Background

While integrated management is widely supported, the spatial information on socioeconomic and biogeophysical processes needed for comprehensive evaluation of integrated resource management plans is not readily accessible to local decision makers. Advances in remote sensing, geographic information systems (GIS), multiple criteria decision making, risk management and biogeophysical simulation make it possible to develop user-friendly, interactive, decision support systems for integrated resource planning and management.

Objectives

The objectives of this study are: (1) develop a user-friendly, interactive watershed management decision support system (WAMADSS) that identifies the relative contribution of sub-watershed areas to agricultural non-point source pollution and evaluates the effects of alternative land use/management activities and practices (LUMAPs) on farm income, soil erosion and surface water quality at the watershed scale; and (2) demonstrate the utility of WAMADSS in identifying and evaluating LUMAPs for controlling soil erosion and surface and ground water pollution in Goodwater Creek watershed.

Methods

WAMADSS is a knowledge-based computer system which integrates data, information, physical simulation and economic analysis to identify alternative LUMAPs for solving specific watershed problems. It has three major components: a graphical user interface (GUI), a GIS and a modeling system.

1. Graphical User Interface (GUI):
   • Composed of visual subjects such as menus, buttons and input fields.
   • Provides access to the GIS and modeling system.
   • Provides interactive interface for developing model inputs and entering parameters.
   • Allows the user to select LUMAPs, parameters and evaluation criteria desired.
2. Geographic Information System (GIS)
   • ARC/INFO software is used.
   • Manages spatial and tabular database, generates model parameters and presents water quality /economic results of baseline and alternative LUMAPs.
   • ARC Macro Language (AML) generates the GUI.
3. Modeling System
   • Agricultural Non-Point Source Pollution Model (AGNPS)
   • Soil and Water Assessment Tool (SWAT)
   • Cost and Return Estimator (CARE)

Major Findings

WAMADSS makes complex and technical information and knowledge available to decision makers in a user-friendly graphical user interface. It allow users to organize information based on existing data and scientific knowledge, design alternatives and assess consequences of new watershed management plans or policies, and evaluate and compare alternative watershed schemes. The use of GIS minimizes the time involved to manually enter or manipulate the large amount of input data required to describe the spatial detail of a watershed. It also minimizes human error and inconsistencies in distinguishing landscape characteristics across a watershed that would otherwise be collected by conventional methods

A Watershed Management Tool Using SWAT and ARC/INFO

While watershed management is gaining wide support, the spatial information on socioeconomic and physical processes needed for evaluating alternative watershed management plans is not readily accessible to local decision makers. With wide adoption of geographic information system (GIS) technology, a user-friendly and interactive decision support system appears to be an efficient tool in watershed management. A graphical user interface is developed to incorporate the Soil and Water Assessment Tool (SWAT) with ARC/INFO. The menu interface provides a tool to identify the relative contribution of sub-watershed areas to agricultural nonpoint source pollution and evaluate the effects of alternative land use management practices on surface and ground water quality at the watershed scale. SWAT is a widely used environmental simulation tool based on continuous daily time-step process. ARC/INFO contains modules for maintaining and analyzing spatial and tabular data in an effective manner. The interface guides the user through a series of menus by: 1) leading the user through the steps in generating model inputs, executing the model and analyzing outputs, 2) allowing for changes in land use and management practices and re-evaluating potential consequences, and 3) viewing graphical and tabular results side-by-side for alternative scenarios. The tool is developed using ARC/INFO Arc Macro Language.

Introduction

Effective watershed management requires an understanding of basic hydrologic and biophysical processes in the watershed. A number of simulation models have been developed to evaluate water quality parameters affected by agricultural land management at both field and watershed scale. Widely used field scale models include CREAMS (Chemicals, Runoff, Erosion from Agricultural Management Systems), EPIC (Erosion-Productivity Impact Calculator), and GLEAMS (Groundwater Loading Effects of Agricultural Management System). Watershed scale models include storm event based AGNPS (Agricultural Non-Point Source Pollution Model) and continuous daily time step model SWRRB (Simulator for Water Resources in Rural Basins). Expansion of SWRRB model’s capacities to facilitate more subbasins and sophisticated routing structure resulted in a new watershed scale model SWAT (Soil and Water Assessment Tool).

