San Carlos Bay Water Quality Monitoring Status and Trends Report

Transcription

1 San Carlos Bay Water Quality Monitoring Status and Trends Report City of Sanibel Natural Resources Department 8 Dunlop Road Sanibel, FL 3397

2 Executive Summary The purpose of this monitoring report is to provide an update on the City s monitoring efforts in San Carlos Bay as part of the Charlotte Harbor National Estuary Program s Water Quality Monitoring Network. The information contained in this report can be used by resource managers, scientists, and policy makers for making informed decisions on how best to manage and protect one of Lee County s most valuable water resources, the Caloosahatchee estuary. This project was made possible by the Sanibel City Council and funding support was provided by the Lee County Tourist Development Council using Beach and Shoreline funds. The report was prepared by the City of Sanibel s Natural Resources Department staff. Data was collected by City staff or the City s subcontractor, Lee County Environmental Laboratory, following the Standard Operating Procedures adopted by the Charlotte Harbor National Estuary Program s () Water Quality Monitoring Network and adhering to Florida Department of Environmental Protection (FDEP) standard protocols. Annual audits were conducted by staff to ensure that the data was collected according to protocol. Lab parameters were analyzed by the Cape Coral Environmental Laboratory from March 22 through March 2 and by the Lee County Environmental Lab from May 2 to November 28. All lab analyses were done following standard methods and adhering to strict National Environmental Laboratory Accreditation Conference (NELAC) quality assurance guidelines. The reporting period runs from March 2, 22 through November 13, 28. Field parameters include temperature, dissolved oxygen, salinity, specific conductance and ph. Laboratory parameters include nitrate, nitrite, ammonia, total nitrogen, total Kjeldahl nitrogen, total phosphorus, ortho-phosphorus, chlorophyll-a, biological oxygen demand, total organic carbon, silica, color, turbidity, total suspended solids, enterococci bacteria, fecal coliform, and streptococci. Median values for each of the parameters were compared to the percentile distributions of other typical Florida estuaries from Hand (28) to determine the status of San Carlos Bay s water quality relative to other areas of the state. Water temperature, salinity, specific conductance, and ammonia were higher than average, while total Kejldahl nitrogen, total nitrogen, ortho phosphorus, total phosphorus, chlorophyll-a, color, biological oxygen demand, total organic carbon, turbidity, entercocci bacteria, fecal coliform, and fecal streptococci were lower than average. Several trends were identified in this report including increasing trends in salinity, specific conductance, and chlorophyll-a as well as decreasing trends in nitrate nitrogen, total nitrogen, and ortho phosphorus. Trends were not identified for the remaining parameters. This study agrees with others that suggest that water quality in San Carlos Bay appears to be greatly influenced by freshwater inflow from the Caloosahatchee River (Doering and Chamberlain 1999; Doering et al 2; Doering et al. 2). Our data also suggest that increased rainfall and regulatory releases from Lake Okeechobee, associated with tropical storm activity in 24 and 2, resulted in elevated chlorophyll-a concentrations in San Carlos Bay

3 Table of Contents Executive Summary 1 Description of the Study Area.. 3 Sampling Protocol... 4 Data Analysis Results and Discussion Rainfall Patterns Caloosahatchee River Flow Data from S Water Quality Status and Trends.. 13 Conclusions References Cited...17 Acknowledgments Appendix A - Output from Statistical Analyses

4 Description of the Study Area San Carlos Bay is situated at the mouth of the Caloosahatchee River in Lower Charlotte Harbor and has an area of approximately 12 mi 2 (3 km 2 ) (Fig. 1). Freshwater is supplied to the estuary by the Caloosahatchee River, Pine Island Sound, and Matlacha Pass, with the majority coming from the Caloosahatchee. The Caloosahatchee River encompasses a watershed of approximately 1,87 mi 2 (4,37 km 2 ). Due to the lack of local water storage in the system, the river also receives regulatory releases from Lake Okeechobee which has a watershed of approximately 4,4 mi 2 (11,39 km 2 ) and includes the Kissimmee River valley reaching as far north as Orlando. The Caloosahatchee River is currently listed as impaired for nutrients by the Florida Department of Environmental Protection (FDEP) under the Florida Impaired Waters Rule (2-33, F.A.C). The river receives nutrients from a number of sources including regulatory releases from Lake Okeechobee, urban and agricultural runoff, wastewater treatment plant discharge, aging septic systems, and atmospheric deposition. In December 28, the FDEP proposed a Total Maximum Daily Load (TMDL) to limit nitrogen in the tidal Caloosahatchee to 8,377,8 pounds per year (3,8 metric tons/yr). The TMDL is based on numeric water quality targets developed by Corbett and Hale (2) that are protective of seagrasses in San Carlos Bay. Figure 1. Map of Lower Charlotte Harbor with study area identified by the polygon at the mouth of the Caloosahatchee River

5 Sampling Protocol Samples were collected as part of the Charlotte Harbor National Estuary Program s () Coastal Charlotte Harbor Water Quality Monitoring Network. The program uses a random-stratified approach that divides the study area into twelve sections or stratum. Each stratum is sampled by staff from state and local governments including the FDEP, the Southwest Florida Water Management (SWFWMD) District, the South Florida Water Management District (SFWMD), the Florida Fish and Wildlife Research Institute (FWRI), Lee County, and the City s of Sanibel and Cape Coral (Fig. 2). The random-stratified approach allows for more robust statistical analysis that is representative of the overall water quality in each stratum. Since it is a random sampling design, as opposed to fixed stations, more samples are required in order to detect trends in the data. Each stratum is divided into equal sections using a grid system. On a monthly basis each agency generates five sampling locations in each stratum using a random number generator. These numbers are then translated to coordinates and are used to locate the sampling site. As partners with the, the City of Sanibel collects data from the San Carlos Bay stratum (Fig. 2). This report covers only those samples collected by the City of Sanibel or their contractor, Lee County Environmental Laboratory. All data analysis was done by staff from the City s Natural Resources Department. Samples were collected by boat using a.2 liter Wildco Alpha Bottle TM or similar sampling device. Water samples were collected.m from the surface and where the overall water depth was greater than 3.m, samples were also taken.m from the bottom. All field measurements were also taken.m from the surface and.m from the bottom and were recorded using a Hydrolab Minisonde 4a with a Hydrolab Surveyor 4a data logger or a YSI XL data sonde with a MDS data logger. Laboratory analyses were completed by either the Cape Coral Environmental Lab (prior to May 2) or the Lee County Environmental Lab (April 2 to present). The field parameters collected include: temperature (ºC), dissolved oxygen (mg/l), salinity (psu), and specific conductance (µmohs). Lab parameters include ammonia nitrogen (mg/l), nitrate nitrogen (mg/l), nitrite nitrogen (mg/l), total Kjeldahl nitrogen (mg/l), total nitrogen (mg/l), ortho phosphorus (mg/l), total phosphorus (mg/l), chlorophyll-a (µg/l), color (cu), biochemical oxygen demand (-day) (mg/l), total organic carbon (mg/l), total suspended solids (mg/l), turbidity (ntu), enterococci (colonies/1ml), fecal coliform (colonies/1ml), fecal streptococci (colonies/1ml), and silica (mg/l). Biological oxygen demand and silica were added in May 2 and fecal streptococci was discontinued in March

