Grantee Research Project Results
2004 Progress Report: Improved Science and Decision Support for Managing Watershed Nutrient Loads
EPA Grant Number: R830654Title: Improved Science and Decision Support for Managing Watershed Nutrient Loads
Investigators: Chapra, Steve , Kirshen, Paul , Vogel, Richard , Hemond, Harold F. , Durant, John
Institution: Tufts University , Massachusetts Institute of Technology
EPA Project Officer: Packard, Benjamin H
Project Period: January 20, 2003 through January 19, 2006 (Extended to January 19, 2007)
Project Period Covered by this Report: January 20, 2004 through January 19, 2005
Project Amount: $749,179
RFA: Nutrient Science for Improved Watershed Management (2002) RFA Text | Recipients Lists
Research Category: Water , Watersheds
Objective:
The objectives of this research project are to: (1) conduct scientific studies of both watershed and aquatic processes relevant to the delivery and impact of nutrients on natural waters; (2) incorporate this science into watershed and receiving water models; and (3) integrate these models into a decision-support framework that can support cost-effective environmental decisionmaking related to nutrient control of eutrophication.
Progress Summary:
This year’s research proceeded on three major fronts: scientific studies, modeling/decision support studies, and stakeholder interactions. The following sections summarize our progress in each of these areas.
Scientific Studies
The scientific studies focused on nutrient dynamics in both watersheds and receiving waters.
Watershed Nutrient Studies (Tufts University and Massachusetts Institute of Technology [MIT])
Three separate studies were conducted by Tufts and MIT to increase understanding of nutrient dynamics within the watershed and how nutrient fluxes are delivered from the watershed to the receiving water.
Nutrient Loading Estimates (Tufts). One component of our study involves determining the major source(s) of nitrogen species for our study watershed, the Aberjona River (Figure 1), which is listed on the section 303d of the Clean Water Act for impairment by unionized ammonia. This work has scientific relevance as it allows us to understand nutrient export from urban and residential watersheds. Further, it will provide data to calibrate and validate our watershed loading models and to drive our lake eutrophication model.
Figure 1. The Aberjona River Watershed. Nutrient sampling locations are marked with stars.
As part of the project, monthly samples and two storm events (sampled intensely through the rise and fall of the hydrograph) were collected at the U.S. Geological Survey (USGS) gaging station (Figure 1) between January 2003 and October 2004 (n = 37). We measured dissolved ammonia, nitrate, total nitrogen, and total phosphorus, as well as dissolved oxygen, conductivity, temperature, pH, and flow.
This dataset has allowed us to investigate several issues within the Aberjona River watershed. First, the data were used to develop regression models to calculate the annual load of dissolved ammonia through the basin (Figure 2). Use of regression modeling and storm sampling for dissolved nutrients in small urban basins has not been widely investigated. Estimations of the average annual dissolved ammonia load were made using a variation of the USGS ESTIMATOR regression model. Our dataset allowed us to compare regression models based on different time steps by applying either instantaneous or daily average flow values to point concentrations. In addition, comparisons were made between models with and without detailed storm data. Without including storm sampling data, little difference was seen in the fit and load estimate using the different time-step models. Smaller time-step models with intensive storm sampling provided more precise load estimates than daily average time-step models or models with fewer storm samples. We concluded from this study that when designing a sampling strategy to make the most precise annual load estimate in a small watershed (< 100 km2), efforts should be made to collect detailed storm samples if flow data is available in intervals smaller than the daily average. Otherwise, collecting just a few samples (i.e., 1-3) per storm is adequate.
Figure 2. Predicted Ammonia Loading Generated With Regression Models Created Using Observations Taken During the National Water-Quality Assessment (NAWQA) Study Period and During the Nutrient Project Study Period
The work completed on regression modeling were presented at and published in the proceedings of the Environmental and Water Resources Institute 2005 Watershed Management Conference in Williamsburg, Virginia (Cutrofello and Durant, 2005).
