , % /21/ A , % 1033 Total or Avg. 79, % 889 * Events prior to installation of sample stations. ** Malfunction of discrete sampling systems No samples taken. Page 156

157 Ch 3. Table 4. (cont d). Start Date Event Number TSS Load Sampled (kg) TP Load Sampled (kg) TDP Load Sampled (kg) PP Load Sampled (kg) TSS Equiv Conc (ppm) TP Equiv Conc (ppb) TDP Equiv Conc (ppb) PP Equiv Conc (ppb) % PP P Content of TSS (mg/kg) 01/25/ A , % /13/ A , % /27/ A , % /27/ A , % /30/ A % /18/ A , % /27/ A , % /18/ A ** ** ** ** ** ** ** ** ** ** 08/04/ A , % /11/ A , % /18/ A , % /26/ A , % /03/ A ** % ** 11/06/ A , % /25/ A ** ** ** ** ** ** ** ** ** ** 12/05/ A % /14/ A , % /21/ A % 2384 Total or Avg. 128, % 1461 ** Malfunction of discrete sampling system No samples taken. Page 157

158 Ch. 3 Table 5. Hydraulic Event statistics for UF9206B. Start Date Event Number Interevent Time (days) Start Decimal Date Finish Decimal Date Duration (hrs) Volume Pumped (m3) Cumulative Volume (m3) Volume Sampled (m3) 01/24/ B /24/ :20 01/24/ : ,996 40,996 * 01/29/ B /28/ :55 01/30/ : , ,498 * 02/09/ B /09/ :45 02/11/ : , ,630 * 03/19/ B /19/ :50 03/23/ : , ,305 * 04/14/ B /13/ :24 04/20/ : ,984 1,022,289 * 05/09/ B /08/ :45 05/09/ : ,693 1,093,982 * 05/17/ B /17/ :54 05/17/ : ,090 1,115,072 * 06/29/ B /29/ :50 06/30/ : ,079 1,146,151 * 07/09/ B /09/ :45 07/13/ : ,485 1,438, ,629 08/03/ B /02/ :39 08/06/ : ,647 1,711, ,130 09/17/ B /17/ :25 09/23/ : ,040 2,031, ,513 09/30/ B /29/ :15 09/29/ : ,416 2,044,739 13,416 10/03/ B /02/ :40 10/14/ : ,135,627 3,180,366 1,119,777 Total or Avg. 1,831, % 03/20/ B /19/ :30 03/23/ : , , ,081 03/30/ B /29/ :50 04/01/ : , ,367 84,587 07/12/ B /11/ :05 07/12/ : , ,370 18,255 07/17/ B /17/ :55 07/18/ : , ,471 31,783 07/ 23/ B /23/ :05 07/30/ : , , ,927 08/02/ B /02/ :15 08/04/ : ,289 1,157, ,067 08/06/ B /06/ :00 08/08/ : ,835 1,290, ,657 09/09/ B /08/ :00 09/17/ : ,899 1,880, ,099 09/27/ B /26/ :35 10/04/ : ,177 2,424, ,861 10/09/ B /09/ :35 10/10/ : ,874 2,500,007 74,457 10/22/ B /22/ :50 10/28/ : ,893 2,812, ,513 11/05/ B /04/ :30 11/07/ : ,148 2,930, ,783 11/09/ B /09/ :00 11/09/ : ,471 2,945,519 ** 11/12/ B /12/ :50 11/14/ : ,698 2,992,217 ** 11/19/ B /19/ :35 11/19/ : ,869 3,003,086 8,137 12/31/ B /31/ :15 01/01/ : ,001 3,050,087 40,992 Total or Avg. 2,304, % 02/10/ B /10/ :10 02/18/ : , , ,027 02/23/ B /22/ :45 02/23/ : , ,703 21,949 07/01/ B /01/ :10 07/02/ : , ,308 42,265 07/09/ B /09/ :35 07/15/ : ,493 1,023, ,136 08/25/ B /25/ :10 09/02/ : ,682 1,373, ,282 09/06/ B /05/ :05 09/06/ : ,260 1,409,743 ** 09/12/ B /11/ :55 09/12/ : ,560 1,448,303 37,968 10/14/ B /14/ :35 10/15/ : ,962 1,536,265 87,396 10/26/ B /25/ :00 10/28/ : ,561 1,692,826 43,646 11/17/ B /16/ :50 11/18/ : ,754 1,777,580 72,052 11/21/ B /21/ :30 11/22/ : ,544 1,867,124 87,841 Total or Avg. 1,400, % * Events prior to installation of sample stations. ** Malfunction of discrete sampling system No samples taken. Page 158

159 Ch. 3 Table 5. (cont d). Start Date Event Number Interevent Time (days) Start Decimal Date Finish Decimal Date Duration (hrs) Volume Pumped (m3) Cumulative Volume (m3) Volume Sampled (m3) 01/08/ B /08/ :00 01/08/ :00 3 7,931 7,931 7,931 01/18/ B /18/ :00 01/21/ : ,231 31,161 23,231 01/25/ B /25/ :00 01/26/ : , ,659 96,498 03/13/ B /13/ :00 03/18/ : , , ,938 03/23/ B /23/ :00 03/24/ : , ,649 53,135 03/27/ B /27/ :00 03/30/ : , , ,114 04/26/ B /26/ :00 05/01/ : , , ,984 05/18/ B /18/ :00 05/23/ : ,844 1,043, ,515 05/27/ B /27/ :00 05/31/ : ,860 1,249, ,170 06/22/ B /22/ :00 06/24/ : ,397 1,367,642 43,924 08/04/ B /04/ :00 08/04/ : ,494 1,395,136 ** 08/08/ B /08/ :00 08/09/ : ,836 1,446,973 51,836 08/13/ B /13/ :00 08/15/ : ,107 1,564, ,107 08/21/ B /21/ :00 08/30/ : ,109 1,972, ,062 09/05/ B /05/ :00 09/06/ : ,590 2,075, ,429 09/11/ B /11/ :00 09/11/ :00 2 9,942 2,085,720 9,942 09/18/ B /18/ :00 09/22/ : ,257 2,224, ,408 09/26/ B /26/ :00 10/05/ : ,456 2,835, ,038 11/05/ B /05/ :00 11/09/ : ,203 3,008, ,203 12/14/ B /14/ :00 12/14/ : ,587 3,022,222 13,587 12/17/ B /17/ :00 12/17/ :00 3 5,909 3,028,132 5,909 12/21/ B /21/ :00 12/23/ : ,272 3,093,403 65,272 Total or Avg. 2,544, % ** Malfunction of discrete sampling system No samples taken. Page 159

160 Ch. 3 Table 6. Physical-Chemical event statistics for UF9206B. Start Date Event Number TSS Load Sampled (kg) TP Load Sampled (kg) TDP Load Sampled (kg) PP Load Sampled (kg) TSS Equiv Conc (ppm) TP Equiv Conc (ppb) TDP Equiv Conc (ppb) PP Equiv Conc (ppb) 01/24/ B * * * * * * * * * * 01/29/ B * * * * * * * * * * 02/09/ B * * * * * * * * * * 03/19/ B * * * * * * * * * * 04/14/ B * * * * * * * * * * 05/09/ B * * * * * * * * * * 05/17/ B * * * * * * * * * * 06/29/ B * * * * * * * * * * 07/09/ /03/ /17/ /30/ /03/ B B B B B Total or Avg ,717 21, , , % 33% 61% 78% 40% 40% % PP P Content of TSS (mg/kg) 03/20/ /30/ /12/ /17/ /23/ /02/ /06/ /09/ /27/ /09/ /22/ B B B B B B B B B B B ,755 3, ,414 2,790 10,377 22,127 2,780 18, % 47% 48% 43% 26% 22% 47% 33% 25% 81% 61% /05/ B , % /09/ B ** ** ** ** ** ** ** ** ** ** 11/12/ B ** ** ** ** ** ** ** ** ** ** 11/19/ B % /31/ B % 2205 Total or Avg. 135, % /10/ B , % /23/ B % /01/ B , % /09/ B , % /25/ B , % /06/ B ** ** ** ** ** ** ** ** ** ** 09/12/ B % /14/ B #N/A #N/A % ** 10/26/ B , % /17/ B , % /21/ B , % 901 Total or Avg. 69, % 1128 * Events prior to installation of sample stations. ** Malfunction of discrete sampling system No samples taken. Page 160

161 Ch 3. Table 6. (cont d). Start Date Event Number TSS Load Sampled (kg) TP Load Sampled (kg) TDP Load Sampled (kg) PP Load Sampled (kg) TSS Equiv Conc (ppm) TP Equiv Conc (ppb) TDP Equiv Conc (ppb) PP Equiv Conc (ppb) % PP P Content of TSS (mg/kg) 01/08/ B % /18/ B , % /25/ B , % /13/ B , % /23/ B , % /27/ B , % /26/ B , % /18/ B , % /27/ B % /22/ B % /04/ B ** ** ** ** ** ** ** ** ** ** 08/08/ B % /13/ B % /21/ B , % /05/ B , % /11/ B % /18/ B , % /26/ B , % /05/ B , % /14/ B % /17/ B % /21/ B % 3304 Total or Avg. 135, % 1899 ** Malfunction of discrete sampling system No samples taken. Page 161

162 Ch. 3 Table 7. Hydraulic event statistics for UF9209A. Start Date Event Number Interevent Time (days) Start Decimal Date Finish Decimal Date Duration (hrs) Volume Pumped (m3) Cumulative Volume (m3) Volume Sampled (m3) 03/20/ A /20/ :25 03/22/ : , , ,539 03/24/ A /24/ :45 03/24/ : , ,227 ** 03/30/ A /30/ :05 03/31/ : , ,723 ** 06/02/ A /02/ :20 06/02/ : , ,642 91,919 06/10/ A /10/ :40 06/10/ : , ,189 ** 06/13/ A /13/ :15 06/13/ : , ,807 49,618 06/16/ A /16/ :55 06/17/ : , ,998 ** 06/19/ A /19/ :59 06/20/ : , , ,370 06/22/ A /22/ :30 06/29/ : ,956 1,278, ,743 07/12/ A /12/ :05 07/13/ : ,926 1,399,563 ** 07/15/ A /15/ :25 07/20/ : ,616 1,840, ,082 07/22/ A /22/ :45 07/25/ : ,047 2,082, ,372 07/27/ A /27/ :30 07/27/ : ,719 2,130,945 48,719 08/02/ A /02/ :50 08/06/ : ,285 2,376, ,639 08/08/ A /08/ :05 08/08/ : ,996 2,416,226 39,996 08/23/ A /23/ :55 08/23/ : ,711 2,450,937 34,711 09/05/ A /05/ :30 09/06/ : ,291 2,537,228 82,227 09/08/ A /08/ :05 09/19/ : ,355,337 3,892,565 1,157,917 09/29/ A /29/ :35 10/04/ : ,648 4,467, ,667 10/24/ A /24/ :59 10/24/ : ,256 4,538,469 59,477 10/26/ A /26/ :15 10/27/ : ,691 4,675, ,647 10/30/ A /30/ :30 10/31/ : ,983 4,747,143 47,957 Total or Avg. 3,243,601 68% 12/31/ A /31/ :35 01/01/ : , , ,701 01/15/ A /15/ :55 01/15/ : , ,476 82,108 01/17/ A /17/ :35 01/17/ : , ,900 86,424 01/20/ A /19/ :45 01/19/ : , ,529 42,629 02/11/ A /11/ :15 02/13/ : , , ,402 02/14/ A /14/ :28 02/15/ : , ,422 86,244 02/19/ A /19/ :20 02/19/ : , ,833 48,621 02/24/ A /24/ :15 02/24/ : , ,703 66,013 03/09/ A /09/ :05 03/09/ : , ,897 70,671 05/20/ A /20/ :35 05/20/ : , ,086 ** 06/09/ A /09/ :45 06/09/ : ,564 1,033,650 43,415 06/13/ A /13/ :20 06/22/ : ,146,795 2,180, ,717 06/24/ A /24/ :35 06/28/ : ,055 2,593, ,017 06/30/ A /30/ :20 06/30/ : ,252 2,649,752 18,627 07/02/ A /02/ :35 07/04/ : ,726 2,876, ,823 07/08/ A /08/ :15 07/15/ : ,177 3,423, ,824 07/22/ A /22/ :30 07/25/ : ,593 3,826, ,889 08/12/ A /12/ :10 08/12/ : ,508 3,876,756 ** 08/14/ A /14/ :45 08/16/ : ,071 4,344, ,951 08/18/ A /18/ :05 08/19/ : ,886 4,575, ,949 08/21/ A /21/ :20 08/22/ : ,650 4,762, ,949 08/28/ A /28/ :00 09/02/ : ,895 5,546, ,480 09/04/ A /04/ :55 09/05/ : ,259 5,672, ,259 09/12/ A /12/ :10 09/12/ : ,813 5,733,330 58,710 09/22/ A /22/ :55 09/22/ : ,148 5,763,478 29,715 09/25/ A /25/ :55 09/26/ : ,630 5,965, ,168 09/30/ A /30/ :10 10/02/ : ,524 6,209, ,342 10/30/ A /30/ :30 10/30/ : ,558 6,249,190 36,234 11/18/ A /18/ :15 11/21/ : ,388 6,508, ,481 11/24/ A /24/ :30 11/24/ : ,012 6,549,590 41,013 11/27/ A /27/ :10 11/28/ : ,950 6,597,540 47,874 Total or Avg. 5,127,250 78% ** Malfunction of discrete sampling system No samples taken. Page 162

163 Ch. 3 Table 7. (cont d). Start Date Event Number Interevent Time (days) Start Decimal Date Finish Decimal Date Duration (hrs) Volume Pumped (m3) Cumulative Volume (m3) Volume Sampled (m3) 01/01/ A /01/ :00 01/05/ : , , ,283 02/21/ A /21/ :00 02/24/ : , , ,311 03/17/ A /17/ :00 03/19/ : , , ,312 03/24/ A /24/ :00 03/31/ : , , ,472 04/28/ A /28/ :00 05/01/ : ,455 1,353, ,267 05/28/ A /28/ :00 05/30/ : ,792 1,646, ,792 06/10/ A /10/ :00 06/10/ : ,586 1,693,029 42,845 06/19/ A /19/ :00 06/25/ : ,520 2,311, ,142 07/16/ A /16/ :00 07/17/ : ,028 2,445, ,997 07/24/ A /24/ :00 07/25/ : ,989 2,566, ,989 08/05/ A /05/ :00 08/06/ : ,739 2,713, ,739 08/11/ A /11/ :00 08/14/ : ,166 3,051, ,225 08/22/ A /22/ :00 08/29/ : ,568 3,893, ,569 09/02/ A /02/ :00 09/03/ : ,500 4,010, ,500 09/28/ A /28/ :00 10/01/ : ,360 4,323, ,360 11/05/ A /05/ :00 11/07/ : ,193 4,486, ,193 12/17/ A /17/ :00 12/17/ : ,961 4,590, ,131 12/23/ A /23/ :00 12/23/ : ,802 4,666,856 76,802 12/27/ A /27/ :00 12/27/ : ,648 4,727,503 60,648 Total or Avg. 4,299,578 91% Page 163

164 Ch. 3 Table 8. Physical-chemical event statistics for UF9209A. Start Date Event Number TSS Load Sampled (kg) TP Load Sampled (kg) TDP Load Sampled (kg) PP Load Sampled (kg) TSS Equiv Conc (ppm) TP Equiv Conc (ppb) TDP Equiv Conc (ppb) PP Equiv Conc (ppb) % PP P Content of TSS (mg/kg) 03/20/ A , % /24/ A ** ** ** ** ** ** ** ** ** ** 03/30/ A ** ** ** ** ** ** ** ** ** ** 06/02/ A , % /10/ A ** ** ** ** ** ** ** ** ** ** 06/13/ A % /16/ A ** ** ** ** ** ** ** ** ** ** 06/19/ A % /22/ A % /12/ A ** ** ** ** ** ** ** ** ** ** 07/15/ A , % /22/ A % /27/ A % /02/ A , % /08/ A % /23/ A % /05/ A % /08/ A , % /29/ A , % /24/ A % /26/ A , % /30/ A % 3341 Total or Avg. 38, % /31/ A , % /15/ A , % /17/ A , % /20/ A , % /11/ A , % /14/ A % /19/ A % /24/ A , % /09/ A , % /20/ A ** ** ** ** ** ** ** ** ** ** 06/09/ A % /13/ A , % /24/ A , % /30/ A % /02/ A , % /08/ A , % /22/ A , % /12/ A ** ** ** ** ** ** ** ** ** ** 08/14/ A % /18/ A , % /21/ A % /28/ A , % /04/ A , % /12/ A % /22/ A % /25/ A , % /30/ A , % /30/ A , % /18/ A , % /24/ A % /27/ A % 605 Total or Avg. 113, % 1171 ** Malfunction of discrete sampling system No samples taken. Page 164

165 Ch. 3 Table 8 (cont d). Start Date Event Number TSS Load Sampled (kg) TP Load Sampled (kg) TDP Load Sampled (kg) PP Load Sampled (kg) TSS Equiv Conc (ppm) TP Equiv Conc (ppb) TDP Equiv Conc (ppb) PP Equiv Conc (ppb) % PP P Content of TSS (mg/kg) 01/01/ A , % /21/ A , % /17/ A , % /24/ A , % /28/ A , % /28/ A , % /10/ A , % /19/ A , % /16/ A , % /24/ A , % /05/ A , % /11/ A , % /22/ A , % /02/ A , % /28/ A , % /05/ A , % /17/ A , % /23/ A , % /27/ A , % 400 Total or Avg. 472, % 372 Page 165

166 Ch. 3 Table 9. Particulate P contributions to P loads. Farm UF9200A 47% 48% 56% 28% UF9206A 26% 36% 36% 27% UF9206B 40% 36% 35% 37% UF9209A NA 67% 68% 80% Ch. 3 Table 10. Annual averages of key parameters from study farms. Farm Year Total Drainage (10 6 m 3 ) TSS Equiv. Conc. (ppb) TP Equiv. Conc. (ppb) TDP Equiv. Conc. (ppb) PP Equiv. Conc. (ppb) Estimated TP Load (kg) Estimated TDP Load (Kg) Estimated PP Load (kg) TSS P Content (mg/kg) UF9200A UF9206A UF9206B UF9209A 2000* *Sampling at Farm UF9209A started in Page 166

167 Ch. 3 Table 11. Summary of annual rainfall and pumping data. Farm Year Total Drainage (10 6 m 3 ) Total Drainage (inches) Rainfall (inches) Pumping to Rainfall Ratio (in/in) UF9200A UF9206A/B UF9209A Table 11 summarizes the rainfall and pumping d ata for each farm. In Ta ble 11 the data for stations UF9206A and UF9206B have been consolidated. This is done because the farm has integ rated cross-connections, which makes any allocation of rainfall to either pump station a dubious v alue. The d ata summarized in these two tables allows some basic comparisons to be made among farms, over time, and with other sources. The normalized or unit area loads (UAL) for particulate P and dissolved P are pre sented in mass per unit farm area (kg P/acre) in Figure 8. Particula te P loads from farm UF9 200A have been fairly constant over the four-year period, averaging about 0.15 kg particulate P/acre (0.33 lb P/acre). Loads from farm UF9206A and B were combined and presented for the farm as an overall entity. Part iculate P loads from this farm have been steadily decreasing during the first three years of the study, reaching loads values similar to those observed in UF9200A. However, in year 2003 particulate P load slightly increased to about 0.24 kg/acre (0.53 lb/acre). The loads for UF9209A were substantially lower than the other two farms, averaging 0.05 kg particulate P/acre (0.11 lb/acre) during the last three years (Figure 8A). Page 167

168 Estimated Annual Unit Area Particulate P Loads Part iculate P Expor t (kg/ acre) A A 06A/B 09A Farm Export (kg/acre) P B Estimated Annual Unit Area TDP Loads TD A 06A/B 09A Farm Ch. 3 Figure 8. A) Annual unit area particulate P loads, and B) annual unit area dissolved P loads from 2000 to Page 168

