Linking Nitrogen Export to Landscape Heterogeneity: The Role of Infrastructure and Storm Flows in a Mediterranean Urban System

Urban ecosystems are often sources of nonpoint source (NPS) nitrogen (N) pollution to aquatic ecosystems. However, N export from urban watersheds is highly variable. Examples of densely urbanized watersheds are not well studied, and these may have comparatively low export rates. Commonly used metrics of landscape heterogeneity may obscure our ability to discern relationships among landscape characteristics that can explain these lower export rates. We expected that differences not often captured by these metrics in the relative cover of vegetation, structures, and impervious surfaces would better explain observed variation in N export. We examined these relationships during storms in residential watersheds. Contrary to expectations, land cover did not directly predict variation in N or water export. Instead, N export was strongly linked to drainage infrastructure density. Our research highlights the role of fine‐scaled landscape attributes, mainly infrastructure, in explaining patterns of N export from densely urbanized watersheds. Changes to hydrologic flow paths by infrastructure explained more variation in N export than land cover. Our findings support further development of landscape ecological models of urban N export that focus on hydrologic modification by infrastructure rather than traditional landscape measures such as land use, as indicators for evaluating patterns of NPS nitrogen pollution in densely urbanized watersheds.     

WATER QUALITY IN AGRICULTURAL,URBAN, AND MIXED LAND USE WATERSHEDS1

ABSTRACT: Water quality and nonpoint source (NPS) pollution are important issues in many areas of the world, including the Inner Bluegrass Region of Kentucky where urban development is changing formerly rural watersheds into urban and mixed use watersheds. In watersheds where land use is mixed, the relative contributions of NPS pollution from rural and urban land uses can be difficult to separate. To better understand NPS pollution sources in mixed use watersheds, surface water samples were taken at three sites that varied in land use to examine the effect of land use on water quality. Within the group of three watersheds, one was predominately agriculture (Agricultural), one was predominately urban (Urban), and a third had relatively equal representation of both types of land uses (Mixed). Nitrogen (N), phosphorus (P), total suspended solids (TSS), turbidity, pH, temperature, and streamflow were measured for one year. Comparisons are made among watersheds for concentration and fluxes of water quality parameters. Nitrate and orthophosphate concentrations were found to be significantly higher in the Agricultural watershed. Total suspended solids, turbidity, temperature, and pH, were found to be generally higher in the Urban and Mixed watersheds. No differences were found for streamflow (per unit area), total phosphorus, and ammonium concentrations among watersheds. Fluxes of orthophosphate were greater in the Agricultural watershed that in the Urban watershed while fluxes of TSS were greater in the Mixed watershed when compared to the Agricultural watershed. Fluxes of nitrate, ammonium, and total phosphorus did not vary among watersheds. It is apparent from the data that Agricultural land uses are generally a greater source of nutrients than the Urban land uses while Urban land uses are generally a greater source of suspended sediment.     

Assessing the Relationship Between Landscape Patterns and Nonpoint‐Source Pollution in the Danjiangkou Reservoir Basin in China1

Abstract: The constrained ordination method from quantitative ecology was utilized to assess the relationship between landscape patterns and nonpoint‐source (NPS) pollution for the purpose of identifying effective water‐quality improvement practices in Danjiangkou Reservoir (DJKR) basin, China. The soil and water assessment tool (SWAT) was applied to simulate NPS pollution and the Fragstats model was applied to calculate the landscape metrics. The study concluded that organic nutrients formed the main NPS pollutant in the DJKR basin and that most of the NPS pollution occurred along with soil loss. Based on partial redundancy analysis, the conclusion that landscape metrics were significantly correlated to NPS pollution indices was obtained. Specifically, the composition of LULC (land use/land cover) was the most effective factor to estimate NPS pollution. Dry cultivated land was identified as the main source of NPS pollution, and paddy fields were characterized with the most intensive soluble nutrients loss. In addition, the reason that fragmented and complex landscape patterns exacerbate NPS pollution was that natural landscape composed most of this area. Moreover, the fragmented natural landscape indicated intensive agricultural activities that were the crucial trigger for NPS pollution. Combined with the economic condition in China, Conversion of Cropland to Forests Program (CCFP) should be conducted selectively and gradually in the DJKR basin.     