Estimates of the average annual dissolved ammonia load from the basin are approximately 29,500 (95% CI: 21,000 to 42,000) kg NH4-N/year and 29,200 (21,000 to 41,100) kg NH4-N/year for models excluding and including storm samples, respectively. The yield or load per square mile corresponding to these load estimates is 4.6 kg NH4-N/year/mi2. Noting that approximately 30 percent of the total nitrogen load in the watershed is in the form of dissolved ammonia (Campo, et. al., 2003; Cutrofello and Durant, 2005), a total nitrogen yield of 15.1 kg N/year/mi2 was estimated. This total nitrogen yield is greater than the total nitrogen yields of 5 local watersheds of similar size and 16 larger watersheds, 5 of which are presented in Table 1, from across the country (Campo, et. al., 2003; Puckett, 1994).
Table 1. Annual Total Nitrogen Load Transported in Select NAWQA Watersheds (Puckett, 1994) Compared With the Aberjona Watershed
Watershed | kg-N/yr/ha |
Willamette River, OR |
7.7 |
White River, IN |
13.9 |
Susquehanna River, PA |
7.7 |
Connecticut River, CT |
5.4 |
Potomac River, DC |
8.9 |
Aberjona River, MA |
15.1 |
Next, we used regression modeling to quantify the amount of dissolved ammonia entering the river from its main tributary, Horn Pond Brook in the Horn Pond subbasin, and its most upstream reach, the Upper Aberjona subbasin (Cutrofello and Durant, 2005). The regression models support the conclusions already drawn using the point estimates of the original longitudinal study—the Upper Aberjona subbasin is the source of the majority of the ammonia flowing through the river, whereas the Horn Pond subbasin contributes only a small percentage of the total (Figure 3). Ignoring denitrification and other processes of the nitrogen cycle, just over 70 percent of the dissolved ammonia in the basin originates in the Upper Aberjona subbasin. Although some of this ammonia will be lost through plant uptake and some will be lost through denitrification after undergoing nitrification to nitrate, a large portion remains in the system. The roughly 7,500 kg NH4-N/year not accounted for in this analysis, could be attributed in part to sewage discharges to the river.
Further investigations into the Upper Aberjona subbasin revealed a possible source in the Halls Brook Holding Area (HBHA) tributary. Sampling completed at the outlet of HBHA and in the Aberjona River just north of the confluence with HBHA, as depicted in Figure 4, reveals loadings 10 to 50 times greater coming from HBHA than from upstream in the Aberjona River. The holding area consists of a shallow pond at the north fed by groundwater and surface water flow from Halls Brook. The pond feeds a long, narrow wetland, which joins the Aberjona River just north of Route 128 in Woburn. Previous studies have identified the groundwater flowing from the Industri-Plex Superfund site as a major contributor of contaminants to the pond at the north of the holding area.
Figure 3. Annual Average Dissolved Ammonia Loads With 95 Percent Confidence Intervals Calculated Using Regression Modeling
Figure 4. The HBHA With Sampling Locations Indicated by Stars. Groundwater flow from the Industri-Plex Superfund site enters the north basin of the pond.
To test the hypothesis that HBHA is the source of ammonia to the river, fixed monitoring stations were set up in the north and south basins of the pond as well as its surface water inlet, Halls Brook, its outlet into the wetlands, and at the outlet of the wetlands where HBHA joins the Aberjona River. Ten months of water column profiles reveal a meromictic pond with a highly saline bottom later containing extremely elevated levels of ammonia (> 500 mg/L). Surface water loadings from Halls Brook do not appear to account for the levels of ammonia in the pond. Sediment profiles from the north basin show pore water concentrations (as deep as 90 cm) similar in magnitude to the water of the meromictic layer. The north basin has been shown to receive groundwater from the Superfund site to the north.
The investigative and mass balance portion of the source detection will be submitted for publication to Environmental Science and Technology in early summer (Cutrofello and Durant 2005).