SWAT offers distributed parameter and continuous time simulation, and flexible watershed configuration. The capacities of SWAT model have yielded growing applications of the model (Arnold and Allen 1994, Heidenreich et al. 1995). On the other hand, a full scale use of the model demands great amount of time, expertise and cost for acquiring input data, running the model and analyzing the results. In recent years, efforts have been made to integrate SWAT model with raster based GRASS (Geographic Resources Analysis Support System) GIS to facilitate development of model input and analysis of model output (Srinivasan and Arnold 1995, Rosenthal et al. 1995).

Objective

The objective of this project is to develop a user-friendly and menu-oriented graphic user interface (GUI) for streamlining modeling processes involving SWAT model and ARC/INFO GIS. Specifically, the tool is composed of three functional divisions: (1) generating model inputs, executing the model and analyzing outputs for baseline land use management, (2) allowing for changes in land use and management practices and re-evaluating potential water quality consequences, and (3) viewing graphical and tabular results side-by-side for baseline and alternative scenarios.

The reasons for using ARC/INFO are threefold: (1) it is a widely used GIS software that contains both raster and vector based modules for effective database management, (2) menu-oriented integration of SWAT and ARC/INFO was not in place, and (3) ARC/INFO is a GIS software being used for SWAT modeling and other watershed scale studies in-house(Heidenreich et al. 1995, Zhou 1996).

SWAT Model Description

The Soil and Water Assessment Tool (SWAT) is a process-based continuous daily time-step model which evaluates land management decisions in large ungauged rural watersheds. It is designed to predict long-term nonpoint source pollution impacts on water quality such as sediment, nutrient and pesticide loads (Arnold et al., 1994).

Model inputs include management inputs such as crop rotations, tillage operations, planting and harvest dates, irrigation, fertilizer use, and pesticide application rates, as well as the physical characteristics of the watershed and its subbasins such as precipitation, temperature, soil type, land slope and slope length, width and slope, Manning’s n values and USLE K factors. Either simulated or measured precipitation and temperature values may be used in SWAT. Measured streamflow and sediment concentrations can be statistically compared with model predictions.

Model processes include calculations of water balance (i.e. surface runoff, return flow, percolation, evapotranspiration, and transmission losses), crop growth, nutrient cycling, and pesticide movement. Model outputs include subbasin and watershed values for surface flow, ground water and lateral flow, crop yields, and sediment, nutrient and pesticide yields.

The SWAT-ARC/INFO Integration

Figure 1 depicts the composition and diagram of three functional divisions in the SWAT-ARC/INFO tool. Each division consists of a series of graphic menu interfaces to accomplish the tasks. Major steps are described below and illustrated through an example application of Goodwater Creek watershed located in north central Missouri

.

Figure 1. Composition and diagram of three functional divisions

1. Model Baseline Establishment

Four basic types of watershed data are required to extract spatial input for SWAT model. These data are hydrologic features, soil distribution, land use information, and contour lines depicting topography.

Major procedures in this division include: (1) develop subbasin database, (2) develop stream database, (3) calculate landuse and soil variations within subbasins, (4) enter physical variable parameters such as land use management, (5) compile model readable input files, and (6) run model and compile outputs.

(1) Develop subbasin database

Figure 2 shows the menu for subbasin database development. The menu contains functions for displaying basic GIS layers, two options for subbasin delineation, and buttons for proceeding directly to next steps. Delineation of subbasins is performed using ARC/INFO GRID module. It generates a hydrologically correct digital elevation model (DEM) from contour line coverage, and determines subbasin outlets and drainage contributing areas to the outlets. The resulting natural subbasins are then grouped into certain number of subbasins based on the criteria specified by the user. In this instance, 5 subbasins are created.

(click on the image to view full version)

Figure 2. Subbasin delineation

Subbasins may be manually edited by the user to better reflect landuse, soil, hydrology, or hypsography in the watershed. Figure 3 shows the menu for editing boundaries and identification numbers of subbasins. Completion of this step proceeds the user to interactively identify outlet subbasin one-by-one. Manual identification of the outlet subbasins ensures correct routing pattern, specially for subbasins that contain no stream segments or are not hydrologically defined. Land slopes of subbasins are automatically calculated by averaging percent-rise slope values of the grid cells within each subbasin.

