AWRA banner
Advancing Water Resources Research and Management

AWRA SYMPOSIUM ON GIS AND WATER RESOURCES
Sept 22-26, 1996
Ft. Lauderdale, FL

-----

Watershed-Scale Nonpoint Source Pollution Modeling and Decision Support System Based on a Model-GIS-RDBMS Linkage

Jaewan Yoon (1)

HTML 2.0 version (No Form, No Table) also available

Table of Contents

Abstract
Introduction
Study Watershed
Geograpphic Information System (GIS)
Agricultural Nonpoint Source Pollution (AGNPS) Model
Direct Linkage between Models and GIS
Methodology
Critical Area Assessment and BMP Scenario Simulation
Summary and Conclusions
References

Abstract

 Methods were developed for directly linking the distributed parameter model, AGNPS with a Geographic Information System (GIS) and a relational database management system (RDBMS) to investigate a nonpoint source pollution problem. AGNPS (Agricultural Nonpoint Source model) is an event-based model that simulates runoff and the transport of sediment and pollutants from mainly agricultural watersheds. Distributed parameter models such as AGNPS are often applied to large problem domains. Linking such models to GIS and database management system can facilitate better data storage, manipulation and analysis than conventional methods. In this study, rather than manually implementing AGNPS, extracted data are integrated in an automatic fashion through a direct linking between the AGNPS model engine and GIS. This direct linkage results in a powerful, up-to-date tool that would be capable of monitoring and instantaneously visualizing the transport of any pollutant that AGNPS can simulate. Thereby, it reduces the time required to analyze the numerical output from AGNPS, and enables users to identify critical areas of nonpoint source pollution (NPS) and furthermore, to perform various "what if" scenarios to support the decision making processes such as Best Management Practices (BMP) for the watershed. A case study was performed on a watershed of 98.1 km2 (=24240 acres) in the upper segment of the Sand Hill River Subbasin in Minnesota by using this linkage implementation. Simulated results showed that the optimal BMP scenario achieved an average reduction of about 26% from current nonpoint source pollutant levels on soluble and sediment-attached nitrogen and phosphorous. Especially, this optimal BMP scenario was most effective in controlling the erosion and sediment yield. As a result, a maximum 50% reduction of sediment yield was observed.

KEYWORDS: Water Quality, Watershed-scale, NPS, GIS, Linkage.

Introduction
Nonpoint source pollution (NPS) is a type of the pollution generated from diffused sources in the public and private domain. The Environmental Protection Agency (EPA) defines the nonpoint source pollution as pollution originating from urban runoff, construction, hydrologic modification, silviculture, mining, agriculture, irrigation return flows, solid waste disposal, atmospheric deposition, stream bank erosion, and individual sewage disposal (Corbitt, 1990). More than fifty percent of the pollution entering the nation’s waters comes from nonpoint sources (Tyler, 1992), and it is responsible for almost two-thirds of the pollution that prevents achievement of water quality standards ( Alm, 1990).

Nonpoint Source Pollution management is highly dependent on hydrologic simulation models. Evaluating alternative management strategies through experiments and a limited amount of field measurements is not feasible, and a modeling study is often the only viable means of providing input to management decisions. The hydrologic system was commonly simplified in the past as a "lumped model." Under this simplification, the spatial distribution of parameters lost their real meaning in hydrologic modeling. In contrast to this simplification, a distributed parameter model maintains the spatial distribution of the parameters. Therefore, the application of distributed parameter models is of practical necessity especially in case of the nonpoint source pollution management. The major disadvantage of distributed parameter models are the large amounts of time required for assembling and manipulating the input data sets. The distributed nonpoint source pollution models used to study pollutant transport and erosion easily generate towering amounts of data for analysis in even a small watershed. On a large non-homogeneous watershed, a complete simulation and analysis can be very time consuming.