6 Figure 2. Coastal Charlotte Harbor Monitoring Network (Corbett 24) The data used in this report extends from March 22 through November 28. Seventy-two monthly sampling events were conducted and approximately 3 samples collected for each of the listed parameters. Some parameters were dropped or added based on utility. Rainfall data was obtained from the Lee County Natural Resources Division hydrological monitoring data website: Data was collected by Lee County staff from eighteen sites throughout the Caloosahatchee and Estero Bay watersheds and provided a good approximation of the rainfall patterns within the study area (Fig. 3). - -

7 Figure 3. Map showing the location of the Lee County rain gauges (Lee County 28). Legend Rain gauge Data Analysis Monthly station profile data were averaged to obtain a single monthly value for each sampling station (n=3). Station data were then averaged to get a single monthly value representative of the mean monthly water quality for the San Carlos Bay stratum (n=72). There was no significant difference between the median values of the station data versus monthly averages (t-test, p >.49) so monthly averages were used for the trend analysis and monthly station profile data were used for comparison with the typical values of other Florida estuaries (Hand 28) and Criteria for Surface Water Quality for Class III waters, F.A.C. One-half of the minimum detection limit (MDL) was used for samples where the values were below the MDL. All equipment blanks and duplicates were excluded from the analyses. Mean, median, range, and standard deviation were calculated for all station data as well as for the monthly average data (Table 1). Water quality status for San Carlos Bay was determined by comparing the median values for each of the parameters to the typical values of other Florida estuaries provided by Joe Hand (28) (Table 2). The values used for comparison are updated values to those used in Hand (1989), Typical Water Quality Values for Florida s Lakes, Streams, and Estuaries. In order to estimate San Carlos Bay s water quality relative to typical Florida estuaries, percentile distributions were grouped in the following categories: 1-3 = below average, 31-9= average, 7-9= above average (Duffey et al. 27) (Table 2). - -

8 Table 1. Descriptive statistics and percentile distributions relative to typical Florida estuaries (Hand 28) n Mean Median Min Max Standard Deviation Median compared to Typical FL Estuary Percentiles (Hand 28) Status Relative to Typical FL Estuaries Trend Results (α=.) Water Temperature Higher than Average No Trend Dissolved Oxygen Average No Trend Salinity Higher than Average Increasing Specific Conductance Higher than Average Increasing ph Average No Trend Ammonia Higher than Average No Trend 1 Nitrate Average Decreasing Nitrite Average No Trend Nitrate + Nitrite Average Decreasing Total Kjeldahl Nitrogen Lower than Average No Trend Total Nitrogen Lower than Average Decreasing 2 Ortho Phosphorus Lower than Average Decreasing Total Phosphorus Lower than Average No Trend Chlorophyll-a Lower than Average Increasing Color Lower than Average No Trend BOD -day Lower than Average No Trend Total Organic Carbon Lower than Average No Trend Total Suspended Solids Average No Trend Turbidity Lower than Average No Trend Silica N/A N/A No Trend Enterococci Lower than Average No Trend Fecal Coliform Lower than Average No Trend Fecal Streptococci Lower than Average No Trend 1 Trends identified using data from the entire period of record were the result of lower minimum detection limits beginning /22/ when the City changed contractors. Data from /22/ through 11/13/8 did not reveal a trend. 2 Trend results from the Seasonal Kendall test were significant (P<.), but results of the Mann-Kendall test were not significant at the 9% confidence interval

9 Table 2. Typical Florida Estuarine Water Quality Percentile Distributions (Hand 28) Percentile Temp DO Salinity Spc Cond ph Ammonia Nitrate Nitrite Nitrate + Nitrite TKN TN Ortho P No. of Waterbodies Percentile TP Chl-a Color BOD-Day TOC TSS Turbidity Silica Enterococci Fecal Coliform Fecal Strep NA NA NA NA NA NA NA NA NA No. of Waterbodies N/A

10 Time series plots and annual box and whisker plots were created for each of the parameters. The time series plots show the change in a parameter over time while box and whisker plots graphically display the range of the data as vertical bars (i.e. whiskers), the 2 th and 7 th percentiles identified as the upper and lower horizontal bars, the median value ( th percentile) identified by the middle horizontal bar, and the mean value identified by the plus sign. A statistical outlier test was done using Dixon s outlier test. This test identifies points that do not appear to fit the distribution of the rest of the data (Sanitas Technologies 27). Since the data used were monthly averages of the station profile data, outliers were not considered to be the result of sampling error. Non-parametric tests were used for the trend analysis, which do not assume equal variance, and therefore all outliers were left in the dataset. Tests for seasonality were done using the methods described in Statistical Analysis of Ground-water Monitoring Data at RCRA Facilities, Interim Final Guidance (U.S. EPA 1989) to identify parameters that exhibited statistically significant seasonal variation. The Mann-Kendall test for temporal trends (Kendall 1938, Hollander & Wolf 1973) and Sen s Slope estimate (Gilbert 1987) were used to identify trends in the data over time. The trend analysis can be used to determine the significance of a trend and estimate its relative magnitude. A Seasonal Kendall test was used to evaluate the influence of seasonal variation on the data. For this test, wet season was defined as June through October and dry season defined as November through May. Seasons were defined by averaging monthly rainfall for the study period (22-28) (Fig. 4). The Mann-Kendall test, Sen s Slope estimate, and the Seasonal Kendall are non-parametric tests and do not require data to be normally distributed. All values below the detection limits are assigned a value of one-half the detection limit. The confidence level for these tests is 9% and p<. is considered significant. Trend analysis was performed using flow adjusted and non-flow adjusted data. Flow adjustment was done following the methods described by Helsel and Hirsch (1992) and was used to evaluate the influence of river flow on water quality trends. Flow data used for flow adjustment was a 3- day average prior to sampling. Statistical analysis was done using Microsoft Excel 23 (Microsoft Corporation) and WQStat Plus (Sanitas Technologies 27). The complete results of each of these tests can be found in Appendix A. Results and Discussion Annual rainfall within the study area averaged. inches during the study period (22-28). On average, 9% of the rainfall was recorded from June through September (wet season) with the remaining 31% recorded between October and May (dry season). Annual rainfall was lowest in 27 (3.34 inches), while the highest annual rainfall was recorded in 2 (9.43 inches) (Fig. & Table 3). The 24 and 2 hurricane seasons were very active with five named storms making landfall in south Florida. Rainfall associated with tropical storm activity increased stormwater runoff within the Lake Okeechobee and Caloosahatchee watersheds and resulted in higher than normal basin inflows and large regulatory releases from Lake Okeechobee