LAND COVER AS A FRAMEWORK FOR ASSESSING RISK OF WATER POLLUTION1

ABSTRACT: A survey of numerous field studies shows that nitrogen and phosphorous export coefficients are significantly different across forest, agriculture, and urban land‐cover types. We used simulations to estimate the land‐cover composition at which there was a significant risk of nutrient loads representative of watersheds without forest cover. The results suggest that at between 20 percent and 30 percent nonforest cover, there is a 10 percent or greater chance of N or P nutrient loads being equivalent to the median values of predominantly agricultural or urban watersheds. The methods apply to environmental management for assessing the risk to increased nonpoint nutrient pollution. Interpretation of the risk measures are discussed relative to their application for a single watershed and across a region comprised of several watersheds.     

GROUPWISE MODELING STUDY OF BACTERIALLY IMPAIRED WATERSHEDS IN TEXAS: CLUSTERING ANALYSIS1

ABSTRACT: Under the Clean Water Act (CWA) program, the Texas Commission on Environmental Quality (TCEQ) listed 110 stream segments in the year 2000 with pathogenic bacteria impairment. A study was conducted to evaluate the probable sources of pollution and characterize the watersheds associated with these impaired water bodies. The primary aim of the study was to group the water bodies into clusters having similar watershed characteristics and to examine the possibility of studying them as a group by choosing models for total maximum daily load (TMDL) development based on their characteristics. This approach will help to identify possible sources and determine appropriate models and hence reduce the number of required TMDL studies. This in turn will help in reducing the effort required to restore the health of the impaired water bodies in Texas. The main characteristics considered for the classification of water bodies were land use distribution within the watershed, density of stream network, average distance of land of a particular use to the closest stream, household population, density of on‐site sewage facilities (OSSFs), bacterial loading from different types of farm animals and wildlife, and average climatic conditions. The climatic data and observed instream fecal coliform bacteria concentrations were analyzed to evaluate seasonal variability of instream water quality. The grouping of water bodies was carried out using the multivariate statistical techniques of factor analysis/principal component analysis, cluster analysis, and discriminant analysis. The multivariate statistical analysis resulted in six clusters of water bodies. The main factors that differentiated the clusters were found to be bacterial contribution from farm animals and wildlife, density of OSSFs, density of households connected to public sewers, and land use distribution.     

AGRICULTURAL LAND USE IMPACTS ON BACTERIAL WATER QUALITY IN A KARST GROUNDWATER AQUIFER1

ABSTRACT: The impact on water quality by agricultural activity in karst terrain is an important consideration for resource management within the Appalachian Region. Karst areas comprise about 18 percent of the Region's land area. An estimated one-third of the Region's farms, cattle, and agricultural market value are on karst terrain. The purpose of this study was to compare fecal bacteria densities in karst groundwater impacted by two primary agricultural land uses in central Appalachia. Fecal bacteria densities were measured in cave streams draining two primary land management areas. The first area was pasture serving a beef cow-calf operation. The second area was a dairy. Neither area had best management practices in place for controlling animal wastes. Median fecal coliform and fecal streptococcus densities were highest in cave streams draining the dairy. Median fecal coliform densities in the dairy-impacted stream were greater than 4,000 CFU/100 ml and the median fecal coliform densities in the pasture-impacted streams were less than 10 CFU/100 ml. Median fecal streptococcus densities in the same streams were greater than 2,000 CFU/100 ml and 32 CFU/100 nil, respectively. A second dairy, with best management practices for control of animal and milkhouse waste, did not appear to be contributing significant amounts of fecal bacteria to the karst aquifer. It was concluded that agriculture was affecting bacterial densities in the karst aquifer. New management practices specifically designed to protect karst groundwater resources may be one way to protect the groundwater resource.     

Effects of Impervious Cover at Multiple Spatial Scales on Coastal Watershed Streams1

Abstract: The spatial scale and location of land whose development has the strongest influence on aquatic ecosystems must be known to support land use decisions that protect water resources in urbanizing watersheds. We explored impacts of urbanization on streams in the West River watershed, New Haven, Connecticut, to identify the spatial scale of watershed imperviousness that was most strongly related to water chemistry, macroinvertebrates, and physical habitat. A multiparameter water quality index was used to characterize regional urban nonpoint source pollution levels. We identified a critical level of 5% impervious cover, above which stream health declined. Conditions declined with increasing imperviousness and leveled off in a constant state of impairment at 10%. Instream variables were most correlated (0.77 ≤ |r| ≤ 0.92, p < 0.0125) to total impervious area (TIA) in the 100‐m buffer of local contributing areas (～5‐km² drainage area immediately upstream of each study site). Water and habitat quality had a relatively consistent strong relationship with TIA across each of the spatial scales of investigation, whereas macroinvertebrate metrics produced noticeably weaker relationships at the larger scales. Our findings illustrate the need for multiscale watershed management of aquatic ecosystems in small streams flowing through the spatial hierarchies that comprise watersheds with forest‐urban land use gradients.     

