DEVELOPMENT OF NUTRIENT TMDLS FOR THE RARITAN RIVER BASIN Thomas Amidon and James F. Cosgrove, Jr., P.E. TRC Omni Environmental Corporation

Transcription

1 DEVELOPMENT OF NUTRIENT TMDLS FOR THE RARITAN RIVER BASIN Thomas Amidon and James F. Cosgrove, Jr., P.E. TRC Omni Environmental Corporation ABSTRACT A large watershed study was performed to develop Total Maximum Daily Loads (TMDLs) as necessary to address phosphorus and other conventional impairments in the Raritan River Basin. The Raritan River Basin encompasses over 1,100 square miles in the north-central portion of New Jersey that drain to the Raritan Bay. Extensive data collection was performed, the purpose of which was to augment presently available information in order to provide data necessary to evaluate nutrient chemistry and use impairment, as well as provide data necessary for the modeling work. Streams and lakes were evaluated to identify critical locations where water column phosphorus is causing, or potentially causing, instream impairment. Nutrient chemistry, diurnal dissolved oxygen and ph, and chlorophyll-a measurements were used to assess nutrient limitation, productivity, and phosphorus impairment. Data collected for this study were also evaluated to determine the nature and cause of other conventional impairments in the Raritan River Basin, namely total suspended solids, ph, temperature, and dissolved oxygen. Hydrologic, hydraulic, and water quality modeling were performed in order to relate point and nonpoint sources to water quality targets at critical locations. The TMDL(s) will be established in order to determine pollutant load reductions that will be necessary to bring each impaired segment into compliance with applicable water quality standards. KEYWORDS Phosphorus, nutrient, TMDL, watershed, diurnal dissolved oxygen INTRODUCTION The Raritan River Basin Nutrient TMDL project was designed to help the New Jersey Department of Environmental Protection (NJDEP) develop Total Maximum Daily Loads (TMDLs) as necessary to address phosphorus and other conventional impairments in the Raritan River Basin. The Raritan River Basin encompasses over 1,100 square miles in the central portion of New Jersey that drain to the Raritan Bay. Predominant land uses and land cover include agriculture, forest, and a variety of urban densities. Sandwiched between Philadelphia and New York City in north-central New Jersey, the Raritan River basin has undergone substantial development and continues to experience enormous development pressure. Water Quality Standards and Impairment Designations The following streams in the Raritan River Basin are designated by NJDEP as impaired for phosphorus because instream concentrations of phosphorus exceed the 0.1 mg/l threshold in the Surface Water Quality Standards (SWQS): Beden Brook ; Bound Brook; Cakepoulin Creek; 2275

2 Lamington River; Manalapan Brook; Matchaponix Brook; McGellairds Brook; Millstone River, Upper and Lower; Neshanic River; Pike Run; Raritan River, North Branch, South Branch, and Mainstem; Rockaway Creek; Six Mile Run; Spruce Run (downstream of reservoir); Stony Brook; Weamaconk Creek; and Wemrock Brook. In addition to phosphorus impairment designations, NJDEP has designated several of these streams as impaired for ph, total suspended solids (TSS), temperature, and dissolved oxygen (DO). The total phosphorus criterion for lakes, ponds, and reservoirs is 0.05 mg/l, 50% of the criterion in streams. The SWQS do not clearly define the distinction between a stream and a lake; however, there are at least ten named lakes that are in-line to streams within the Raritan River basin, meaning they are formed by dams that form impoundments along streams. None of these lakes are routinely monitored by NJDEP for phosphorus. The Surface Water Quality Standards (SWQS) also contain Nutrient Policies, including the qualitative requirement that: nutrients shall not be allowed in concentrations that cause objectionable algal densities, nuisance aquatic vegetation, abnormal diurnal fluctuations in dissolved oxygen or ph, changes to the composition of aquatic ecosystems, or otherwise render the waters unsuitable for the designated uses. Both the qualitative Nutrient Policies and quantitative phosphorus criteria are subject to two important qualifications. The first qualifier is that background conditions replace the criteria for all water quality characteristics that do not meet the promulgated water quality criteria as a result of natural causes. The second qualifier in the SWQS is that NJDEP, in order to protect uses, may establish watershed or site-specific water quality criteria for nutrients that replace the generic phosphorus criteria for both lakes and streams. FLOW AND WATER QUALITY SAMPLING SUMMARY Extensive monitoring was performed from May to November of 2003 and 2004 in accordance with the Quality Assurance Sampling Plan prepared by TRC Omni Environmental Corporation (TRC Omni) and approved by NJDEP (TRC Omni, May 24, 2004). Sampling under specified flow and weather conditions was performed at 88 stream, lake, and sewage treatment plant (STP) locations throughout the Raritan River basin. TRC Omni began the sampling phase of this project in May 2004 and completed the final low-flow sampling event on August 4-5, Sampling for this project was designed to complement phosphorus evaluation studies performed in 2000, 2001, and 2003 for substantial portions of the Lower Millstone River and mainstem Raritan watersheds (TRC Omni, May 6, 2004; TRC Omni, December 8, 2004) as well as phosphorus evaluations performed concurrently in 2004 in segments of the South Branch Raritan River (TRC Omni, March 8, 2005) and Matchaponix Brook (TRC Omni, April 11, 2005). The purpose of these phosphorus evaluations was to evaluate nutrient limitation and use impairment in order to determine the applicability of the phosphorus stream criterion. Data from both the phosphorus evaluations as well as the sampling phase will be utilized for the TMDL study. Stream sampling stations were selected so that eutrophication impacts could be evaluated and a water quality model could be calibrated and verified within the phosphorus-impaired segments of the Raritan River Basin. Six locations that drain small watersheds were selected for both baseflow and stormwater stations in order to characterize prevalent land use types for which 2276