An important aside to this research relates to the U.S. Environmental Protection Agency’s investigation into the contamination originating on the Industri-Plex Superfund site. The recently released Multiple Groundwater Source Response Plan Remedial Investigation Report looks at metals, especially arsenic, benzene, and volatile organic compounds, with no attention paid to inorganic nitrogen in the groundwater. Using the results of the current research, a comment will be prepared to submit to the Feasibility Study and Proposed Plan for the site, which is scheduled to be completed in May 2005.
The Nitrogen Budget of the Aberjona Watershed (MIT). An essential part of the background for this research is a good knowledge of the fluxes of nitrogen that occur on the Aberjona watershed. The watershed, and the subwatersheds, for which fluxes are individually measured, are shown in Figure 5.
Integrated annual flux of dissolved inorganic nitrogen for 2003 to 2004 was calculated from 18 sampling dates where nitrate, ammonium, and flow were measured. The watershed as a whole exported approximately 74 metric tons of inorganic nitrogen, or 11 kg N/ha year, ranking it among the most heavily polluted North American basins. The relatively industrialized subbasin draining through Montvale Avenue (30% or 714 of 2304 ha classified as industrial by the USGS land use [LU] 21 code) exported approximately 61 metric tons of inorganic nitrogen, corresponding to 26 kg N/ha year; however, one-half of this nitrogen load apparently originates in the confined area of the HBHA (described in the previous section) and more nearly resembles a point source than nitrogen from nonpoint runoff. The average area loading of nitrogen in this northern subbasin of the Aberjona is thus probably closer to 13 kg N/ha year. The relatively forested and residential subbasin (only 6% industrial LU) draining through Wedge Pond and included Horn Pond exported about 11 metric tons of nitrogen, most of which is nitrate. On average ammonium dominates the Halls Brook source (700 mg/s compared to 220 mg/s nitrate), but by the time the river empties into the Upper Mystic Lake, nitrate is dominant (approximately 1700 vs. 700 mg/s ammonium). The flux of ammonium decreases by about 320 mg/s from Montvale Avenue downstream to the USGS sampling site. This is a gross indication of in-situ nitrification occurring against the backdrop of several sources and sinks of nitrogen species in the water column. Figure 6 shows the annual pattern of inorganic nitrogen fluxes at the USGS gage just upstream of the Mystic Lakes.
Figure 5. The Aberjona Watershed. Sites at which annual nitrogen fluxes have been determined are those indicated with bars representing total metric tons of dissolved inorganic nitrogen between March 2003–March 2004.
Figure 6. Fluxes of Inorganic Nitrogen for the Aberjona Watershed as a Function of Time of Year, January 2003–June 2004
For the Route 128, Montvale, Wedge Pond, and USGS sampling sites, flux is correlated with flow, with more scatter at higher flows. Concentration of nitrate and ammonium is correlated with flow at Route 128 but not downstream at Montvale or USGS. This may indicate that high concentrations of nitrogen at the Halls Brook sediment interface are getting pushed out of the basin during high flow events, whereas nitrogen scavenged from impermeable surfaces is diluted in runoff by precipitation in the rest of the watershed.
Nitrification Studies in the Aberjona River (MIT). As noted above, the nitrogen budget has strongly implicated nitrification as an important process in the Aberjona River. Figure 7 illustrates the concurrent increase in nitrate flux and decrease in ammonium flux observed as one goes downstream from Montvale to USGS.
Figure 7. Average Change in Flux of Ammonium and Nitrate Between Montvale and the USGS Gaging Station in the Aberjona River. Loss of ammonium, accompanied by gain in nitrate, suggests that much of the ammonium flux in the river is nitrified before entering the Mystic Lakes. This in turn is expected to have consequences for nutrient regeneration in the lake.