A logical step in improving the quality of the hydrologic modeling would be the integration of spatially distributed parameter models with practical data management scheme such as the geographic information system (GIS) and database management system (DBMS). This integration of distributed hydrologic models-GIS-DBMS can be defined as a tool to collect, manage, analyze, simulate and display spatially varying information ( Parker, 1990). This paper describes direct AGNPS-GIS-DBMS linkage to: (1) effectively pinpoint critical areas where resources are threatened, (2) give the necessary information instantaneously based on BMP scenario simulations for further remediation and conservation efforts, (3) provide quality information to decision-makers cost effectively, and (4) help impartially distribute incentives and regulations used for water quality management.

Table of Contents Table of Contents

Study Watershed
The Sand Hill River Watershed District of the North Central Minnesota had proposed building a dam which will create a permanent pool of 5 km2 (=1,217 acres) near the northwestern Minnesota town of Winger for flood protection on the main stem of the Red River of the North basin. The feasibility of the dam project has been controversial since it was first proposed more than 10 years ago due to concerns regarding low annual average flow rate and water quality problems. No definite solutions or alternatives for alleviating these problems have been investigated ( Polk County Water Planning Task Force, 1990). This paper will investigate the application of a distributed parameter hydrologic model, AGNPS (Agricultural Nonpoint Source Pollution Model) and GIS-DBMS direct linkage to the upper segment area of 98.1 km2 (=24,240 acres) which has an outlet at the proposed dam site.

The upper segment has a complex topographic pattern which consists of moraines, potholes, small lakes, wetlands and marshes that affects its hydrologic characteristics substantially as shown in Figure 1.

Figure 1
 Figure 1. Topographic Patterns of the Study Watershed
(12.2Kbytes, 645x386 pixels)

For the analysis of such complex topographic features, a spatially distributed model such as AGNPS is more suitable than a conventional lumped model. Also nonpoint source pollution management on such complex topography is difficult to plan and highly dependent on simulation models because nonpoint source pollution is site specific. Evaluating alternative management plans through field experiments is usually not feasible, and thus simulation study is often the only practicable means of providing input to management decisions. In this sense, AGNPS is an ideal distributed parameter model for the study area to pinpoint "critical" areas of sediment and nutrient production in a watershed, and to estimate the potential benefits of implementing Best Management Practices (BMP) to alleviate current water quality problems.

Table of Contents Table of Contents

Geographic Information System (GIS)
The initial attempts to apply computer technology to the reduction of the substantial spatial data handling problems were associated with military purposes and produced useful results only after the massive application of computing resources. This was a function not only of the computer technology in the late '50s and early '60s, but also an early demonstration of the special problems encountered in digital spatial data handling (Dangermond, 1987; Parent and Church, 1987). Early digital mapping systems in the 1960's were mainly automated map drafting systems. Their application was to produce conventional maps with the use of a computer that became more interactive with the introduction of software and hardware systems. In the 1970's database management systems (DBMS) conjugated with the improvements in computers' software and hardware capabilities had a great influence on the management and analysis of digital geographic information. In the seventies the term GIS became well established, and its application and areas of research expanded beyond the military use and spread to planning and management of natural resources, land records, utilities, etc.

A GIS is an assembly of knowledgeable users and computer hardware and software designed to collect, manage, analyze and display spatially referenced data (Nystrom et al., 1986; Petersen et al., 1987; Eash, 1994), wildlife (Giles, 1990), and stormwater management (Holbert et al., 1990) among many others.

The State of Virginia has implemented a system, VirGIS (Flagg et al., 1990; Shanholtz et al., 1990) to develop spatially represented databases. Spatially referenced data, and the use of NPS (non-point source) models allow the identification, prioritization, and targeting of nonpoint source pollution problems. For example, using GIS-based data on topography, precipitation, land cover, soils, runoff, rainfall erosivity, sediment production, land use, livestock density and animal loading indices, an Agricultural Pollution Potential Index (APPI) was developed for each of 104 watersheds in the State of Pennsylvania. The development of these indices allowed prioritizing watersheds to deal with problems concerning nitrogen, phosphorous and loss of sediment (Hamlett et al., 1990).