3 stormwater data have not already been collected during previous studies. Lake inlet, lake outlet, and in-lake stations were selected for nine run-of-the-river lakes in order to characterize their impact on the phosphorus budget as well as the impact of phosphorus on these critical locations. Six tributary stations were selected in order to characterize substantial inputs into the major streams within the study area. Wastewater treatment plants within the study areas were chosen for sampling if their permitted flow exceeds 1.0 mgd (classified as major according to EPA) or if their actual flow exceeds 0.1 mgd. As shown in Figure 1, sampling locations were selected to complement and extend the study areas of Phosphorus Evaluations. Figure 1 shows the locations of sampling stations for both the sampling phase of the TMDL study and the phosphorus evaluation studies. The sampling phase of the TMDL study consisted of the following networks: 9 baseflow locations, 6 of which also serve as stormwater locations; 6 stormwater locations, which also serve as baseflow locations; 33 stream characterization stations; 9 lake inlet locations; 9 lake outlet locations; 9 lake stations; 6 tributary locations; and 13 wastewater treatment plant outfalls. The sampling phase of the TMDL study consisted of the following sampling events: Three low-flow sampling events were performed at 66 stations, each event consisting of two consecutive days of sampling for a total of six samples per station. Sampling networks that were sampled during the low-flow events include baseflow, characterization, lake inlet, lake outlet, STP (one 24-hour composite per event), and tributary network stations. Three high-flow sampling events were performed at 57 stations, each event consisting of two consecutive days of sampling for a total of six samples per station. Sampling networks that were sampled during the high-flow events include characterization, lake inlet, lake outlet, STP (one 24-hour composite per event), and tributary network stations. Eight ambient sampling events were performed at 42 stations, each event consisting of one day of sampling for a total of eight samples per station. Sampling networks that were sampled during the ambient events include characterization and lake inlet network stations. Three diurnal monitoring events were performed at 42 stations, each event consisting of continuous monitoring every 15 minutes during at least three consecutive days coincident with the low-flow events. Sampling networks that were sampled during the diurnal events include characterization and lake network stations. 2277

4 Figure 1: Sampling Locations for Raritan River Basin TMDL Study 2278

5 Three periphyton sampling events were performed at 42 stations, each event consisting of a single composite sample or observation for a total of three composite samples and/or observations per station. Periphyton sampling events were performed coincident with the low-flow events. Sampling networks that were sampled during the periphyton events include baseflow and characterization network stations. Three stormwater events were performed at the six stormwater network stations, each event consisting of approximately five samples per storm for a minimum total of 15 samples per station. Flow and weather condition requirements, specific parameters monitored during each event, and sampling procedures are specified in the Quality Assurance Sampling Plan (TRC Omni, May 24, 2004). Figures 3 and 4 below show the flow and precipitation conditions at a representative location during sampling in 2004 and 2005, respectively. Figure 2: Flow and Precipitation Conditions During Sampling (2004) SB Raritan River near High Bridge /01/04 06/15/04 06/29/04 Flow (cfs) 07/13/04 07/27/04 08/10/04 08/24/04 09/07/04 09/21/04 10/05/04 10/19/04 11/02/04 11/16/04 11/30/ Precipitation (inches) 0 Clinton STP USGS Gage ( ) Low Flow Event High Flow Event Ambient Event Periphyton Event Stormwater Event D