Landscape Planning for Agricultural Non–Point Source Pollution Reduction. II. Balancing Watershed Size,Number of Watersheds,and Implementation Effort

Agricultural non–point source (NPS) pollution poses a severe threat to water quality and aquatic ecosystems. In response, tremendous efforts have been directed toward reducing these pollution inputs by implementing agricultural conservation practices. Although conservation practices reduce pollution inputs from individual fields, scaling pollution control benefits up to the watershed level (i.e., improvements in stream water quality) has been a difficult challenge. This difficulty highlights the need for NPS reduction programs that focus efforts within target watersheds and at specific locations within target watersheds, with the ultimate goal of improving stream water quality. Fundamental program design features for NPS control programs—i.e., number of watersheds in the program, total watershed area, and level of effort expended within watersheds—have not been considered in any sort of formal analysis. Here, we present an optimization model that explores the programmatic and environmental trade-offs between these design choices. Across a series of annual program budgets ranging from $2 to $200 million, the optimal number of watersheds ranged from 3 to 27; optimal watershed area ranged from 29 to 214 km²; and optimal expenditure ranged from $21,000 to $35,000/km². The optimal program configuration was highly dependent on total program budget. Based on our general findings, we delineated hydrologically complete and spatially independent watersheds ranging in area from 20 to 100 km². These watersheds are designed to serve as implementation units for a targeted NPS pollution control program currently being developed in Wisconsin.     

A GEOGRAPHIC INFORMATION SYSTEM TO PREDICT NON-POINT SOURCE POLLUTION POTENTIAL1

ABSTRACT: Bacterial densities (total coliform, fecal coliform, and fecal streptococci) and suspended solids in runoff from a feedlot, pasture, and corn field were measured. Densities of fecal coliform were highest from the feedlot but were 1000 to 10,000 times greater than the water quality standard for swimmable waters from all three land uses. Densities of fecal streptococci were highest from the corn field, which suggests that wildlife are the source of bacteria. Fecal coliform/fecal streptococci ratios distinguished cattle from wildlife as the source of bacterial pollution both among land uses and among seasons of the year. Suspended solids concentrations in runoff ranged from 423 to 925 mg/l and were highest from the corn field. A Geographic Information System (GIS), which utilizes a raster or grid-cell format, was developed to include algorithms associated with non-point source pollution. The system accepts digitally mapped information on soil type, topography, and land use. It calculates characteristics such as slope and slope length, and relates these characteristics to soils and land use parameters in order to produce three dimensional maps of runoff potential, sediment pollution potential, and bacterial pollution potential. It offers the advantages of retaining the geographic character of pollution potential information and of conveying in three-dimensional graphical terms the effects of topography, soil type, land use, and land management practices.     

A Water Quality Model for Regional Stream Assessment and Conservation Strategy Development

Non-point-source (NPS) pollution remains the primary source of stream impairment in the United States. Many problems such as eutrophication, sedimentation, and hypoxia are linked with NPS pollution which reduces the water quality for aquatic and terrestrial organisms. Increasingly, NPS pollution models have been used for landscape-scale pollution assessment and conservation strategy development. Our modeling approach functions at a scale between simple landscape-level assessments and complex, data-intensive modeling by providing a rapid, landscape-scale geographic information system (GIS) model with minimal data requirements and widespread applicability. Our model relies on curve numbers, literature-derived pollution concentrations, and land status to evaluate total phosphorus (TP), total nitrogen (TN), and suspended solids (SS) at the reach scale. Model testing in the Chesapeake Bay watershed indicated that predicted distributions of water quality classes were realistic at the reach scale, but precise estimates of pollution concentrations at the local scale can have errors. Application of our model in the tributary watersheds along Lake Ontario suggested that it is useful to managers in watershed planning by rapidly providing important information about NPS pollution conditions in areas where large data gaps exist, comparisons among stream reaches across numerous watersheds are required, or regional assessments are sought.     