11 Figure 4. Average monthly rainfall Lee County, FL Rainfall (inches) 4 2 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure. Average annual rainfall Lee County, FL Rainfall (Inches)

12 Table 3. Monthly rainfall Lee County, FL Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Total Flow data recorded by the U.S. Army Corps of Engineers at S-79, Franklin Lock and Dam, were obtained from the SFWMD DB-Hydro water quality database. Average annual flows were highest in 2 and lowest in 27 (Appendix A; Fig. 32.2) and corresponded with local rainfall patterns (Fig. 7). Average monthly flows generally followed local rainfall patterns and exhibited seasonal variation, except during times when regulatory releases from Lake Okeechobee occurred outside of the typical rainy season (Fig.). There was an inverse relationship between mean monthly flow and salinity (Fig. 8). Figure. Hydrograph of average daily flow for the Caloosahatchee River recorded at S Flow (cfs) 1 Jan-2 Oct-2 Jul-3 Apr-4 Jan- Oct- Jul- Apr-7 Jan-8 Oct

13 Figure 7. Hydrograph of average monthly rainfall and average monthly flow at S Average Monthly Rainfall Monthly Average Flow at S Rainfall (inches) S-79 Flow (cfs) Jan-2 Oct-2 Jul-3 Apr-4 Jan- Oct- Jul- Apr-7 Jan-8 Oct-8 Figure 8. Hydrograph of average monthly salinity and flow at S-79 the day of sampling Salinity Flow at S-79 day of sampling Rainfall (inches) S-79 Flow (cfs) Mar-2 Dec-2 Sep-3 Jun-4 Mar- Dec- Sep- Jun-7 Mar

14 Water Quality Status and Trends Water temperature exhibited seasonal variation typical of southwest Florida (Appendix A; Fig. 9.1). Average annual temperatures varied less than ºC with lowest average annual temperatures recorded in 24 and highest in 22. Compared with other Florida estuaries, water temperature was higher than average (Table 1), which is likely do to the geographic location of San Carlos Bay relative to other Florida estuaries. Dissolved oxygen varied seasonally with higher concentrations recorded during the winter and lower concentrations in the summer. This pattern is expected since colder water holds more oxygen than warmer water. Nine of the samples (n=3) had values below 4 mg/l and was below the Florida Surface Water Criteria for Class III marine waters 2.% of the time. Four out of 72 months had monthly averages below mg/l (Appendix A; Fig. 1.1). Compared to other Florida estuaries dissolved oxygen concentrations were average (Table 1). Mean annual salinity was greatest in 27, corresponding with the drought, and was lowest in 23. Average monthly salinity in the bay fell below 2 psu five out of 72 months sampled (Appendix A; Fig. 11.1), corresponding with high freshwater flows recorded at S-79 (Fig. ). Salinity also exhibited an inverse relationship with flow (Fig. 8). There were significant increasing trends identified by both Mann-Kendall and Seasonal Kendall tests using data that was not adjusted for flow (p<.) (Appendix A; Fig. 11. & 11.7). No trends were identified in flow adjusted data indicating that the decreasing trend in salinity is the result of decreasing flow to the estuary over the period of record (Appendix A; Fig. 11. & 11.8). Compared to other Florida estuaries, salinity was higher than average (Table 1) and is likely the product of the project boundaries, which do not include the river portion of the estuary where lower salinities often occur. Specific conductance followed nearly identical patterns to salinity (Appendix A; Figs & 12.1). Since conductance is a measure of the water s ability to transmit an electrical current and is proportional to the amount of dissolved solids, such as chlorides in the water, it is commonly used as a proxy for salinity. Conductance, like salinity, had significant increasing trends in both Mann-Kendall and Seasonal Kendall tests using data that was not adjusted for flow (p<.) (Appendix A; Fig. 12. & 12.7). No trends were identified in flow adjusted data indicating that the decreasing trend in conductance is the result of decreasing flow to the estuary over the period of record (Appendix A; Fig. 12. & 12.8). Compared to other Florida estuaries, specific conductance was higher than average (Table 1) and is probably the product of the project boundaries. Mean monthly ph values exhibited significant seasonal variation (p<.), but varied little between years. There were no significant trends identified at the 9% confidence interval and values were average compared to other Florida estuaries (Table 1). Ammonia nitrogen was below the MDL (<.1) from March 22 through March 2 (Appendix A; Fig. 14.1). It is apparent that this was an equipment limitation since samples analyzed by the Lee County lab, where the MDL was substantially lower (<.14), were able