6 Figure 3: Flow and Precipitation Conditions During Sampling (2005) SB Raritan River near High Bridge Flow (cfs) Precipitation (inches) /01/05 04/15/05 04/29/05 05/13/05 05/27/05 06/10/05 06/24/05 07/08/05 07/22/05 08/05/05 08/19/05 0 Clinton STP USGS Gage ( ) Low Flow Event Periphyton Event Stormwater Event D70 IDENTIFICATION OF CRITICAL LOCATIONS Based on the extensive data collected for this study, as well as a rich inventory of previous studies performed in the basin, we evaluated to what degree water column phosphorus appears to be controlling productivity and whether phosphorus is causing, or potentially causing, instream impairment. Nutrient chemistry, diurnal dissolved oxygen and ph, and chlorophyll-a measurements were used to assess nutrient limitation, productivity, and phosphorus impairment. Critical locations and associated water quality targets at those locations were established through an evaluation of the existing and potential water quality impacts from phosphorus. Diurnal dissolved oxygen, phytoplankton, and periphyton were used to evaluate use impairment. Assessment results for the North Branch Raritan River Watershed are described below. Figure 4 shows the sampling locations in the North Branch Raritan River, which can be assessed as four distinct subwatersheds: the Lamington River subwatershed upstream of its confluence with Rockaway Creek; the Rockaway Creek subwatershed, including the portion of the Lamington River from the Rockaway Creek confluence to the North Branch Raritan River Confluence; the upper North Branch Raritan River subwatershed upstream of the Lamington River confluence; and the lower North Branch Raritan River downstream of the Lamington River confluence. 2280

7 Figure 4: Sampling Locations in the North Branch Raritan River Watershed 2281

8 Phosphorus Phosphorus is generally much lower than 0.1 mg/l (average TP = mg/l) and appears to be a limiting nutrient in the headwaters and tributaries of all four subwatersheds (LR1, NBRC1, SBRC1-CLi, BuB1, IBB, IB1, PeB1, MiB1, NBRR1). The Lamington River exceeds 0.1 mg/l TP in Chester downstream of Roxbury Township STP, but returns to a low phosphorus condition downstream in Pottersville (LR3) and near Whitehouse (LR4). Under ambient and low-flow conditions, stations LR3 and LR4 are very similar to the headwater condition at station LR1. Under high-flow conditions, stations LR3 and LR4 exceed 0.1 mg/l TP occasionally, coincident with high iron and high TSS concentrations. Such exceedances may be related to instream erosion and/or other runoff inputs rather than productivity. The Lamington River downstream of Roxbury Township STP (LR2) appears to remain phosphorus-limited when total phosphorus is near 0.1 mg/l, although there are many times when phosphorus levels are quite high and therefore do not appear to be limiting productivity. The upstream stations in the Rockaway Creek subwatershed (NBRC1 and SBRC1-CLi) are very similar to the Lamington River just upstream of the Rockaway River confluence: phosphorus is generally lower than 0.1 mg/l (0.034 and mg/l, respectively), but exceeds 0.1 mg/l occasionally, coincident with high iron and high TSS concentrations. However, station SBRC- CLi is near the inlet to Cushetunk Lake, and frequently exceeds the lake criterion of 0.05 mg/l TP (average TP = mg/l). Interestingly, phosphorus concentration increases in Cushetunk Lake outlet (SBRC3-CLo) compared to its inlet (SBRC1-CLi), indicating either substantial inlake phosphorus sources or direct sources to the lake. Phosphorus appears to remain the limiting nutrient in Cushetunk Lake even when TP exceeds 0.1 mg/l, as is common in lake environments. Phosphorus concentrations exceed 0.1 mg/l commonly in the mainstem Rockaway Creek and Lamington River downstream of the Rockaway Creek confluence (Readington-Lebanon Sewerage Authority STP discharges to South Branch Rockaway Creek just upstream of the confluence with the North Branch Rockaway Creek). However, phosphorus no longer appears to limit productivity in the mainstem Rockaway Creek and lower Lamington River when TP exceeds 0.1 mg/l. In the North Branch Raritan River, phosphorus occasionally exceeds 0.1 mg/l. It is difficult to assess whether phosphorus continues to control productivity when TP exceeds 0.1 mg/l, except in the lower North Branch Raritan River (NBRR7) where it does not appear to be limiting productivity. While several municipal STPs discharge to the North Branch Raritan River and its tributaries, these STPs perform a high degree of phosphorus treatment. Consequently, the concentration of phosphorus in the North Branch Raritan River during low-flow events is low and fairly constant upstream to downstream (Figure 5). On the other hand, phosphorus concentration during high-flow events increases noticeably from upstream to downstream (Figure 5), coincident with high TSS values. This pattern is consistent with stream bank erosion. The lake criterion of 0.05 mg/l is frequently exceeded at the inlet to Ravine Lake (NBRR2-RLi), especially during high-flow events. Phosphorus at the outlet of Ravine Lake (NBRR4-RLo) is higher than the inlet, indicating substantial in-lake sources or direct sources to the lake. 2282