(click on the image to view full version)

Figure 3. Subbasin editing

(2) Develop stream database

Stream data parameters are required for two types of channels in the model: the main stream channel flowing from each subbasin inlet to the subbasin outlet, and the longest stream channel extending from each subbasin outlet to the most distant point in the subbasin. Functionality for the channel-type identification is shown in Figure 4.

(click on the image to view full version)

Figure 4. Stream channel identification

Once the two types of stream channels are identified, lengths and slopes of the channel segments are calculated automatically for subbasins. The slope of a channel segment is calculated by dividing the channel length with the elevation difference between the "starting" and "ending" nodes. Elevation of two end-nodes of a stream segment is determined with TIN (triangulated irregular network) module.

(3) Calculate Landuse and Soil Variations within Subbasins

A natural subbasin is usually composed of several land uses (or crops) and soil types. In SWAT modeling, a subbasin is required of representing a unique landuse (or crop rotation) and soil type. A straight forward approach is to use predominant landuse (or crop rotation) and soil. SWAT model also allows for non- spatial subdivision of subbasins into smaller sub-units based on landuse and soil variations, a concept of virtual subbasins. Virtual subbasins represent percentages of the larger subbasin area.

Methods of using either predominant or all land uses and soils are facilitated through the SWAT-ARC/INFO tool (Figure 5). Combination of all land uses and soils can result in enormous number of virtual subbasins for a watershed. To reduce un- necessary computation, the user is provided options to select certain land uses for breaking down by soils and/or specify a minimum percentage to exclude minor land uses in subbasins.

(click on the image to view full version)

Figure 5. Calculating land use and soils variation

In this example application, all land uses occupying 10% or more of subbasins are accounted for. Areas of the minor land uses (<10% of a subbasin) are re-allocated to major land uses to reflect 100% subbasin areas. Predominant soil type of a big subbasin is used for all virtual subbasins within the subbasin. Thus, a total of 15 virtual subbasins are generated for the watershed in this example.

(4) Enter Physical Variable Parameters

The user needs to collect and enter several physical variables such as soil group number, curve number, hydrologic variables, and pesticide selection. A series of graphic input menus guide the user to go through the process of parameter entry painlessly. Figure 6 shows the menu for entering hydrologic parameters. Hydrologic parameters may vary from subbasin to subbasin.

(click on the image to view full version)

Figure 6. Entering hydrologic parameters

Management input files include data for management practices such as planting, harvest, tillage operations, and pesticide and fertilizer application. Inputs include dates, operation type ode and application amounts. These input parameters are compiled through management input menu (see A in Figure 6). A management file can have any number of year of crop rotation. For each management file, data can be entered through 9 individual operation practices including planting, irrigation, fertilizer, pesticide, , harvest and kill, tillage, harvest only, kill only, and grazing respectively (see B in Figure 6). An spreadsheet input menu (see C in Figure 6) is also facilitated in the actual management file format. It is intended for SWAT- experienced users to enter all data quickly.

(click on the image to view full version)

Figure 7. Management input file development

Upon the completion of physical parameter input, the SWAT-ARC/INFO tool compiles the databases into SWAT readable input files and executes the model. A 5 year simulation is run for this example application.

SWAT simulates on a wide range of water quality parameters. This tool compiles the following parameters for watershed and subbasin level output:

average annual crop yields in bushels

average monthly pesticide concentration in runoff in PPB (parts per billion)

average annual nutrient concentration in surface and ground water including nitrate and phosphorus in PPM (parts per million)

average annual and monthly sedimentation in tons/acre

2. Evaluation of Alternative Land Use and Management Practices (LUMP)

Three alternative scenarios can be evaluated in addition to baseline land use management impact on water qualities. Each new scenario is based on an existing setting such as baseline or a previously created scenario. LUMP evaluation consists of assessing consequences of changes in landuse activities and/or management practices such as fertilizer, pesticide, and tillage operations. Figure 8 shows the menu for landuse activity change in the watershed. The user can change land uses for certain crop fields, or switch crops from one to another across watershed. In this example, all corn fields are changed to forest to generate Scenario #1.