15 to detect ammonia. Compared to other Florida estuaries, ammonia levels were higher than average (Table 1), which is also likely the result of equipment limitations and the higher detection limit used by the Cape Coral lab. No significant trends were detected from May 22, 2 through November 13, 28. Mean annual nitrate nitrogen values were highest in 2 and lowest in 27, corresponding with years that received the highest and lowest rainfall, respectively (Appendix A; Fig. 1.2). Compared to other Florida estuaries, nitrate values were average (Table 1). There were significant decreasing trends identified by both Mann-Kendall and Seasonal Kendall tests using data that was not adjusted for flow (p<.) (Appendix A; Fig. 1. & 1.7). No significant trends were identified in flow adjusted data indicating that the decreasing trend in nitrate is the result of decreasing flow to the estuary over the period of record (Appendix A; Fig. 1. & 1.8). Mean monthly nitrite nitrogen concentrations were below the MDL (<.1) from March 22 through March 28 (Appendix A; Fig.1.1). Similar to ammonia, this was likely the result of an equipment limitation since samples analyzed by the Lee County lab, where the MDL was substantially lower (<.2), were able to detect nitrite. Compared to other Florida estuaries, nitrite values were average (Table 1) and no significant trends were detected (Appendix A; Fig ). Total Kjeldahl nitrogen varied seasonally and between years with mean annual concentrations lowest in 2 and highest in 23 (Appendix A; Fig. 18.2). Median concentrations were lower than average compared to other Florida estuaries (Table 1) and no significant trends were detected (Appendix A; Fig ). Total nitrogen was lower than average compared to other Florida estuaries (Table 1). The Seasonal Kendall test, using non-flow corrected values, identified a significant decreasing trend (p<.) (Appendix A; Fig. 19.7); however trends were not significant using the Mann- Kendall with non-flow adjusted values or for either test using flow adjusted values (Appendix A; Figs. 19., 19. & 19.8). This suggests that seasonal variability and decreases in freshwater inflow are influencing the trend. Mean monthly ortho phosphorus values were below the MDL (<.) from 22 through 24 (Appendix A; Fig 2.1). This was likely the result of an equipment limitation since samples analyzed by the Lee County lab, where the MDL was <.4, frequently detected ortho phosphorus. When we initially did the trend analysis we discovered that the higher MDL used by the Cape Coral lab was influencing the trend. In order to remove the influence of the MDL, the trend analysis was done using only data from May 2 through November 28. There was a significant decreasing trend using non-flow adjusted values (p<.) for both Mann- Kendall and Seasonal Kendall tests (Fig. 2. & 2.7). There was also a trend identified using flow adjusted values for the Seasonal Kendall (Appendix A; Fig. 2.8), but not the Mann-Kendall (Appendix A; Fig. 2.). This indicates that decreasing trends in ortho phosphorus may be attributed to lower flows from 2 to 27 (Fig. 2. & 2.8). Compared to other Florida estuaries, ortho phosphorus values were lower than average (Table 1)

16 Total phosphorus mean annual concentrations were highest in 23 and lowest in 24. Compared to other Florida estuaries, total phosphorus was lower than average (Table 1) and no significant trends were detected (Appendix A; Fig ). Chlorophyll-a exhibited a high degree of interannual variability with average annual values in 2 over five times greater than those recorded in 2, the next highest year (Appendix A; Fig. 22.2). Mean monthly concentrations exceeded 114 µg/l on August 3 th 2 lagging approximately one month behind average monthly flows at S-79 that exceeded 11, cubic feet per second (Fig. 7 & Appendix A-Fig. 22.1).The spike in chl-a is most likely associated with a blue-green algae bloom (Microcystis spp.) that occurred in the lower estuary following high rainfall within the watershed and regulatory releases from Lake Okeechobee. Mean monthly values exceeded the proposed TMDL light target in San Carlos Bay (<4 µg/l) 17 of the 72 months sampled. Average chlorophyll-a values for the entire study period were 8.12 µg/l (Table 1). Despite elevated concentrations in 2, compared to other Florida estuaries, median chlorophyll-a values were lower than average (Table 1). A significant increasing trend was identified by both Mann-Kendall and Seasonal Kendall tests using non-flow adjusted and flow adjusted values (Appendix A; Fig ). Mean annual color values were highest in 2 and lowest in 27, corresponding with the wettest and driest years. Mean monthly values were greatest in August 28 and July 2, respectively (Appendix A; Fig. 23.1). Compared to other Florida estuaries, color was lower than average (Table 1) and no significant trends were detected (Appendix A; Fig ). Biochemical oxygen demand was added as a parameter in May 2. Mean annual values were highest in 2, while the lowest values were recorded in 27. Compared to other Florida estuaries, values were lower than average (Table 1) and no significant trends were detected (Appendix A; Fig ). Total organic carbon concentrations were greatest in 27; lowest in 24 and exhibited a high degree of interannual variability (Appendix A; Fig. 2.2). Total organic carbon was lower than average (Table 1) when compared to other Florida estuaries and no significant trends were detected (Appendix A; Fig ). Mean annual concentrations of total suspended solids were highest in 22 and lowest in 24. Compared to other Florida estuaries, total suspended solids were average (Table 1) and no significant trends were detected (Appendix A; Fig ). Mean annual turbidity concentrations were highest in 2 and lowest in 27, corresponding with the wettest and driest years recorded during the period of record (Appendix A; Fig. 27.2). Compared to other Florida estuaries, turbidity was lower than average (Table 1) and no significant trends were detected (Appendix A; Fig ). Silica was added as a parameter in May 2 and exhibited only slight variation between 2 and 28. Values from other Florida estuaries were not available from Hand (28) for this parameter. No significant trends were detected (Appendix A; Fig ). - 1-