9 Figure 5: Average Phosphorus Concentrations in North Branch Raritan River Total Phosphorus (mg/l) instream criterion lake criterion NBRR1 NBRR2-RLi NBRR5 NBRR6 NBRR7 NBRR1 NBRR2-RLi NBRR4-RLo NBRR5 NBRR6 NBRR7 NBRR1 NBRR2-RLi NBRR4-RLo NBRR5 NBRR6 NBRR7 Ambient High-Flow Low-Flow Productivity Indicators Phytoplankton concentrations are quite low throughout the North Branch Raritan River watershed (Figure 6), especially during the summer months of July, August and September. The average of all phytoplankton observations at all stream stations in the watershed was 1.9 µg/l chlorophyll-a. Only the three most downstream stations, the North Branch Raritan River and Lamington River at Burnt Mills just upstream of their confluence (LR5 and NBRR6) and the North Branch Raritan River near the outlet of the watershed (NBRR7), were observed to have single-sample phytoplankton concentrations of 8 µg/l or higher. While lake water quality was generally assessed using outlet stream data (NBRR4-RLo and SBRC3-CLo), a few additional in-lake phytoplankton grab samples were taken; these samples were near-surface grab samples taken near the center of the lakes laterally and in the deeper area upstream of the dam longitudinally. These near-surface grabs within the lakes were generally higher than the outlets that were sampled more frequently, which is not surprising since they were not depth-composited. Phytoplankton chlorophyll-a concentrations measured within Ravine Lake ranged from non-detect during low-flow dry weather in August 2005 to 18.7 µg/l on October 4, 2005 during fall turnover. The outlet of Ravine Lake was sampled 12 times and showed a maximum chlorophyll-a concentration of less than 7 µg/l. Phytoplankton chlorophylla concentration measured within Cushetunk Lake was observed to be as high as 23.4 µg/l on 2283

10 August 22, The outlet of Cushetunk Lake was sampled 12 times and showed a maximum chlorophyll-a concentration of 4 µg/l. Periphyton densities are also fairly low throughout the watershed (Figure 7). Only two sites, both in the Lamington River (LR3 and LR5), were observed to have maximum periphyton densities over 50 mg/m²: 58 and 64 mg/m², respectively. None of the phytoplankton or periphyton values exceeded the thresholds in the NJDEP Technical Manual (NJDEP, 2003). These thresholds were developed for streams; they are compared with measurement from all sampling locations as a frame of reference. Figure 6: Phytoplankton Concentrations in North Branch Raritan River North Branch Raritan River Study Area Phytoplankton (Water Column Algae) LR1 LR2 LR3 LR4 LR5 NBRC1 NBRR1 NBRR4- RLo NBRR5 NBRR6 Chlorophyll-a (mg/m³) NBRR7 RC1 SBRC3- CLo Summer Average NJDEP Seas. Avg. Threshold Summer Maximum NJDEP Max. 2-week Avg. Threshold 2284

11 Figure 7: Periphyton Concentrations in North Branch Raritan River North Branch Raritan River Study Area Periphyton (Attached Algae) LR1 LR2 LR3 LR4 Chlorophyll-a (mg/m²) LR5 NBRC1 NBRR1 NBRR5 NBRR6 NBRR7 RC1 Summer Average NJDEP Seas. Avg. Threshold Summer Maximum NJDEP Max.threshold The Lamington River subwatershed (upstream of Rockaway Creek) generally exhibits substantial, but not apparently unhealthy, diurnal DO and ph swings. The exception is the Lamington River in Chester downstream of Roxbury Township STP (Figure 8). Figure 8: Diurnal Dissolved Oxygen at Lamington River downstream of Roxbury STP LR2 : Lamington River at Ironia Road in Chester Dissolved Oxygen (mg/l) % 175% 150% 125% 100% 75% 50% 25% DO as Percent of Saturation 0.0 8/3 6 AM 8/3 12 PM 8/3 6 PM 8/4 12 AM 8/4 6 AM 8/4 12 PM 8/4 6 PM 8/5 12 AM 8/5 6 AM 0% 2005 Diurnal DO Grab DO Winkler DO DOSAT 2285