169 Dissolved P loads from farm UF9200A followed the same pattern observed with the particulate P values, averaging about 0.16 kg dissolved P/acre during the first three years of the study (Figure 8B). However, in year 2003 total load due to dissolved P increased to an average of 0.34 kg P/acre. Dissolved P loads from farm UF9206A and B have been steadily decreasing during the first three years of the study, reaching an average load value of 0.21 kg P/acre in But, in 2003 dissolved P load values increased, averaging 0.48 kg P/acre. The loads for UF9209A were substantially lower than the other two farms, averaging 0.02 kg dissolved P/acre during the last three years. Figure 9 shows the volume of drainage water pumped during the last four years of the study, expressed as inches to normalize to farm area. Figure 10 shows the annual pumping-torainfall ratios, in inches of water pumped per inch of rain for each farm. Farm UF9206A and B are combined and shown as a single farm. 35 Annual Drainage Pumping: 2000 to Annual Pumping (in) UF9200A UF9206A/B UF9209A Year Ch. 3 Figure 9. Annual drainage pumping from 2000 to Page 169

170 For the years 2000 and 2001 the normalized volumes pumped from UF9206A/B were % higher than those from the other two farms (Figure 9). In 2002 the normalized volumes were almost identical for all three farms. But in 2003, volumes pumped from UF9206A/B increased to about the same values observed during the first two years. Farm UF9200A also showed a slight increase in water pumped, while farm UF9209A showed a decrease to about the same values observed in 2000 and A similar relationship is observed in the relative volume ratios presented in Figure 10. Data from Figures 9 and 10 shows the existing variability of these parameters over the study period for all three farms. Farm UF9206A/B started with high volumes and ratios in 2000 and decreased through 2002, but it went back again to about the volumes and ratios of the first two years. Farm UF9200A started with low volumes and ratios and steadily increased for the rest of the study, with a slight decrease in the pumping to rainfall ratio in Farm UF9209A started with low volumes and ratio during the first years, but those values decreased to about the original levels for year Pumping to Rainfall Ratios: 2000 through 2003 Inches Pumped per Inch Rainfall - in/in UF9200A UF9206A/B UF9209A Year Ch. 3 Figure 10. Annual pumping to rainfall ratios from 2000 to Page 170

171 It is expected that UF9206A/B, with its large acreage planted to vegetables, would have a higher pumping ratio than the sugarcane farms. However, it is interesting to see the ratios of all three farms converge to a similar value in However, in 2003 pumping to rainfall ratios came back to the same pattern observed in the first two years of the study. The equivalent concentrations of TSS and particulate P are shown in Figures 11 and 12 respectively. Equivalent concentrations for both TSS and particulate P showed a decline from 2000 to 2001 at UF9200A and at UF9206A and B and then remained relatively constant or showed a slight increase from 2001 to In 2003, average TSS concentrations from these three farms remained constant. Equivalent concentrations for TSS and particulate P at UF9209A showed a steady increase from 2001 to 2002, but TSS values for 2003 increased from 22 mg/l to 110 mg/l. Equivalent TSS Conc. - mg/l Equivalent TSS Concentration vs. Year UF9200A UF9206B UF9206A UF9209A Year Ch. 3 Figure 11. Equivalent TSS concentrations from 2000 to Page 171

172 Particulate P changes observed during the last four years are important to this study. Equivalent particulate P concentrations at UF9200A and UF9206A/B significantly declined from 2000 to 2001, and then remained almost constant for all three farms from 2001 to 2002 (Figure 12). In 2003, average particulate P concentrations at UF9200A decreased from 79 ppb to 64 ppb, however, farm UF9206A and B showed a notable increase, with the highest value observed at UF9206B. Average particulate P concentrations from farm UF9209A have been steadily increasing during the last three years, but they are still lower than the other three farms. The P content of TSS gives us an indication to the nature of the particulate P as illustrated in Figure 13. Phosphorus concentrations of TSS from farm UF9200A averaged 2462 mg/kg for year 2000 and 2894 mg/kg for After year 2001, P concentrations decreased to 1778 mg/kg for 2002 and 1311 mg/kg for A similar pattern was observed at farm UF9209A, where P concentrations from TSS steadily decreased from 1756 mg/kg in 2001 to 373 mg/kg in In Year 2003, this farm pumped the canals significantly lower and for longer periods of time than previous years increasing the amount of bottom soils or sediments, which are low in P content compared to biologically produced particulate matter, to be exported from this farm. Equivalent PP Conc. - ppb Equivalent PP Concentration vs. Year UF9200A UF9206B UF9206A UF9209A Year Ch. 3 Figure 12. Equivalent particulate P concentration from 2000 to 2003 Page 172

173 The mixed crop farm, UF9206A and B showed P concentrations of TSS ranging from 752 to 1253 mg/kg from 2000 to In 2003, P concentrations increased to 1461 mg/kg for UF9206A and 1899 mg/kg for UF9206B. In discussions of the Biological Contribution Mechanisms presented earlier in this study, it was noted that biologically sourced particulates would be expected to have P content in the range of mg/kg for plant material, and mg/kg for plant detritus. On an annual average basis these conditions are satisfied for 2000 and 2001 for the sugarcane farms, UF9200A and UF9209A, but not in 2002 and The mixed-crop farm UF9206A/B did not satisfy this condition from 2000 to 2002, and only marginal in These results show that a given population of TSS during a pumping event is highly heterogeneous with a wide range in P content and transport properties. Also farm UF9206A/B has implemented a consistent aquatic weed control program that minimizes the amount of biologically produced particulate matter. P Content of TSS vs. Year 3500 P Content of TSS - mg/kg UF9200A UF9206B UF9206A UF9209A Year Ch. 3 Figure 13. Average P content of TSS from 2000 through Page 173

174 Particulate Phosphorus Load Distribution Analysis In addition to suspended solids and P analysis, every sample taken during the study has an associated set of supporting data, which includes sample time, sample duration, instantaneous flows, instantaneous levels, cumulative time since event start, and cumulative flow since event start. This data may be used to calculate derived parameters such as loads, load rates, and velocities at sections of known configuration. This was done with all samples taken in this program. For the purpose of analysis, the parameter particulate P load rate, defined as the kg of particulate P exported per hour, was of special importance. The use of load rate causes normalization among samples that might have had different sampling time durations. The load rate of a sub-event (or pack of water) defines its levels of importance based on the contribution to the overall annual particulate P load. The higher the load rate, the more the particular sub-event or packet of water contributed to the annual load. The data in each annual location data set were ranked by particulate P load rate, from lowest to highest. Once this was done, the cumulative hydraulic and particulate P loads of the data points as ranked, were determined. Figures 14 to 17 show the results of this analysis, with the cumulative loads expressed as a fraction of the total load. Data in Figures 14 to 17 may be interpreted as follows. Moving from left to right along the X- axis traces the accumulation of packets of water that contributed to the overall hydraulic load for the year. Moving from bottom to top of the Y-axis traces the cumulative contribution to the overall particulate P load of the particulate P contained in the corresponding packet of water. The slope of the curve is a direct indicator of the relative contribution to the particulate P load of a given packet of water. The shallower the slope, the less particulate P a given packet of water contributes, the steeper the slope, the more a specific packet contributes. Adjacent data points (packets of water) on the curve may be widely separated in time; what they have in common is a similar load rate, or contribution priority. This is important, because the data (water packets) are now sorted by priority of importance relative to particulate P export. Page 174

175 Cumulative Hydraulic and Particulate Phosphorus Load Distributions UF9200A through 2003 Cumulative Fractional PP Load Distribution Cumulative Fractional Hydraulic Load Distribution Ch. 3 Figure 14. Cumulative hydraulic and particulate P load distributions for UF9200A. Page 175

176 Cumulative Fractional PP Load Distribution Cumulative Hydraulic and Particulate Phosphorus Load Distributions UF9206A through Cumulative Fractional Hydraulic Load Distribution Ch. 3 Figure 15. Cumulative hydraulic and particulate P load distributions for UF9206A. Page 176

177 Cumulative Fractional PP Load Distribution Cumulative Hydraulic and Particulate Phosphorus Load Distributions UF9206B through Cumulative Fractional Hydraulic Load Distribution Ch. 3 Figure 16. Cumulative hydraulic and particulate P load distributions for UF9206B. Page 177

178 Cumulative Fractional PP Load Distribution Cumulative Hydraulic and Particulate Phosphorus Load Distributions UF9209A through Cumulative Fractional Hydraulic Load Distribution Ch. 3 Figure 17. Cumulative hydraulic and particulate P load distributions for UF9209A. Page 178

179 All curves presented in Figures 14 to 17 show similar results. Taking farm UF9200A, year 2003 (Figure 14) as an example, the lowest 25% of the load rate ranked hydraulic load contributed about 7% of the annual load, the lowest 50% of the load rate ranked hydraulic load contributed about 23% of the annual particulate P load, the lowest 75% contributed about 43%. The most important point of this analysis is the upper end of the curves. For the example of UF9200A, year 2003, the last 20% of the hydraulic load contributed 52% of the annual particulate P load. In all cases, 50% of the annual particulate P load was contributed by less than 25% of the hydraulic load. For farm UF9200A, 18-22% of the annual hydraulic load during the last four year of the study, contributed to 50% of the annual particulate P load (Figure 14). For farm UF9206A, 16-26% of the annual hydraulic load contributed to 50% of the annual particulate P load (Figure 15). For farm UF9206B, 13-19% of the annual hydraulic load contributed to 50% of the annual particulate P load (Figure 16), and for farm UF9209A, only 8-14% of the annual hydraulic load contributed to 50% of the annual particulate P load (Figure 17). This analysis now allows the study of particulate P loading to be concentrated on the event subelements that made the most contribution. Once the distribution were analyzed, the data points that represented contributions in the top 50% were identified and tagged. The top contributors were then analyzed for unique characteristics and processes that would contribute to elevated particulate P transport. This analysis is intimately tied to the farm management practices characteristics of each farm, so a summary of those characteristic practices is appropriate. Farm Management Practices that May Impact Particulate P export UF9200A Sugarcane is the main crop. Average canal sediment dredging program. Cleans canals on as-needed basis. No major canal-work over last four years. Controls aquatic weeds with herbicides on a periodic basis when weed build-up is extensive. Has weed boom within 50 meters (164 ft) of pump station. Page 179

180 Has encountered upward of 50% coverage of canal surface by aquatic plants as measured by aerial survey (Daroub et al., 2003). Reduces discharge by pumping to fallow fields when available. Practices flow reduction during dry season, which leads to long periods of no discharge. Pumping is done with three fix speed electric pumps, two high and one low capacity. Flow control achieved by choice of high or low capacity pump. On-off level controllers control all pumps. Maintains a minimum canal depth at pump station of 0.72 meters (2.4 ft), however, average canal depths have ranged from m ( ft) during the last four years. Encounters frequent occasions of on-off pump cycling because of level control. Pump cycles typically have a period of 30 to 60 minutes. No major operational changes over the study period, but the pumping to rainfall ratio steadily increased from 2000 to 2002 and slightly decreased in 2003 (Figure 10). The use of the small pump during drainage events increased from 52% in 2002 to 71% in UF9206A and B Mixed crop farm including sugarcane, sod, vegetables, and rice. Regular canal maintenance and improvement program. Contains multiple control structures that allow extensive flexibility in water management. Regular aquatic weed control, preventing extensive build-up. Has encountered on the order of 20% coverage of canal surfaces by aquatic weeds (Daroub et al., 2003). Has weed booms within 50 meters (164 ft) of pump stations, but also has upstream structures that impound aquatic plants. Page 180

181 Reduces discharge by redistribution to fallow or planted fields, primarily rice. Discharge reduction typically practiced during wet season, leading to discharges in dry season. Pumping is at two pump stations, each with two, variable speed, high capacity diesel pumps. Level control is manual by speed reduction or pump shut down. Canal level is reduced to the bottom of the canal on occasion. No major changes in operation from 2000 through UF9209A Sugarcane is the main crop. Regular canal dredging and maintenance program. Canals are larger in size relative to pump capacity compared to the other farms, so typical velocity in canals is lower than other two farms. Regular weed control program in main canals to prevent aquatic plant build-up. Weed booms within 50 meters (164 ft) of pump station. Not a part of aerial survey program but visual observation indicated that aquatic coverage is equal to or less than UF9206A/B for main canals, with occasional extensive build-up in field canals. Pumping is done with three variable speed high capacity diesel pumps. Discharge control is primarily by number of pumps operating. No automatic level control, but levels were manually controlled to a minimum canal depth at the pump station of 1.1 meters (3.6 ft) in 2001, 0.4 meters (1.3 ft) in 2002 and 0.8 meters (2.6 ft) for The change in minimum canal depth has constituted a major change in the operational mode of this farm during the last two years. Operational mode typically includes shutting down at night, so there is a typically pump cycling of 8 hours on and 16 hours off. Page 181

182 Process Distribution Analysis of the Top 50% of Particulate Phosphorus Loads Throughout this discussion the emphasis will be on the top 50% of the annual load and on the elements that contributed to parts of that load. Table 12 shows the total annual measured particulate phosphorus load per site from 2000 to The table also shows the percentage point distribution of the top 50% of the particulate P load, as defined in Figures 14-17, among the contributing processes. These processes were defined in the introduction. It should be emphasized that the percentage numbers shown in this table and throughout this discussion are percentage points from the top 50%, so 25 percentage points from the top 50 % represents half of the top 50% of the annual load. Ch. 3 Table 12. Distribution of processes contributing to top 50% of measured loads at study farms. Farm Year Measured First Prior Pump PP Load Flush High Restart (kg) Velocity Pump Cycling Spike Other UF9200A % 10.7% 10.1% 6.5% 2.1% 0. 0% % 9.2% 3.3% 15.0% 6.9% 0.0% % 15.0% 5.0% 12.2% 0.0% 0.0% UF9206A % 9.8% 7.2% % 2.0% % 28.9% 6.5% % 1.8% % 30.3% 6.7% % -- UF9206 B % 9.6% *40.6% % 27.4% % % 30.9% 5.0% % -- UF9209 A 2000 #N/A #N/A #N/A #N/A #N/A #N/A #N/A % 31.0% 0.9% % % 25.2% 12.2% % -- * One Event - Low canal level at start of storm It should be noted here that the year 2000 at UF9206B is dominated by the Other process category. There was one event at this farm that made a large contribution to annual par ticulate P load at both pump stations, primarily because of low canal levels at the start of a major storm. Its elements were assigned to the defined categories at UF9206A, but the response was so pronounced and illustrative at UF9206B that the entire process was Page 182

183 as signed to the Other category. The reasons for this will become evident in the next section. Figures 18a-d show the contents of Table 12 in graphical form. Each pump station exhibits its own distinctive characteristics because of the water management philosophies of the individual grower. Year-to-year variations exist for all farms but some conclusions are evident. Spikes contribute on the order of 5% to the top 50% of the particulate phosphorus load. Pump cycling, present only at UF9200A, contributes to about 10% of the top 50% load at that farm. If the effect of the Other ca tegory at UF9206B in 2000 is discounted, both pump stations at that farm show similar dis tributions over the three-year period. Th e two sugarcane fa rms do not show similar patterns. P ump cycling and pump restart together account for around 17% of the top 50% at UF9200A. At UF9209A where pump cycles are h ave long periods, pump restart is variable, accounting for less than 1% in 2001 and abo ut 12% in First flush accounts for more ( 15-20%) at UF9200A than at the other sta tions (7-15 %) if the effects of 2000 at UF9206A and B are discounted. The distrib ution between transient and steady hydraulic conditions ma y be evalu ated by lumping First Flush, Pump Restart, and Pump Cycling into one category identified as Pump Transients. This is shown in Figures 19a-d. Under this analysis, UF9206A/B and UF9209A exhibit some similarities f or the time period where comp arison is possible. Prior high velocity accounts for 25-30% at all three stations. Pump transients accoun t for around 15% at UF9206A/B, and average around 18%, with some variation, at UF9209A. On the other hand pump transients consistently account for around 35% at UF9200A, while prior high velocity accounts for around 10%. A general conclusion drawn from this is that high particulate phosphorus export from UF 9200A tends to be more influenced by pump transients, while it tends to be more influenced by velocity considerations at other farms. Page 183

184 45% UF9200A Process Type for Top 50% PP Load UF9200A 45% UF9200A Process Type for Top 50% PP Load UF9206A 40% 40% Per Cent of Annual Loa d 35% 30% 25% 20% 15% 10% Per Cent of Annual Loa d 35% 30% 25% 20% 15% 10% % 5% A 0% First Flush Prior High Pump Restart Pump Cycling Spike Other Velocity C 0% First Flush Prior High Velocity Pump Restart Pump Cycling Spike Other 45% UF9200A Process Type for Top 50% PP Load UF9209A 45% UF9200A Process Type for Top 50% PP Load UF9206B 40% 40% Per Cent of Annual Loa d 35% 30% 25% 20% 15% 10% Series Per Cent of Annual Loa d 35% 30% 25% 20% 15% 10% % 5% B 0% First Flush Prior High Pump Restart Pump Cycling Spike Other Velocity D 0% First Flush Prior High Velocity Pump Restart Pump Cycling Spike Other Ch. 3 Fi gure 18. Percentage point distribution of contributing processes for top 50 % of particulate P load for A) UF9200 A, B) UF9209A, C) UF9206A, and D) UF9206B. Page 184

185 45% UF9200A Process Type for Top 50% PP Load UF9200A 45% UF9200A Process Type for Top 50% PP Load UF9206A 40% 40% Per Cent of Annual Load 35% 30% 25% 20% 15% 10% Per Cent of Annual Load 35% 30% 25% 20% 15% 10% % 5% A 0% Pump Transients Prior High Velocity Spike Other C 0% Pump Transients Prior High Velocity Spike Other 45% UF9200A Process Type for Top 50% PP Load UF9209A 45% UF9200A Process Type for Top 50% PP Load UF9206B 40% 40% Per Cent of Annual Load 35% 30% 25% 20% 15% 10% Per Cent of Annual Load 35% 30% 25% 20% 15% 10% % 5% B 0% Pump Transients Prior High Velocity Spike Other D 0% Pump Transients Prior High Velocity Spike Other Ch. 3 Figure 19. Percentage point distribution of consolidated processes for top 50% of particulate P load for A) UF9200A, B) UF9209A, C) UF9206A, and D) UF9206B. Page 185

186 These analyses give a general insight into the process that affect particulate P transport at each of the farms and the relative role those processes play at each location. The next step is to evaluate events at each location that were major contributors to the top 50% of the particulate P loads and determine the characteristics of these events, or the sub-events within them, that caused especially increased particulate P transport. Analysis of Major Event Contributing to Top 50% Particulate P Loads The objective of this analysis is to identify conditions that give rise to the increased particulate P transport events. Elevated P load events have been defined as those sub- events that had a particulate P load rate contributing 50% of the annual particulate P load in only 10-25% of the annual hydraulic load. It must be emphasized that throughout this analysis what is being studied is the precursor conditions leading to the event and the subevents within the event that contribute to the top 50% particulate P loads. The percentage points referred to for an event do not represent the total contribution of that event to the annual particulate P load. Rather they refer to the percentage of the annual particulate P load that was contained in the event that fell into the top 50% of the annual particulate P load. As an example, a large, long hydraulic event that ran at low particulate P loads rates could contributes substantially to the total annual particulate P load, while having few sub-events that qualified for inclusion into the top 50%. In this case the event would not appear as a significant contributing event even though, by its duration, it contributed considerably to the total annual load. A short event that contained numerous sub-events that were excursions into increased particulate P load rates might be ranked higher than the larger event, even though its total load was much lower than the larger event. Table 13 shows the events that contained sub-events that were in the top 50% and the percentage point distribution by each event to the top 50%. Totals for each farm-year do not add up to exactly 50% because the sub-events chosen did not add up to exactly 50%. The most distinctive pattern observed from this data is the number of farm-years that were dominated by few events. Data from this table shows that six of the 15 farm-years sampled had a single event that contributed 30% or more to the top 50% particulate P load. Three farm-years had two events that contributed a total of 30% or more. Three farm-years had three events that contributed a total of 30% or more. Only two of the 15 farm-years had their load rates distributed such that it took more than three events to contribute a total of 30% or more to the top 50%. Page 186

187 Ch. 3 Table 13. Percentage point distributions of top 50% load events. Farm Event Number Percentage Points Event Number Percentage Points Event Number Percentage Points Event Number 2003 Percentage Points UF9200A 00A A A A A A A A A A A A A A A A A A A A A UF9206A 06A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A UF9206B 06B B B B B B B B B B B B B B B B B B B B B B B B B UF9209A 09A A A A A A A A A A A A A A A A A A A A A A A A A Page 187