To address this question quantitatively, we have developed a novel incubation scheme that allows sediment or other materials from the river to be incubated at in-situ conditions, in flowing river water. Briefly, native sediments are transferred to channels made of polyvinyl chloride (PVC), within which they are allowed to equilibrate with natural river conditions for several weeks. The tubes are then capped, placed on a temperature-controlled circulation system, and incubated under conditions of flowing water under closed-system conditions in which fluxes from the sediment to the water can be closely followed. Tracers or metabolic inhibitors such as nitrapyrin can also be introduced. Importantly, the system allows long-term contact to occur, in the river, between the incubated sediment and the river water prior to incubation yet allows rigorous mass-balance measurements to be conducted. Initial trials to determine reproducibility among replicate tubes have provided promising results.
In-Lake studies (MIT)
In earlier work on this project, we demonstrated that the release of phosphorus from sediments in Upper Mystic Lake, which receives the drainage of the Aberjona Watershed, closely paralleled that of arsenic during summer anoxia. Thus, the coupling that was previously demonstrated between arsenic and nitrate also appears to hold true between phosphate and nitrate, as hypothesized at the beginning of this project. Specifically, this means that lake redox status as it may be controlled at times by nitrogen transformations is explicitly coupled to the cycle of phosphorus—a fact of likely importance to improved future lake nutrient models.
Because this stratified lake, located in Medford, Massachusetts, has a large amount of anaerobic metabolism of organic matter occurring in the sediments, it experiences accumulation of methane (CH4) caused by fermentation, in addition to accumulation of carbon dioxide (CO2) from respiratory reactions such as denitrification. To gain a quantitative understanding of the relative roles of these various processes, during the last year we followed the arsenic-phosphorus-nitrate study with a more complete redox balance of the Upper Mystic Lake hypolimnion. The summer of 2004 was of particular interest in the sense that oxygen depletion, and then nitrate depletion, occurred early in the season. During this period, we measured methane and carbon dioxide, as well as profiles for nitrate (NO3–), sulfate (SO43-), and iron (Fe2+) used to estimate the rates of the various anaerobic decomposition reactions.
The equivalent electron and carbon flow of the reactions were also calculated to obtain material balances within the hypolimnion of unified modeling language (UML). From the calculations, the approximate organic carbon decomposition rate, measured as carbon dioxide accumulation, was 5.5 mmol m–2 d–1. The amount of decomposition from the reactions involving nitrate, sulfate, iron and methane formation together accounted for only 50 percent of the total organic carbon decomposition. Therefore, 50 percent of the carbon dioxide accumulation in UML could not be accounted for. Possible explanations for the excess carbon dioxide production could be the formation of reduced iron minerals and/or the loss of methane caused by ebullition and oxidation. Such explanations suggest future studies of UML to better resolve the electron budget.
As part of this work, we also developed a novel and very effective gas-tight lake sampler, which was necessary because we needed to precisely sample dissolved gases in the water column as well as ionic electron acceptors. Previously, limnological sampling for dissolved gases involved filling glass bottles with water pumped from depth using a peristaltic pump; however, such methods introduce the potential for gas exchange with the atmosphere. Therefore, there was a need for a dissolved gas sampler that could be used to obtain samples at precise depth intervals while at the same time isolating the samples from outside influences.
This new in-situ sampling device is both inexpensive and simple in operation, making it ideal for routine use in collecting water samples for either monitoring or research. It captures 10s of ml of water per operation, from any depth in the water column, eliminates human and environmental contact with the sample, and allows samples to be analyzed by headspace equilibration if desired without transfer to other glassware. The sampler can be built from a series of three glass syringes, which when wetted are essentially gastight and operate with very low friction. Two of the syringes are connected end to end in a PVC setting and lowered to depth, whereas the third syringe is at the surface, connected to the others by a water-filled tube (Figure 8).
Figure 8. A Sampler Used to Monitor Dissolved Gas Concentrations in Lake Water
Samples are collected by flushing and pumping approximately 25 mL directly into the bottom sampling glass syringe and analyzed by headspace equilibration, followed by gas chromatography using a thermal conductivity detector. Figure 9 shows several methane and carbon dioxide profiles respectively, for Upper Mystic Lake during the sampling period. Because of the low cost and effectiveness of this sampler, we plan to publish the design, and we expect that other user groups involved in water sampling may find this design to be very helpful.