12 The DO observations suggest that sediment oxygen demand may be substantial both upstream (LR1) and downstream (LR2) of the discharge; however, the extreme diurnal DO swings as high as 10 mg/l exacerbate the low DO, causing the stream to violate the 4.0 mg/l minimum DO criterion and even become apparently anoxic during the night under extreme low-flows. It should be noted that this stream segment is the only one in the Raritan River basin designated by NJDEP as impaired for DO (NJDEP, 2004). Macrophytes are responsible for much of the productivity in the Lamington River (Figure 9) and most other streams in the Raritan River Basin. Figure 9: Rooted Macrophytes at LR2 The North Branch Rockaway Creek (NBRC1), mainstem Rockaway Creek (RC1) and lower Lamington River (LR5) all experience substantial diurnal DO swings driven by rooted macrophytes. During critical low-flow periods, the North Branch Rockaway Creek experienced DO swings up to 8 mg/l (Figure 10), while the mainstem Rockaway Creek and lower Lamington River experienced DO swings as high as 6 mg/l. These three segments of the stream are remarkably similar in terms of productivity, although their phosphorus concentrations are dramatically different. It is also important to note that diurnal ph swings observed at all three locations were as high as 1 s.u., meaning a ten-fold diurnal swing in hydrogen ion concentration. Also, the peaks of these ph swings exceeded the maximum ph criterion of 8.5 s.u. occasionally at all three locations. The South Branch and mainstem Rockaway Creek and lower Lamington River are occasionally affected by releases from Round Valley Reservoir through its Whitehouse Release in the South Branch Rockaway Creek. Summer 2005 diurnal monitoring at stations RC1 and LR5 was impacted by releases from Round Valley Reservoir; it its likely that the diurnal variation of DO and ph at these locations would have been greater in the absence of the increased flow from Round Valley during August and September of Finally, Cushetunk Lake (SBRC2-CL) routinely experiences 8 mg/l and 1 s.u. diurnal fluctuations of DO and ph, respectively. The ph during these periods of high productivity frequently exceeds the maximum criterion of 8.5 s.u. 2286

13 Figure 10: Diurnal Dissolved Oxygen at North Branch Rockaway Creek NBRC1 : North Branch Rockaway Creek at Route 523 near Whitehouse Station Dissolved Oxygen (mg/l) % 175% 150% 125% 100% 75% 50% 25% 0.0 7/29 6 AM DO as Percent of Saturation 7/30 6 AM 7/31 6 AM 8/1 6 AM 8/2 6 AM 8/3 6 AM 8/4 6 AM 8/5 6 AM 8/6 6 AM 0% 2005 Diurnal DO Grab DO Winkler DO DOSAT The North Branch Raritan River upstream of Ravine Lake (NBRR1) shows no apparent signs of excessive productivity. In fact, Ravine Lake itself (NBRR3-RL) experiences only mild diurnal DO swings in its epilimnion. However, the hypolimnion of Ravine Lake was observed to be anoxic from the bottom (21 feet deep) all the way to the beginning of the thermocline around 10 feet deep. Overall productivity appears to be fairly low in Ravine Lake; during two peak summer observations, in-lake near-surface phytoplankton chlorophyll-a was measured nondetect and 10.7 µg/l. It is likely that the deep anoxic thermocline is caused more by the circulation patterns and temperature stratification than decomposition of detrital matter. Diurnal DO swings at the two most downstream locations in the North Branch Raritan River (NBRR6 and NBRR7) are substantial, but do not appear to be excessive or causing obvious impairment. During critical conditions, DO ranges from 6 to 12 mg/l at these locations. On the other hand, the North Branch Raritan River in Bedminster (NBRR5) experiences extreme DO and ph swings, observed as high as 14 mg/l and 1.5 s.u., respectively. Furthermore, the daily DO troughs during August 2005 at NBRR5 were observed to reach the minimum DO criterion of 4 mg/l (Figure 11). 2287