Recent Changes in Stream Flashiness and Flooding,and Effects of Flood Management in North Carolina and Virginia

The southeastern United States has undergone anthropogenic changes in landscape structure, with the potential to increase (e.g., urbanization) and decrease (e.g., reservoir construction) stream flashiness and flooding. Assessment of the outcome of such change can provide insight into the efficacy of current strategies and policies to manage water resources. We (1) examined trends in precipitation, floods, and stream flashiness and (2) assessed the relative influence of land cover and flow‐regulating features (e.g., best management practices and artificial water bodies) on stream flashiness from 1991 to 2013. We found mean annual precipitation decreased, which coincided with decreasing trends in floods. In contrast, stream flashiness, overall, showed an increasing trend during the period of study. However, upon closer examination, 20 watersheds showed stable stream flashiness, whereas 5 increased and 6 decreased in flashiness. Urban watersheds were among those that increased or decreased in flashiness. Watersheds that increased in stream flashiness gained more urban cover, lost more forested cover and had fewer best management practices installed than urban watersheds that decreased in stream flashiness. We found best management practices are more effective than artificial water bodies in regulating flashy floods. Flashiness index is a valuable and straightforward metric to characterize changes in streamflow and help to assess the efficacy of management interventions.     

AVGWLF‐Based Estimation of Nonpoint Source Nitrogen Loads Generated Within Long Island Sound Subwatersheds1

Abstract: The Generalized Watershed Loading Functions (GWLF) model and its ArcView interface (AVGWLF) were used to estimate and examine the components of the total nitrogen (TN) nonpoint source (NPS) load generated within New York and Connecticut (CT) watersheds surrounding Long Island Sound (LIS, the Sound). The majority of data used as model inputs were generally available from online sources, and the work involved an overall calibration to streamflow and TN data in accordance with generic guidelines recommended in the GWLF manual. The GWLF model performance for three calibration and two validation watersheds in CT was compared with results of a detailed model, Hydrological Simulation Program in Fortran, developed in a previous study. The results of the application illustrate the usefulness of the relatively simpler, less parameter‐intensive GWLF model in performing exploratory loading analysis in preparation for adaptive nutrient management in the LIS watersheds. The presented methodology is valuable for identification of priority watersheds for NPS pollution reduction and also for planning‐level evaluation of best management practices to achieve the desired reductions. It is estimated that ground‐water base flow may be the largest pathway for NPS TN to the Sound, contributing about 54% of the total NPS TN load, a finding with significant implications for LIS total maximum daily load reduction scenarios. In addition to ground water, septic systems are estimated to contribute about 17% of the total load, with the remaining TN load being mostly runoff from urban (17%), agricultural (5%), and low impact (e.g., forest) areas (6%).     

MODELING THE DISTRIBUTION OF DIFFUSE NITROGEN SOURCES AND SINKS IN THE NEUSE RIVER BASIN OF NORTH CAROLINA,USA1

This study quantified nonpoint source nitrogen (NPS‐N) sources and sinks across the 14,582 km² Neuse River Basin (NRB) located in North Carolina, to provide tabular data summaries and graphic overlay products to support the development of management approaches to best achieve established N reduction goals. First, a remote sensor derived, land cover classification was performed to support modeling needs. Modeling efforts included the development of a mass balance model to quantify potential N sources and sinks, followed by a precipitation event driven hydrologic model to effectively transport excess N across the landscape to individual stream reaches to support subsequent labeling of transported N values corresponding to source origin. Results indicated that agricultural land contributed 55 percent of the total annual NPS‐N loadings, followed by forested land at 23 percent (background), and urban areas at 21 percent. Average annual N source contributions were quantified for agricultural (1.4 kg/ha), urban (1.2 kg/ha), and forested cover types (0.5 kg/ha). Nonpoint source‐N contributions were greatest during the winter (40 percent), followed by spring (32 percent), summer (28 percent), and fall (0.3 percent). Seasonal total N loadings shifted from urban dominated and forest dominated sources during the winter, to agricultural sources in the spring and summer. A quantitative assessment of the significant NRB land use activities indicated that high (greater than 70 percent impervious) and medium (greater than 35 percent impervious) density urban development were the greatest contributors of NPS‐N on a unit area basis (1.9 and 1.6 kg/ha/yr, respectively), followed by row crops and pasture/hay cover types (1.4 kg/ha/yr).     