188 The timing of these clusters is also interesting. In 2000, all three stations under study at the time (UF9209A did not start sampling until 2001) were dominated by one event. In 2001, two stations out of four were dominated by one event. In 2002, one out of four stations was dominated by one event, and in 2003, three out of four stations were dominated by one event. The focus of the event analysis is by priority. First priority is annual-dominant events and second priority the lesser events that still contributed 10% or more to the top 50%. UF9206A and UF9206B both operate under the same control policy and, because of intensive interconnections, do not have a watershed within the farm that stays constant over time. In the aggregate, analysis of these two stations yields the same conclusions. The objective of this analysis is to identify operating parameters and conditions that give rise to increased particulate P load rates. The event statistics used are referenced from Tables 1-8. Where it is appropriate for illustration purposes selective graphical presentations are made. Event Analysis for UF9200A UF9200A - Year 2000 There was one dominant event in 2000, Event 00A (00A ; Oct. 2, 2000), which contributed 42.9 percentage points to the top 50% (Table 13). This is the event that was used as the example of the various transport processes in the introduction. Its profiles are shown in Figures 5 and 6. The event itself had an inter-event time of 7 days, but the preceding pumping events had been at low levels for an extended duration. There had been a pumping event in April, followed by 75 days of no pumping. There were then two small events in July 3-11 where only the small pump was used. After this there followed a period of 70 days of no pumping. On August there were again two small events where the small pump was used. At the start of the second of these events the large pump was run for two hours, then pumping reverted to the small pump. The large pump was finally used extensively during event 276. The hydraulic history prior to this event was long periods of no flow, punctuated with several small events with low velocity. Event 276 was essentially the first time in over five months that the canal system had seen relatively high velocities on a sustained basis. Page 188

189 During the event about 458,000 m 3 of water were pumped. Of this volume, 136,000 m 3, or about 30%, contained the sub-events that contributed 42.7 percentage points to the top 50% of the annual particulate P load. The breakdown of processes was: First Flush %, Restart Flush - 8.6%, Pump Cycling - 6.5%, and Continued High Velocity %. UF9200A - Year 2001 There was one dominant event in 2001, Event 00A (00A , Aug. 2, 2001), which contributed 30.0 percentage points to the top 50% (Figure 20). The hydraulic history of this event is different from that of , but it has some commonality. In 2001 there was no pumping for the first six months of the year. Pumping started on July 1, and there were one small event (01.192) and two large events ( and ) prior to event During all these events the large pump was used exclusively. Events and had total volumes similar to , but contained far less of the top 50% sub-events. The difference was the amount of continuous run time undergone in Event For the three prior events, pumping had often been stopped at night and restarted the following morning. The maximum run time for Event was 8 hours, for , 32 hours, and for , 40 hours. Each of these events contributed modestly to the top 50%. In contrast, during Event the large pump was run continuously for 89 hours before it went into shortperiod cycle mode for 12 hours. Most of the contribution from Event came after 36 hours of run time, and during the short-period cycle time. During the event about 645,000 m 3 of water were pumped. Of this volume, 217,000 m 3, or about 34%, contained the sub-events that contributed 30.0 percentage points to the top 50% of the annual particulate P load. The breakdown of processes was: First Flush - 4.1%, Restart Flush 0% (No restarts), Pump Cycling 12.3%, and Continued High Velocity 11.3%, and Spikes 2.3%. Page 189

190 UF9200A Hydraulic History 7/1/01-8/7/ Event Event 192 Event 204 Event 214 /s Pumping Rate - m Decimal Date Ch. 3 Figure 20. Hydraulic history for event UF9200A (August 2, 2001). UF9200A - Year 2002 There were no dominant events in The distribution of contributors to the top 50% was relatively even over the top five events. It is important to note that in 2002 the longest sustained run time of any event was 30 hours, far less than the 89 hours seen in Event The top contributor in 2002 was Event 00A (00A , June 21, 2002). This was the first pumping event of the wet season. It had an inter-event time of 108 days, lasted for 14 days, contributed 47% of the annual hydraulic load, but yet only contributed 9.5 percentage points to the top 50% particulate phosphorus load. The pumping pattern for this event provides some interesting insight. It started with the large pump going through two 8- hour long-period cycles, followed by a 32-hour cycle then an 18-hour cycle, then eight hours of short-period cycles. At that point, about 478,000 m 3 of the event total of 1,248,000 m 3 of water had been pumped. From this point on, pumping was primarily with the small pump, with the large pump being started only for short periods (4-6 hours) on occasion. There were no top 50% contributing sub-events in the final 770,000 m 3 of water discharged. Page 190

191 The flow, velocity, and particulate P load rate profiles of the first 478,000 m 3 for Event are shown in Figure 21. The X axis crosses the Y axis at the cut-off point for the top 50% load rate. Points on or above this line are included in the top 50% load rate. This event very clearly exhibits the patterns of start-up flush, continued high velocity, and pump cycling described in the introduction. It has been discussed in some detail, because it is representative of the events encountered at UF9200A in Event UF9200A-172 PP Load Rate, Flow, and Velocity PP Load Rate PP Load Rate-Top 50% Flow Velocity x PP Load Rate - kg/hr Flow - m3/s Velocity x 2 - m/s Decimal Date 0.00 Ch. 3 Figure 21. Profiles for event UF9200A (June 21, 2002). The events encountered in 2002 exhibited various combinations of large and small pumps and of pump recycling. In general they behaved as expected, showing increased load rates at start-up after long inter-event times, increased load rates after continued high velocity, and increased load rates during short-period pump cycling. In the absence of a dominating event in 2002 the behavior of the farm system was typical of that shown by Figure 21. UF9200A Year 2003 There were no dominant events in 2003 contributing more than 30 percentage points to the top 50%. However three major events contributed most of the percentage points to the top Page 191

192 50% during this year (Table 13). The top contributor in 2003 was Event 00A (Sept. 4, 2003). This event had an inter-event time of 5 days, lasted for 116 hours, but only contributed 5% of the total hydraulic load, and yet contributed 23.9 percentage points to the top 50% particulate P load. Event 00A (Nov. 5, 2003) was also a top contributor in This event had an inter-event time of 35 days, lasted for 132 hours, but only contributed 4% of the total hydraulic load and yet contributed 14.3 percentage points to the top 50% particulate P load. This was a typical event observed during this year, and some of its profiles are shown in Figure 22. Flow, velocity and canal level profiles for this event are shown in Figure 22A, TSS, TDP and particulate P concentrations and loads are presented in Figures 22B and C, respectively. This event started November 5, 2003 in response to some rainfall that occurred around the same time. At the beginning of the event, only the small pump was operated for several hours (flow rate for that period was lower than 0.5 m 3 /s). Figure 22B and C show the start-up flush of particulate P at the beginning of the event. On November 6, the larger pump was used for several hours and the effect on flow velocity and suspended solids during that period are well evident. Flow velocity increased from about 0.12 m/s to 0.35 m/s, similarly, TSS concentrations increased from about 40 mg/l to more than 300 mg/l. Particulate P concentrations experienced the same increase until the large pump was turned off and only the small pump was operating, at which point concentrations started to decrease because of the decrease in average canal velocity. The events observed in 2003 exhibited various combinations of small and large pumps cycling due to the combination of the large and small pumps used. In general, the majority of the events in 2003 showed increased load rates at start-up after long inter-event times, increased load rates after continued high velocity, and relative increased load rates during short-period pump cycling. Page 192

193 A 0.35 Level, Flow, Rain ( m, m 3 /s, in) Canal level Flow Rain Velocity /s) Velocity (m / 0 5/0 3 11/ 0 6 / / 0 7/ / 0 8 / / 0 9 / / 10 / / 11/ / 12 / PP Conc TDP Conc TSS Conc B TDP, PP Conc (mg/l) TSS Conc (mg/l) / 0 5/ / 0 6 / / 0 7/ / 0 8 / / 0 9 / / 10 / / 11/ / 12 / 0 3 ate (kg/hr) C PP Load Rate TDP Load Rate TSS Load Rate e (kg/hr) TDP, PP Load r TSS Load Rat / 0 5/ / 0 6 / / 0 7/ / 0 8 / / 0 9 / / 10 / / 11/ / 12 / 0 3 Ch. 3 Figure 22. Profiles for event UF9200A (November 5, 2003). Page 193

194 Event Analysis for UF9206B UF9206B - Year 2000 There was one dominant event in 2000, Event 06B (06B , Oct. 2, 2000) (Table 13). This is the event that was classed in the Other category in the earlier discussion on process distribution. Figure 23 shows the details of the critical portion of At the start of the event, the grower pulled the canal levels down in anticipation of a large rainfall. At Decimal Date the canal level was about 0.6 m. At this low level, approximately 50% of the bottom surface of the canal is exposed (Figure 23). Rainfall started around DD and continued at a relatively steady rate until DD At that time, the canal level had risen to about 0.8 m, where approximately 30% of the bottom surface of the canal is still exposed. From that point, the rainfall intensity (as indicated by the slope of the cumulative rainfall curve) increased significantly over the next six hours. That period coincides exactly with the dramatic rise in particulate P load rate. It is surmised that the intense rainfall on the exposed canal bottom induced major dislodgement and mobilization of sediment particulate P. The heavy mobilization continued after the rainfall stopped, and was exacerbated from DD through DD , when the canals were pulled down to low levels at high pumping rates, causing high canal velocities. These increased velocities, as high as 0.4 m/s, coincide almost exactly with the peak in particulate P load rates exhibited at the same time. It appears that the low canal levels, coupled with intense rainfall, caused a massive mobilization of sediment, which was amplified by the subsequent high velocities encountered when the canal levels were brought down under high pumping rates. The eighteen-hour period included in Figure 23 contributed 40.5 percentage points to the top 50%. There were additional periods of increased load rate later in this event, arising from continued high velocity, which contributed an additional 9.2 percentage points, bringing the total to The total discharge for the event was 1,135,000 m 3. Flow contributing to the top 50% was 199,000 m 3, or about 18% of the total event discharge. Page 194

195 Event 9206B Load Profiles PP Load Rate 12.0 Canal Level Velocity Rainfall 2.4 Cumulative Rain 10.0 PP Load Rate 2.0 PP Load Rate - kg/hr Canal level (m) ) in) Velocity (m/s Cum. Rainfall ( Level 2.0 Velocity Decimal Date Ch. 3 Figure 23. Profiles for event UF9206B (October 2, 2000). Page 195

196 UF9206B - Year 2001 There was no dominant event at UF9206B during Contributions to the top 50% were concentrated in three events, 06B (06B ), 295 (06B ), and 308 (06B ), which together accounted for 39.8% of the contribution to the top 50% (Table 13). The remaining contributions were spread among nine events. During the year there were 16 events, nine of which pumped 100,000 m 3 or more of water. Event (06B , Sep. 26, 2001), the top contributor, had a volume of 543,000 m 3, an inter-event time of 9 days, and was preceded by an event (01.251) of larger volume (590,000 m 3 ). What differentiated from its predecessors was canal velocity. Event (06B ) had average canal velocities of 0.19 m/s while averaged 0.24 m/s. In addition, because of canal draw down, velocities in excess of 2 m/s were encountered during Event 9206B PP Load Rate, Flow, and Velocity PP Load Rate PP Load Rate - Top 50% 2. 5 PP Load Rate - kg/hr Flow Velocity Flow - m3/s Velocity - m/s Decimal Date Ch. 3 Figure 24. Profiles for event UF9206B (September 26, 2001). Page 196

197 Figure 24 shows the profile of Event There is some contribution to the top 50% from start-up flush, but the predominant contribution comes in the period between Decimal Date and At DD pumping rate was almost tripled in response to heavy rainfall. This transient gave rise to an increased velocity and increased load rate. Load rates then generally decreased until about DD when velocities started to increase due to increased pumping and decreased canal level. The load rates increased dramatically in a response that was very similar to that seen in Event , and then dropped off as soon as velocity was reduced. Flow that contained load rates in the top 50% totaled 185,000 m 3, or about 34% or the event total. Supply exhaustion is seen around DD when velocities exceeded those seen at the start of the event or at DD 272.4, but load rates remained much lower than those seen at the start of the event or at DD The other two major contributors, (06B ) and (06B ), also had periods of low canal level and high velocity and exhibited essentially the same behavior as Event , including exhibiting characteristics of supply exhaustion. UF9206B - Year 2002 There were two events, 06B (06B ) and (06B ) that together contributed a total of 35.4 percentage points to the top 50% (Table 13). The rest of the top 50% contributions were spread among seven other events. Event 06B (06B , Feb. 10, 2002), which contributed 24.4 percentage points, came close to being a dominant event. This event, which had an inter-event time of 40 days, came early in February as a result of a dry season storm. It was characterized by high velocities and supply exhaustion. Figure 25 shows the profiles for Event There was first flush, then a load rate response to the elevation of velocities to 0.5 m/s, followed by an additional increase when there was a velocity excursion to greater than 2 m/s. There was a second velocity excursion to above 2 m/s, which did not produce the same response as the first. Later in the event velocity increases to 0.3 m/s did not produce load rates as great as those seen at the start of the event at velocities of 0.2 m/s. There were some additional load rate excursions near the end of the event due to spikes. Total discharge during the event was 558,000 m 3 of water, of which 119,000 m 3 or about 21% of the flow, contained load rates in the top 50%. This event is an excellent example of Page 197

198 Event 9206B PP Load Rate, Flow, and Velocity PP Load Rate PP Load Rate-Top 50% Flow 1.60 ad Rate - kg/hr o PP L Velocity m 3 /s Flow - V elocity - m/s Decimal Date 0.00 Ch. 3 Figure 25. Profiles for event UF9206B (February 10, 2002). the build-up over long inter-event times being discharged early in the event because of high canal velocities. The other major contributing event was 06B (06B , Aug. 25, 2002). Like Event , this event had a long inter-event time of 41 days. Figure 26 shows the profiles for this event. There was a first flush and pump restart response, then a steadily increasing velocity when pumping was resumed after 22 hours. This higher velocity produced several increased load rates, but after about 24 hours, continued high velocity produced no additional increased load rates. At the end of the event, a pump transient and velocities up to 0.6 m/s mobilized some additional material that had not been transported at lower velocities, but continued operation at these higher velocities produced steadily reducing load rates, indicating supply exhaustion of this more dense, less transportable material. Page 198

199 Event 9206B PP Load Rate, Flow, and Velocity PP Load Rate PP Load Rate-Top 50% 1.20 PP Load Rate - kg/hr Flow Velocity Flow - m3/s Velocity - m/s Decimal Date Ch. 3 Figure 26. Profiles for event UF9206B (August 25, 2002). Total discharge during the event was 350,000 m 3, of which 49,000 m 3 or about 14% of the flow, contained load rates in the top 50%. Event , a late August event, was preceded by Event (06B ), an early July event. Event had similar pumped volume to and an average velocity of 0.22 m/s compared to the average of 0.24 m/s for Event , however, contributed only 2 percentage points to the top 50%, compared with 11 percentage points for This comparison highlights a difficulty of data analysis for UF9206A and B, which has a sophisticated water management system and where large volumes of water, with some of its associated particulate material, can be moved to various locations and impoundments on the farm. UF9206B Year 2003 There were also no dominant events in 2003 contributing more than 30 percentage points to the top 50% at this station. However three major events contributed most of the percentage points to the top 50% during this year (Table 13). The top contributor in 2003 was Event 06B (March 27, 2003). This event had an inter-event time of 3 days, lasted for 64 Page 199

200 hours, but only contributed 7% of the total hydraulic load, and yet contributed 24.7 percentage points to the top 50% particulate P load. Total discharge during the event was 211,400 m 3. This event was responsible for the largest particulate P load of the year, accounting for 20% of the annual load. Figure 27 shows the profiles of event 06B (March 27, 2003). Flow, velocity and canal level profiles for this event are shown in Figure 27A, TSS, TDP and particulate P concentrations and loads are presented in Figures 27B and C, respectively. This event started in the late afternoon of March 27, 2003 in response to a heavy rainfall in the area that lasted for a few hours. Canal levels rapidly increased from about 1.2 m to 1.7 m (3.9 to 5.6 ft) in a couple of hours. In response, the grower started pumping with the flow rates rapidly increasing up to 1.75 m 3 /s. There was a particulate P first flush and a load rate response as flow rate and velocity increased. Figure 27B and C show the start-up flush of particulate P concentration and load at the beginning of the event. Particulate P concentration rapidly increased from about 0.1 mg/l to 1.0 mg/l, then as pumping intensity decreased particulate P concentration and load steadily decreased for the remaining of the event. This was a long continuous pumping event that lowered the canal levels from 1.7 m (5.6 ft) early in the event (March 27) to about 0.8 m (2.6 ft) in the last 24 hours of the event (March 29). Flow rate and velocity were maintained during that period, as pumping continued, canal level considerably decreased and velocity in the canal rapidly peaked up at about 0.55 m/s before it came down as the event ended. This is a good example of what can happen to flow velocity when canal levels are brought down too low. This event is a good example of high water velocities and supply and exhaustion conditions. Most of the easily transportable particulate material accumulated during the inter-event period was transported out of the farm during the first 24 hours of the event. Although water velocity drastically increased during the last few hours of the event, the supply of easily transported material has been already exhausted, so no more significant amounts of particulate P was transported out of the farm. Total dissolved P concentrations also rapidly increased at the beginning of the event from <0.05 mg/l to about >1.0 mg/l, then steadily decreased for the remaining of the event. As stated earlier, increases in TDP at the beginning of the event are difficult to explain because they respond differently than particulate P and will not be discussed as it is beyond the scope of this study. Page 200

201 Level, Flow, Rain ( m, m 3 /s, in) Canal level Flow Rain Velocity A Ve loci ty (m /s) /27/03 03/28/03 03/29/03 03/30/03 03/31/ PP Conc TDP Conc TSS Conc B TDP, PP Conc (mg/l) TSS Conc (mg/l) /27/03 03/28/03 03/29/03 03/30/03 03/31/ C PP Load Rat e TDP Load Rate TSS Load Rate TDP, PP Load rate (kg/hr) TSS Load Rate (kg/hr) /27/03 03/28/03 03/29/03 03/30/03 03/31/03 Ch. 3 Figure 27. Profiles for event UF9206B (March 27, 2003). Page 201

202 Event Analysis for UF9209A U F9209A - Year 2000 Discharge events were monitored for UF9209A for the years There is no sampling data for U F9209A - Year 2001 There was one dominant event during the year, Event 09A (09A , Sep. 8, 2001). Water management at UF9209A has some differences from the other study farms. Typical pumping procedure is to pump during the day and shut off at night, giving an approximate 8-hour on, 16-hour off cycle. Figure 28 illustrates this including an approximately 90-day period preceding event What is evident from this graphic is that was the only event in the period to have overnight pumping. The 12 events preceding had average velocities ranging from.0.04 to 0.13 m/s, maximum velocities ranging from 0.06 to 0.19 m/s, and no overnight pumping. Average velocity during was 0.13 m/s; maximum velocity was 0.25 m/s, and there was an extended pumping period that lasted for 103 hours. Figure 29 shows the profiles for There were several system failures during this event so not all of the event was sampled. Presumably if it had, the event would have contributed even more to the top 50%. The data that are available show infer a pattern. Each of the 8-hour pump cycles shows one top 50% sub-event, attributable to either pump transients or continuing high velocity. When the extended pumping is started at Day there is an initial increased load rate arising from start-up flush, then load rates reduce well before the velocity reduces. Beyond the data gap load rates continue at relatively low levels until the velocity increase at Day Shortly thereafter load rates show periodic increases and these periodic increases last well beyond Day when the velocity is reduced. Page 202

203 Selected Hydrauli c Profiles - UF9209A Flow Rate - m3/s Decimal Date Ch. 3 Figure 28. Selected pumping profile for UF9209A for year Page 203