14 Figure 11: Diurnal Dissolved Oxygen at North Branch Raritan River NBRR5 : North Branch Raritan R. at Rte 202/206 just upstream of small trib in Bedminster % % Dissolved Oxygen (mg/l) % 175% 150% 125% 100% 75% 50% /3 6 AM 8/4 6 AM 8/5 6 AM 8/6 6 AM 8/7 6 AM 8/8 6 AM 8/9 6 AM 8/10 6 AM 8/11 6 AM DO as Percent of Saturation 25% 0% 2005 Diurnal DO Grab DO Winkler DO DOSAT TSS, ph, and Temperature None of the streams in the North Branch Raritan River watershed were designated by NJDEP as impaired for TSS (NJDEP, 2004). However, exceedances of the applicable maximum TSS criteria (25 mg/l for trout waters and 40 mg/l for non-trout waters) were observed during highflow sampling events. The TSS increases dramatically during high-flow events from upstream to downstream in all three subwatershed, as illustrated in Figure 14. This pattern points to stream bank erosion as the primary cause, as seen in Figure 15 (North Branch Raritan River in Bridgewater). TSS near the outlets of Cushetunk Lake (SBRC3-CLo) and Ravine Lake (NBRR4-RLo) may also be exacerbated by flushing of sediments stored in the lakes. 2288

15 Figure 12: TSS in North Branch Raritan River Subwatershed TSS (mg/l) NT TM, TP NBRR2-RLi NBRR5 NBRR6 NBRR7 IB1 BuB1 MiB1 PeB1 NBRR2-RLi NBRR4-RLo NBRR5 NBRR6 NBRR7 IBB IB1 BuB1 MiB1 PeB1 NBRR2-RLi NBRR4-RLo NBRR5 NBRR6 NBRR7 Ambient High-Flow Low-Flow Solids, Total Suspended (TSS) - Maximum Solids, Total Suspended (TSS) - Average Figure 13: Stream Bank Erosion in North Branch Raritan River 2289

16 None of the streams in the North Branch Raritan River watershed were designated by NJDEP as impaired for ph (NJDEP, 2004). While the peak ph of the diurnal ph swings at several locations did exceed the maximum ph criterion of 8.5 s.u., this occurred during mild swings at lower temperatures as well. It is possible that the naturally high ph and somewhat low alkalinity throughout the North Branch Raritan River watershed cause the ph to exceed the maximum ph criterion occasionally, even during mild ph swings. The Lamington River at Route 523 (near LR3) is designated by NJDEP as impaired for temperature. Figure 16 shows the average and maximum temperatures during the summer (June to September) at locations throughout the watershed (upstream to downstream). Not only do many of the maximum temperatures exceed applicable criteria, but the summer average temperature at all of the locations classified as trout waters is near or above the applicable maximum criterion of 20 degrees C: LR3, NBRC1, SBRC1-CLi, SBRC3-CLo, IBB, IB1, BuB1, NBRR1, NBRR2-RLi, and NBRR4-RLo. TRC Omni is not aware of any thermal discharges causing these exceedances of the trout temperature criterion. On the contrary, these temperatures appear to be naturally occurring. The streams are wide and flat, and therefore very shallow under low flows. Despite having excellent riparian canopy, these trout streams are exposed to sunlight due to their width. This condition is best illustrated in the Lamington River, which even exceeds the applicable temperature criterion of 30 degrees C in its downstream non-trout segment (LR4). Figure 17 shows the Lamington River with dozens of geese wading in its shallow water. Trout streams downstream of Cushetunk Lake and Ravine Lake are further impacted by solar heating of the epilimnion prior to discharge, another natural cause of increased temperature. Figure 14: Summer Temperatures in the North Branch Raritan River Watershed Other NT Bass/Perch LR1 LR2 LR3 LR4 NBRC1 SBRC1-CLi SBRC3-CLo RC1 LR5 IBB IB1 BuB1 NBRR1 NBRR2-RLi NBRR4-RLo PeB1 MiB1 NBRR5 Temperature (C) NBRR6 NBRR7 Trout Temperature - Average Temperature - Maximum 2290