Table of Contents Table of Contents

Agricultural Nonpoint Source Pollution ( AGNPS) Model

Nonpoint source pollution from agricultural and urban areas affects both surface and subsurface water quality, and accounts for more than one-half of the biological oxygen demand (BOD) and most of the suspended solids, phosphorus, nitrogen and toxic substances entering waterways.

The AGNPS model was developed by the U.S. Department of Agriculture, Agricultural Research Service in cooperation with the Minnesota Pollution Control Agency (MPCA) and the Soil Conservation Service (Young et al., 1987, 1995) for the analysis of large agricultural watersheds ranging in size between 500 and 23,000 acres. The model was originally developed (1) to analyze and provide estimates of runoff with primary emphasis upon sediment and nutrients transport, from agricultural watersheds, and (2) to compare the effects of various conservation alternatives upon implementation on the management practices of the watershed. Though AGNPS was developed for the State of Minnesota, several investigations in various-sized agricultural watersheds throughout many regions of United States have been reported (Bingner et al., 1989; Lee et al., 1990; Yoon and Disrud, 1994).

AGNPS is a cell-based distributed parameter hydrologic model which requires 22 categories of information such as landuse, surface condition, channel data, fertilization, etc. for each cell, and is capable of generating both estimates of quantity and quality of runoff from the watershed for a given storm event (Young et al., 1987, 1995). Cell segmentation of 2.5-, 10-, and 40-acre parcels in the study watershed is shown in Figure 2. Analysis of pollutant loads from feedlots, investigations into the effects of implementing various conservation practices including impoundment terraces, and the ability to output water quality characteristics at intermediate points throughout the watershed network are all within the model's capabilities. The erosion component in the AGNPS model assumes overland erosion, channel erosion and deposition in impoundments. The model uses the modified universal soil loss equation (USLE) by Wischmeier and Smith (1978). Field topography is the major factor in determining the type of flow, i.e. overland, overland-channel, overland-channel-impoundment, etc.

Figure 2
 Figure 2. Cell Segmentation of 2.5-, 10-, and 40-acre Parcels in the Study Watershed
(18.4Kbytes, 645x503 pixels)

The model uses distributed parameter inputs and operates on a cell basis. The distributed parameter approach of the model is most appropriate to preserve the spatial characteristics of the watershed and to obtain more accurate results. By itself, AGNPS is data-intensive model. In general, as the area or the resolution of the watershed increases, data requirements also increase proportionally. But with tools such as a geographic information system and a database management system, AGNPS has an advantage over other lumped models. Several distributed parameter hydrologic models have been developed in recent years based on this quantized grid-cell concept. Models that are structured in this way include Areal Nonpoint Source Watershed Environment Response Simulation model (ANSWERS) (Beasley and Huggins, 1982), the very detailed Systeme Hydrologique Europeen model (SHE) (Abbott et al., 1986), and the Water Erosion Prediction Project model (WEPP) (Foster et al., 1987).

Various output options are available with the AGNPS. Primary output includes estimates of runoff volume, peak flow rate at the watershed outlet, area-weighted erosion for both upland and channel areas. Also given are estimates of the sediment delivery ratio, sediment enrichment ratio, mean sediment concentration, and total sediment yield for each of five sediment particle size classes. Also available is a nutrient analysis, which includes N, P and chemical oxygen demand (COD) mass per unit area for both soluble and sediment absorbed nutrients, and N, P and COD concentrations in the runoff. Output parameters can be requested for a specific cell or for all cells if desired.

Table of Contents Table of Contents

Direct Linkage Between Model and GIS
There have been several attempts to integrate GIS with existing hydrologic models for the watershed modeling and simulation. In their attempt to integrate GIS with a finite element storm hydrograph model (FESHM), Wolfe and Neale (1988) had to manually process the data to meet the FESHM's data format requirement. Improvement in handling similar procedures was presented by Ross et al. (1990) through external data handling utilities for a hydrologic simulation model. Ozbilgin et al. (1991) used GIS to generate hydroconductivites from predefined cross-sections of the San Fernando Valley, California for a three-dimensional groundwater model (MODFLOW). Panuska et al. (1991), integrated two terrain-enhancer programs, TAPES-C and TAPES-G ( Moore, 1988), into AGNPS to estimate the flow path parameters.