Modeling/Decision Support Studies (Tufts)
Models and decision-support software were developed for both the watershed and the receiving water.
Watershed Modeling (Tufts)
We refined the watershed loading function model by modifying internal calculation structure and by adding a module for a new best management practice (BMP) based on stakeholder feedback regarding their wish to explore the effect of fertilizer reduction within the watershed. This brought the total number of simulated BMPs to six, including porous pavement, street sweeping, detentions ponds, swales, bioretention storage, and fertilizer reduction programs. The model is intended as a user-friendly screening level tool for municipalities and decisionmakers to explore management alternatives and optimal approaches to meeting watershed nutrient reduction. The interface is spreadsheet-based to make it accessible to wide range of technical and nontechnical users. Screen shots of the user interface are shown in Figure 10.
Figure 9. Hypolimnetic (a) Methane and (b) Carbon Dioxide Concentrations in Upper Mystic Lake During the 2004 Sampling Period
Figure 10. Sample Screenshots of Spreadsheet-Based User Interface
Figures 11a-b show results of the nutrient model applied to the Aberjona River for the no-BMP base case, with the observed loads split into two intervals for model calibration and validation.
Figure 11. Monthly (a) Dissolved Nitrogen Load and (b) Particulate Phosphorus Load
BMP cost and effectiveness input data were refined and results for basin-wide nutrient load reduction are shown in Figure 12, with an example decision model result for the allocation of swales to various land uses shown in Figure 13. Details are provided in Limbrunner, et al. (2005)
Work on a more complex decision model is continuing, with the completion of a fully distributed storm water management model that preserves spatial connectivity of 120-m cells in the watershed. We are presently working on a similar event based model approach to study the effect and optimal location of sediment trapping BMPs, and a linear system analog for comparison of performance between linear and nonlinear optimization models. Figure 14 shows a result from the stormwater management model, detailing optimal location of infiltration-based BMPs. Details are provided by Perez-Pedini, et al. (2005).
Figure 12. Long-Term Basin-Wide Nutrient Load Reduction
Figure 13. Optimal Allocation of Swales Among Land Uses for Various Remedial Budgets
Figure 14. Optimal Entry of Stormflow-Reducing, Infiltration-Based BMPs for Allocations of (a) 25, (b) 100, (c) 150, and (d) 400 BMPs
Our intent is to compare optimal distributed solutions to the storm flow problem with those of the sediment yield problem to determine whether certain BMP plans are effective for both water quantity and water quality management.
In-Lake Modeling (Tufts)
The primary goal for Year 2 of the project was to test the lake model that we had developed in Year 1. We did this by applying it to our study lake (the Upper Mystic Lake) as well as other lakes for which adequate datasets exist. For the latter, we chose an oligotrophic system (Platte Lake, Michigan), a highly eutrophic lake (Onondaga Lake, New York), and a lake that experienced a wide range of trophic states over the period of record (Lake Washington, Washington).
The model performed adequately in all cases; however, the exercise resulted in the identification of several model refinements that we will implement in Year 3. These include: (1) the incorporation of luxury uptake in the phytoplankton; (2) the simulation of pH, alkalinity, and calcite precipitation; (3) refinement of the transparency submodel; and (4) the inclusion of phosphorus sorption on inorganic particles. In addition, we will incorporate denitrification in a manner consistent with results of the in-lake scientific studies conducted by MIT as described above.
Stakeholder Interactions (Tufts and MIT)
There have been two direct stakeholder meetings during the project. These are summarized below. In addition, stakeholder interaction and coordination have occurred through meetings of the Mystic Watershed Collaborative (MWC) and the biannual Mystic Watershed Research Conferences. We have presented the decision support model to staff at the Massachusetts Office of Coastal Zone Management and to a community group within the watershed, called the Friends of Winter Pond. These additional meetings were conducted as informal seminars to discuss our approach, the interface, and possible applications of the model.