17 Figure 15: Lamington River Community of Geese Waders These geese are standing in the water, not swimming! Accordingly, based on the extensive data collected, TRC Omni assessed the major watersheds throughout the Raritan River Basin. Careful attention was given to identifying critical locations where phosphorus has the potential to cause water quality impairment. In addition, the nature and cause of impairments (or perceived impairments) due to TSS, ph, and temperature was evaluated. Critical locations identified within the North Branch Raritan River watershed, for instance, include the following. The Lamington River in Chester (LR2) experiences extreme dissolved oxygen swings and DO depletions driven by macrophytes and perhaps sediment oxygen demand. Cushetunk Lake often exceeds the lake criterion of 0.05 mg/l total phosphorus and exhibits high diurnal DO and ph swings. The Rockaway Creek and lower Lamington River (NBRC1, RC1, LR5) all exhibit substantial diurnal DO swings driven by rooted macrophytes. Ravine Lake often exceeds the lake criterion of 0.05 mg/l total phosphorus, although productivity does not appear to be excessive. The North Branch Raritan River in Bedminster (NBRR5) experiences extreme diurnal DO and ph swings, and the minimum DO decreases as low as 4 mg/l. NUTRIENT AND SEDIMENT TMDLS FOR RUN-OF-THE-RIVER LAKES This study in the Raritan River Basin is used to demonstrate the analysis and explicit consideration of run-of-the-river lakes in a TMDL context. Morphological characteristics of the run-of-the-river lakes in the Raritan River basin are presented in Table 1, based on bathymetry 2291

18 surveys performed by TRC Omni in 2004 and Most important are the shallow depths, the dominance of macrophytes among primary producers, and the short residence times. Use protection is presented in terms of diurnal dissolved oxygen swings and depletions, average total phosphorus concentration, as well as sedimentation (fill-in) rates. Acceptable annual loading rates of phosphorus and suspended solids are calculated for individual lakes and used to establish water quality targets for the basin-wide nutrient TMDL. This approach focuses the nutrient TMDL on preventing and restoring use impairments at locations that are most susceptible to nutrient enrichment and sedimentation. Table 1: Morphological Characteristics of Run-of-the-River Lakes Study Area Lake Area (acres) Max. Depth (feet) Avg. Depth (feet) Volume (acre-feet) North Branch Cushetunk Lake Raritan Ravine Lake South Branch Raritan Matchaponix Upper Millstone Solitude Lake Duhernal Lake Weamaconk Lake Gordon Pond Grovers Mill Pond Peddie Lake Plainsboro Pond HYDROLOGIC AND WATER QUALITY MODELING The purpose of the watershed modeling was to relate point and nonpoint sources to water quality targets at critical locations in order to provide the basis for a watershed-based TMDL. Development and calibration of flow simulations were performed in order to provide the hydrologic basis for watershed-wide water quality models, since dynamic water quality models require stream flow and velocity calculations as their foundation. The Raritan River Basin model requires continuous surface runoff, baseflow and stream flow for determining the pollutant loads. The use of a hydrologic model for the simulation of surface runoff, baseflow and stream flow provides a basis for estimating pollutant loads and calibrating a water quality model. In addition, it allows for TMDLs scenarios that account for the impact of land use changes on hydrology and pollutant loading. Hydrologic models mimic the land phase of the water cycle. They use climatologic data, land use, soil types, and topography as a basis for simulating the surface runoff and baseflow from contributing areas. Some of the popular hydrologic models, which are included within the BASINS modeling framework (USEPA, 2001) are the Hydrologic Simulation Program-Fortran (HSPF) (Bicknell, et al., 1997) and the Soil and Water Assessment Tool (SWAT) (Arnold, et al., 1998). In addition to the models available within the BASINS framework, the General Watershed Loading Function (GWLF) (Haith et al., 1997; Haith and Shoemaker, 1987) and HEC-Hydrologic Modeling System (HEC-HMS) (USACE, 2000) have been applied in many areas throughout the United States. 2292