These hydrologic modelings were performed with the extracted data from the GIS to execute available models. In most cases, either the data manipulations were done manually before running models, or all necessary data manipulation schemes such as transformation of data formats and extraction of usable data for existing models had to be provided by specially written software ( Fisher, 1989; Piwowar and LeDrew, 1990). These procedural redundancies have created bottleneck phenomena and even hindered the ideal integration of GIS and water resources tools. Also, most of the literature admit that many linking problems still exist when integrating models with a vast amount of spatial data in unorganized formats. Problems remain with the search for the optimal schema for spatial data handling, especially where adequate transformations of data within a GIS environment in order to feed the models directly are absolutely necessary. Methods suggested to improve the efficiency of the linkage between GIS and models usually include implementing various database management strategies or interfaces within the environment for these problems (Wood et al., 1988; Cline et al., 1989).

In summary, the main focus in the GIS usage with existing models has been one-directional data extraction to the model rather than direct data exchange to and from the model. It is imperative to develop data handling procedures to prepare inputs to the model, and to effect changes in watersheds, and procedures to analyze and display corresponding model results within the GIS environment. This bilateral linkage approach through effective use of DBMS will contribute to better hydrologic modeling and water resources management.

In this study, methods were developed for directly linking the distributed parameter model, AGNPS with the vector-based GIS, Geo/SQL ( Generation 5 Technology, 1990) and the relational database management system (RDBMS), ORACLE in terms of pre and postprocessing of data, in applying the GIS to the ongoing water resources planning and decision making process in the study area. This direct linkage results in a powerful, up-to-date tool that would be capable of monitoring and instantaneously visualizing the transport of any pollutant that AGNPS can simulate. Thus, the linkage is truly bilateral rather than the conventional one-way data extraction setup for input preparation. The advantage of linkage is that because the distributed parameter model such as AGNPS is often applied to large problem domains, a linkage to GIS and RDBMS may be a more appropriate tool for data storage, manipulation and analysis than a collection of many input and output files.

Visualization of the spatially varying data and time-dependent data such as runoff or sediment yield at the outlet will greatly enhance the ability of conservation managers to make further analyses and to make proper decisions. Furthermore, the graphical display may provide an indication of problems due to erosion and pollutant movement on a watershed and help pinpoint critical locations for further study and/or control action (Yoon and Padmanabhan, 1995). In a broad sense, graphical representation of a watershed under consideration and the simulation results from a pollutant movement and erosion prediction model would facilitate the researcher's and user's quick understanding of processes that produce the erosion and pollutant movement ( Srinivasan and Engel, 1991).

Although this study emphasizes the methodology to the upper segment of the Sandhill River subbasin in Minnesota, the methodology is applicable to the investigations and analysis of watersheds in general.

Table of Contents Table of Contents

Methodology
The basic data layers for the study area were developed by digitizing the relevant maps, including USGS-10 foot interval contour quadrangle maps of 1:24000 scale and SCS Soil maps. The spatial land cover/land use data were compiled from LANDSAT thematic maps as shown in Figure 3. Prepared data were inserted into ORACLE database as separate characteristic data groups so that the maximum flexibility can be assured when various scenarios are applied later. Then data from the database and digital base layers were georeferenced to create a spatial database of study area. Sets of programmed SQL (structured query language) of the ORACLE database management system were implemented to extract necessary information in a usable data format whenever a specific area is selected in GIS for AGNPS executions.