(click on the image to view full version)

Figure 8. Landuse activity change

Alteration of management practices are facilitated through the management input file menu (Figure 6). When an exiting management file is selected, the menu input fields are populated with base scenario data. The user can edit and add data as desired.

3. Input and Output Display

The SWAT-ARC/INFO tool facilitates side-by-side display of two scenarios in both graphic and tabular format. Spatial input data accessible through the tool include: landuse, soils, hydrography, DEM, average pesticide application for each month, tillage operation, nitrate and phosphorous applications. SWAT input files are displayed in text format. All output components compiled by the tool are presented in graphic format at subbasin level and tabular format at watershed level. Figure 9 shows subbasin level sedimentation in for baseline and scenario #1.

(click on the image to view full version)

Figure 9. Output display of sedimentation for baseline and scenario #1

Conclusions and Discussions

The SWAT-ARC/INFO tool seamlessly links the SWAT model with ARC/INFO through a series of user-friendly graphic menus. The tool streamlines the whole process of model input development and output analysis, and thereby greatly speeds up a complex modeling process. In the meantime, the tool offers various flexible options for user-defined applications. This tool has been used for several SWAT related watershed analyses in-house.

Like any other software, the SWAT-ARC/INFO tool has its limitations. The tool does not check meaningfulness of the values entered by the user. The user is responsible to ensure that any physical parameters entered are correct and meaningful.

Both SWAT and ARC/INFO are very complex products, and demand highly professional knowledge to ensure reliable output. While such sophisticated system is needed for certain watershed studies, a watershed management tool using simpler models and PC-based GIS software such as ArcView is in greater demand.

Economic and Environmental Impact Assessment Using a Watershed Management Decision Support Tool

A Watershed Management Decision Support System (WAMADSS) is used to evaluate the economic and environmental consequences of alternative land use/management practices (LUMPs). WAMADSS consists of three components: (1) a GIS, (2) an economic model, and (3) two environmental simulation models. Specifically, ARC/INFO serves as the engine for integrating the Cost and Return Estimator (CARE) program, the AGricultural NonPoint source (AGNPS) pollution model and the Soil and Water Assessment Tool (SWAT) in a seamless interactive decision support system. The three components are accessed through a graphical user interface which enables decision makers to generate scenarios. Specifically, the user changes LUMPs and WAMADSS generates model input files, executes the programs and imports the output into the GIS. Tabular and spatial results are then viewed side-by-side for different scenarios.

Introduction

There is a growing consensus that an effective way to control nonpoint source pollution and enhance the long-term sustainability of agriculture and rural communities is through locally-based planning and management at the watershed scale. Watersheds are commonly defined as a geographic area in which water, sediments, and dissolved materials drain to a common outlet or as a series of ecosystems linked spatially and temporally by the downward flow of water.

Total watershed management provides a framework for integrating knowledge and perspectives of the social and natural sciences into planning, policy and decision making. Such an interdisciplinary framework is required to simultaneously address the socioeconomic and environmental impacts of natural resource policy and management. While knowledge about the interactions among socioeconomic and physical processes in a watershed is essential for improving sustainability of agricultural production, the mere generation of such knowledge is insufficient. The knowledge must be delivered to potential users or stakeholders in a way that maximizes its usefulness in watershed planning and management.

Specifically, long-term agricultural sustainability can be achieved by: (a) increasing knowledge regarding the spatial and temporal interactions between economic and environmental processes and how these interactions are altered by changing land use and/or management practices (LUMPs) at the watershed scale and (b) developing decision support systems (DSS) which make this knowledge accessible to and usable by individuals and groups involved in watershed planning and management. This research contributes to both goals by developing and implementing a Watershed Management Decision Support System (WAMADSS).

The objectives of this paper are to: (1) Provide the context for establishing an interdisciplinary approach to problem solving in natural resources; and (2) discuss the functionality of each component of WAMADSS and how it contributes to the decision support tool. The paper is limited to discussing the methodology for integrating models with ARC/INFO and does not, therefore, examine 'what if' scenarios generated by WAMADSS or how their results compare to empirical data.