PREDICTABILITY OF SURFACE WATER POLLUTION LOADING IN PENNSYLVANIA USING WATERSHED‐BASED LANDSCAPE MEASUREMENTS1

ABSTRACT: We formally evaluated the relationship between landscape characteristics and surface water quality in the state of Pennsylvania (USA) by regressing two different types of pollutant responses on landscape variables that were measured for whole watersheds. One response was the monthly exported mass of nitrogen estimated from field measurements, while the other response was a GIS‐modeled pollution potential index. Regression models were built by the stepwise selection protocol, choosing an optimal set of landscape predictors. After factoring out the effect of physiography, the dominant predictors were the proportion of “annual herbaceous” land and “total herbaceous” land for the nitrogen loading and pollution potential index, respectively. The strength of these single predictors is encouraging because the marginal land cover proportions are the simplest landscape measurements to obtain once a land cover map is in hand; however, the optimal set of predictors also included several measurements of spatial pattern. Thus, for watersheds at this general hierarchical scale, gross landscape pattern may be an important influence on instream pollution loading. Overall, there is strong evidence that using landscape measurements alone, obtained solely from remotely sensed data, can explain most of the water quality variability (R²= approx. 0.75) within these watersheds.     

STATEWIDE EMPIRICAL MODELING OF BACTERIAL CONTAMINATION OF SURFACE WATERS1

ABSTRACT: Bacterial contamination of surface waters is attributed to both urban and agricultural land use practices and is one of the most frequently cited reasons for failure to meet standards established under the Clean Water Act (CWA) (P.L. 92–500). Statewide modeling can be used to determine if bacterial contamination occurs predominantly in urban or agricultural settings. Such information is useful for directing future monitoring and allocating resources for protection and restoration activities. Logistic regression was used to model the likelihood of bacterial contamination using watershed factors for the state of Maryland. Watershed factors included land cover, soils, topography, hydrography, locations of septic systems, and animal feeding operations. Results indicated that bacterial contamination occurred predominantly in urban settings. Likelihood of bacterial contamination was highest for small watersheds with well drained and erodible soils and a high proportion of urban land adjacent to streams. The number of septic systems and animal feeding operations and the amount of agricultural land were not significant explanatory factors. The urban infrastructure tends to “connect” more of the watershed to the stream network through the creation of roads, storm sewers, and wastewater treatment plants. This may partly explain the relationship between urbanization and bacterial contamination found in this study.     

Spatial Variations in the Relationships between Land Use and Water Quality across an Urbanization Gradient in the Watersheds of Northern Georgia, USA

A spatial statistical technique, Geographically Weighted Regression (GWR) is applied to study the spatial variations in the relationships between four land use indicators, including percentages of urban land, forest, agricultural land, and wetland, and eight water quality indicators including specific conductance (SC), dissolved oxygen, dissolved nutrients, and dissolved organic carbon, in the watersheds of northern Georgia, USA. The results show that GWR has better model performance than ordinary least squares regression (OLS) to analyze the relationships between land use and water quality. There are great spatial variations in the relationships affected by the urbanization level of watersheds. The relationships between urban land and SC are stronger in less-urbanized watersheds, while those between urban land and dissolved nutrients are stronger in highly-urbanized watersheds. Percentage of forest is an indicator of good water quality. Agricultural land is usually associated with good water quality in highly-urbanized watersheds, but might be related to water pollution in less-urbanized watersheds. This study confirms the results obtained from a similar study in eastern Massachusetts, and so suggest that GWR technique is a very useful tool in water environmental research and also has the potential to be applied to other fields of environmental studies and management in other regions.     

Identification of the source of nitrate contamination in ground water below an agricultural site, Jeungpyeong, Korea

This study applied hydrogeological characterization and isotope investigation to identify source locations and to trace a plume of ground water contaminated by nitrate. Most of the study site is agricultural fields with the remainder being residential. A poultry farm is also within the study area, so that potential point and nonpoint sources were present. Estimates of seasonal ground water recharge from irrigation and precipitation, leakage of sewage, and the regional ground water flow were linked to the seasonal changes in isotopic values. Ground water recharge largely occurred in spring and summer following precipitation or irrigation, depending on the locations. Natural and fertilized soils were identified as nonpoint sources of nitrate contamination in this area, while septic and animal wastes were identified as small point sources. The seasonal changes in the relative impact of these sources on ground water contamination were related to such factors as source distribution, the aquifer confining condition, precipitation rate, infiltration capacity, recharge rate, and the land use pattern.     