204 Event 9209A PP Load Rate, Flow, and Velocity PP Load Rate PP Load Rate - kg/hr PP Load Rate Top 50% Flow Velocity x Flow - m 3 /s Velocity x 10 - m/s Decimal Date Ch. 3 Figure 29. Profiles for event UF9209A (September 8, 2001). In general this event showed an expected response but, point for point, the response was less well correlated with velocity than had been seen in the analyses of UF9200A and UF9206B events. The continuation of periodic elevated load rates after the velocity was reduced was the opposite of what would be predicted from supply exhaustion considerations, and in fact what is seen may be due to delayed supply. UF9209A has a total length from the upper end of the main canal to the pump station of about 4700 m, compared with about 3900 m at UF9200A and about 2700 m at UF9206B. Average velocity at UF9209A is about 0.11 m/s, compared with about 0.34 m/s at UF9200A and about 0.26 m/s at UF9206B. These numbers can be used to calculate a hypothetical transit time from one end of the canal to the other. This number has little physical meaning because of variations in upstream canal widths and changes in flow rates in the canal as you move upstream due to flow from tributary canals, but it does allow some comparison to be made. The calculated transit time is 11.9 hours for UF9209A, compared with 3.2 hours for UF9200A and 2.9 hours for UF9206B. The characteristic time for UF9209A is about four times what it is for the other farms. This could well explain the long times required for the Page 204

205 system to show the effects of continued high velocity, especially in view of the typical operating mode of only pumping for 8 hour periods. UF9209A - Year 2002 There were no dominant events in 2002, but three events, 09A (09A ), (09A ), and (09A ) combined to contribute 33.5 percentage points to the top 50%. An additional eight events contributed to the rest of the top 50% (Table 13). During 2002 the water management policies were modified at UF9209A. During 2001 and up until September of 2002, pumping was controlled such that the canal level at the pump station never fell below 1.65 m, which corresponds to a canal depth of 1.31 m at the pump station. From September 2002 on, the canal level was allowed to fall to a minimum of 0.78 m, which corresponded to a canal depth of only 0.44 m, about 34% of the previously allowed minimum depth. This change gave the opportunity for significantly higher velocities than had been seen previously. The general policy of 8-hour on, 16-hour off cycles was practiced during the year, however in mid-to-late August there were several events that ran more than 24 hours. Figure 30 illustrates the pumping profile for this period and for a period of approximately a month and a half earlier. Event (09A , Aug. 28, 2002) contributed 10.4 percentage points to the top 50%. It was the fourth of a sequence of four events that pumped continuously for 24 hours or more. The three preceding events were , , and Prior to Event , the first of the series, there had been a two-month period during which pumping duration had been limited to 8 hours or less. Events through were large events with volumes ranging from 186,000 m 3 to 486,000 m 3, but none of the three made any contribution to the top 50%. Page 205

206 Selected Hydraulic Profiles - UF9209A Flow Rate - m3/s Decimal Date Ch. 3 Figure 30. Selected pumping profile for UF9209A for year Page 206

207 Event 9209A PP Load Rate, Flow, and Velocity PP load Rate PP load Rate-Top 50% Flow Velocity PP Load Rate - kg/hr Flow - m3/s Velocity x 10 - m/s Decimal Date 0.0 Ch. 3 Figure 31. Profiles for event UF9209A (August 29, 2002). The profiles for Event are shown in Figure 31. The contributions to the top 50% came from first flush and a response to continued high velocity. The response started 32 hours after the velocity increased. What distinguished from its predecessors was the timing of the high velocity. Event lasted for 56 hours but maintained a velocity in the range of 0.1 m/s for all but the last four hours. Events 230 and 233 lasted for 28 and 24 hours respectively, and had velocities in the range of m/s for all but the last 4-6 hours. It appears from this analysis that the events preceding either had low velocity or stopped pumping before the effects of continued high velocity could be seen. Page 207

208 Event 9209A PP Load Rate, Flow, and Velocity PP Load Rate PP Load Rate-Top 50% Flow 3.5 PP Load Rate - kg/hr Velocity x Flow - m3/s Velocity x 10 - m/s Decimal Date Ch. 3 Figure 32. Profiles for event UF9209A (September 30, 2002). Events and occurred after the minimum canal depth was lowered, and show the expected response to velocity excursions. Figure 32 shows the profiles for (09A , Sep. 30, 2002). This event, which had two continuous pumping segments of 6 hours and 20 hours, had an average velocity of 0.22 m/s, which was greater than the maximum velocities of the events described previously. The contributions to the top 50% came in the second half of the 20-hour segment. This short event contributed 11.6 percentage points to the top 50%. Event (09A ) had a duration similar to , and velocity profiles that showed responses similar to Its contribution to the top 50% was 11.5 percentage points. At these higher velocities, both events showed evidence of supply exhaustion (Table 13). Page 208

209 UF9209A Year 2003 There was one dominant event during the year, Event 09A (June 19, 2003) that contributed 33.3 percentage points to the top 50% (Table 13). There were two other events, 09A (Feb. 21, 2003) and 09A (March 17, 2003), which combined contributed 27.9% to the annual top 50% particulate P load. Event 09A had an interevent time of 9 days, lasted for 151 hours, and contributed 13% of the total hydraulic load. Total discharge during the event was 211,400 m 3. This event was responsible for the largest particulate P load of the year, accounting for 23% of the annual load. Figure 33 shows the profiles of event 09A Flow, velocity and canal level profiles for this event are shown in Figure 33A, TSS, TDP and particulate P concentrations and loads are presented in Figures 33B and C, respectively. This event started early the morning of June 19, 2003 in response to some rainfall that fell during the previous 24 hours. Canal levels steadily increased from about 1.9 m (6.2 ft) to about 2.2 m (7.2 ft) during the 24 hours before pumping started. On June 19, the grower pumped for a few hours and stopped. On June 21 there was another rainfall event that made the canal level to steadily increase from about 2.0 m (6.6 ft) to 2.6 m (8.5 ft). This prompted the grower to turn on the pumps, with flow rate rapidly increasing from about 0.75 m 3 /s to 4.6 m 3 /s in a short period of time (Figure 33A). This continuous pumping event lasted more than 24 hours with flow velocities increasing up to 0.25 m/s. Figures 33B and C show the start-up flush of particulate P concentration and load at the beginning of the event. Although, pumping on June 19 lasted for a few hours, particulate P concentration increased from about 0.26 mg/l to 0.58 mg/l. In the second pumping period of the event, particulate P concentrations increased from about 0.05 mg/l to about 0.28 mg/l, then as pumping intensity decreased, particulate P concentrations and loads steadily decreased for the remaining of the event. There was one last peak of particulate P concentrations at the end of event that coincides to several hours of pumping. This event showed evidence of supply and exhaustion, especially at the beginning, when particulate P concentration rapidly increased as a result of particulate P material accumulated during the nine days of inter-event period prior to the pumping event. Page 209

210 L evel, Flow, Rain (m, m 3 /s, in ) Canal level Flow Rain Velocity A Velocity ( m/s ) /17/03 06/18/03 06/19/03 06/20/03 06/21/03 06/22/03 06/23/03 06/24/03 06/25/03 06/26/ PP Conc TDP Conc TSS Conc B TDP, PP Conc (mg/l) S Conc (mg/l) TS /17/03 06/18/03 06/19/03 06/20/03 06/21/03 06/22/03 06/23/03 06/24/03 06/25/03 06/26/03 P Load rate (kg/hr) TDP, P PP Load Rat e TDP Load Rate TSS Lo ad Rat e C TSS Load Ra te (k g/hr) /17/03 06/18/03 06/19/03 06/20/03 06/21/03 06/22/03 06/23/03 06/24/03 06/25/03 06/26/03 Ch. 3 Figure 33. Profiles for event UF9209A (June 19, 2003). Page 210

211 Data Synthesis for Key Operating Characteristics There is additional data that is pertinent to the synthesis process. This data will be presented in the next sections. The synthesis will be conducted in the Discussion section. Concentrated Suspended Solids (Bulk Samples) Analysis Concentrated discharge suspended solids samples (Bulk Samples) were obtained for 14 events at UF9200A, 15 events at UF9206B and 20 events at UF9209A over the time period 3/20/01 through 12/30/02. Tables shows a summary of key parameters associated with those samples. In general the solids are roughly 50% organic matter and 50% inorganic matter. The specific gravities of the solids are in the range of The specific gravity of cellulose is 1.7; calcium carbonate specific gravity is 2.7. A 50/50 mix of cellulose and calcium carbonate would have a specific gravity of 2.3, which corresponds roughly to the observed specific gravities. The average bulk densities and solids content of the concentrated suspended solids may be used to estimate the character of the suspended solids in their settled state in the canals. Using UF9200A as an example, the average bulk density of the settled solids was kg/l and the dry matter content was g dry mass/liter of settled solids. The average estimated annual suspended solids load from UF9200A was about 89,000 kg. Taking this load and converting to volume using the typical values of bulk density and solids content gives a total sediment volume of 1413 m 3. If it were very conservatively assumed that all the solids came from the lower 2/3 of the main canal, the effective canal bottom area would be about 18,300 m 2. Dividing the volume by the area gives a calculated depth of m (7.7 cm) or about 3 inches. Prorating this amount over a typical value of six large pumping events per year gives an estimated value of about 1.3 cm or about one-half inch of flocculant sediment exported per event. This illustration shows that the exported suspended solids volume is relatively small when present in its settled state. This must be considered when solids removal and/or control plans are being developed. Page 211

212 Ch. 3 Table 14. Selected physical and chemical characteristics of concentrated suspended solids samples UF9200A. Farm Date Bulk % Dry Specific % % Density Dry Density Gravity Ash Organic (g/cm3) (g/cm3) Matter Phosphorus Content (mg/kg) UF9200A 7/13/ /19/ /20/ /17/ ** /25/02 ** 5.0 ** ** /27/02 ** 7.2 ** ** /9/ /9/ /15/ ** /17/ /28/ /19/ /25/ ** /17/ Average Ch. 3 Table 15. Selected physical and chemical characteristics of concentrated suspended solids samples UF9206B. Farm Date Bulk Density (g/cm3) % Dry Dry Density (g/cm3) Specific Gravity % Ash % Organic Matter Phosphorus Content (mg/kg) UF9206B 3/19/ /19/ /17/ ** /29/ /30/ /4/ /6/ /12/02 ** 5.0 ** ** /15/ /28/ /19/ /25/ /11/ /13/ /17/ Average Page 212

213 Ch. 3 Table 16. Selected physical and chemical characteristics of concentrated suspended solids samples UF9209A. Farm Date Bulk Density (g/ cm3) % Dry Dry Density (g/ cm3) Specific % Ash % Gravity Organic Matter P hosphorus Content (mg/kg) UF9209A 3/20/ /19/ /17/01 ** ** /29/ /28/02 ** 6.1 ** ** /9/ /19/ /30/ /4/02 ** 7.9 ** ** /5/02 ** 5.7 ** ** /1/ /10/ /31/ /25/ /2/ /11/ /13/ /17/ ** /23/ /30/ Average Particle Size Distribut ion and Sedimentation Velocities Particle size distribution s were determined f or the bulk samples, and also for some samples that were collected by centrifug ation in the s econd half of Particle sizes were measured using standard techniques w ith a Coulter LS130 particle counter. Settling velocity determinations were done u sing a bottom withdr awal sedimentation co lumn, following Method of the U.S. Bureau of Reclamation (USBR, 1989). Calculations of velocity d istributions were done by the method of tangents described by Vanoni (1975). The settling velocity determi nations requ ired more material than was available from t he bulk sample collections, so materia l from the cen trifuge studies was used for the sedimentation studies. Centrifuged samples were resuspended and split for particle size and settling velocity determinations so the velocity distributions coincided with particle size distribution determination. Page 213

214 Figure 34 shows particle size distributions for UF9200A. These samples covered the time period from mid-july, 2001 through mid-november, There appear to be two families of distributions, those with a pe ak (Mode) of around 30 micrometers, and th ose with a peak of around 300 micrometers. All three farms exhibited this type behavior. The break from one mode to another is t ime depen dent, and takes place in the summ er of Particl e Size Distributions - UF9200A 7/ 19/01-11/19/02 4.5% 4.0% 3.5% 3.0% Volume Percent 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% Particle Diameter - micrometer Ch. 3 Figure 34. Particle size distributions for bulk samples from UF9200A. In 2001, bulk samples were taken over the period 7/19/01 through 10/29/01. In 2002 bulk samples were taken over the period 6/19/02 through 12/13/02. If the data are divided into three time-period groups the clustering into particle-size families is very evident. Figures 35a-c show the average particle size distribution for each farm for three time periods. The periods are: a) July-October, 2001, b) June-August, 2002, and c) September-December, Page 214

215 4.0% Average Particle Size Distributions July-October 2001 Volume Fraction 3.5% 3.0% 2.5% 2.0% 1.5% 1.0% UF9200A UF9206B UF9209A 0.5% 0.0% A Particle Diameter - micrometer 4.0% Average Particle Size Distributions June-August 2002 Volume Fraction 3.5% 3.0% 2.5% 2.0% 1.5% 1.0% UF9200A UF9206B UF9209A 0.5% 0.0% B Particle Diameter - micrometer 4.0% Average Particle Size Distributions September-December 2002 Volume Fraction 3.5% 3.0% 2.5% 2.0% 1.5% 1.0% UF9200A UF9206B UF9209A 0.5% 0.0% C Particle Diameter - micrometer Ch. 3 Figure 35. Particle size distributions during various time periods. Page 215

216 The distributions for all three farms were almost identical for the first time period, with average peaks in the micrometer range. In the second period, the average peaks had shifted upward to a range of micrometers for UF9206B and UF9209A, and 320 micrometers for UF9200A. In the third period, the peaks had all shifted back to the previous range of micrometers. Figure 36 shows the time sequence of the peaks of the individual curves that made up the averages. Peak (Mode) Particle Size vs Time for Study Farms Particle Siz e - micrometer UF2900A UF9206B UF9209A /1/01 7/31/01 9/30/01 11/30/01 1/30/02 4/1/02 6/1/02 7/31/02 9/30/02 11/30/02 1/30/03 Date Ch. 3 Figure 36 Time sequence of peaks of particle dize distribution curves. It is clear from these curves that through the summer and fall of 2001 the three farms had similar distributions, with characteristically low particle size as represented by the peak, or mode. In the summer of 2002, however, the particle size distributions at all three farms shifted upward, as represented by an increase in their peaks by factors of four to ten. From late summer through the fall of 2002, the particle size distributions returned to approximately their 2001 character, with about the same peak but with a slightly flatter distribution. It is interesting that the particle size distributions were similar for all three farms in 2000 given the differences in operating conditions at the farms. What is equally interesting is the apparent basin-wide shift in particle sizes in the summer of 2002, followed by a return to Page 216

217 previous conditions in the fall of The possible causes for this phenomenon are unclear at this point, but the phenomenon serves to highlight the temporal variability of particulate properties in EAA farms. The transport properties of particulate matter exported from EAA farms are of interest. Of particular interest is the settling velocity distribution of the particles being transported. In theory settling velocities can be calculated from the particle size distribution using hydraulic formulas such as Stokes Law, which applies to creeping flow that occurs at low Reynolds numbers. The problem in using such formulas to estimate velocities is that they require knowledge of the density differential between the particle and water. For a solid particle, the density is equal to the particle specific gravity, but for flocculant particles, which are porous, the effective specific gravity of the flocculated particle is less than the specific gravity of the solid material that makes up the particle. The typical specific gravity of the bulk samples was about 2.2, however using this value for flocculated material would give settling velocities that are too high. An effective density may be calculated if actual settling data are compared with calculated settling velocities and the particle density adjusted to match the two data sets. A preliminary assessment is presented here. For this study, samples were collected by continuous centrifugation to obtain sufficient material to do replicate settling column runs and particle size determinations. The example presented is from samples obtained from UF9209A in early September 2002 (the third time period discussed above). Figure 37 shows the results of matching settling column data with settling velocity particle size calculations. In this example an effective particle density of 1.8 g/cm 3 gave the best match of the two curves. The fit was reasonably good up to a settling velocity of about 140 m/day, which cumulatively represented about 70% of the system particle mass. Beyond this point the settling column showed lower velocities than Stokes Law calculations. Page 217

218 Settling Velocity Distribution - Specific Gravity = 1.8 UF9209A 10/1/ Vol. Fraction < V Stokes Law Calculation Settling Column Measurement V (Settling Velocity) - m/day Ch. 3 Figure 37. Settling velocity distributions Stokes Law vs. settling column. The departure beyond 140 m/day is relatively unimportant for several reasons. First, the accuracy range of the settling column data is determined by the spacing of sample times. In this case sample times are set to take more samples later in the run to increase accuracy in the slower settling range, so the fast settling fractions are less accurate. Stokes Law, with the appropriate density, is probably a better estimate of the rapid settling velocities than the column. Second, inaccuracy in the fast settling fractions is less important than inaccuracy in the slower settling fractions because the slower settling fractions control what is discharged. A simple example of the use of this data is to estimate the trajectory of a specific fraction of the particle mass. Consider the median velocity. Half the particle mass has settling velocities less than the median, half is greater. In our example, the median velocity is about 60 m/day, or about 0.07 cm/s. At UF9209A in 2002 the average canal depth near the pumps was 1.7 m. The annual average canal velocity was m/s. At 0.07 c/s, it would take 41 minutes for the median particle to sink from the top of the canal to the bottom. During this time the particle would have moved downstream 388 m, or almost a quarter of a mile. If the process Page 218

219 is repeated for the 25 th percentile particle, with a settling velocity of 13 m/day, the distance increases to 1785 m, or over a mile. This simplified example illustrates why retention systems close to the pump station may not be effective. Farm Sediment Surveys The objective of the farm sediment surveys was to determine the P contained in inventory in the main canal sediments, to evaluate the properties of the farm sediments, and to attempt to monitor changes in sediment character and inventory over time. Average Sediment Depth versus Time rage Se diment Depth (m) Ave UF9200A UF9206B UF9209A /1/00 1/9/01 4/19/01 7/28/01 11/5/01 2/13/02 5/24/02 9/1/02 12/10/02 Date Ch. 3 Figure 38. Average sediment depth over time. The surveys reported here cover the 22-month period from November 2000 through August 2002 for UF9200A, UF9206B, and UF9209A. Figure 38 shows the average sediment depth for each farm over that period. It appears from the graph that there was a trend toward Page 219

220 sediment accumulation over the study period at UF9206B and UF9209A, while at UF9200A sediment depth remained relatively constant. Samples from sediment cores were typically taken in 2 cm-thick layered sections at various depths from the top 12 cm of core. Occasionally deeper layers were taken. Depth is referenced to the midpoint of the section. Figures 39a-c show the P content (mg P/kg dry sediment) of the various layers over time. There were no major trends discernable in the data; oscillation appeared to be the trend over the time period studied. Comparison of the average sediment P content displayed in the figures with the data for discharged suspended solids found in Table 10 shows that the average P content of the discharged suspended solids from UF9200A for all three years, and from UF9209A for 2001, was substantially higher than that of the respective farm sediment. On the other hand, at UF9206B for all three years the average P content of the discharged suspended solids approached that of the sediment. This will receive more attention in the discussion section. The P content (on a dry basis) typically decreases as depth increases, but the bulk density of the sediment increases as depth increases. The result is that P concentration, expressed as kg P/m 3 wet sediment, typically stays constant or increases as depth increases. Figure 40 illustrates this with all farms plotted together. This relative constancy of P concentration allows extrapolations to be made with some confidence beyond the typical 12 cm sampling depth. This extrapolation was used to estimate the TP inventory in the main canals of the study farms. Page 220

221 Sediment P Content UF9200A Sediment P Content (mg P/kg Dry Sed.) cm 6 cm 11 cm /01/00 01/09/01 04/19/01 07/28/01 11/05/01 02/13/02 05/24/02 09/01/02 12/10/02 Date Average Sediment P Content UF9206B Sediment P Conten (mg P/kg Dry Sed.) t cm cm cm Average 0 10/1/00 1/9/01 4/19/01 7/28/01 11/5/01 2/13/02 5/24/02 9/1/02 12/10/02 Date Sediment P Content UF9209A 1200 Sediment P Content (mg P/kg Dry Sed.) cm 3 cm 5 cm 0 Average 10/1/00 1/9/01 4/19/01 7/28/01 11/5/01 2/13/02 5/24/02 9/1/02 12/10/02 Date Ch. 3 Figure 39. Sediment P content vs. time and depth at A) UF9200A, B) UF9206B, and C) UF9209A.. Page 221