This study is best illustrated by using WAMADSS in an interactive computer session. However, given the medium required to present this research, the authors are limited to presenting a dynamic interactive process in a static format. Refer to "A Watershed Management Decision Support System (WAMADSS): Economic and Environmental Impacts of Land Use Activities for Reducing Nonpoint Source Pollution" which minimizes this limitation by providing the reader with a graphical presentation of a WAMADSS session and its functionality (Fulcher, 1996).

Relevance of the Study

Economics may be necessary to understanding the use and abuse of natural resources, but it is not sufficient because natural resource issues are inherently interdisciplinary in nature. Information and knowledge from engineering, chemistry, physics, sociology, and biology, to name a few disciplines, are needed to effectively address resource-based issues. According to Kneese (1989), the study of resource economics has "required and motivated researchers to reach out beyond their own disciplines and to integrate ideas from other fields".

Gottfried (1992) asserts that few economists have addressed the interrelated nature of ecological goods and services, that is, the relationship among spatial units. Those that have addressed this issue have not been able to rigorously apply their methodologies as the tools were not available to communicate and model these relationships effectively. The natural resource economist must take an interdisciplinary approach to dealing with resource management issues by incorporating tools not traditionally used in the profession. Emerging computer technologies including GIS, remote sensing, expert systems, distributed-parameter and mathematical modeling, are the tools needed to effectively analyze different components of a landscape and their interactions.

An ecological economic framework for evaluating environmental and economic impacts of changing LUMPs is adopted in this research. Ecological Economics is an emerging branch of applied economics that deals with studying ecosystems as integral components of the landscape. According to Costanza (1993), the integration of ecology and economics is necessary as conceptual and professional isolation have led to environmental and economic policies that are mutually destructive.

In the United States, many agencies are adopting 'ecosystem management' watershed approaches to carry out their mandates (Born and Sonzogni 1995). For example, the U.S. Environmental Protection Agency has adopted the Watershed Approach Framework (Brady 1995, USEPA 1996). The Natural Resources Conservation Service has embraced the Watershed Approach and Ecosystem-Based Assistance (EBA). 'EBA' was chosen by NRCS rather than 'ecosystem management' to stress that NRCS assists stakeholders in developing plans that address LUMPs, their interactions and the interactions between LUMPs and humans. The U.S. Forest Service and the U.S. Bureau of Land Management have adopted Management by Watersheds (Maxwell 1993) and Ecosystem Management (Wise 1993). These holistic-oriented initiatives and mandates fall under the auspices of Integrated Environmental Management which is a new paradigm being adopted by numerous nations to address resource management (Born and Sonzogni 1995).

The common elements running through all the above initiatives are the need to: (1) identify the problem, (2) involve stakeholders and (3) implement integrated actions. These actions include integrating economic, social, ecological, and cultural characteristics into the watershed assessment.

The drive to understand the interrelationships among landscape elements in a holistic, ecological framework is taking federal agencies by a storm. Therefore, "the ecosystems approach is guided as much by moral conviction as by empirical evidence" (Armitage 1995). Although broadsweeping statements regarding these interdisciplinary approaches are pointing managers, practitioners and researchers in the right direction, there is little indication of how to do research concerning environmental issues viewed as a whole.

An interactive decision support system is necessary to effectively educate and inform stakeholders of current and potential watershed conditions. WAMADSS enables stakeholders to generate scenarios, change LUMPs by geographically selecting those areas in the watershed to be modified, and evaluate, in graphical and tabular format, how these changes influence economic returns and water quality. These actions are conducted in a facilitated decision making process

Decision Support Systems for Watershed Management

A decision support system is an integrated approach for helping people make better decisions. DSSs are typically computer programs that are used by individuals or a group in a facilitated decision making process. These tools are becoming more prevalent with the advent of GIS, remote sensing and sophisticated computer technologies because they improve the objectivity of the decision making process, particularly where complex interactions of LUMPs on the landscape are involved.