17 Enterococci bacteria was added as a parameter in May 2. The highest recorded value was 4 cfu, which is well within the good category (-3 cfu) for the Florida Healthy Beaches monitoring criteria. Compared to other Florida estuaries, enterococci levels were lower than average (Table 1) and there were no significant trends detected (Appendix A; Fig ). Fecal coliform values did not exceed 8 cfu during the period of record and were well within the good category (-199) for the Florida Healthy Beaches monitoring criteria. Compared to other Florida estuaries, fecal coliform values were lower than average (Table 1) and no significant trends were detected (Appendix A; Fig ). Fecal streptococci bacteria was only measured from Highest monthly averages were recorded in 23 and lowest in 22 and 2. Compared to other Florida estuaries, values were lower than average (Table 1) and no significant trends were detected (Appendix A; Fig ). Conclusions From 22-28, a high degree of interannual variability in precipitation and freshwater inflow was experienced throughout the study area. Annual average rainfall was greatest in 2 and lowest in 27. Hurricane activity in 24 and 2 was above average, with five named storms passing in close proximity to Lake Okeechobee. Strong winds stirred up bottom sediments increasing nutrient concentrations in the lake (Havens 2; Rogers and Allen 28). In an effort to protect the Herbert Hoover Dike from potential breeches, the U.S. Army Corps of Engineers began high volume regulatory releases to the St. Lucie and Caloosahatchee estuaries in March 2. High rainfall within the watershed coupled with regulatory releases from the lake resulted in average monthly flows at S-79 exceeding 11, cfs in July and 8, cfs in August of 2. The results of this study are consistent with other research suggesting that water quality in San Carlos Bay is greatly influenced by freshwater inflow from the Caloosahatchee River (Doering and Chamberlain 1999; Doering et al 2; Doering et al. 2). Results from flowadjusted Seasonal Kendall trend analyses indicated that trends for salinity, conductivity, nitrate, total nitrogen, and ortho phosphorus were related to changes in freshwater inflow. These data also suggest that increased rainfall and regulatory releases from Lake Okeechobee resulted in elevated chlorophyll-a concentrations in San Carlos Bay. Following the period of highest flows (July-August 2), average chlorophyll-a concentrations peaked in San Carlos Bay at 114 µg/l in August 2 and 92 µg/l in September 2. Additionally, this study identified a decreasing trend in flow at S-79 associated with a decreasing trend in total nitrogen in the bay, suggesting that observed increases in chlorophyll-a are likely related to increased nutrient loads as a result of increased flows. Doering and Chamberlain (1999) also reported on the relationship between flow and chlorophyll-a and found that the location of the chlorophyll-a maximum is determined by the flow velocity. At low flows chlorophyll-a decreased with distance from S-79, at intermediate flows chlorophyll-a peaked in the upper - 1-

18 estuary and decreased further downstream, and at high flows chlorophyll-a was lower at the head of the estuary and higher downstream. Resource-based numeric water quality targets were developed for Charlotte Harbor by Corbett and Hale (2) and have recently been adopted by the Florida Department of Environmental Protection to establish a Total Maximum Daily Load (TMDL) for nutrients in the tidal Caloosahatchee. These targets suggest that in order to maintain a healthy seagrass population in San Carlos Bay, a subsurface irradiance of 2% at 2.2 meters below the surface must be maintained. In order to achieve this target, chlorophyll-a concentrations should be maintained at less than 4 µg/l (Corbett and Hale 2). However, from the median concentration for chlorophyll-a was 8.12 µg/l, over twice the concentration recommended to meet the light target for the bay. Through implementation of the TMDL and the Basin Management Action Plan process, it is expected that reductions in nutrient loading will result in lower chlorophyll-a concentrations in the bay. These reductions will likely increase light available to seagrasses and improve fisheries productivity. References Cited Corbett C. A. 24. Coastal Charlotte Harbor Monitoring Network Description and Standard Operating Procedures. Charlotte Harbor National Estuary Program Technical Report 2-3. Corbett, C.A. and J.A. Hale. 2. Development of water quality targets for Charlotte Harbor, Florida using seagrass light requirements. Florida Scientist 9:3-. Doering, P.H., R.H. Chamberlain Water quality and source of freshwater discharge to the Caloosahatchee Estuary, Florida. Journal of the American Water Resources Association 3: Doering, P.H., R.H. Chamberlain, and K.M. Haunert. 2. Spatial variability in the response of chlorophyll a to nutrient loading and freshwater discharge in the Caloosahatchee Estuary. Charlotte Harbor Watershed Summit. February 2, Punta Gorda, FL. Doering, P. H., R.H. Chamberlain, K.M. Haunert. 2. Chlorophyll A and its use as an indicator of Eutrophication in the Caloosahatchee estuary, Florida. Florida Scientist 9: Duffey, R., R.E. Leary, J. Ott. 27. Charlotte Harbor and Estero Bay Aquatic Preserves Water Quality Status & Trends for Final Report - Charlotte Harbor National Estuary Program. Gilbert, R.O Statistical Methods for Environmental Pollution Monitoring. Van Nostrand Reinhold

19 Hand, J. and M. Friedman Typical Water Quality Values for Florida s Lakes, Streams and Estuaries. Florida Department of Environmental Regulation, Bureau of Surface Water Management,Technical Report. Hand, J. 28. Typical Water Quality Values for Florida s Lakes, Streams and Estuaries. In Press. Havens, K.E. 2. Lake Okeechobee: hurricanes and fisheries. LakeLine 2:2-28. Helsel, D.R. and R.M. Hirsch Statistical Methods in Water Resources. Elsevier, Amsterdam. Hollander, M. and D.A. Wolf Nonparametric Statistical Methods. John Wiley & Sons. Kendall, M.G A new measure of rank correlation: Biometrika 3: Rogers, M.W. and M.S. Allen. 28. Hurricane impacts to Lake Okeechobee: altered hydrology creates difficult management trade offs. Fisheries 33(1) Sanitas Technologies. 27. WQStat Plus User s Guide. Shawnee, KS. U.S. EPA. April Statistical Analysis of Ground-water Monitoring Data at RCRA Facilities, Interim Final Guidance. Office of Solid Waste Management Division, U.S. Environmental Protection Agency, Washington, DC. Acknowledgements We would like to thank Sanibel City Council for supporting the project and the Lee County Tourist Development Council for funding the last three years through the Beach and Shoreline Funding Program. We would also like to thank Rob Johnson and his staff at the Cape Coral lab and Keith Kibbey and his staff at the Lee County Environmental lab for processing the samples and the South Florida Water Management District and the Lee County Natural Resources Division for providing flow and rainfall data

20 Appendix A Output from Statistical Analyses - 19-

21 4 TIME SERIES 4 BOX & WHISKERS PLOT 3 3 Concentration (Deg_C) /2/2 11/2/3 1/3/3 11/19/4 8/3/ /2/ 4/1/7 1/28/8 11/13/ obs obs. 2 9 obs obs. 28 Constituent: Temperature (Deg_C) Date: 1//9 Time: 9:3 AM Constituent: Temperature (Deg_C) Date: 1//9 Temperature (Deg. C) Time: 9:37 AM Figure 9.1 Figure 9.2 Concentration(Deg_C) OUTLIER ANALYSIS 3/2/22 8/2/23 12/3/24 /22/2 1/1/27 11/2/22 4/22/24 8/3/2 2/14/27 /11/28 Log-transformed: Mean = 3.19 Std. Dev. =.27 Critical Tn = 3.92 No statistical outliers are present. Concentration (Deg_C) SEASONALITY: vs. Deseasonalized Data Mar 22 Jul 2 Nov 28 For the data shown, the Kruskal-Wallis test indicates SEASONALITY at the % significance level. Because the calculated Kruskal-Wallis statistic is greater than the Chi-squared value, we conclude that at least one season has a significantly different median concentration of this constituent than any other season. Calculated Kruskal-Wallis statistic = Tabulated Chi-Squared value = with 1 degrees of freedom at the % significance level. There were 1 groups of ties in the data, consequently the Kruskal-Wallis statistic (H) was adjusted. The adjusted statistic (H') was utilized to determine if the medians were equal. Kruskal-Wallis statistic (H) = Adjusted Kruskal-Wallis statistic (H') = Concentration (Deg_C) Constituent: Temperature (Deg_C) Date: 1/12/9 Time: 3:4 PM Constituent: Temperature (Deg_C) Date: 1/12/9 Time: 11:21 AM Figure 9.3 Figure 9.4