222 Wet Sediment Phosphorus Concentration versus Depth Wet Sediment P Concentration (kg P/m 3 Wet Sediment) UF9200A UF9206B UF9209A Depth (cm) 12 Ch. 3 Figure 40. Wet sediment P concentration vs. depth. Table 17 shows the estimated average P inventories over the study period in the main canal sediments for the three study sites. Reach Distance refers to the distance upstream from the pump station over which sediment inventorying and sampling were done. Total Inventory refers to the estimated TP content in the main canal sediments. Top 12 cm Inventory refers to the measured P content in the top 12 cm of sediment, which was the typical core depth that was sampled. The inventories were divided by the total reach length in order to put them on a basis where they might be compared among farms. This is called the Normalized P Inventory, and is expressed in units of kg P/km canal length. Page 222

223 Ch. 3 Table 17. Estimated average sediment P inventories. Farm Reach Distance (km) P Inventory Total (kg) P Inventory Top 12 cm (kg) Normalized P Inventory Top 12 cm (kg/km) UF9200A UF9206B UF9209A If the total canal P inventories are compared with the annual TP loads shown in Tables 1-8, it becomes clear that there is a substantial reservoir of P in the canal sediments. The availability of this P has yet to be determined, but the total reservoir is equivalent to many years worth of P export at each farm. Normalized Sediment Top 12 cm Phosphorus Concentration versus Time 350 ion (kg/k m) Normalized Sediment P Concentrat UF9200A UF9206B UF9209A 0 10/1/00 1/9/01 4/19/01 7/28/01 11/5/01 2/13/02 5/24/02 9/1/02 12/10/02 Date Ch. 3 Figure 41. Normalized top 12 cm sediment P content variation with time. Figure 41, which shows the change in the normalized P content of the top 12 cm with time, gives a qualitative estimation of the status of the sediment P over the sampling period at the Page 223

224 three farms. It appears that the canal sediments at UF9200A and UF9206B, which are within fiv e miles of each other on the East side of the EAA, were in a P release mode, while the sediments at UF 9209A, which is on the west side o f the EAA fifteen miles from the other two farms, were in a P accumulation mode initially, followed by a P release mode. At this point in the study there are no ready explanations for this, but tracking this phenomenon over several years may yield some insight into the longer-term dynamics of P exchange between the sediments and the overlying water column. Figures 42a-c show the geometric mean top 12 cm P concentration over the sampling period at each sampling location for each farm. It is interesting to note that each farm appears to have a region of high sediment P concentration well upstream of the pump station. These locations stayed relatively constant from 2001 to Within the context of the Biological Contribution Mechanism, these locations may correspond to areas where biological material accumulates. Page 224

225 r Wet Sediment P Content - kg/m^ Geometric Mean Sediment Phosphorus Concentration versus Distance Upstream UF9200A Distance upstream - miles /m^3 Wet Sediment P Content - kg Geometric Mean Sediment Phosphorus Concentration versus Distance Upstream UF9206B Distance upstream - miles Wet Sediment P Content - kg/m^ Geometric Mean Sediment Phosphorus Concentration versus Distance Upstream UF9209A Distance upstream - miles Ch. 3 Figure 42. Sediment P concentration in top 12 cm vs. location in farm canal. Page 225

226 Discussion and Summary There is a substantial deal of heterogeneity when dealing with particulate P transport in agricultural systems. The supply of particulate P in a drainage event is generally heterogeneous, thus, transport of suspended solids is not always proportional to the transport of particulate P. The light and easily transportable material, which is generally high in P concentration, is easily mobilized under relatively mild hydraulic conditions. Continued mobilization and export leads to supply exhaustion of the material high in P concentration, although it may not lead to supply exhaustion of suspended solids low in P concentration. Biological activity in the water column can play a significant role in increasing the amount of transportable particulate P in canals, especially after long inter-event periods. However, sedimentation processes can work to slow down transportability through consolidation and mineralization of fresh deposited particulate matter. Hydraulic conditions may vary substantially over the course of a pumping event and also from event to event. Seasonal conditions can affect the physiological status of the biological population. There are some indications that changes in climatic and/or hydraulic conditions from year to year may give rise to annual changes in the physical-chemical properties of the transportable particulate P. One of the primary goals of this study was to identify conditions that cause increased particulate P load rates, and analyze those conditions to determine operating procedures that might reduce particulate P export. Load rate is the product of flow and concentration over a given unit time period. Elevated particulate P loads may occur from transport of moderate amount of high P content material. This condition is more likely to occur whenever there is a large supply of readily transportable biological material close to the pump station. This light material can be transported at moderate flow rates, for example at pump start-up after long inter-event time periods. Elevated particulate P load rates may also occur from transport of large amounts of lower P content sediment material over a short period of time. This type of elevated particulate P load rate could occur during high pumping rate events, that causes canal level to drop close to the bottom, increasing flow velocity, resulting in the dislodge and transport of base sediment material in the canal. There are obviously intermediate conditions with combination of flow and concentration that can cause increased particulate P load rates. This study searched to isolate specific conditions at each farm that could be associated with increased particulate P load rates. Page 226

227 Farms Summary Farm UF9200A, is a sugarcane operation that has shown some of the highest average canal velocities during the study period (Table 18). This farm has an average canal improvement and aquatic weed management program, with some sections of the main canal covered with aquatic weeds. However, the grower makes an effort to keep aquatic weeds away from the pump station. Discharge is controlled by selection of either high or low capacity pumps. Level control is practiced by automatic shut down and start-up of the chosen pump at canal level set points. This lead to a minimum allowable canal level, but also causes short periods (less than one hour) of pump cycling. The imposition of level control also insures that canal velocities will not exceed a certain maximum, which can be detrimental to the annual P load of the farm. Farm UF9206A/B is a mixed crop operation that has shown average canal velocities that fall in between the other two study farms (Table 18). This farm has an extensive canal improvement program and a regular aquatic weed management control. The elaborate and complex canal system of this operation allows the farmer to impound and transfer large volumes of water throughout the farm. The number and speed of pumps running at a particular station, controls the volume of water discharged. This farm does not have automatic level control, which sometimes leads to the canals being pumped close to the floor. The fact that this is a mixed crop operation that includes different kind of vegetables, require the grower to keep a closer management of the water table across the farm. Pumping is generally more extensive, frequent, and often of longer duration. This is indicated by the high pumping-to-rainfall ratios observed in Figure 10. Farm UF9209A is mainly a sugarcane operation that has shown the lowest average canal velocities of the three farms during the study (Table 18). This farm has a regular canal management program, and an aquatic weed control that falls in between the other two farms. The main canals are kept reasonably clean of aquatic weeds, but some secondary canals and field ditches have extensive weed coverage. Typical procedure is to run the pumps on a long period cycle of 8 hours on, 16 hours off. Level is controlled manually, with the pumps being turned off when a predicted canal level is reached. However, during the last 15 months, this level has been reduced, resulting in a reduction of the minimum level in the canal and the increase in the maximum allowable canal velocity. Page 227

228 Ch. 3 Table 18. Average velocity and canal depths of study farms. Farm Year Average Velocity (m/s) Average Depth (m) UF9200A UF9206B UF9209A (3.7 ft) 1.08 (3.5 ft) 1.01 (3.3 ft) 0.97 (3.2 ft) 0.91 (3.0 ft) 0.84 (2.8 ft) 0.65 (2.1 ft) 0.77 (2.5 ft) 2.08 (6.8 ft) 1.71 (5.6 ft) 1.86 (6.1 ft) There were notable variations in some measured farm key parameters over the study period. Normalized annual average particulate P loads exported from farms UF9200A and UF9209A have been fairly constant, averaging about 0.15 and 0.05 kg P/acre, respectively (0.33 and 0.11 lb/acre) (Figure 8A). Particulate P loads from farm UF9206A/B have shown a notable decrease during the first-three years of the study, declining from an average value of 0.41 kg P/acre (0.90 lb/acre) in 2000 to 0.12 kg P/ acre (0.26 lb/acre) in However, in 2003 the average particulate P load values went back to 0.24 kg P/acre (0.53 lb/acre). This increase in particulate P load may be related to an increase in total drainage volume (45% increase) and number of drainage events observed at both stations (UF9206A and B) in 2003 (Table 10). Annual average TSS and particulate P concentrations rapidly decreased from 2000 to 2001 at farms UF9200A and UF9206A/B, then remained relatively constant or showed a slight increase from 2001 to 2002 (Figures 11 and 12). In 2003, average TSS concentrations for these three farms remained constant. Annual TSS and particulate P concentration for farm UF9209A have behaved in the opposite direction. Average TSS concentration at this farm increased from 12 mg/l in 2001 to 22 mg/l in 2002, but in 2003 the concentration significantly increased to 110 mg/l. Annual particulate P concentrations also showed a Page 228

229 steady increase during the last three years, but the change was not as drastic as that observed in the TSS concentration. Increases in TSS and particulate P concentrations at this farm du ring the last three years are mainly the result of changes in operations that started in the last quarter of Pumping records shows that during the last 15 months, canal levels have been pumped closer to the bottom and the average duration of each event increased from 41 hours in 2002 to 61 hours in 2003 (Table 8). These longer drainage events combined with low canal levels in 2003, resulted in the transport of large quantities of low P bottom canal sed iments. Ph osphorus Content The annual mass average P content of discharged suspended solids was lower than anticipated in some cases. It was noted in the Introduction that there is a fixed amount of transportable high P-content m aterial in a canal system at any given time. This material is transported, along with lower P-content material from the sediment, under flow conditions in the canals. The relative rates of transport of the various fractions depend on the hydraulic conditions and the relative availability of each fraction. The size of the various inventories is dependent on antecedent hydraulic and biological conditions in the water column and in the sediment. When the inventory of readily transported high P-content material is exhausted, transport does not stop. Material of lower P-content from the sediment will continue to be transported as long as there is adequate supply and adequate mobilizing conditions. It was seen in the farm sediment survey section that there is ample sediment inventory at all farms to supply multiple years of particulate P export. Figure 43 shows the P content distribution of the top 50% loads for UF9200A in The X axis is the cumulative fraction of the top 50% loads (which totaled about 74 kg), the Y axis is the suspended solids P content corresponding to a particular load in that distribution. The first 3% (0.03 fraction) of the distribution represents sub-events where there were very high transport rates of very low P content material. This possibly resulted from bed transport (see Figure 4) arising from high local velocities or prolonged shear stress on the base sediment that resulted in rupture of the underlying low-p-content sediment on several occasions. At UF9200A the sediment P content averaged about 1200 mg/kg in 2002 (see Figure 39a). The distribution from 3% to about 47% bracketed this value. Suspended solids transported Page 229

230 in these sub-events most likely contained a preponderance of material that was sourced from base sediment erosion. At the 47% point the curve takes a sharper upward slope. From this point on the sub-events most likely contained suspended solids that were sourced primarily from relatively fresh biological material, which has a characteristically higher P content. Top 50% Load Phosphorus Content Distribution UF9200A ,000 Suspe n ded Solids Phosphorus Content - mg / kg 7,000 6,000 5,000 4,000 3,000 2,000 1, Cumulative Fraction of Top 50% of Load Ch. 3 Figure 43. Phosphorus content distribution for top 50% loads UF9200A The interpretation of this analysis is that about half of the top 50% load sub-events at UF9200A in 2002 resulted from transport of relatively freshly generated biological material, and about half resulted from transport of the underlying sediment after the supply of readily transportable material was exhausted. The concept of supply exhaustion can be used to postulate an explanation for the wide variation in values of P content observed for the various farms. Assuming that biological formation proceeds at a relatively steady pace for any given season, the accumulation of readily transportable, high P-content material will depend on how rapidly that material is transported off the farm. Conceptually, the supply of high P-content material should increase as the discharge rates decrease. Page 230

231 Phosphorus Content vs Pumping to Rainfall Ratio 3500 kg Annual Average TSS P Content - mg/ UF9200A UF9206B UF9209A Pumping to Rainfall Ratio - in/in Ch. 3 Figure 44. Correlation of pumping-to-tainfall ratio with P content. If this is the case, then P content would be negatively correlated with some measure of intensity of discharge. Figure 44 shows a graph of annual average P content versus a measure of pumping intensity, the annual pumping-to-rainfall ratios, for the farms studied. UF9206 is a composite of the A and B stations. There appears, in fact, to be a correlation, and one in the expected direction. This hypothesis helps resolve the discrepancy between expected P content of discharged suspended solids based on the Biological Contribution Hypothesis and the actual average P content observed in some cases, especially at UF9206A/B. Lower discharge rates at UF9200A in 2000 and 2001 and at UF9209A in 2001 gave high P content. Increases in those rates in 2002 gave suspended solids with lower P content. At UF9206A/B, where the sediment P content is in the range of 700 mg/kg, the discharge rates are consistently high. This farm practices regular weed control and water impoundment, which reduces the available supply of high P-content material. Because of the high Page 231

232 discharge, most of the limited available supply is exhausted, and the majority of the transported suspended solids are sourced from the base sediment. Mobilization of the base sediment at UF9206A/B is accelerated by the practice of occasionally drawing the canals down to extremely low levels with the accompanying velocity excursions. Dominant Events, Velocity, and Response Times Event-percentage point distribution showed that six of the 15 farm-years studied were dominant events that contributed 30 percentage points or more to the top 50% of the annual particulate P load. In seven farm-years two or three events contributed 30 or more percentage points. In only two of the 15 station-years there were no dominant events. This implies that a large fraction of the particulate P transport was extremely periodic in nature. These periods or episodes typically started when pumping operation deviated from typical practices, but these deviations were characteristics to each particular farm. At UF9200A, dominant events started from high pumping velocity after long inter-event times. At UF9206B, dominant events started from canal levels that were too low at the beginning of the event, or were allowed to get so low that extreme velocities were encountered. At UF9209A, dominant events started from extended pumping, and from and a deviation from the normal pattern of 8 hours on, 16 hours off. During the last 15 months of the study, this farm started to pump for longer period of time, which resulted in reduction of minimum canal depths and increases in canal velocity. The impact in canal velocity was different at each study farm. The concept of supply exhaustion applies to the transportable supply. At a given farm, if a specific maximum velocity is not exceeded, there will be material that is not transportable because it requires a velocity greater than the maximum to be mobilized. This residual supply will vary from farm to farm depending on the maximum velocity. The higher the velocity, the greater the mobilization, so farms with greater velocity would be expected to have higher amounts of solids transported. Figure 45 shows the annual average particulate P concentration versus annual average canal velocities for the three farms from 2001 to The year 2000 was excluded because there was no data from farm UF9209A, and also because the other two farms had dominant events that contributed large amounts to particulate P discharge loads. Figure 45 shows the expected response of increase in particulate P concentration with increase in canal velocity. Farm UF9209A showed the lowest average velocities and particulate P Page 232

233 concentrations of the three farms. Farm UF9206B showed a steady increase in both parameters from 2001 to Velocity and canal depth have an impact on the response times for continued high velocity. In the presentation of farm events, it was noted that the lower velocities and greater canal depths at UF9209A, caused to have a longer response time than the other two farms, before the effects of continued high velocities were observed. Thus, a velocity higher than normal will be able to mobilize previously un-mobilized material, regardless of the absolute value of the velocity. 120 Particulate Phosphorus vs. Velocity Annual Averages UF9200A Equivalent Particulate Phosphorus Concentration (ppb) UF9206B UF9209A Velocity (m/s) Ch. 3 Figure 45. Annual average particulate P vs. annual average velocity. Key Processes Demonstrated at Each Farm The diversity of the farms has allowed a number of observations to be made regarding the importance of various operating parameters in multiple contexts. Following there is a discussion of key points demonstrated at each farm. UF9200A This farm showed the least consistent aquatic weed control program and some of the highest velocities of the three farms. This combination generally increases start-up after long inter-event times. This farm had two dominant events during the four year Page 233

234 study that were characterized for their long duration and high velocities. Responses of the same magnitude were not seen at the other two farms. Short-period pump cycling continues to contribute far more than its hydraulic share to increased particulate load rates, but their frequency was slightly reduced in Short period pump cycling leads to increased particulate P loads and it is discouraged. Canal level control, when practiced without pump cycling, interrupted continued high velocity and reduced P load rates. Higher velocities and shallower canals give this farm a shorter response time than farm UF9209A. However, this farmer is reducing the average canal velocities of the farm by increasing the use of the small pump during the year. UF9206A/B This farm practiced the most regular aquatic weed control and more complex water management program of the three farms. Results from this farm suggest that the supply-exhaust mode is a very frequent occurrence during the year. Part of this is due to the fact that it has a higher rainfall-to-pumping ratio than the other two farms, but part of it is also probably due to the fact that this farm has achieved a true reduction of its highly transportable particulate P inventory. Particulate P contribution at this farm ranges between 25 to 40% of the TP load, which is a considerable reduction from its historical value of around 50%. This farm does not practice level control, and average water velocities have been steadily increasing during the four years of the study. These high velocities have resulted in the mobilization of large quantities of low P content suspended solids, resulting in an increase in the overall particulate P load rate in This farm showed the lowest annual average canal depth (UF9206B, 0.77 m or 2.5 ft) of all three farms. Elimination of low canal level practice, couple with continued optimized weed and water management practices, could categorize this farm in a most favorable condition with respect to particulate P load reduction. Page 234

235 UF9209A This farm has the advantage of having few aquatic weed in the main canals combined with wide and deep canals, which results in the lower velocities observed in the three farms. Its relative particulate P contribution is high (65-80% of TP), but its absolute contribution is the lowest of the three farms. The practice of long-period pump cycling appears to be beneficial at this farm because of the long response times. Because of its low velocities, the farm has theoretically a reservoir of readily transportable material stored along the main canal. This has been observed during the last 15 months of the study. The farm has deviated from its normal operations, increasing the pump run time and lowering minimum canal levels, causing the increase in the amount of low P sediment material to be exported out of the farm in This farm appears to be more sensitive to moderate changes in operating conditions than the other two farms. Page 235

236 Conclusions and General Recommendations The results of this study show that there are certain operating procedures that when implemented could lead to reductions in the transport of particulate P and therefore overall P export. Increased particulate P loads may occur from transport of moderate amounts of high P content material. This condition is more likely to occur whenever there is a large supply of readily transportable biological material close to the pump station. This light material can be transported at moderate flow rates, for example at pump start-up after long inter-event time period s. Increased particulate P load rates may also occur from transport of large amounts of lower P content sediment material over a short period of time. This type of increased particulate P load rate could occur during high pumping rate events, that causes canal level to drop close to the bottom, increasing flow velocity, and resulting in the dislodging and transport of base sediment material in the canal. There are obviously intermediate conditions when the combination of flow and concentration can cause increased particulate P load rates. Reducing velocity in the canals will reduce the loads of particulate P transported off the farm especially the lower P content sediment material. Reducing velocity in canals can be achieved by pumping at a slower rate for a longer period of time and practicing canal level control. In the EAA, pumping rates may be easily doubled or tripled by running multiple pumps or switching from small to large capacity pumps. Velocities may also change rapidly when canals are drawn down to low levels. Floating aquatic weeds contribute to the readily transportable biological materials close to the pump station, but the best approach for aquatic weed management may not be straight forward. Weed booms are recommended to keep the floating aquatic weeds away from the pump station. Spot chemical treatment is also recommended to control aquatic weeds and prevent major infestation. Chemical treatment of major aquatic weed infestations, however, will lead to the death and accumulation of highly transportable sediments. Following is more detailed description of general recommendations for reducing particulate P transport: Velocity Velocity is a key control parameter for reducing particulate P export. Recommended velocities are relative, in that they must be within the operating framework of the configuration of the farm. Velocities should be as low as possible, and velocity excursions should be avoided, regardless of the average or typical velocity of the canal system. Velocities greater than 0.4 m/s (1.3 ft/sec) have been associated with greater transport rates at the study farms. Given the parabolic relationship between velocity and Page 236

237 erosion, slow and long periods is preferred than fast and short periods for pumping a given volume of water. Pump Cycling and Reduced Run Times Long-run period cycling of about 8-16 hours, which reduces continuous pumping duration, has been shown to be beneficial in interrupting continued high velocity transport. This was evidenced on farms where the response time of the farm hydraulic system (i.e., the time required from pump start-up to the time when the equivalent of one volume of farm canal water is exported) is greater than the pump cycling period. Short period cycling of one hour or less is detrimental and should be avoided. Level Control Control of canal water levels is critical in avoiding major velocity excursions, and also to stay away from large deviations of the normal farm canal velocities. Lack of level control or major changes in minimum canal levels have resulted in dominant events at the two farms that did not practice strict canal water level control. Canal levels should be controlled to give minimum canal depths that do not exceed the maximum velocity recommendation. Aquatic Weed Control Weed control programs in the main canals is one of the most productive techniques in reducing the supply of high P content biomass. Physical removal along the entire length of the main canals is expensive to implement and not practical. For that reason, installation of weed-retention booms is recommended to be located at a distance >300 m (984 ft) upstream the main pump station. Spot spraying of weeds closest to the pump station is also recommended. Chemical treatment of major weed infestations will lead to the accumulation of transportable material into the bottom of the canal. Page 237