The integration of hydrology and water quality (H/WQ) models with GIS is well documented in the literature particularly with respect to land use planning, water quality protection and natural resource management. Integrating models with a GIS overcomes the limitations of stand-alone or disjointed approaches to decision making. Also, increasingly powerful computer hardware coupled with lower processing costs and more sophisticated GIS software permit analyses of complex environmental issues in a more timely fashion.

Tim and Jolly (1994) classified the integration of H/WQ models into three levels: (1) ad hoc integration, (2) partial integration, and (3) modeling within GIS. This research extends this framework to include the integration of economic models as well. Ad hoc integration involves developing the model and GIS separately. Specifically, the GIS is only used to extract input data required by the model. Partial integration involves using the GIS to supply the models with input data and view their results. Partial integration, as defined by Tim and Jolly, is modified in this research to reflect whether it is non-interactive or interactive. Specifically, a non-interactive decision support system (NIDSS) involves the manual or semi-automatic extraction of input parameters from the GIS that are required by the models. Similarly, a manual or semi-automatic procedure is used to import the model output into the GIS for display purposes.

A manual process requires a GIS/modeling expert to enter the appropriate actions on the computer command-line prompt in order to pass input and output between the GIS and model while a semi-automated process employs graphical user interfaces or menus to streamline the manual process. Even with a semi-automated process the model and GIS remain disjointed. While a non-interactive level of integration generates the modeling results, it is not suitable for watershed planning and management because of the excessive time and expertise it takes to manually or semi-automatically manipulate the data and generate scenarios.

An Interactive decision support system (IDSS), on the other hand, enables a decision maker to generate scenarios and view results based on changing LUMPs in a timely fashion. The technical expertise required to manually or semi-automatically manipulate the data is now imbedded into in the DSS. This approach is the most difficult in terms of developmental efforts because it requires knowledge of GIS, the models being integrated and the GIS programming language needed to seamlessly link the models together. WAMADSS is an IDSS.

WAMADSS Development

In this study WAMADSS is developed for evaluating environmental and economic impacts of changing LUMPs. In order to address these impacts, economic and environmental simulation models are required to simulate changes. The Cost and Return Estimator (CARE) program is a crop budget generator and the AGricultural NonPoint source (AGNPS) pollution model and the Soil and Water Assessment Tool (SWAT) simulate sediment, runoff, and nutrient and pesticide transport from agricultural watersheds. These models are integrated into a seamless DSS that is driven by a geographic information system (GIS).

CARE, AGNPS and SWAT require information about tillage practices, application rates of fertilizers or pesticides, the number of acres in a given activity and other factors. However, as stand-alone programs, they are limited due to their narrow functionality. Limitations include: inability to modify input parameters in a timely fashion, an inferior capability to view results, intensive data input requirements and PC memory constraints for executing the programs.

WAMADSS is used to overcome these limitations by interfacing the economic and environmental simulation components with the GIS component on the same computing platform. These components are accessed through a graphical user interface (GUI) or series of menus in order to facilitate analyses. Specifically, the economic and environmental models are linked to ARC/INFO GIS via its programming language - ARC Macro Language (AML). The interface enables the decision maker to manipulate LUMPs, execute the models, and view results within the GIS. Specifically, AMLs, menus and C programs are used to (1) generate CARE and AGNPS (or SWAT) input parameters in ARC/INFO; (2) generate and export the input files to the models; (3) execute the models; (4) import the output files into the GIS; and (5) view model results in graphical and tabular format in ARC/INFO.

As stated previously, WAMADSS consists of three components: a GIS, an economic model, and two environmental simulation models. A discussion of each component's contribution to WAMADSS is in order.

Geographic Information Systems

ARC/INFO is used as a method to monitor and significantly improve the decision maker's ability to manipulate spatial and non-spatial data in order to evaluate alternative management practices. This approach enhances the "best judgment" decisions offered by conventional simulation models. Specifically, ARC/INFO serves as the glue for WAMADSS; it provides a programming language that links the components together and furnishes a menu interface for the end user.