22 SEN'S SLOPE ESTIMATOR SEN'S SLOPE ESTIMATOR (Alt. Values) 4 4 Concentration(Deg_C) Slope = No. 1.9 No No No Concentration(Deg_C) Slope = No. 1.9 No No No Mar 22 Jul 2 Nov 28 Mar 22 Jul 2 Nov 28 Constituent: Temperature (Deg_C) Date: 1//9 Time: 1:9 AM Constituent: Temperature (Deg_C) Date: 1/7/9 Time: 2:1 PM Figure 9. Figure 9. Temperature (Deg. C) Flow Adjusted 4 SEASONAL KENDALL SLOPE ESTIMATOR SEASONAL KENDALL SLOPE ESTIMATOR (Alt. Values) 4 3 Slope = Slope =.318 Concentration(Deg_C) 2 1 Z=-1.9 8% No 9% 1.4 No 9% 1.9 No Concentration(Deg_C) 2 1 Z= 8% No 9% 1.4 No 9% 1.9 No Mar 22 Jul 2 Nov 28 Mar 22 Jul 2 Nov 28 Constituent: Temperature (Deg_C) Date: 1/12/9 Time: 11:22 AM Constituent: Temperature (Deg_C) Date: 1/12/9 Time: 11:24 AM Figure 9.7 Figure 9.8

23 1. TIME SERIES 1 BOX & WHISKERS PLOT Target = /2/2 11/2/3 1/3/3 11/19/4 8/3/ /2/ 4/1/7 1/28/8 11/13/ obs obs. 2 9 obs obs. 28 Constituent: Dissolved Oxygen (mg/l) Date: 1/12/9 Time: 3: PM Constituent: Dissolved Oxygen (mg/l) Date: 1//9 Dissolved Oxygen (mg/l) Time: 11:2 AM Figure 1.1 Figure OUTLIER ANALYSIS 3/2/22 8/2/23 12/3/24 /22/2 1/1/27 11/2/22 4/22/24 8/3/2 2/14/27 /11/28 Log-transformed: Mean = 1.87 Std. Dev. =.17 Critical Tn = 3.92 No statistical outliers are present SEASONALITY: vs. Deseasonalized Data Mar 22 Jul 2 Nov 28 For the data shown, the Kruskal-Wallis test indicates SEASONALITY at the % significance level. Because the calculated Kruskal-Wallis statistic is greater than the Chi-squared value, we conclude that at least one season has a significantly different median concentration of this constituent than any other season. Calculated Kruskal-Wallis statistic = Tabulated Chi-Squared value = with 1 degrees of freedom at the % significance level. There were groups of ties in the data, consequently the Kruskal-Wallis statistic (H) was adjusted. The adjusted statistic (H') was utilized to determine if the medians were equal. Kruskal-Wallis statistic (H) = Adjusted Kruskal-Wallis statistic (H') = Constituent: Dissolved Oxygen (mg/l) Date: 1/12/9 Time: 3:7 PM Constituent: Dissolved Oxygen (mg/l) Date: 1/12/9 Time: 11:33 AM Figure 1.3 Figure 1.4

24 SEN'S SLOPE ESTIMATOR SEN'S SLOPE ESTIMATOR (Alt. Values) Slope = No No No No Slope = No No No No Mar 22 Jul 2 Nov 28 Mar 22 Jul 2 Nov 28 Constituent: Dissolved Oxygen (mg/l) Date: 1//9 Time: 12:33 PM Constituent: Dissolved Oxygen (mg/l) Date: 1/7/9 Dissolved Oxygen (mg/l) Time: 1:42 PM Figure 1. Figure 1. Flow Adjusted 1 SEASONAL KENDALL SLOPE ESTIMATOR SEASONAL KENDALL SLOPE ESTIMATOR (Alt. Values) 1 8 Slope = Slope = Z=-.3 8% No 9% 1.4 No 9% 1.9 No 4 2 Z= % No 9% 1.4 No 9% 1.9 No Mar 22 Jul 2 Nov 28 Mar 22 Jul 2 Nov 28 Constituent: Dissolved Oxygen (mg/l) Date: 1/12/9 Time: 11:34 AM Constituent: Dissolved Oxygen (mg/l) Date: 1/12/9 Time: 11:3 AM Figure 1.7 Figure 1.8

25 4 TIME SERIES 4 BOX & WHISKERS PLOT 3 3 Concentration (psu) /2/2 11/2/3 1/3/3 11/19/4 8/3/ /2/ 4/1/7 1/28/8 11/13/ obs obs. 2 9 obs obs. 28 Constituent: Salinity (psu) Date: 1//9 Time: 12:42 PM Constituent: Salinity (psu) Date: 1//9 Time: 12:43 PM Figure 11.1 Figure 11.2 Salinity (psu) Concentration(psu) OUTLIER ANALYSIS 3/2/22 8/2/23 12/3/24 /22/2 1/1/27 11/2/22 4/22/24 8/3/2 2/14/27 /11/28 Log-transformed: Mean = 3.3 Std. Dev. =.232 Critical Tn = 3.92 Statistical outliers are shown as solid squares. SEASONALITY: For the data shown, the Kruskal-Wallis test indicates NO SEASONALITY at the % significance level. Because the calculated Kruskal-Wa statistic is less than or equal to the Chi-squared value, we conclude that no season has a significantly different median concentration of this constituent than any other season. Calculated Kruskal-Wallis statistic =.1 Tabulated Chi-Squared value = with 1 degrees of freedom at the % significance level. There were 2 groups of ties in the data, consequently the Kruskal-Wallis statistic (H) was adjusted. The adjusted statistic (H') was utilized to determine if the medians were equal. Kruskal-Wallis statistic (H) =.1 Adjusted Kruskal-Wallis statistic (H') =.1 Concentration (psu) vs. Deseasonalized Data Mar 22 Jul 2 Nov 28 Concentration (psu) Constituent: Salinity (psu) Date: 1//9 Time: 12:44 PM Constituent: Salinity (psu) Date: 1/12/9 Time: 11:38 AM Figure 11.3 Figure 11.4