238 References Atkinson, B Biochemical Engineering and Biotechnology Handbook, 2 ed., Macmillan, New York. Behrendt, H The chemical composition of phytoplankton and zooplankton in a eutrophic shallow lake. Arch. Hydrobiol., 118: Daroub, S.H., J.D. Stuck, T.A. Lang, O.A. Diaz, and M. Cheng Implementation and verification of BMPs for reducing particulate phosphorus transport. Phase 11: Annual Report. Submitted to the Everglades Agricultural Area and the Department of Environmental Protection, April, Engle, D.L., and J.M. Melack Floating meadow epiphyton: Biological and chemical features of epiphytic material in an Amazon floodplain lake. Freshwater Biology, 23: EREC-Standard Operating Procedure Standard Operating Procedures of Water and Sediments for the Water Quality laboratory. Everglades Research and Education Center, Belle Glade, FL. Fiskell, G.A., and I.K. Nicholson Organic phosphorus content of Pahokee Muck and Spodosols in Florida. Soil and Crop Sci. So. Fla. Proc., 45:6-11. Izuno, F.T. and A.B. Bottcher The effects of on-farm agricultural practices in the organic soils of the EAA on nitrogen and phosphorus transport. Final Project Report submitted to the South Florida Water Management District, West Palm Beach, FL. Izuno, F.T., R.W. Rice, J.D. Stuck, and T.A. Lang Everglades Agricultural Area Sediments and Effectiveness of Soil Sediment Trapping in Rock Pit Diversions. Final Project Report submitted to the Florida Department of Environmental Protection and the Everglades Agricultural Area Environmental Protection District, Tallahassee, FL. Izuno, F.T. and R.W. Rice Implementation and verification of BMPs for reducing P loading in the EAA. Final Project Report submitted to the Florida Department of Environmental Protection and the Everglades Agricultural Area Environmental Protection District, Tallahassee, FL. Reddy, K.R, and W.F. DeBusk Decomposition of water hyacinth detritus in eutrophic lake water. Hydrobiologia, 211: Sharpley, A.N., T.C. Daniel, and D.R. Edwards Phosphorus movement in the landscape. J. Prod. Agric. 6: Sharpley, A.N., S.J. Smith, O.R. Jones, W.A. Berg, and G.A. Coleman The transport of bioavailable phosphorus in agricultural runoff. J. Environ. Qual. 21: Sharpley, A.N., S.J. Smith, and J.W. Naney The environmental impact of agricultural nitrogen and phosphorus use. J. Agric. Food Chem. 36: Stuck, J.D Particulate Phosphorus Transport in the Water Conveyance Systems of the Everglades Agricultural Area. Ph.D. Thesis submitted to the University of Florida, Department of Agricultural and Biological Engineering, Gainesville, FL. Stuck, J. D., F.T. Izuno, K.L. Campbell, and A.B. Bottcher Farm-level studies of particulate phosphorus transport in the Everglades Agricultural Area. Trans. ASAE 44(5): Page 238

239 Tubea, B., K. Hawksby, and R. Mehta The effects of nutrient, ph, and herbicide levels on algal growth. Hydrobiologia, 79: U.S. Environmental Protection Agency Methods for chemical analysis of water and wastes. Environ. Monit. Support Lab., Cincinnati, OH. USBR Earth manual, part 2, a water resources technical publication. U.S. Department of the Interior, Bureau of Reclamation, Washington, DC. Vanoni, V.A Sedimentation Engineering. American Society of Civil Engineering, New York. Page 239

240 CHAPTER 4 Floating Aquatic Weeds and EAA Farm P Loads List of Figures Chapter 4 Ch. 4 Figure 1. Map of BMP demonstration farm at the EREC Ch. 4 Figure 2. A masked section of main farm canal prior to image analysis Ch. 4 Figure 3. A masked section of main farm canal with open water areas delineated Ch. 4 Figure 4. Phosphorus in main farm canals at UF9200A Ch. 4 Figure 5. Phosphorus in main farm canals at UF9206A&B Ch. 4 Figure 6. Comparison of P in main farm canals at UF9200A and UF9206A&B Ch. 4 Figure 7. Unit area total, particulate, and dissolved P loads Ch. 4 Figure 8. Cumulative hydraulic and particulate P load distributions Ch. 4 Figure 9. Profiles for event Ch. 4 Figure 10. Profiles for event Ch. 4 Figure 11. Profiles for event Ch. 4 Figure 12. Profiles for event Ch. 4 Figure 13. Profiles for event List of Tables Chapter 4 Ch. 4 Table 1. Drainage statistics for particulate P demonstration farm Ch. 4 Table 2. Physiochemical statistics for particulate P demonstration farm Ch. 4 Table 3. Summary of drainage event constituent data Ch. 4 Table 4. Summary of rainfall and pumping data Ch. 4 Table 5. Canal sediment physical and chemical characteristics Page 240

241 Introduction To achieve additional reductions in the Everglades Agricultural Area (EAA) farm phosphorus (P) exports through improvements in BMP implementation, the processes of P cycling, especially particulate P production in farm canals require better elucidation. In EAA farm canals the predominant floating or emergent aquatic plant species are water lettuce (Pistia stratiotes), water hyacinth (Eichhornia crassipes), and water pennywort (Hydrocotyle verticillata). These three fast-growing aquatic plants are capable of quickly covering a farm s entire system of canals and ditches, effectively inhibiting drainage of farm fields and increasing canal drainage velocities. To achieve additional P load reductions from EAA farms, the influence of these floating aquatic weeds on farm P loads required assessment. In theory the elimination of emergent aquatic weeds should provide conditions that optimize P co-precipitation with calcium carbonate from the canal water column, a process that occurs during active photosynthesis by submerged aquatic plants growing in waters saturated with calcium carbonate (DeBusk and Dierberg, 2003). Photosynthesis-induced calcium carbonate precipitate contains P that is of low bio-availability and relatively low transportability. Optimizing P co-precipitation in main farm canals was identified as a means to encourage the sequestering of P in less mobile canal sediments and to allow for eventual recycling of canal sediments back to farm fields. The impacts of controlling emergent aquatic weed and drainage flow velocity were assessed and demonstrated at the Everglades Research and Education Center s (EREC) BMP Demonstration Farm. Aquatic weeds are key contributors to an EAA farm s P load. They serve as storage sites for P, as sources for P-laden detritus in the drainage stream, as well as sources of P returning to the water column during decay. By determining the actual magnitude of the floating aquatic weed P mass of two EAA farms over two growing seasons via aerial photography coupled with ground-truth surveys, a reliable estimate of the contribution of floating aquatic weeds to farm P loads was calculated. Aerial Survey of Emergent Aquatic Weeds The contribution of in-stream growth of aquatic flora and fauna to EAA farm P loads has been of interest since the late 1990 s (Izuno et al., 1999). Stuck et al. (2001) collected bulk drainage water samples at three pump stations serving two farms in the EAA. A comparison between the total suspended solids (TSS) and P concentrations in the TSS loads leaving the farms determined that, at high TSS concentrations (>50 mg TSS/L) P concentrations in Page 241

242 the solids asymptotically approached values of 1,000 to 2,000 mg P/kg. At lower TSS concentrations (<50 mg TSS/L; representative of most EAA drainage water samples) the P concentration in the solids approached values from 7,000 to 16,000 mg P/kg. Additionally, the P contents of field soil samples and established ditch sediment samples were less than 800 mg P/kg. The above findings are significant in that they identify the primary sources of particulate P occurring in farm drainage water. Most national research and particulate transport mitigation practices target the reduction of field soil erosion and channel sediment and bank erosion. The EAA research results show that the movement of soil particles from field surfaces and canal bed sediment to pump stations is not the primary factor contributing to high particulate P drainage water loads and, ultimately, relatively high P loads. The data show that when TSS increases, indicative of higher canal velocities and the transport of soil particles and bottom sediment, the P concentration of the particulates is less than when TSS concentrations are lower. When canal velocities are low, and minimal soil, bank, and bottom sediment erosion occurs (low TSS concentrations), P concentrations of the particulates are more indicative of biological components within the canal system, and about an order of magnitude greater than the P concentrations of soils and canal bed sediment. These findings led to the project s assertion that that particulate P and total P concentrations are greatly influenced by a biological contribution or control mechanism (BCM) (Stuck et al., 2001). Essentially, this means that elevated particulate P concentrations in the drainage stream are as dependent on channel velocities as they are on the amount of light, flocculent, organic particulates available for transport at most canal velocities. These particulates origins are from the surficial layer of the bottom sediment, from suspended aquatic macrophytic detritus and viable algae and planktonic species, and from the dislodgement of particulates from the aquatic macrophytes during flow induced agitation. Given that the BCM is the driving force between elevated particulate P and total P concentrations and loads during the five to 10 major pumping events that occur on farms during the course of a year, it was hypothesized that significant reductions in P loads could be achieved, if the aquatic macrophyte growth and senescence cycle could be managed or eliminated. It should be pointed out however that the viable aquatic plants and animals are simply part of the P cycling that is present in farm canal systems in the EAA. In gross terms, these organisms remove dissolved P from the water column and concentrate it in tissue. The P is held in the living tissue that serves as a reservoir or temporary holding area. Page 242

243 During growth, living tissue senesces and sloughs off into the water column. This flocculent detritus represents particulates of extremely high P content (7,000 to 16,000 mg/kg) until it settles out of the water column, decays, and recycles much of its P back to the water column or accretes it in sediments on the canal bed. During the year, entire plants will also die and settle to the canal bottom. Removing the aquatic plants from the canal could potentially remove large amounts of P available to entrainment during drainage. Removal of the aquatic plants as P sinks could also have the reverse effect of increasing the dissolved P concentration in the drainage water. To begin the process of garnering the necessary information to enable the determination of the feasibility of controlling the growth and senescence of aquatic weeds, main canals and field ditches, weed coverage over the course of two weed cycles was determined. Additionally, the P content of the weeds was determined in order to calculate P removal or P re-introduction potentials. To accomplish the above, two activities were undertaken. Aerial surveys of the farm main canals, coupled with ground surveys of the field ditches, were undertaken for a two-year period. These surveys yielded the total area of surface waters covered by aquatic weeds. The P mass estimate was determined through aerial surveys of the farm main canals coupled with ground surveys and physical sample collection. Aerial photographs of the main farm canals and field ditches were taken once per month from October through March and twice per month during April through September. The aerial reconnaissance cycle began in July 2000 and ended in June BMP Demonstration Farm Research has shown there are still several drainage events during the year that contribute a substantial percentage (10 to 50%) to a farm s total P load (Stuck et al, 2001). These high P loading drainage events are due to increased particulate P concentrations that are associated with lower canal levels, high flow velocities, and/or long duration quiescent periods prior to the drainage event. After scrutinizing the results accumulated from eight years of BMP implementation and assessment, it was evident that efforts should focus intensively on those specific drainage events that occur less frequently, but contribute heavily to the total P load leaving a farm. A three-year study conducted on three EAA farms investigated conditions on-farm that result in large P load events (Daroub et al., 2003). The study revealed that approximately one-half of a farm s particulate P load was comprised of recently deposited, less dense, Page 243

244 organic matter of aquatic plant origin and one-half consisted of denser, higher mineral content, canal sediments. Aquatic plants and plant detritus are key components in the development of an EAA farm s P load. They serve as sinks for soluble reactive P (SRP), as sources for P-laden detritus in the drainage stream, as well as sources of soluble P (SRP and dissolved organic P) returning to the water column during decay. Limiting the growth of the predominant floating weeds found in EAA farm canals should result in a reduction in easily transportable particulate P floc and subsequent particulate P loads. In light of these findings on particulate P sources and transport on EAA farms, the University of Florida s Everglades Research and Education Center created a BMP demonstration sugarcane farm at the research center (Figure 1). Two hydraulically isolated sugarcane blocks of 125 and 200 acres each were created and equipped with identical drainage pumps and monitoring instrumentation to record rainfall, flow, canal levels and to collect discrete hourly drainage water samples. The BMP farm was established to demonstrate to growers the operational differences between an optimized BMP (referred to as BMP in illustrations) sugarcane farm and a conventional BMP (referred to as CONTROL) sugarcane farm. Demonstration farm data, i.e., drainage volume, P species concentrations, total suspended solids (TSS), canal levels, flow velocities, and rainfall were analyzed to assess the effectiveness of optimized BMPs. The first two years of the study compared the effects of velocity control and floating aquatic weed management on particulate, dissolved, and total P farm loads. The main objectives of the demonstration farm were to apply flow velocity and floating aquatic weed controls and effectively demonstrate the resultant P load reductions. Page 244

245 N Main Pump (HC17.9TS) Control Pump 47-OP-3 47-A-10N 47-CD-10N 47-A-10S Control 47-E-10 N Block Trees 47-C-10SW 47-G-10N 47-GH-10N 47-CD-10S 47-H-10N 47-E-10S 47-F-10S 47-G-10S BMP Pump 47-GH-10S 47-I J-10N 47-KL-10N 47-J-10S 47-M-10NW 47-M-10NE 47-J-10NN 47-J-10NS 47-N-10NN 47-N-10NS VIRGIN LAND 47-KL-10S BMP Block 47-OP-10N 47-M-10SW 47-N-10S 47-OP-10S Ch. 4 Figure 1. Map of BMP demonstration farm at the EREC. Page 245

246 Objectives The project objectives targeted by this work were to demonstrate the effects of BMPs designed to control growth and remove floating and suspended aquatic plants on farm level drainage water P concentration and loads and to demonstrate farm water-management systems that could lead to greater levels of particulate matter retention. The specific tasks identified and conducted to meet the objectives were: 1) enhance the performance of sediment control BMPs that are currently employed in the EAA by growers through assessment, education, and demonstration programs that emphasize the importance of limiting farm canal drainage velocities through pump speed (flow) and canal level management, 2) assess and demonstrate the load reducing effects of controlling the predominant emergent aquatic weeds found in main farm canals throughout the EAA (Pistia stratiotes, Eichhornia crassipes, and Hydrocotyle verticillata), and 3) conduct additional BMP assessments at the sugarcane demonstration farm at the EREC and simultaneously provide a hands-on, educational venue for EAA growers that directly show the effective implementation of particulate P BMPs. Materials and Methods Aerial Survey of Emergent Aquatic Weeds The two farm sites used in the study were UF9200A, a sugarcane monoculture farm, and UF9206A, a mixed cropping system farm growing primarily sugarcane, vegetables, sod, and rice. The sites were selected because of their contrasting existing aquatic weed management practices and cropping patterns. Site UF9200A typically did very little to control aquatic weeds and rarely pumped long and vigorously enough to clear the main canal system of weeds. At UF9206A&B, however, there was a consistent and general effort to keep the main canal system clear of aquatic weeds, except for a stretch of main canal that was primarily used as a channel for extraordinary farm drainage. Additionally, practices at the site included the cultivation of rice which would clear farm ditches during the frequent flood pumping and harvest drainage periods. Page 246

247 Aerial photographs of the main farm canals were taken once per month from October through March and twice per month during April through September. Gaps in the schedule did occur due to inclement weather conditions and the grounding of flights in the Belle Glade area due to the 9/11 incident. The aerial reconnaissance cycle began in July 2000 and ended in June Digital photographs were taken with an Olympus C-3030 camera at a resolution of 2048x1536 pixels from an altitude of between 1900 and 2000 feet. Camera settings were ISO 400 at a shutter speed of 1/800th of a second. The approximate coverage of a single photograph was about 2000x1500 ft, yielding a ground resolution of about one pixel per square foot. Each photograph was taken such that two field widths (1320 ft) of main canal were covered by each frame. Using this procedure, 18 photographs were required during each event at UF9200A and 33 were needed at UF9206A&B. A total of 31 sets of photographs were taken at each farm during the study, resulting in 1581 images to analyze. Digital near infrared and multi-spectral camera units were also investigated for potential use, but they were found to be unnecessary and/or less appropriate for the needs of this study. The digital images were originally recorded as JPEG files. Each frame was opened in Corel PhotoPaint for processing. Frames were rotated as necessary and then cropped to reduce file size. The frames were then imported into CorelDraw to outline or mask the water surface in the canal section (Figure 2). At this point, some judgment calls were necessary in determining how to deal with irregular canal banks. After checking with a professional aerial photographer, and with persons experienced in these types of analyses, it was agreed that a certain amount of judgment or human intervention would be necessary during the analysis process. It was reinforced that there was a need to mask the water surface area of interest to limit the image analysis program to the area within the marked boundary that averaged the uneven canal bank. After preparing all frames, the cropped and masked frames were exported back to individual JPEG files. Page 247

248 Ch. 4 Figure 2. A masked section of main farm canal prior to image analysis. ESRI s ArcView GIS version 3.2 with the Image Analysis extension was used to determine weed coverage in a section of canal. The program converted the images to red, blue and green band composites. The total area of the masked area was determined in square units by the program. It was then decided that it would be easier and more reliable to have the program select the total area of the exposed water surface in the image rather than trying to select the weeds (Figure 3). Shadows and greatly varying color values of the weedy Ch. 4 Figure 3. A masked section of main farm canal with open water areas delineated. portions of the image resulted in unreliable and unrepeatable results. As the water surface areas were delineated, areas in square units were generated by the program and summed. The percentage of the masked area that was open water was calculated. By subtraction, the percentage of weed coverage was determined. Ground truth data from each event yielded Page 248

249 the width of the water surface and the length was determined as a multiple of field widths. Multiplying the percentage of weed cover by the surface area of the canal section yielded the area of weed coverage in square feet and square meters for main canals, field ditches, and total farm open waterways at UF9200A and UF9206A&B. Corresponding to 13 out of the 31 photographic events during the study period, mats of floating aquatic weeds were harvested from main canals and field ditches. Mat densities were variable, and at times mats with areas less than one m 2 were collected. During collection, detritus loss invariably occurred. The average aquatic weed P content at UF9200A was 1.07 g P/m2 of weeds. At UF9206A&B, the P content of aquatic weeds was 1.41 g P/m2. Over 85% of the weeds in the samples were water lettuce. It was verified that the average P content in 1 m2 of weeds from the two sites was significantly different (t-test, P=0.01). Hence, calculations for total P in weeds at each site were completed using the different values. All data from the aerial weed biomass assessment can be found in files in Appendix D on the accompanying CD. BMP Demonstration Farm Materials and Methods Two similar blocks of sugarcane fields at the Everglades Research and Education Center were identified and hydraulically segregated via canal and culvert installations (Figure 1). The first research block of 125 acres was designated a Control block and employs the array of BMP practices that are found on the EREC s BMP permit for the South Florida Water Management District (SFWMD). These practices include nutrient application control, nutrient spill prevention, soil testing, one-inch rainfall retention, and four sediment controls - level fields, field ditch sumps, discharge barrier, and slow field ditch drainage. The second research block of 200 acres was designated the optimized BMP block. This block employs the same set of BMPs as the Control block plus two additional practices of critical velocity control and floating aquatic weed control. In addition all field ditches and canals were cleaned of bottom sediments with a backhoe prior to the initiation of the study. Maximum velocity of the BMP block s main drainage canal during pumping does not exceed 0.12 m/s. This critical velocity value is an approximate critical flow velocity derived from recent on-farm particulate phosphorus load studies (Daroub et al, 2003). Velocity in this block s main canal is controlled via pump drainage flow rate (engine RPM adjustment) and canal level monitoring. Page 249