After each AGNPS execution, an output file containing estimates of runoff volume, peak flowrate at the watershed outlet, area-weighted erosion for both upland and channel areas, sediment delivery ratio, sediment enrichment ratio, mean sediment concentration, and total sediment yield for each of five sediment particle size classes (clay, silt, small and large aggregates, sand), and a nutrient analysis, which includes N, P and COD mass per unit area for both soluble and sediment absorbed nutrients, and N, P and COD concentrations in the runoff are created and loaded into temporary spatial database tables. This transition of original data with output results will maintain the valid georeferences between data tables in the DBMS and graphic entities in GIS. As a result, graphical representation in GIS facilitates a more effective and efficient way of interpreting AGNPS' distributed parameter model results and in making decisions than dealing with a plethora of numerical output. By using predefined SQL sets programmed for the study, a new input data for AGNPS can be constructed based on user definable scenario by specifying desired input parameters and degrees of change for each parameter based on previous results.

Figure 3
 Figure 3. Spatial Database Compilation
(13.3Kbytes, 558x432 pixels)

Table of Contents Table of Contents

Critical Area Assessment and BMP Scenario Simulation
By using AGNPS-GIS-RDBMS linking methodology, assessment of nonpoint sources from the study watershed was performed using a 25-year, 24-hour storm event. The SCS synthetic rainfall type II distribution was applied to simulate rainfall distribution. Critical areas of high volume of flow, erosion and sediment, and nutrient yields were assessed within the watershed.

Sensitivity analysis of AGNPS model parameters as shown in Table 1 was performed to construct "what if," or alternative management scenarios to improve current water quality problems by using Model-GIS-RDBMS linkage. Input parameter matrix for the sensitivity analysis is shown in Figure 4. Results from the sensitivity analysis show that seven AGNPS parameters out of twenty-two were relatively effective in controlling nonpoint source pollutants from this particular watershed. They were SCS curve number, Manning's roughness coefficient, cover and management factor (C-factor), land slope, channel sideslope, practice factor (P-factor) and fertilization availability factor in descending order of significance.

Table 1. Twelve Input Parameters used in the Sensitivity Analysis

Parameter Description Nature of Variable
RA Storm rainfall lumped
EI Storm Energy-Intensity value lumped
CS Channel slope spatially varying
CSS Channel sideslope spatially varying
MA Manning's roughness coefficient spatially varying
KF Soil erodibility factor spatially varying
CF Cover and management factor spatially varying
PF Practice factor spatially varying
FA Fertilization availability factor spatially varying
LS land slope spatially varying
SL Field slope length spatially varying
SCS SCS curve number spatially varying

Figure 4
 Figure 4. Input Parameter Matrix for the Sensitivity Analysis
(7 Kbytes, 367x341 pixels)

Based on the sensitivity analysis, parameters were prioritized to minimize nonpoint source pollutants. After seven simulations by using AGNPS-GIS-RDBMS linkage, a best management practice (BMP) scenario was formulated and then simulated to evaluate potential improvements on current water quality problems (Figure 5). Simulated results showed that the BMP scenario achieved an average reduction of about 26% from current nonpoint source pollutant levels. Especially, the final BMP scenario No. 7 was most effective in controlling the erosion and sediment yield and a maximum 50% reduction of sediment yield was observed. Various rainfall frequencies including 1-, 2-, 5-, 10-, 25- and 100-year return periods were also investigated to determine the effect of various return periods on the current watershed response.

Figure 5
 Figure 5. Bilateral Linkage of Model-GIS-RDBMS for "What-If" Scenarios
(10.1Kbytes, 588x321 pixels)

The direct AGNPS-GIS-RDBMS linkage developed in this study is a powerful tool that would be capable of monitoring the pollutant and sediment transports within the watershed. In addition, this system will be capable of effectively pinpointing areas of concern. The overall long range benefit will be easier information management for storing, retrieving, analyzing, updating and maintaining of various watershed related data, and more effective and flexible methods for water resources planning and management. Furthermore, the GIS would be advantageous when study areas are large, analysis and modeling are performed repeatedly, or if alternative "what if" landuse/landcover scenarios are explored. Further diversified modeling and simulation activities can be done based on the established database.