Fecal coliform loadings and stocks in buttermilk bay,Massachusetts, USA,and management implications

Abundance of fecal caliform bacteria is a weak index of the presence of human pathogens in wastewater entering coastal waters. In spite of this, use of fecal caliform indices for management purposes is widespread. To gain insight into interpretation of fecal coliform data, we evaluated size of stocks of fecal coliforms in water, sediments, and sea wrack, in Buttermilk Bay, a coastal embayment in Massachusetts. Sediments contained most of the fecal coliforms. Fecal coliforms in sediments were as much as one order of magnitude more abundant than in the water column or in sea wrack. The fecal coliforms in sediments of Buttermilk Bay were so abundant that resuspension of fecal coliforms from just the top 2 cm of muddy sediments could add sufficient cells to the water column to have the whole bay exceed the federal limit of fecal coliforms for shellfishing. The major sources of fecal coliforms to the bay were water-fowls, surface runoff, groundwater, and streams. Waterfowl were the largest source of fecal coliforms during cold months; surface runoff, streams, and groundwater were most important during warm months. Redirection of surface runoff pipes is unlikely to be a very successful management action since contributions via this source are insufficient to account for the measured increases in concentrations of fecal coliforms in water. Removal of waterfowl is also unlikely to be useful, since fecal coliform concentrations leading to closures of shellfish beds and swimming areas are most frequent during warm months when waterfowl are rarest. Rates of loss of fecal caliform cells from the water column by death and tidal exchange were high. Mortality of cells was about an order of magnitude larger than losses by tidal exchange. The amounts of fecal coliforms brought into the bay by waterfowl, surface runoff, groundwater, and streams are an order of magnitude smaller than the losses by mortality and tidal removal. This implies that there is an additional source of fecal coliforms within the bay. We suggest that resuspension of the upper layers of sediments can easily account for the fecal coliforms present in the water. Fecal coliform content of water and shellfish were not correlated. In contrast, sediment and shellfish fecal coliform abundances were significantly related. Monitoring of fecal coliforms in sediments may provide a better assessment of shellfish than sampling of water. The large fecal coliform stock in sediments should be the first priority for management. Efforts ought to be directed toward the reduction of sediment fecal coliform stocks. Lowering nutrient additions to coastal water bodies may be one practical approach.     

WATERSHED BASED,INSTITUTIONAL APPROACH TO DEVELOPING CLEAN WATER RESOURCES1

ABSTRACT: Access to clean and sufficient amounts of water is a critical problem in many countries. A watershed approach is vital in understanding pollution pathways affecting water resources and in developing participatory solutions. Such integration of information with participatory approaches can lead to more sustainable solutions than traditional “crisis‐to‐crisis” management approaches. This study aims at applying a watershed based joint action approach to manage water resources. Since most watersheds have urban and rural sources of pollution and a wide disparity in access to and use of water, alternative solutions need to take an integrated approach through cooperative actions. An institutional model was applied to seven subwatersheds in Honduras to evaluate various sources and effects of water contamination and water shortages. Two specific pathways of water resources degradation were studied (contamination from coffee pulp manufacturing and urban nonpoint sources) to develop alternative solutions that mitigate downstream impacts of access to clean water. A locally driven joint mechanism to reuse coffee pulp in farming systems is proposed. Such an institutional solution can maximize benefits to both farms and the coffee pulp industry. A combination of education and investment in sanitary facilities in urbanizing areas is proposed to minimize urban sources of water contamination.     

COMPARATIVE EVALUATION OF LAND COVER DATA SOURCES FOR EROSION PREDICTION1

ABSTRACT: A fundamental problem in protecting surface drinking water supplies is the identification of sites highly susceptible to soil erosion and other forms of nonpoint source (NPS) pollution. The New York City Department of Environmental Protection is trying to identify erodible sites as part of a program aimed at avoiding costly filtration. New York City's 2,000 square mile watershed system is well suited for analysis with geographic information systems (GIS); an increasingly important tool to determine the spatial distribution of sensitive NPS pollution areas. This study used a GIS to compare three land cover sources for input into the Modified Universal Soil Loss Equation (MUSLE), a model estimating soil loss from rangeland and forests, for a tributary watershed within New York City's water supply system. Sources included both conventional data (aerial photography) and Landsat data (MSS and TM images). Although land cover classifications varied significantly across these sources, location-specific and aggregate watershed predictions of the MUSLE were very similar. We conclude that using Landsat TM imagery with a hybrid classification algorithm provides a rapid, objective means of developing large area land cover databases for use in the MUSLE, thus presenting an attractive alternative to photo interpretation.