250 Each research block was equipped with a discharge pump and was drained at a rate proportional to the area of the block. Both blocks were monitored for discharge flow, canal water level, and discrete hourly sampling of discharge water over the duration of the activity. Water samples were analyzed for total suspended solids (TSS), total P (TP), total dissolved P (TDP), and particulate P. Automatic water samplers were fitted with carousels containing 24-1 L sample collection bottles. Water samples were iced and retrieved daily during pumping events. Consistent with published protocols, retrieved water samples were filtered and then ph adjusted to < 2 with H 2 SO 4 prior to storage at 4 o C and then analyzed for P. Samples were analyzed within 28 days of collection. Particulate P was calculated as the difference between total P and TDP. Water samples for TDP analysis were immediately filtered though a 0.45 µm filter-membrane, samples for TP analysis were not filtered. Analysis for TP and TDP were performed using the mercury oxide digestion method (Method 365.4, EPA 1993). Total suspended solids analysis was conducted following Method (EPA, 1993). The pump stations of each block were electronically instrumented to monitor discharge flows and collect water samples during farm drainage events. An electric relay was mounted on the pumps which triggered Dataloggers (Campbell Scientific Inc., model CR10X) to instruct autosamplers (ISCO, model 3700) to collect water samples from the drainage stream. Dataloggers also recorded rainfall measured with a tipping bucket rain gage (Campbell Scientific, Inc model TE525), canal water levels measured with submersible pressure transducers (WICA Instrument Corp.), and temperature (Campbell Scientific, Inc., model 107). Each block has an identical drainage pump that is a trailer mounted, 14 Standard Axial Flow belt driven pump powered by a 60 HP diesel engine. The pumps have a drainage capacity range of 200 to 5000 gal min -1. Flow is measured by an inline DeltaForce Magnetic Flowmeter mounted into the discharge pipes. The DeltaForce flow units consist of a DeltaMag Flowtube and a DeltaPulse monitor. This technology includes self-contained sensors that mount flush with flowtube internals from side standpipes. Each compact sensor contains a coil that generates a strong magnetic field inside the flowtube. The sensors also contain the electrodes and circuitry needed to measure the resulting flow voltage signal. The use of multiple sensors produces a flowtube that is highly insensitive to piping effects, and an inherent redundancy that permits the measurement to continue should a single sensor fail. Page 250

251 Farming practices at the two sugarcane blocks are typical of adjacent sugarcane farms. Sugarcane is the main crop with some acreage planted in rotation to sweet corn or beans. Since 1996 a standard sugarcane fertilization program based on IFAS soil test laboratory results has been followed and was applied to both blocks. Discharge total P concentrations from the EREC whole farm have averaged mg/l since 1994; the average unit area P load for the EREC whole farm has averaged 0.76 kg P Acre/yr. These unit area loads reflect the history of the EREC farm, which originally was a mixed commodity research farm that conducted vegetable, dairy, cattle, and sugarcane research. Since 1997, approximately 85% of the total cultivated acreage has been planted to sugarcane and the remaining cultivated acreage has been planted to vegetable research field trials. Floating weed retention booms were installed in May 2003 across the surface of the main canals of both blocks. The weed retention boom at the optimized BMP site is located at 400 ft from the pump intake. The purpose of this device is to trap floating aquatic plants, sequestering them downstream from the scouring zone of the pump structure. The weed retention boom at the Control block is located 30 ft from the pump intake. Since this canal does not receive any herbicide treatment, the main reason for the installation of the weed boom in this canal was to prevent clogging at the drainage pump intake. Growth of floating aquatic weeds was uninhibited in the main canal of the Control block. Floating aquatic weeds in the main canal of the BMP block were manually removed from the retention boom at project startup. Thereafter floating aquatic weeds were controlled bi- weekly as needed via manually spot spraying with glyphosate. Field ditches of both blocks are sprayed annually with glyphosate after cane harvest and before onset of the rainy season. Main canal sediment depth measurements and samples were taken after canals in the BMP block had been cleaned but prior to the study s first drainage event. Four evenly spaced transect locations were measured upstream of the pump station at each block. Canal sediment surface elevation and depth were determined at each location using a neutrally buoyant disc pad on the end of a calibrated pole. Sediment depth was determined using a calibrated penetrometer probe that was driven through the sediment to the canal bottom. Both depths were referenced to the then-current canal water surface elevation. At the same time, core samples of the sediment were taken at each transect. These core samples were then sectioned and analyzed for key physical and chemical parameters, including bulk density, solids content, specific gravity, volatile organic matter content, and P content. Page 251

252 Results Aerial Survey of Emergent Aquatic Weeds Using the average aquatic weed phosphorus contents of 1.41 g P/m2 and 1.07 g P/m2 for UF9200A and UF9206A&B, respectively, the total P stored in the aquatic weeds in the main canals at UF9200A and UF9206A&B was plotted over the duration of the project (Figures 4 and 5). As suggested by Stuck (1996), there is a definite pattern of higher weed coverage, evidenced by higher total P, at UF9200A between September and February. The cyclical pattern is clearly demonstrated in Figure 4. The fact that the pattern is evident at UF9200A suggests that this is either a natural or farming practice phenomenon since little was done to control aquatic weed populations at the site. P in Aquatic Weeds in Main Farm Canals, kg and g/m 2 UF9200A als, kg rm Can P in Aquatic Weeds in Main Fa m 2 als, g/ Pin Aquatic Weeds in Main Farm Can /6/2000 8/6/2000 9/6/ /6/ /6/ /6/2000 1/6/2001 2/6/2001 3/6/2001 4/6/2001 5/6/2001 6/6/2001 7/6/2001 8/6/2001 9/6/ /6/ /6/ /6/2001 1/6/2002 2/6/2002 3/6/2002 4/6/2002 5/6/2002 6/6/2002 Date Total P, kg Total P, g/m^2 Ch. 4 Figure 4. Phosphorus in main farm canals at UF9200A. March through August, the period of low weed coverage, also coincides nicely with the rainy season when many of the larger pumping events occur and canal levels are kept low in anticipation of heavy rains. However, this same time period is also coincident with the grand Page 252

253 growth phase of the sugarcane crop. The active growth of the sugarcane could limit nutrient availability for weed growth, the more complete ground cover could inhibit incident sunlight required for growth, and weed control practices for sugarcane growth and ditch bank maintenance could all be contributing factors along with the natural growth cycle of the weeds. Conversely, activities associated with harvest from September through February could simply preclude proper attention being paid to weed control. At UF9206A&B (Figure 5), the farmer practiced more active aquatic weed control. Although the farm has nearly twice the length of main canals as UF9200A (43,720 ft versus 23,600 ft), it can be seen that weed cover is about the same during the March through August period. This means that the percentage of weed cover at UF9200A, and hence the total P stored in aquatic weeds, was about twice as great as at UF9206A&B (Figure 6). P in Aquatic Weeds in Main Farm Canals, kg and g/m 2 UF9206A&B P in Aquatic Weeds in Main Farm Canals, kg /6/2000 8/6/2000 9/6/ /6/ /6/ /6/2000 1/6/2001 2/6/2001 3/6/2001 4/6/2001 5/6/2001 6/6/2001 7/6/2001 8/6/2001 9/6/ /6/2001 P in Aquatic Weeds in Main Farm Canals, g/m 2 11/6/ /6/2001 1/6/2002 2/6/2002 3/6/2002 4/6/2002 5/6/2002 6/6/ Date Total P, kg Total P, g/m^2 Ch. 4 Figure 5. Phosphorus in main farm canals at UF9206A& B. Generally, aquatic weeds at UF9206A&B were controlled by herbicide application, except around the main pumps and water control structures where mechanical removal occurred at times. Nonetheless, aquatic weed coverage at UF9206A&B was more constant, and Page 253

254 significantly less (percentage basis), than at UF9200A. During peak, weed growth periods, approximately 45 kg of phosphorus was held in the weeds in the approximate 2.5 miles of mains at UF9200A. While this does not appear to be a significant amount, if one were to take 45 kg P and calculate its concentration dissolved in water from a drainage event of one Total P in Aquatic Weeds in Main Farm Canals, g/m 2 UF9200A and UF9206A&B Total P, g/m /6/2000 8/6/2000 9/6/ /6/ /6/ /6/2000 1/6/2001 2/6/2001 3/6/2001 4/6/2001 5/6/2001 6/6/2001 7/6/2001 8/6/2001 9/6/ /6/ /6/ /6/2001 1/6/2002 2/6/2002 3/6/2002 4/6/2002 5/6/2002 Date UF9200A Total P, g/m^2 UF9206A&B Total P, g/m^2 Ch. 4 Figure 6. Comparison of P in main farm canals at UF9200A and UF9206A&B. inch of water from the 2-section (1280 ac) farm, the resulting concentration is 0.34 mg P L Further, if we use numbers from Stuck (1996), which showed that nearly 38% of the mass of water lettuce (dry weight basis) was sloughed off viable plants when agitated, we can calculate that the weed mass, at its peak while viable, could account for 0.13 mg P L -1 during a 1-inch drainage event. This does not include the P in the flocculent sediments in the top layer of the bed sediment, nor does it include the transport of the more consolidated bed sediments that mobilize under higher flow rates. -1. Page 254

255 BMP Demonstration Farm Results Drainage data from the two sugarcane blocks were collected from five drainage events that occurred from July 28 through December 17, A summary of the hydraulic data for each of the events for each block is presented in Table.1. Approximately 80% of the total monitored flow was sampled at the Control block and 95% of the flow at the BMP block. Physical and chemical drainage water data from each event is presented in Table 2. The data presented includes equivalent P concentrations and content for each event. The equivalent concentrations are calculated numbers and represent the total sampled mass of the entity of interest, e.g. total suspended solids, divided by the total pumping volume (during sampling) of the event or year. They represent a characteristic or mass average concentration of the sampled portions of the event or year. Similarly, the P content is calculated as the total sampled mass of particulate P divided by the total sampled mass of suspended solids, and represents the mass average P content for the sampled portion of each event. The characteristic concentrations may be used to estimate the total loads, compensating for un-sampled periods. The P content of TSS for the Control Ch. 4 Table 1. Drainage statistics for particulate P demonstration farm. Start Date Event Number Interevent Time (days) Start Decimal Date Finish Decimal Date Duration (hrs) Volume Pumped (m 3 ) Cumulative Volume (m 3 ) Volume Sampled (m 3 ) Control Farm 07/28/2003 CON /28/ :00 07/31/ : ,048 4,048 3,839 08/04/ /11/2003 CON CON /04/ :00 08/11/ :00 08/07/ :00 08/13/ : ,238 8,267 12,286 20,553 5,648 5,104 08/25/2003 CON /25/ :00 08/26/ : ,454 27,007 6,454 12/16/2003 CON /16/ :00 12/17/ : ,222 30,230 3,222 Total or Avg. 24, % BMP Managed Farm 07/28/2003 BMP /28/ :00 07/31/ : ,230 5,230 5,230 08/04/ /11/ /25/ /16/2003 BMP BMP BMP BMP /04/ :00 08/11/ :00 08/25/ :00 12/16/ :00 08/06/ :00 08/14/ :00 08/28/ :00 12/17/ : ,894 15,691 17,691 4,960 17,124 32,815 50,506 55,467 Total or Avg. 8,860 15,691 17,691 4,960 52, % Page 255

256 Ch. 4 Table 2. Physiochemical statistics for particulate P demonstration farm. Start Date Event Number TSS Load Sampled (kg) TP Load Sampled (kg) TDP Load Sampled (kg) PP Load Sampled (kg) TSS Equiv Conc (ppm) TP Equiv Conc (ppb) TDP Equiv Conc (ppb) PP Equiv Conc (ppb) % PP P Content of TSS (mg/kg) Control Farm 07/28/ /04/2003 CON CON % 83% /11/ /25/ /16/2003 CON CON CON Total or Avg , % 47% 38% 61% BMP Managed Farm 07/28/2003 BMP % 08/04/2003 BMP % 08/11/2003 BMP % 08/25/2003 BMP % 12/16/2003 BMP % Total or Avg % and BMP blocks were 1531 and 6108 mg kg -1,respectively, indicating that the solids exported from the Control block had lower P content than the BMP block TSS. A summary of the drainage event data from the two blocks is presented in Table 3. The Control block average TSS concentration of 49.3 mg L -1 was eight times greater than the TSS concentration of the BMP block (6.2 mg L -1 ). Phosphorus concentrations of drainage waters differed greatly for the two blocks. Drainage water concentrations of TP, TDP, and particulate P for the BMP block were 54, 58 and 51 % lower than Control block concentrations. It is interesting that the greatest percent reduction was observed in the TDP concentration of drainage waters. Both blocks were drained similarly on a volume per acre basis and in ratios consistent with rainfall amounts received (Table 4). The pumping to rainfall ratio for the Control block was 0.12; the ratio for the BMP block was calculated to be Average canal velocities during drainage events for the Control and BMP block were calculated to be 0.12 and 0.04 m s -1, respectively. Page 256

257 Ch. 4 Table 3. Summary of drainage event constituent data. Farm Total Drainage (m 3 ) TSS Equiv Conc (mg/l) TP Equiv Conc (ppb) TDP Equiv Conc (ppb) PP Equiv Conc (ppb) Estimated TSS Load (kg) Estimated TP Load (kg) Estimated TDP Load (kg) Estimated PP Load (kg) TSS P Content (mg/kg) Control (125 acres) 30, BMP (200 acres) 55, Ch. 4 Table 4. Summary of rainfall and pumping data. Farm Year Total Total Drainage Drainage Rainfall Pump:Rain Average Ratio Velocity (in) (m 3 ) (in/acre) (in/in) (m/s) Control , BMP , Graphical presentation of the P loads exported on a unit area basis from the two demonstration blocks is presented in Figure 7. BMP block unit area loads for TP, TDP, and particulate P were 28, 21, and 32 % lower than corresponding loads from the Control block. Load Distribution Analysis Every drainage water sample collected and analyzed has an associated set of supporting data, which includes sample time, sample duration, instantaneous flows, instantaneous levels, cumulative time since event start, and cumulative flow since event start. This data was used to calculate derived parameters such as loads, load rates, and velocities near the pump station. This was done for all samples presented in this study. For the purpose of analysis, the parameter P load rate is defined as the kg of P exported per hour. The use of load rate causes normalization among samples that might have had different sampling time durations. The load rate of a sub-event (or packet of water) defines its level of importance in Page 257

258 0.06 Estimated Unit Area Phosphorus Loads P Export (lbs/acre) Control BM P 0.00 TP TDP PP Phosphorus Species Ch. 4 Figure 7. Unit area total, particulate, and dissolved P loads. contributing to the overall annual P load. The higher the load rate, the greater the contribution of that particular sub-event or packet of water to the annual load. The data in each block s data set were ranked by P load rate, from lowest to highest. Once this was done, the cumulative hydraulic and phosphorus (TP, TDP, and particulate P) loads of the data points as ranked, were determined. Figure 8 shows the results of this analysis, with the cumulative loads expressed as a fraction of the total load. For the Control block, the curve of TSS and particulate P cumulative load rate vs. cumulative hydraulic load rate is skewed, indicating an uneven export of these two species relative to their corresponding hydraulic load. For the Control block approximately 80% of the TSS was exported in only 20% of the hydraulic load and approximately 65% of the particulate P load was exported in only 20% of the hydraulic load. The relationship between cumulative TDP load and cumulative hydraulic load was more consistent and uniform indicating relatively uniform export of TDP load across and within drainage events. All cumulative load curves from the BMP block show similar consistent relationships (trends) with respect to cumulative hydraulic load. This indicates that for the BMP block there was much less variation of P loads among sub-events when compared to the Control block, i.e. loads were exported evenly across hydraulic loads. The skewed distribution of the cumulative load rate to hydraulic load rate from the Control block indicates that demonstration efforts should be Page 258

259 directed towards exposing and investigating the Control block drainage event sub-events (hourly loads) that contributed the most to Control block TSS and particulate P loads. Event Analysis As an aid in determining and assessing the factors affecting P loads of sub-events at the Control block, the five drainage events that were sampled between July 28 and December 17, 2003 are represented in a series of graphs (Figures 4 thru 13). Each figure is comprised of two graphs, one graph each for the Control and BMP blocks. The figures provide a direct comparison of the drainage details from each block during the each event s time frame. Each graph shows the relationship between TP, TDP, and particulate P concentrations and flow velocity and flow rate over time for a specific block (Control or BMP). Drainage Event (July 31, 2003; Figure 9) is the first event sampled upon completion of installation of drainage pumps and flow monitoring equipment. Drainage flow rates -1-1 between 400 and 500 gal hr acre were imposed upon both blocks for the duration of the event. During the event canal velocities averaged 0.03 and 0.08 m sec -1 for the BMP and Control blocks; maximum hourly average canal velocities at the BMP and Control sites were 0.04 and 0.10 m sec -1. Concentrations of TP, TDP, and particulate P for the BMP block were 46, 74, and 3 % less than the respective Control block concentrations. The relatively higher TDP concentrations of the Control block are interesting to note and wil be addressed in the discussion section. Also, it is interesting to note that even though the average velocity of the Control block was more than twice that of the BMP block, no increase in particulate P concentration was observed. This would indicate that canal velocities for this event were below velocities required for substantial sediment and recently settled solids transport. Drainage Event (August 4, 2003; Figure 10) occurred over the course of three days starting on August 4, Drainage flow rates were set at 600 gal hr -1 acre -1 at the initiation of drainage and as the event progressed drainage rates fell slightly over the duration of the event. During the event canal velocities averaged 0.04 and 0.15 m sec -1 for the BMP and Control blocks, respectively. The maximum hourly average canal velocities at the BMP and Control sites were 0.06 and 0.46 m sec -1. Concentrations of TP, TDP, and particulate P in drainage water from the BMP block were 74, 49 and 79 % less than the respective Control block P concentrations. The effect of velocity on particulate P concentration can be clearly Page 259

260 bution istri Cumulative Fractional Load D TSS load PP load TDP load TP load Control Block Cumulative Fractional Hydraulic Load Distribution onal Load Distribution Cumulative Fracti TSS load PP load TDP load TP load BMP Block Cumulative Fractional Hydraulic Load Distribution Ch. 4 Figure 8. Cumulative hydraulic and particulate P load distributions. Page 260

261 CONTROL CON TP CON PP CON TDP CON V elocity CON Flow Concentration (mg/l) Velocity (m/s) Flow Rate (gal/hr/a) /31/03 06:00 07/31/03 12:00 07/31/03 18:00 08/01/03 00:00 0 BMP BMP TP BMP PP BMP TDP BMP Velocity BMP Flow Concentration (mg/l) Velocity (m/s) Flow Rate (gal/hr/a) /31/03 06:00 07/31/03 12:00 07/31/03 18:00 08/01/03 00:00 0 Ch. 4 Figure 9. Profiles for event Page 261

262 observed in the graph labeled Control in Figure 10. On August 4 and 5, 2003 as the velocity in the main canal at the Control block increased an immediate response is observed in the particulate P concentration in drainage waters. Drainage Event (August 8, 2003; Figure 11) event occurred as one 43-hour continuous pumping event at the Control block. At the BMP block the event was partitioned into three pumping periods of 30, 14, and 5 hours each. A drainage flow rate of 500 gal hr -1 acre -1 was applied to each block at the initiation of the event. Event averaged canal velocities for the BMP and Control blocks were 0.03 and 0.11 m sec -1 ; maximum hourly average canal velocities for the BMP and Control blocks were 0.04 and 0.21 m sec -1. From the Control block the effects of increased velocity on drainage water particulate P concentration is clearly evident on August 13, 2003; with the rapid increase in canal velocity a sharp increase in particulate P concentration is observed. The drainage water average TP, TDP, and particulate P concentrations from the BMP block were 30,40, and 20 % less than P concentrations from the Control block. Drainage Event (August 25, 2003) is presented in Figure 12. The Control block event was one continuous 25-hr pumping period. The BMP block partitioned the event into two long duration cycles of 33 and 19 hours each. The second cycle of pumping at this block was added to address research farm manager concerns of high water levels in cane fields in the BMP block as well as water levels in research fields adjoining the BMP block. The south border of the BMP block is adjacent to a large farm canal and at times the high water level in the canal causes seepage into the south areas of the BMP block. Event average canal velocities for the BMP and Control blocks were 0.04 and 0.10 m sec -1 ; maximum hourly average canal velocities for the BMP and Control blocks were calculated to be 0.06 and 0.19 m sec -1. The effects of increased velocity on particulate P concentration of drainage water are evident during this event at the Control site. There also appears to be spikes in particulate P concentrations at the BMP site that are unrelated to canal velocity. These spikes most likely are the result of particulates entering the drainage stream through random dislodgement of particulates from field ditches. The pumping cycling observed at the control block appeared to reduce particulate P export concentrations from those one would expect from the velocity observed. The drainage water average TP, TDP, and particulate P concentrations from the BMP block were 16, 26, and 8 % less than P concentrations from the Control block. Page 262