Table of Contents Table of Contents

Summary and Conclusions
Methods were developed for directly linking the distributed parameter model, AGNPS with a GIS and a relational database management system (RDBMS) to investigate a nonpoint source pollution problem. AGNPS is an event-based model that simulates runoff and the transport of sediment and pollutants from mainly agricultural watersheds. The model operates on a cell basis so that the spatial variation in parameters of each cell can be accounted for in the analysis throughout the entire watershed. Distributed parameter models such as AGNPS are often applied to large problem domains. Linking such models to GIS and relational database management system can facilitate better data storage, manipulation and analysis than conventional methods. In this study, a vector-based GIS and a relational database management system are implemented to create and manipulate the watershed database for AGNPS. Rather than manually implementing AGNPS, extracted data are integrated in an automatic fashion through a direct linking between the AGNPS model engine and GIS. This direct linkage results in a powerful, up-to-date tool that would be capable of monitoring and instantaneously visualizing the transport of any pollutant that AGNPS can simulate. Thus, the linkage is truly bilateral rather than the conventional one-way data extraction setup for input preparation. Thereby, it reduces the time required to analyze the numerical output from AGNPS, and enables users to assess critical areas of nonpoint source (NPS) pollution and furthermore, to perform various "what if" scenarios to develop best management practices (BMP) for the watershed. Rather than limiting AGNPS' capabilities just to the critical area assessment, this "what-if" scenario manipulation can be used to generate various AGNPS simulation input scenarios that contain a single parameter or multiple parameter variations of user specified AGNPS parameters. Provision is made to check each scenario parameters' logical upper and lower boundary conditions so that generated AGNPS scenarios inputs will be realistic.

A case study was performed on a watershed of 98.1 km2 (=24,240 acres) in the upper segment of the Sand Hill River Subbasin in Minnesota by using this linkage implementation. Simulated results showed that the optimal BMP scenario achieved an average reduction of about 26% from current nonpoint source pollutant levels on soluble and sediment-attached nitrogen and phosphorous (Figure 6). Especially, this optimal BMP scenario was most effective in controlling the erosion and sediment yield. As a result, a maximum 50% reduction of sediment yield was observed. For example, visual comparisons between before and after the optimized BMP scenario implementation are shown in Figures 7 and 8 for soluble nitrogen loading rate and nitrogen concentration at each cell level.

Although this study emphasizes the methodology to the upper segment of the Sand Hill River Subbasin in Minnesota, the methodology is applicable to the future investigations and analysis of watersheds in general.

Figure 6
 Figure 6. Magnitude of Reductions (%) made by the Final BMP Scenario to Base Values at Watershed Outlet
(15.9Kbytes, 588x384 pixels)

Figure 7
 Figure 7. Soluble Nitrogen Reduction Based one Optimal BMP Scenario
(12.1Kbytes, 613x402 pixels)

Figure 8
 Figure 8. Nitrogen Concentration Reduction Based one Optimal BMP Scenario
(11.7Kbytes, 631x415 pixels)