19 The main structure of hydrologic models can be classified into continuous time, event based, spatially lumped and spatially distributed models. The first two classifications are related to the temporal structure, while the last two classifications are related to the spatial structure of the models. Continuous simulation models generally mimic the complete land phase of the hydrologic cycle and provide long-term hydrologic simulations. Event based models focus on specific events. The spatial structure of hydrologic models varies according to the resolution of the simulation elements. Lumped models consider a unique drainage area. Thus, distinct features present within the simulated area are represented by an average value parameter. Spatially distributed models can simulate distinct areas with site-specific parameters that can potentially represent characteristics such as soil permeability, land cover and precipitation patterns. The resolution of the spatial distribution of hydrologic models can vary. For example, HSPF uses pervious and impervious areas as hydrologic simulation units, SWAT uses multiple combinations of soil and land use within a watershed, and GWLF uses the land use types as simulation units. The simulation of hydrology is important to support the simulation of transport and fate of pollutants in the Raritan basin. The water quality aspect of the TMDL requires sophisticated water quality modeling tools for the simulation of eutrophication and other water quality variables. A modeling framework using the Water Quality Analysis Program 7.0 (WASP7) was successfully developed for the Passaic River TMDL (TRC Omni, 2005). Because of WASP7 modeling capabilities, and because it provided good results for the large scale water quality simulations in the Passaic Basin, it was chosen as the water quality modeling tool for the Raritan TMDL. WASP7 requires flow inputs. Therefore, the development of a hydrologic model that could be easily linked to WASP was necessary to provide an efficient modeling structure for the Raritan River Basin. HydroWAMIT is the hydrologic model developed for the Raritan River Basin TMDL study. It operates within the WAMIT structure, developed by TRC Omni Environmental for the Passaic River Basin study. HydroWAMIT has similar structure and components of GWLF and HSPF. Although HSPF, SWAT, GWLF and HEC-HMS could potentially be applied for the Raritan River Basin, HydroWAMIT was developed in order to provide a structure that can easily exchange information with WASP. In addition, HydroWAMIT provides a structure that can represent the impervious fractions of the land cover while preventing an overly time consuming calibration process. The most popular hydrologic models available, HSPF, SWAT, GWLF and HEC-HMS either have not been systematically linked to WASP or the versions with linkage are not available in the public domain. A structure for creating flow and load inputs for WASP based on HSPF or SWAT could involve a significant effort. Due to the spatially distributed nature of these models, there are multiple sources of flows and loads, which need to be aggregated or disaggregated to provide the necessary inputs for WASP. A hydrologic model was therefore developed within WAMIT, which is already linked to WASP. The development of a robust hydrologic model was essential. GWLF is one of the most robust models amongst the hydrologic models mentioned in this report, with a simple conceptual structure of the land phase of the hydrologic cycle. Conversely, HSPF, which is able to depict many processes of the hydrologic cycle and theoretically provide a better representation of the 2293

20 flows, is generally over-parameterized. As a result, HSPF calibration can become very complex and time consuming for large areas, even when automated calibration tools are used. Therefore, a model was developed to provide simulation results with similar quality as HSPF and have the robustness of GWLF for large-scale watersheds such as the Raritan River Basin. Figure 16 shows the land phase of the hydrologic cycles as conceptualized in HydroWAMIT. Figure 16: Land Phase of the Hydrologic Cycle HydroWAMIT (Hydrologic Watershed Model Integration Tool) was developed in order to provide a GIS-based graphical interface that: 1) performs daily hydrologic modeling in multiple subwatersheds; 2) manages inputs and outputs of hydraulic model (DAFLOW); 3) performs a hydrograph separation in order to apportion flow among runoff, baseflow, and point sources; 4) calculates nonpoint source (NPS) loads; and 5) converts DAFLOW s stream network into a structure compatible with WASP. Rooted macrophytes are important in many areas throughout the watershed and exert a substantial influence on the dissolved oxygen and nutrient dynamics. Given the importance of rooted macrophytes in streams throughout the study area, WASP 7.0 was used to simulate streams within the basin. WASP 7.0 is a modification of the original EPA WASP code that includes a subroutine to simulate periphyton (or macrophytes) and its effects on nutrient and dissolved oxygen dynamics. The water quality constituents that were simulated in WASP include ammonia, nitrate, organic nitrogen, orthophosphate, organic phosphorus, dissolved oxygen, phytoplankton chlorophyll, ultimate biological oxygen demand, and benthic algae (or 2294

21 macrophytes). Point source loads from sewage treatment plants were entered on a daily time scale based on average daily flow records and interpolations of all available effluent quality data for the calibration periods. Nonpoint source loads from runoff and baseflow are calculated for each contributing subwatershed by attaching constituent concentrations to the runoff and baseflows within HydroWAMIT. Constituent concentrations of runoff and baseflow vary by subwatershed according to the land uses within each subwatershed, based on stormwater and baseflow monitoring of individual land use areas. TMDL SOLUTIONS The integration of in-line lake studies into a basin-wide nutrient TMDL study is presented for a major drainage basin in New Jersey with multiple and significant point and nonpoint sources of phosphorus. The impact of reductions of point and nonpoint sources of phosphorus on productivity at critical locations will be explored within a modeling framework. The impact of present and future development on flow and water quality is also explored. An integrated GIS modeling framework is presented in order to calculate phosphorus reductions necessary to satisfy water quality targets at critical locations in the Raritan River basin. 2295