Arc Macro Language (AML): AML serves as the foundation on which WAMADSS is built. This programming language is used to program and tailor applications such as WAMADSS by automating frequently performed operations, accessing external programs such as C programs and developing graphical or menu-driven user interfaces for end users. AML programs handle all modeling-related activities, from interfacing CARE, AGNPS and SWAT to the GIS, generating model input files, executing the models, to viewing results in Arcplot.

Menu Interface Development: AML not only provides programming tools as stated above, but also provides tools for generating graphical user interfaces (GUI). Menus are AML programs that provide structure to complex processes or operations by organizing actions through widgets or menu objects. The end user interacts with the GUI by moving a mouse cursor over the desired widget and clicking the mouse button to invoke the action. Menus offer choices to the end user. These choices include running an AML, invoking another menu, executing an external program, or entering data in menu text boxes.

Cost and Returns Estimator Model (CARE)

The CARE model is a crop budget generator that was developed by USDA Natural Resource Conservation Service (NRCS) for estimating the costs and returns for crop enterprises. CARE output includes a budget listing gross receipts from production, production activities, inputs and outputs to the land, costs, gross returns, and net returns.

Twelve crop enterprise budgets are created for WAMADSS representing typical cropping systems in Goodwater Creek watershed. Specifically, net returns for four crops - corn, soybeans, sorghum and wheat - and three tillage practices - conventional, conservation and no till - are evaluated. These tillage practices are based on the amount of soil residue remaining on the field surface at planting time. The goal of reducing tillage is to leave more residue to reduce erosion.

Since WAMADSS serves as a template for analyses in other watersheds, the four by three matrix may be expanded or reduced depending on the cropping systems present. New budgets can either be generated within WAMADSS or created on a PC and imported into WAMADSS. All CARE menus in WAMADSS are designed to appear similar to the original menus in order to minimize the learning curve for end users familiar with CARE. The WAMADSS menus are an enhancement over the original CARE menus because input fields that are updated in one menu are automatically updated in other menus.

Once the economic baseline is established for the watershed, the end user can generate scenarios to evaluate the economic impacts of changing LUMPs. Another advantage of using the GIS is its ability to discern variability in crop yields across the watershed. Holding all other management factors constant, the yield potential for crops depends on soil type. By knowing which crop is grown on which soil type, an average yield potential can be determined for that crop at the watershed level. In other words, the spatial distribution of a crop over the watershed will dictate the yield potential for that crop. WAMADSS is able to automatically calculate the yield potential for all crops and pass those values to the CARE data bases.

Agricultural Nonpoint Source pollution model (AGNPS)

AGNPS is a distributed parameter model that can be used to simulate nonpoint source pollution loads in surface runoff from agricultural watersheds (Young and Onstad 1987, 1990). Specifically, soil erosion rates, sediment yield and nutrient (nitrogen and phosphorus) and pesticide transport are simulated for all cells in the watershed, including its outlet. The storm-event-based model operates on a uniform grid cell basis that subdivide the watershed based on its physical properties (i.e., soils type, land use, hypsography). The distributed parameter model lends itself well to a GIS because it preserves the spatial variability of LUMPs in the watershed.

A one-acre cell size was determined to be most appropriate from a land use/management perspective. The AGNPS grid layer is comprised of 19,113 cells or acres. Since AGNPS is a grid-based model, the grid cell layer serves as a cookie cutter, pressing the other layers into discrete grids thereby creating rasterized products (i.e., gridded land use, soils, hydrology, and hypsography).

AGNPS requires three types of input parameters: (a) initial watershed-level parameters, (b) cell parameters, and (c) optional indicator parameters based on which indicators are flagged for inclusion in that cell. Traditionally, the input parameters are entered into AGNPS via its spreadsheet utility For each cell, at least 22 parameters and potentially over 100 optional parameters are manually entered to create an AGNPS input file. For a 19,113 cell watershed this would involve manually entering at least 420,486 records into the spreadsheet file. To change LUMPs, the user would have to update each affected parameter in each record or cell being changed. The potential for human error is therefore quite high. The need for a GIS-based DSS to automatically populate and manipulate this input file is readily apparent in light of the manual effort required.