Table of Contents Table of Contents

References
Abbott, M. B., Bathurst, J. C., Cinge, J. A., O'Connell, P. E., and Hasmussen, J., 1986.
An Introduction to the European Hydrological System-Systeme Hydrologique Europeen, 'SHE,' 2: Structure of a Physically-based, Distributed Modeling System, Journal of Hydrology, 87:61-77.
Alm, A.L., 1990.
Nonpoint Sources of Pollution, Environmental Sciences and Technology, 24(7):967.
Beasley, D.B. and Huggins, L.F., 1982.
ANSWERS (Areal Nonpoint Source Watershed Environment Response Simulation): User's Manual, U.S. Environmental Protection Agency, Chicago, Illinois.
Bingner, R. L., Murphree, C. E., and Mutchler, C. K., 1989.
Comparison of sediment yield models on watersheds in Mississippi, Transactions of the ASAE, 32(2):529-534.
Cline, T. J., Molinas, A. and Julien, P. Y., 1989.
An Auto-CAD-Based Watershed Information System for the Hydrologic Model HEC-1, Water Resources Bulletin, AWRA, Bethesda, Maryland, 25(3):641-652.
Corbitt, R.A., 1990.
Standard Handbook of Environmental Engineering, McGraw-Hill, New York, New York, pp. 7.50-7.57.
Dangermond, J., 1987.
Introduction to Geographic Information Systems Technology, Proceedings in NCGA's Mapping and Geographic Information Systems, November 9-12, San Diego, CA, pp. 32- 40.
Eash, D.A., 1994.
A Geographic Information System Procedures to Quantify Drainage-Basin Characterisitcs, Water Resource Bulletin, 30(1):1-8.
Fisher, G. T., 1989.
Geographic Information System/Watershed Model Interface, Proceedings in the 1989 National Conference on Hydraulic Engineering, ASCE, Hydraulics Division, New Orleans, Louisiana, pp. 851-856.
Flagg, J. M., W. C. Hession and V. O. Shanholts, 1990.
GIS and Water Quality Models as State Level Nonpoint Source Pollution Control Management Tools, Proceedings in Application of Geographic Information Systems, Simulation Models, and Knowledge-Based Systems for Land Use Management, Blacksburg, VA.
Foster, G. R., Lane, L. J., Schertz, D. L., Nordin, J. O., and Wingate, G. D., 1987.
User Requirements: USDA-Water Erosion Prediction Project (WEPP), American Society of Agricultural Engineers, ASAE Paper No. 87-2539, St. Joseph, MI.
Generation 5 Technology, 1990.
Geo/SQL Geographic Information System, User's Guide, Westminster, CO.
Giles, R. H. Jr., 1990.
A Computer-aided Prescription System for Wildlife-Management and Related Areas, Proceedings in Application of Geographic Information Systems, Simulation Models, and Knowledge-Based Systems for Land Use Management, Blacksburg, VA, pp. 11-22.
Hamlett, J. M., Petersen, G. W., Russo, J., Miller, D. A., Baumer, G. M.and Day, R. L., 1990.
GIS-Based Watershed Rankings for Nonpoint Pollution in Pennsylvania, ASAE Paper No. 90-2619, ASAE, St. Joseph, MI.
Holbert, S. B., 1990.
Development of a Geographic Information System Based Hydrologic Model for Stormwater Management and Land Use Planning, Proceedings in Application of Geographic Information Systems, Simulation Models, and Knowledge- Based Systems for Land Use Management, Blacksburg, VA, pp. 197-210.
Lee, T. C. and Zhang, G. T., 1989.
Developments of Geographic Information Systems Technology, Journal of Survey of Engineering, 115(3):304-323.
Moore, I. D., 1988.
A Contour-based terrain analysis program for the environmental sciences (TAPES), Transaction of American Geophysical Union, 69(16):345.
Nystrom, D. A., Wright, B. E., Prisloe, M. P. Jr. and Batten, L. G., 1986.
U.S. Geological Survey, Connecticut Geographic Information Systems Project, Technical papers, ACSM-ASPRS Annual Convention, Washington, DC, pp. 210-219.
Ozbilgin, M. M., Chieh, S. H., Swanson, W. R., Shuter, K. A., Hoye, W. W. and Blevins, M. L., 1991.
Management of a Groundwater Basin using an Integrated Numerical Model and a Geographic Information System, Proceedings in the 18th Annual Conference and Symposium, ASCE, Water Resources Planning & Management Division, New Orleans, Louisiana, pp. 80-85.