26 SEN'S SLOPE ESTIMATOR SEN'S SLOPE ESTIMATOR (Alt. Values) 4 4 Concentration(psu) Slope = No. 1.9 Up Up Up Concentration(psu) Slope = No No No Down Mar 22 Jul 2 Nov 28 Mar 22 Jul 2 Nov 28 Constituent: Salinity (psu) Date: 1//9 Time: 12:4 PM Salinity (psu) Constituent: Salinity (psu) Date: 1//9 Time: 12:47 PM Figure 11. Figure 11. Flow Adjusted 4 SEASONAL KENDALL SLOPE ESTIMATOR SEASONAL KENDALL SLOPE ESTIMATOR (Alt. Values) 4 3 Slope =.82 3 Slope = -.39 Concentration(psu) 2 1 Z=2.3 8% Yes 9% 1.4 Yes 9% 1.9 Yes Concentration(psu) 2 1 Z= % Yes 9% 1.4 No 9% 1.9 No Mar 22 Jul 2 Nov 28 Mar 22 Jul 2 Nov 28 Constituent: Salinity (psu) Date: 1/12/9 Time: 11:38 AM Constituent: Salinity (psu) Date: 1/12/9 Time: 11:39 AM Figure 11.7 Figure 11.8

27 TIME SERIES BOX & WHISKERS PLOT 4 4 Concentration (umohs/cm) /2/2 11/2/3 1/3/3 11/19/4 9/27/ 7/11/ /1/7 3/12/ obs obs. 2 9 obs obs. 28 Constituent: Sp Conductance (umohs/cm) Date: 1/14/9 Time: 2:41 PM Constituent: Sp Conductance (umohs/cm) Date: 1/14/9 Specific Conductance (µmohs) Time: 2:43 PM Figure 12.1 Figure 12.2 Concentration(umohs/cm) OUTLIER ANALYSIS 3/2/22 8/2/23 1/27/2 9/11/2 2/21/28 12/1/22 /13/24 12/29/2 /13/27 11/13/28 Log-transformed: Mean = 1.7 Std. Dev. =.213 Critical Tn = 3.92 Statistical outliers are shown as solid squares. SEASONALITY: For the data shown, the Kruskal-Wallis test indicates NO SEASONALITY at the % significance level. Because the calculated Kruskal-Wa statistic is less than or equal to the Chi-squared value, we conclude that no season has a significantly different median concentration of this constituent than any other season. Calculated Kruskal-Wallis statistic =.34 Tabulated Chi-Squared value = with 1 degrees of freedom at the % significance level. There were groups of ties in the data, so no adjustment to the Kruskal-Wallis statistic (H) was necessary. Concentration (umohs/cm) vs. Deseasonalized Data Mar 22 Jul 2 Nov 28 8 Concentration (umohs/cm) Constituent: Sp Conductance (umohs/cm) Date: 1/14/9 Time: 2:42 PM Constituent: Sp Conductance (umohs/cm) Date: 1/14/9 Time: 2:43 PM Figure 12.3 Figure 12.4

28 SEN'S SLOPE ESTIMATOR SEN'S SLOPE ESTIMATOR (Alt. Values) Concentration(umohs/cm) Slope = No. 1.9 Up Up Up Concentration(umohs/cm) Slope = No No No Down Mar 22 Jul 2 Nov 28 Mar 22 Jul 2 Nov 28 Constituent: Sp Conductance (umohs/cm) Date: 1/14/9 Time: 2:44 PM Constituent: Sp Conductance (umohs/cm) Date: 1/14/9 Specific Conductance (µmohs) Time: 2:4 PM Figure 12. Figure 12. Flow Adjusted SEASONAL KENDALL SLOPE ESTIMATOR SEASONAL KENDALL SLOPE ESTIMATOR (Alt. Values) Concentration(umohs/cm) Slope = 1174 Z=2.3 8% Yes 9% 1.4 Yes 9% 1.9 Yes Concentration(umohs/cm) Slope = -2. Z= % Yes 9% 1.4 No 9% 1.9 No Mar 22 Jul 2 Nov 28 Mar 22 Jul 2 Nov 28 Constituent: Sp Conductance (umohs/cm) Date: 1/14/9 Time: 2:4 PM Constituent: Sp Conductance (umohs/cm) Date: 1/14/9 Time: 2:47 PM Figure 12.7 Figure 12.8

29 1. TIME SERIES 1 BOX & WHISKERS PLOT Concentration (units) /2/2 11/2/3 1/3/3 11/19/4 8/3/ /2/ 4/1/7 1/28/8 11/13/ obs obs. 2 9 obs obs. 28 Constituent: ph (units) Date: 1/12/9 Time: 11:43 AM ph Constituent: ph (units) Date: 1/12/9 Time: 11:4 AM Figure 13.1 Figure 13.2 Concentration(units) OUTLIER ANALYSIS 3/2/22 8/2/23 1/27/2 /22/2 1/1/27 11/2/22 /13/24 9/27/2 2/14/27 /11/28 Log-transformed: Mean = 2.73 Std. Dev. =.14 Critical Tn = 3.87 No statistical outliers are present. Concentration (units) SEASONALITY: vs. Deseasonalized Data Mar 22 Jul 2 Nov 28 For the data shown, the Kruskal-Wallis test indicates SEASONALITY at the % significance level. Because the calculated Kruskal-Wallis statistic is greater than the Chi-squared value, we conclude that at least one season has a significantly different median concentration of this constituent than any other season. Calculated Kruskal-Wallis statistic = 7.89 Tabulated Chi-Squared value = with 1 degrees of freedom at the % significance level. There were 23 groups of ties in the data, consequently the Kruskal-Wallis statistic (H) was adjusted. The adjusted statistic (H') was utilized to determine if the medians were equal. Kruskal-Wallis statistic (H) = Adjusted Kruskal-Wallis statistic (H') = Concentration (units) Constituent: ph (units) Date: 1/12/9 Time: 3:8 PM Constituent: ph (units) Date: 1/12/9 Time: 11:47 AM Figure 13.3 Figure 13.4