2003 Stakeholder Meeting. A 2-hour meeting was held on November 17, 2003, with some of the project team and four stakeholders from the watershed. They represented the Mystic River Watershed Association (MyRWA) and several knowledgeable private citizens. The federal, state, and local officials; an engineering consultant; and a hydrologic scientist who previously had committed to participating were unable to attend.
The initial project goals and methodology were described during the meeting. Comments included:
- Eventually the output of the decision support system should be on a scale helpful to actually identify the particular location of a BMP.
- BMPs should include behavioral actions an individual property owner can undertake such as reducing fertilizer and installing porous pavement.
- Education of all basin property owners is important in BMP implementation.
- Need to understand trade-offs between costs and water quality.
2005 Stakeholder Meeting. A formal meeting was held with the project stakeholder advisory committee on March 8, 2005, to present project results to date and stakeholder needs and feedback. This meeting included 16 people representing the research team, MyRWA, Friends of the Mystic River, Friends of Upper Mystic Lake, Friends of Winter Pond, Winchester Conservation Commission, Massachusetts Department of Environmental Protection, and some interested citizens. As a result of this meeting, more BMPs were added to the modeling efforts, and some new local data sources on BMP performances were supplied to the research team. A future meeting is to be held in early 2006.
Mystic Watershed Collaborative
The MWC is a formal partnership between Tufts University and the MyRWA to restore the Mystic River. Tufts’ role is to provide expertise to the effort through involvement of faculty and student research projects. The MWC meets bimonthly where presentations are made on ongoing Tufts-related research (e.g. the Improved Science and Decision Support for Managing Watershed Nutrient Loads Project), joint Tufts-MyRWA projects and the needs of the community. The MWC started in 1999. Through the MWC, this research project has been discussed several times each year since its inception. The result has been constant updating of MyRWA on the status of the research and its findings. This has resulted in more opportunities for future use of the tools we are developing. For example, in 2005 the Massachusetts Department of Environmental Protection expressed possible interest in using our modeling tools for a nutrient total maximum daily load study of the Mystic River, and the Tri-Town Upper Mystic Stormwater Management Committee may be interested in using the tool for stormwater management.
Research Symposia
Many federal, state, local, citizen, and university organizations are conducting water-related research in the Mystic Watershed. Biannually, there is a Research Symposium held at Tufts University where activities are updated and results shared and discussed. Project members made two presentations on April 3, 2004, at the Mystic 2010: Research Supporting Watershed Goals symposium, which was attended by approximately 75 citizens, scientists, and government officials (including several members of our Stakeholder Group) with interest in our study watershed. Project members will also be making presentations at the next symposium in Spring 2006.
Future Activities:
Along with continuing to publish and present our results, the major focus areas for our research for Year 3 will be:
Watershed Nutrient Studies (Tufts University)
Future plans include calculating mass balances around the pond to provide quantitative support to the hypothesis of a groundwater dissolved ammonia source. We are also planning a nitrogen tracer (δ15N) study of the groundwater, sediment pore water, and the water column to determine the source(s) of ammonia to the pond. Lastly, storm sampling around the pond will be conducted to test the hypothesis that mechanical mixing during large storm events may wash the dissolved ammonia from the bottom, saline water layer out of the pond, through the wetlands, and into the Aberjona River.
Watershed and Lake Nutrient Studies (MIT)
We will carry out the nitrification study. Now that the incubation apparatus has been constructed and tested, this work is ready to proceed with the incubation of river samples under various conditions, including the addition of nitrapyrin to block nitrification in a portion of samples and thus help separate the role of nitrification from that of other sources and sinks.
We will complete the verification of organic nitrogen measurements, as part of completing the nitrogen budget of the watershed. As discussed in our previous annual report, organic nitrogen appears to be a significant fraction of total nitrogen at certain times and locations. Because of the known uncertainties in total nitrogen measurement (refractory materials, etc.) we will replicate our wet-digestion-based organic nitrogen measurements with measurements made independently on a CHN analyzer using combustion techniques.