Panuska, J. C., Moore, I. D. and Kramer, L. A., 1991.
Terrain Analysis: Integration into Agricultural nonpoint source (AGNPS) pollution model, Journal of Soil and Water Conservation, pp. 59-64.
Parent, P. and Church, R., 1987.
Evolution of Geographic Information Systems as Decision Making Tools, Proceedings in GIS '87, American Society for Photogrammetry and Remote Sensing (ASPRS) and American Congress on Surveying and Mapping (ACSM), Falls Church, Virginia, pp. 63-70.
Parker, H.D. 1990.
Introduction, The GIS Source Book, GIS World, Inc., Ft. Colins, Colorado, pp. 1-5.
Petersen, G. W., Miller, D. A., Day, R. L., Sasowsky, K. C. and Evans, B. M., 1987.
An Introduction to Geographic Information Systems and Their Role in Soil and Hydrologic Studies, Paper No. 8252, Journal Services of the Pennsylvania Agricultural Experiment Station.
Piwowar, J. M. and LeDrew, E. F., 1990.
Integrating Spatial Data. A User's Perspective, Photogrammetric Engineering and Remote Sensing, 56(11):1497-1502.
Polk County Comprehensive Local Water Planning Task Force, 1990.
Polk County Comprehensive Water Plan.
Ross, M. A., Fielland, C. F. and Tara, P. D., 1990.
An Integrated GIS/Hydrologic Model for Phosphate Mining Reclamation Design, Proceedings in the 26th Annual Conference of American Water Resources Association, Denver, Colorado, pp. 31-39.
Shanholtz, P. D., Flagg, J. M., Metz, C. D., Desal, C. J. and Kleen, J. W., 1990.
Information Support Systems Laboratory (and VirGIS)-Overview, Proceedings in Application of Geographic Information Systems, Simulation Models, and Knowledge-Based Systems for Land Use Management, Blacksburg, VA.
Srinivasan, R. and Engel, B. A., 1991.
GIS: A Tool for Visualization and Analyzation, American Society of Agricultural Engineers, ASAE Paper No. 91-7574, St. Joseph, MI.
Tyler, E.L. 1992.
Re-Authorization of the Clean Water Act, The Universities Council on Water Resources, Water Resources Update, Spring 1992, Issue No. 88, pp. 7-15.
Wischmeier, W. H. and Smith, D. D., 1978.
Predicting Rainfall Erosion Losses, Agricultural Handbook, no. 537, U.S. Department of Agriculture, Washington, D.C.
Wolfe, M. L. and Neale, C. M. U., 1988.
Input Data Development for a Distributed Parameter Hydrologic Model (FESHM), Proceedings in Modeling Agricultural, Forest and Rangeland Hydrology, International Symposium, ASAE, Chicago, Illinois, pp. 462-463.
Wood, E. F., Sivapalan, M., Beven, K. and Band, L., 1988.
Effects of Spatial Variability and Scale with Implications to Hydrologic Modeling, Journal of Hydrology, 102:29-47.
Yoon, J. and Disrud, L.A. 1994.
Fates of Nonpoint Source Pollutants based on Cascade Routing Model Formulation, Proceedings in National Symposium on Water Quality, 30th Annual AWRA Conference November, 6-10, 1994, Chicago, IL.
Yoon, J. and Padmanabhan, G. 1995.
Multiobjective Agricultural Nonpoint Source Pollution Management on Water Quality, Proceedings in Soil Conservation and Water Quality Symposium, the Minnesota Academy of Science (MNAS), April 27~28, 1995, Morris, MN.
Young, R. A., Onstad, C. A., Bosch, D. D., and Anderson, W. P., 1987.
AGNPS, Agricultural Nonpoint Source Pollution Model; A Large Watershed Analysis Tool, Conservation Research Report 35, Agricultural Research Service, U.S. Department of Agriculture, Washington, D.C.
Young, R. A., Onstad, C. A., Bosch, D. D., and Anderson, W. P., 1995.
AGNPS User's Guide, Version 5.00, 1995, Agricultural Research Service, U.S. Department of Agriculture, Morris, MN.

Table of Contents Table of Contents

-----

Author

Jaewan Yoon
Department of Civil and Environmental Engineering
KDH 135
Old Dominion University
Norfolk, VA 23529-0241
yoon@cee.odu.edu

-----

GIS Symposium | AWRA home page
Maintainer: AWRA Webserver Team
Last modified: 24Nov99 gaw
Copyright(c) 1996, American Water Resources Association