30 SEN'S SLOPE ESTIMATOR SEN'S SLOPE ESTIMATOR (Alt. Values) 1 1 Concentration(units) n = 71 Slope = No. 1.9 No No No Concentration(units) n = 71 Slope = No. 1.9 No No No Mar 22 Jul 2 Nov 28 Mar 22 Jul 2 Nov 28 Constituent: ph (units) Date: 1/12/9 Time: 11:48 AM ph Constituent: ph (units) Date: 1/12/9 Time: 11: AM Figure 13. Figure 13. Flow Adjusted 1 SEASONAL KENDALL SLOPE ESTIMATOR SEASONAL KENDALL SLOPE ESTIMATOR (Alt. Values) 1 8 n = 71 Slope = n = 71 Slope =.82 Concentration(units) 4 2 Z= % Yes 9% 1.4 No 9% 1.9 No Concentration(units) 4 2 Z=1. 8% No 9% 1.4 No 9% 1.9 No Mar 22 Jul 2 Nov 28 Mar 22 Jul 2 Nov 28 Constituent: ph (units) Date: 1/12/9 Time: 11:49 AM Constituent: ph (units) Date: 1/12/9 Time: 11: AM Figure 13.7 Figure 13.8

31 .2 TIME SERIES.2 BOX & WHISKERS PLOT /2/2 11/2/3 1/3/3 Constituent: Ammonia (mg/l) Date: 1//9 11/19/4 8/3/ /2/ 4/1/7 1/28/8 11/13/8 Time: 12: PM Ammonia (mg/l) 22 1 obs. 1%nds 23 1%nds Constituent: Ammonia (mg/l) Date: 1// obs. 1%nds 2 9 obs. 1%nds 2 27%nds obs. 2%nds Time: 12: PM Figure 14.1 Figure %nds.1.1. OUTLIER ANALYSIS. 3/2/22 8/2/23 1/27/2 /2/2 11/27/27 11/2/22 /13/24 9/27/2 3/14/27 7/28/28 Log-transformed: Mean = Std. Dev. =.722 Critical Tn = 3.92 No statistical outliers are present SEASONALITY: vs. Deseasonalized Data Mar 22 Jul 2 Nov 28 For the data shown, the Kruskal-Wallis test indicates NO SEASONALITY at the % significance level. Because the calculated Kruskal-Wa statistic is less than or equal to the Chi-squared value, we conclude that no season has a significantly different median concentration of this constituent than any other season. Calculated Kruskal-Wallis statistic =.2 Tabulated Chi-Squared value = with 1 degrees of freedom at the % significance level. There were 7 groups of ties in the data, consequently the Kruskal-Wallis statistic (H) was adjusted. The adjusted statistic (H') was utilized to determine if the medians were equal. Kruskal-Wallis statistic (H) =.211 Adjusted Kruskal-Wallis statistic (H') =.2.1 Constituent: Ammonia (mg/l) Date: 1/12/9 Time: 2:8 PM Constituent: Ammonia (mg/l) Date: 1/12/9 Time: 1:34 PM Figure 14.3 Figure 14.4

32 SEN'S SLOPE ESTIMATOR SEN'S SLOPE ESTIMATOR (Alt. Values) Slope = Down Down Down Down.1. Slope = Down Down Down Down. Mar 22 Jul 2 Nov 28. Mar 22 Jul 2 Nov 28 Constituent: Ammonia (mg/l) Date: 1//9 Time: 12:8 PM Constituent: Ammonia (mg/l) Date: 1//9 Ammonia (mg/l) Time: 1:1 PM Figure 14. Figure 14. Flow Adjusted.1 SEASONAL KENDALL SLOPE ESTIMATOR SEASONAL KENDALL SLOPE ESTIMATOR (Alt. Values).1.1. Slope = Z=-.37 8% Yes 9% 1.4 Yes 9% 1.9 Yes.1. Slope = Z= % Yes 9% 1.4 Yes 9% 1.9 Yes. Mar 22 Jul 2 Nov 28. Mar 22 Jul 2 Nov 28 Constituent: Ammonia (mg/l) Date: 1/12/9 Time: 1:3 PM Constituent: Ammonia (mg/l) Date: 1/12/9 Time: 1:3 PM Figure 14.7 Figure 14.8

33 .4 TIME SERIES.4 BOX & WHISKERS PLOT /2/2 11/2/3 1/3/3 Constituent: Nitrate (mg/l) Date: 1//9 11/19/4 8/3/ /2/ 4/1/7 1/28/8 11/13/8 Time: 4:3 PM Nitrate (mg/l) 22 1 obs. %nds 23 4%nds Constituent: Nitrate (mg/l) Date: 1// obs. 7%nds 2 9 obs. 2 9.%nds obs. 33%nds Time: 4:37 PM Figure 1.1 Figure %nds OUTLIER ANALYSIS. 3/2/22 8/2/23 12/3/24 /22/2 1/1/27 11/2/22 4/22/24 8/3/2 2/14/27 /11/28 Log-transformed: Mean = Std. Dev. = 1.27 Critical Tn = 3.92 No statistical outliers are present. SEASONALITY: For the data shown, the Kruskal-Wallis test indicates NO SEASONALITY at the % significance level. Because the calculated Kruskal-Wa statistic is less than or equal to the Chi-squared value, we conclude that no season has a significantly different median concentration of this constituent than any other season. Calculated Kruskal-Wallis statistic =.231 Tabulated Chi-Squared value = with 1 degrees of freedom at the % significance level. There were groups of ties in the data, consequently the Kruskal-Wallis statistic (H) was adjusted. The adjusted statistic (H') was utilized to determine if the medians were equal. Kruskal-Wallis statistic (H) =.228 Adjusted Kruskal-Wallis statistic (H') = vs. Deseasonalized Data Mar 22 Jul 2 Nov Constituent: Nitrate (mg/l) Date: 1//9 Time: 4:39 PM Constituent: Nitrate (mg/l) Date: 1/12/9 Time: 1:37 PM Figure 1.3 Figure 1.4