We are writing up two papers on the in-lake work (one on the redox balance, which will also include the new sampler design) and another on the nitrogen-phosphorus-iron-arsenic link.
We intend to run at least two more river sediment incubations to determine if these results support our inferences about nitrification rates in the river.
Watershed Modeling (Tufts)
Work on a more complex decision model is continuing, with the completion of a fully distributed storm water management model that preserves spatial connectivity of 120-m cells in the watershed. We are presently working on a similar event based model approach to study the effect and optimal location of sediment trapping BMPs. Figure 5 shows a result from the stormwater management model, detailing optimal location of infiltration-based BMPs. Details are provided by Perez-Pedini, et al. (2005). Our intent is to compare optimal distributed solutions to the storm flow problem with those of the sediment yield problem, to determine whether certain BMP plans are effective for both water quantity and water quality management. Work is also continuing on a comparison of the use of nonlinear distributed decision-support systems based on Genetic Algorithms, which require very long computing times, with a very efficient linear programming based distributed support system.
Lake Modeling (Tufts)
Incorporation and testing of new model mechanisms as outlined above.
Development of a decision-support system linking the watershed model with the lake model. This will provide a tool that will allow our research to be applied for environmental problem solving.
Stakeholder Interactions (Tufts and MIT)
There will be at least one more direct stakeholder meeting to present our final results and solicit feedback.
References:
Campo KW, Flanagan SM, Robinson KW. Water quality of selected rivers in the New England Coastal Basins in Maine, Massachusetts, New Hampshire, and Rhode Island, 1998-2000. U.S. Geological Survey Water-Resources Investigations Report 03-4210, Pembroke, NH, 2003.
Massachusetts Year 2004 Integrated List of Waters. Division of Watershed Management - Watershed Planning Program, Worcester, MA, 2004. (January 7, 2005).
Puckett LJ. Nonpoint and point sources of nitrogen in major watersheds of the United States. U.S. Geological Survey Water-Resources Investigations Report 94-4001, Reston, VA, 1994.
Journal Articles on this Report : 1 Displayed | Download in RIS Format
Other project views: | All 36 publications | 5 publications in selected types | All 1 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Perez-Pedini C, Limbrunner JF, Vogel RM. Optimal location of infiltration-based best management practices for storm water management. Journal of Water Resources Planning and Management 2005;131(6):441-448. |
R830654 (2003) R830654 (2004) |
Exit Exit |
Supplemental Keywords:
land, sediments, nutrients, eutrophication, ecological effects, heavy metals, decisionmaking, community-based, cost benefit, environmental chemistry, biology, physics, engineering, ecology, hydrology, mathematics, limnology, modeling, northeast, Aberjona River, Mystic Lake, GIS, decision support, stakeholders, anthropogenic processes, aquatic biota, aquatic ecosystems, nitrate concentrations, phosphorus, ammonia, oxygen, Secchi depth, clarity, suspended solids, nutrient flux, nutrient transport, restoration planning, watershed assessment, watershed management,, RFA, Scientific Discipline, INTERNATIONAL COOPERATION, Water, ECOSYSTEMS, Ecosystem Protection/Environmental Exposure & Risk, Aquatic Ecosystems & Estuarine Research, Water & Watershed, Aquatic Ecosystem, Water Quality Monitoring, Biochemistry, Environmental Monitoring, Terrestrial Ecosystems, Ecology and Ecosystems, Watersheds, anthropogenic stress, bioassessment, anthropogenic processes, watershed classification, nutrient transport, ecosystem monitoring, watershed management, biodiversity, nutrient flux, conservation, diagnostic indicators, ecosystem indicators, biota diversity, Mystic Lake, aquatic ecosystems, bioindicators, watershed sustainablility, water quality, biological indicators, ecosystem stress, watershed assessment, conservation planning, ecosystem response, aquatic biota, restoration planning, watershed restoration, biological impairmentProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.