263 The last event of 2003, Drainage Event (December 16, 2003), is presented in Figure 13. For both demonstration blocks the event was comprised of two cycles of approximately 8 hrs each that occurred on December 16 and 17, The first and second cycle drainage -1-1 rates for both blocks were targeted at 400 and 550 gal hr acre, respectively. Event average canal velocities for the BMP and Control block were 0.03 and 0.07 m sec -1 ; maximum average hourly canal velocities for the BMP and Control blocks were calculated to be 0.04 and 0.11 m sec -1. The drainage water average TP, TDP, and particulate P concentrations from the BMP block were 57, 61, and 47 % less than P concentrations from the Control block. The largest difference in drainage water P concentration was for TDP concentrations. The difference in TDP concentrations may be the result of the complete coverage of the canals surface by water lettuce, which leads to anaerobic conditions and likely increased TDP concentrations through increased flux from sediments and plant detritus. The first cycle of canal sediment inventory was completed in June of A summary of the sediment data is presented in Table 5. The sediments from the Control block contained higher P concentrations than the BMP block sediments. At both sites, the highest TP concentrations were at the cm depth. In general, there were no major differences in bulk density or organic matter content. Page 263

264 CONTROL CON TP 0.90 CON PP CON TDP CON V elocity 800 Concentation (mg/l) Velocity (m/s) CON Flow Flow Rate (ga l/hr/ A) /04/03 00:00 08/04/03 06:00 08/04/03 12:00 08/04/03 18:00 08/05/03 00:00 08/05/03 06:00 08/05/03 12:00 08/05/03 18: /06/03 00:00 Concentration (mg/l) Velocity (m/s) BMP BMP TP 0.90 BMP PP 900 BMP TDP 0.80 BMP Velocity 800 BMP Flow Flow Rate (Gal/hr/A) /04/03 00:00 08/04/03 06:00 08/04/03 12:00 08/04/03 18:00 08/05/03 00:00 08/05/03 06:00 08/05/03 12:00 08/05/03 18: /06/03 00:00 Ch. 4 Figure 10. Profiles for event Page 264

265 CONTROL CON TP CON PP CON TDP CON V elocity CON Flow Concentration (mg/l) Velocity (m/s) Flow Rate (gal/hr/a) /11/03 00:00 08/11/03 12:00 08/12/03 00:00 08/12/03 12:00 08/13/03 00:00 08/13/03 12:00 08/14/03 00:00 08/14/03 12: /15/03 00:00 BMP BMP TP BMP PP BMP TDP BMP Velocity BMP Flow Concentration (mg/l) Velocity (m/s) Flow Rate (gal/hr/a) /11/03 00:00 08/11/03 12:00 08/12/03 00:00 08/12/03 12:00 08/13/03 00:00 08/13/03 12:00 08/14/03 00:00 08/14/03 12: /15/03 00:00 Ch. 4 Figure 11. Profiles for event Page 265

266 CONTROL CON TP CON PP CON TDP CON V elocity CON Flow Concentration (mg/l) Velocity (m/s) Flow Rate (gal/hr/a) /25/03 00:00 08/25/03 12:00 08/26/03 00:00 08/26/03 12:00 08/27/03 00:00 08/27/03 12:00 08/28/03 00: /28/03 12:00 BMP BMP TP BMP PP BMP TDP BMP Velocity BMP Flow Concentration (mg/l) Velocity (m/s) Flow Rate (gal/hr/a) /25/03 00:00 08/25/03 12:00 08/26/03 00:00 08/26/03 12:00 08/27/03 00:00 08/27/03 12:00 08/28/03 00: /28/03 12:00 Ch. 4 Figure 12. Profiles for event Page 266

267 CONTROL CON TP CON PP CON TDP CON V elocity CON Flow Concentration (mg/l) Velocity (m/s) Flow Rate (gal/hr/a) /16/03 06:00 12/16/03 12:00 12/16/03 18:00 12/17/03 00:00 12/17/03 06:00 12/17/03 12:00 12/17/03 18: /18/03 00:00 BMP BMP TP BMP PP BMP TDP BMP Velocity BMP Flow Concentration (mg/l) Velocity (m/s) Flow Rate (gal/hr/a) /16/03 06:00 12/16/03 12:00 12/16/03 18:00 12/17/03 00:00 12/17/03 06:00 12/17/03 12:00 12/17/03 18: /18/03 00:00 Ch. 4 Figure 13. Profiles for event Page 267

268 Ch. 4 Table 5. Canal sediment physical and chemical characteristics. Site Section (cm) Bulk Densty (g cm -3 ) Ash Content (%) O.M. (%) TP Content (mg kg -1 ) Sample Size (n) Control Block BMP Block cm cm cm The critical velocity (0.12 m s -1 ) imposed on the BMP block was obtained from recent on- P load research results (Daroub et al, 2003). A more precise value for critical velocity farm for the block may be identified by analyzing P concentrations and loads over time and over a range of canal velocities. Currently the study determines and records flow velocity and canal level as hourly averages. To improve the assessment and determination of critical velocity effects, flow and canal levels for both blocks in the second year of the study will be measured and averaged over a five minute as well as hourly intervals. From the BMP farm demonstration data set, P loads resulting from TDP and particulate P concentration differences were reviewed and examined. Although the study was established to measure particulate P load reductions and the largest contributor to the difference in P load between the control block and the BMP block was particulate P concentration (Table 3), there was an unexpected increase in the TDP concentrations and loads from the Control block relative to the BMP block. The fact that TDP concentrations were appreciably higher in the Control block drainage waters than the BMP drainage waters underscores the need to continue assessing the effects of floating weeds and canal velocities on farm P loads. Researchers will need to sample canal waters and monitor canal conditions between drainage events to further differentiate P species observed in the drainage waters. Two of the many questions that need to be answered include, Are there differences in P species that contribute to the TDP concentration differences?, and What conditions in the canals are causing the differences in TDP concentrations between the Control block and BMP block drainage waters?. By assessing conditions and P species concentrations in the Control block and BMP canals between drainage events, a clearer picture of P cycling within Page 268

269 farm canals may be drawn. In addition, by studying the exported solids for physical and chemical composition differences and conducting P-flux studies on the solids, we will better understand the dynamics of particulate P generation in farm canals with and without floating aquatic weeds. The results from these investigations will assist in determining the effects of the exported solids on P cycling in downstream receiving waters. Conclusions Aerial reconnaissance and simple digital photography were sufficient to determine weed coverage. Additional, sophisticated methods were explored, but did not warrant field trials since digital photography and software analysis programs have become quite sophisticated. It was demonstrated that an aquatic weed crop serves as a substantial reservoir for P during a season. A process by which the weeds could be grown as a crop and incorporated into a management practice to mitigate total P loading is not currently foreseeable. It is apparent that controlling the growth of weeds by conscientious and consistent spraying will reduce particulate production and subsequent particulate P loads. Equally apparent from this study and past studies, is that aquatic plants and animals can significantly affect the P cycle, and hence P loading, in EAA farm canals. The physical removal of weeds from open channels should help limit the available P in the water column. However, a system must be devised that minimizes major dislodgment of detritus during weed removal. There are many factors that contribute to the complicated P cycle in an EAA farm canal and ditch system. From this limited study, there appears to be no clear indication that a simple and sustainable management practice can be devised for incorporating the intentional growth and removal of aquatic weeds into a total P load reduction BMP. One goal of this task was to assess and demonstrate the combined effects of drainage flow velocity and floating aquatic weeds on the P loads exported in the drainage waters from sugarcane fields. From one partial season of drainage data, initial results confirm the hypothesis that particulate P source control (removal of floating aquatic weeds) and application of critical velocity limits lead to measurable P load reductions (Figure 2). The observed P load reduction in the BMP block most likely is a result of decreases in easily transportable particulate P as well as the absence of conditions that allow export of less transportable P sources (canal sediments). Under the study s present arrangement it is Page 269

270 difficult to determine what fraction of the P load reduction is due to source control and what fraction is due to critical velocity control. The efficacy of each practice may be determined separately. At that time both blocks could be operated with identical velocity controls and differing levels of aquatic weed control. Contributions to P Load from controlled aquatic weed growth vs. uncontrolled weed growth could then be compared. Conversely, both blocks could be operated with identical levels of control of floating aquatic weeds and different drainage velocities to compare the effect of velocity on P load. Page 270

271 References Daroub, S., J.D. Stuck, T.A. Lang, O.A. Diaz, and M. Chen Implementation and verification of BMPs for reducing P loading in the EAA, and Everglades Agricultural Area BMPs for reducing particulate Phosphorus transport. Annual Project Report submitted to the Florida Department of Environmental Protection, Tallahassee, FL. and the Everglades Agricultural Area Environmental Protection District, Belle Glade, FL. DeBusk, T.A. and F. Dierberg Capabilities and limitations of Stormwater Treatment Areas for reducing Phosphorus loads to the Water Conservation Areas: a biogeochemical perspective, GEER Program and Abstracts of the Joint Conference on the Science and Restoration of the Greater Everglades and Florida Bay Ecosystem, April Dierberg, F.E., T.A. DeBusk, S.D. Jackson, M.J. Chimney, and K. Pietro Submerged aquatic vegetation-based treatment wetlands for removing phosphorus from agricultural runoff: response to hydraulic and nutrient loading. Water Res. 36: Izuno, F.T., C.A. Sanchez, F.J. Coale, A.B. Bottcher, and D.B. Jones Phosphorus concentrations in drainage water in the Everglades Agricultural Area. J. Env. Qual. 20(3): Izuno, F.T. and R.W. Rice, eds Implementation and verification of BMPs for reducing P loading in the EAA. Final Project Report submitted to the Florida Department of Environmental Protection and the Everglades Agricultural Area Environmental Protection District, Tallahassee, FL. Sievers, P., Pescatore, D., Daroub, S., Stuck, J.D., Vega, J., McGinnes, P., and Van Horn, S Performance and Optimization of Agricultural Best Management Practices, Everglades Consolidated Report, Water Year 2002, Chapter 3, South Florida Water Management District, West Palm Beach, FL (peer reviewed). Stuck, J.D Particulate phosphorus transport in the water conveyance systems of the Everglades Agricultural Area. Ph.D. dissertation submitted to the University of Florida, Department of Agricultural and Biological Engineering, Gainesville, FL. Stuck, J. D., F.T. Izuno, K.L. Campbell, and A.B. Bottcher Farm-level studies of particulate phosphorus transport in the Everglades Agricultural Area. Trans. ASAE 44(5): Stuck, J.D., F.T. Izuno, K.L. Campbell, A.B. Bottcher, and R.W. Rice Farm-level studies of particulate phosphorus transport in the Everglades Agricultural Area. ASAE Trans. 44(5): Stuck, J.D. and T.A. Lang Everglades Agricultural Area Sediments and Effectiveness of Sediment Trapping in Rock Pit Diversions. Report submitted to Florida Department of Environmental Protection, Tallahassee, FL. U.S. Environmental Protection Agency Methods for chemical analysis of water and wastes. Environ. Monit. Support Lab., Cincinnati, OH. Page 271

272 CHAPTER 5 Effects of BMPs on Crops and Soil List of Figures Chapter 5 Ch. 5 Figure 1. Aerial view of the lysimeter field Ch. 5 F igure 2. Diagram of lysimeter plumbing and instrumentation Ch. 5 Figure 3: Vegetable/rice lysimeter planted to rice Ch. 5 Figure 4. Sugarcane lysimeter Ch. 5 F igure 5. Average 1999 monthly water table levels of vegetable/rice lysimeters Ch. 5 F igure 6. Average 1999 monthly water table levels of sugarcane lysimeters Ch. 5 Figure 7. Average soil-test Pw levels across vegetable/rice lysimeters Ch. 5 Figure 8. Average soil-test Pw levels across sugarcane lysimeters List of Tables Chapter 5 Ch. 5 Table 1. Summary of lysimeter total P budgets for a 34-month period (12/97 09/00) Page 272

273 Introduction Accelerated legislative agendas during the early 1990s produced the EAA Regulatory Program, which required growers to implement P-reduction BMPs by January Time was not available to fully study the efficacy of various proposed BMP strategies and little was known regarding potential impacts of BMP implementation on long-term soil fertility and crop production trends. To investigate these issues, a large-scale lysimeter demonstration project was designed in 1996, construction proceeded in 1997, and the lysimeter field was brought on line in December The lysimeter field included 25 lysimeters, 11 smaller units dedicated to vegetable (crisphead lettuce) and rice cropping systems and 14 larger units planted to sugarcane. The sugarcane lysimeter study was designed to demonstrate the effects of higher than traditional water table (WT) levels (that occur under BMP implementation) on 3 popular sugarcane cultivars as well as the effects of delivering nutrient-rich drainage waters (P-fertigation) to sugarcane. The vegetable/rice lysimeter study was designed to demonstrate short- and long-term soil fertility and crop nutrient uptake trends for different vegetable/rice/flooded fallow crop rotations. Drainage waters from the vegetable/rice lysimeters served as the P- fertigation source into select sugarcane lysimeters. An initial sugarcane drainage water P study conducted in the Everglades Agricultural Area (Coale et al., 1994) reported that although P concentrations of drainage waters from a sugarcane field experiment could be assumed representative of the individual field plots, the variability of measured drainage volumes from plots was too great to allow comparison of the drainage volumes and resultant P loads. The variability of drainage volumes from field plots was reportedly thought due to the variation in the underlying marl and caprock layers, which was assumed to heavily influence the lateral and vertical drainage of the research fields. There is a known and substantial body of evidence on the effect of water table depth on crop yield and its variation by crop and crop variety. Sugarcane was found to be adversely affected by a 32 cm depth water table, but unaffected by 61 and 84 cm water table depths (Gascho and Shih, 1978). Sugarcane variety yield was affected differently by water table depth. Reductions in sugarcane yield were attributed to N deficiency caused by decreased N availability in the higher water table soil. A similar EAA water table study conducted using sweet corn found that yield was not affected by imposing 60 and 85 cm water tables but Page 273

274 yield was significantly reduced in the 30 cm water table treatment (Shih, 1985). Irrigation requirements were also significantly higher for the high water table treatment. A similar study conducted earlier with sugarcane reported a similar higher water requirement for sugarcane grown under high water tables compared to sugarcane grown under standard water table depth (Shih and Gascho, 1980). Although crops may be grown at higher water tables there may some disadvantages of higher water table use with respect to overall water consumption (irrigation needed to maintain elevated water tables). Recent studies on sugarcane responses to elevated water table have reported significant varietal differences in physiological parameters such as photosynthesis, transpiration, and stomatal conductance as well as sugar yield (Glaz et al, 2004a; Glaz et al., 2004b). Time and duration of elevated water table has become the focus of current sugarcane research (Glaz, 2005). Early harvest season flood of 21 days was found to enhance sugar yield but later harvest season (November through February) flood reportedly to reduced yields of both the current crop and the succeeding ratoon crop. Low water tables may occur when irrigation water use is restricted or unavailable. The yield loss due to reduced irrigation has been reported for lettuce, celery, and sweet corn grown in the EAA (Shih, 1986). Yield losses were reportedly greatest for sweet corn and celery and less for lettuce. Lettuce suffered less from reduced irrigation water due to its relatively longer root system. In another study (Shih, 1986) sugarcane was reported to suffer no yield loss from imposition of a two month reduced irrigation water treatment. The author speculated that the long growing season and the deep root system of the sugarcane may have helped overcome any deleterious effects of reduced irrigation. Phosphorus sorption by EAA Histosols has been significantly correlated with ash content, ph, carbonates, and total and extractable Ca and P (Porter and Sanchez, 1992). Reported P sorption values on the same set of 18 EAA soils ranged from 20 to 1800 mg P kg, indicating a wide response of the soils of the EAA to added P fertilizers. Materials and Methods The lysimeter field was constructed in the SE corner of farm site UF9206A&B (Fig. 1). This area offered a roughly 3-ft soil depth profile, a convenient irrigation source, and was secluded from heavy farm traffic. Lysimeters were constructed of high density polyethylene geomembrane liner (2.5-mm thick) material used in landfill applications. The material was Page 274

275 cut and fused into box shapes on-site and then positioned in the field within sand-lined holes that were excavated to bedrock. A drainfield network (4-inch slotted polyethylene drainpipe) was positioned on each lysimeter floor, connected to 3 vertical PVC well pipes, and soil was backfilled back into the lysimeter. The drainfield network was designed to provide uniform WT responses to drainage and/or irrigation events. The central PVC well in each lysimeter housed a pressure transducer that continuously recorded WT levels. The PVC well also housed plumbing associated with drainage and irrigation events. Ch. 5 Figure 1. Aerial view of the lysimeter field. Page 275

276 A wide array of instrumentation (dataloggers, pressure transducers, flowmeters, pumps, automated water samplers) allowed the automatic tracking of drainage and irrigation event volumes, continuous WT level monitoring, and acquisition of irrigation and drainage water samples for TP analysis. From an instrumentation perspective, lysimeters were arranged in banks of either 3 or 4 lysimeters, sharing a single irrigation tank (manually filled from the adjacent field perimeter canal) with all electronic monitoring equipment linked to a single Solar panel Instrument shed Tank inlet Irrigation tank - 1 CR-10-5 water samplers with bottle coolers Discharge pump Field surface FIELD SOIL PROFILE Polyethylene liner UNDERLYING CAPROCK Sand pad Water sampler inlet with Tygon tubing Pressure transducer Pump intake Irrigation inlet Ch. 5 Figure 2. Diagram of lysimeter plumbing and instrumentation. CR-10 datalogger (Fig. 2). However, each lysimeter was equipped with a dedicated drainage event water sampler. Field research efforts also included calculations that determine P inputs as fertilizer and P exports contained within harvested crop biomass. The combination of automated lysimeter data acquisition and field assessment efforts allowed for the calculation of overall P budgets (P inputs in irrigation, fertigation, and fertilizer; P exports in drainage and harvested crop biomass) for all sugarcane and vegetable/rice lysimeter cropping systems. Page 276

277 Treatments Vegetable/Rice Lysimeters Three crop rotation treatments (TRT) allow an evaluation of the effects of vegetable (crisphead lettuce) and rice production (Fig. 3), as well as flooded fallow practices, on soil fertility and drainage water total phosphorus (TP) parameters. Drainage waters during the vegetable season were diverted as fertigation inputs into TRT_3 sugarcane lysimeters. Drainage during the rice production season served as fertigation inputs into TRT_4 sugarcane lysimeters. The 3 crop rotation treatments were: TRT_1: TRT_2: vegetables (1 or 2 crops) ---> flooded fallow ---> return to vegetable rotation vegetables (1 or 2 crops) ---> rice (1 harvest) ---> flooded fallow ---> return to vegetable rotation TRT_3: vegetables (1 or 2 crops) ---> rice (2 harvests, including the ratoon) ---> flooded fallow ---> return to vegetable rotation The crop rotation for TRT_1 lysimeters reflected a popular practice whereby fields are maintained under flooded fallow conditions prior to preparation for vegetables in the fall. The TRT_2 rotation replaced part of this flooded fallow period with a single rice crop while TRT_3 continued the rice production through the first ratoon harvest. Ch. 5 Figure 3: Vegetable/rice lysimeter planted to rice. Page 277

278 Sugarcane Lysimeters Four water-related treatments allowed an evaluation of the effects of different average WT levels and fertigation practices on sugarcane production (Fig. 4), soil fertility, and drainage water TP parameters. The 4 water-related treatments were: TRT_1: TRT_2: TRT_3: TRT_4: traditional WT levels (ranging inches below soil surface), water inputs are rainfall + irrigation (farm canal) higher-than-traditional WT levels (ranging inches below soil surface), water inputs include rainfall + irrigation (farm canal) higher-than-traditional WT levels (ranging inches below soil surface), water inputs include rainfall + irrigation (farm canal) + fertigation (lettuce drainage from vegetable/rice lysimeters) higher-than-traditional WT levels (ranging inches below soil surface), water inputs include rainfall + irrigation (farm canal) + fertigation (rice drainage from vegetable/rice lysimeters) To approximate typical field systems, sugarcane lysimeter WT levels were not simply set at a specific level, but were allowed to fluctuate over time, generally within the operational target range. Given prevailing weather and/or management goals (dry weather, heavy rains, different fertigation volumes, etc.), target boundaries were periodically exceeded, which was a reasonable reflection of WT scenarios that can occur on EAA farms. Ch. 5 Figure 4. Sugarcane lysimeter. Page 278