Soil and Water Assessment Tool (SWAT)

The Soil and Water Assessment Tool (SWAT) is a process-based continuous daily time-step model which evaluates land management decisions in large ungauged rural watersheds. It is designed to predict long-term nonpoint source pollution impacts on water quality such as sediment, nutrient and pesticide loads (Arnold et al., 1994).

Model inputs include management inputs such as crop rotations, tillage operations, planting and harvest dates, irrigation, fertilizer use, and pesticide application rates, as well as the physical characteristics of the watershed and its subbasins such as precipitation, temperature, soil type, land slope and slope length, width and slope, Manning's n values and USLE K factors. Either simulated or measured precipitation and temperature values may be used in SWAT. Measured streamflow and sediment concentrations can be statistically compared with model predictions.

Model processes include calculations of water balance (i.e. surface runoff, return flow, percolation, evapotranspiration, and transmission losses), crop growth, nutrient cycling, and pesticide movement. Model outputs include subbasin and watershed values for surface flow, ground water and lateral flow, crop yields, and sediment, nutrient and pesticide yields.

The purpose of interfacing AGNPS and SWAT with WAMADSS is to take advantage of the spatial operations offered by the GIS. AGNPS and SWAT are well suited for interfacing with a GIS because its parameters are landscape-based. Depending on the nature of the analysis, an end user may select one simulation model over the other. Once the physical baseline is established for the watershed, the end user can generate scenarios to evaluate the environmental impacts of changing LUMPs.

For example, if the end user decides to place a 100 foot riparian buffer strip on crop production along a stream, a number of input parameters need to be changed. Specifically, input parameters such as the curve number, surface condition constant, C factor and chemical oxygen demand need to be updated as the land use is changed from crop production to riparian buffer strip. WAMADSS is able to automatically calculate most input parameters, thereby substantially reducing manual data entry.

Three disadvantages of using AGNPS or SWAT as a stand-alone models are: (1) the time involved to manually enter or manipulate the large amount of input data required to describe the spatial detail of a watershed; (2) human error in interpreting landscape characteristics such as land slope, slope shape factor, field slope length, or channel slope from topographic maps; and (3) inconsistency in discerning these characteristics across the entire watershed. WAMADSS is used to minimize these disadvantages by: (1) significantly reducing the time and labor needed to process and manipulate the required input parameters; and (2) generating a surface model from the watershed hypsography (DEM) layer to address the second and third disadvantages stated above.

Data Layer Development

The study area used to implement WAMADSS is the Goodwater Creek watershed, a highly productive claypan area located in Audrain and Boone counties in the north central part of Missouri (Figure 1). This watershed, which is 77.43 square kilometers, is selected for two reasons: (1) atrazine and alachlor concentrations in the surface runoff are 10 to 100 times higher than the current drinking water standards during the late spring and early summer period following chemical application; and (2) extensive water quality monitoring has been conducted in Goodwater Creek watershed which serve to validate results obtained from environmental simulation models.

Figure 1. WAMADSS main menu  (click on the image to view full version)

In order to analyze changes in LUMPs at the watershed scale, a LUMP baseline is established for the watershed. This baseline serves as the foundation for generating scenarios in WAMADSS. The DSS uses five layers to generate the required input parameters for CARE, AGNPS and SWAT - the watershed boundary, soils, land use, hydrography, and hypsography . Additional layers such as road networks, railroads, and a digital elevation model (DEM) are used to visualize the watershed.

An Interactive WAMADSS Session

Decision making at the watershed level begins with a holistic conceptualization of issues affecting the watershed, and an understanding of interconnections between issues. The decision making process is then broken into three phases. In the first phase, a management scenario is formulated by the decision maker. Second, the user is prompted through a series of menus to select the appropriate LUMP options (menu buttons or fill-in text boxes) based on the management scenario. In the third phase, the decision maker evaluates the impacts of the scenario by viewing and comparing economic and environmental results in tabular and graphical format. After viewing the results, a new management scenario can be created based on the current scenario or the baseline; and the cycle repeated.

Scenario Formulation

WAMADDS is targeted toward a specific type of decision maker or end user: a natural resource-based agency such as the Natural Resource Conservation Service or an institution involved in natural resource research. This DSS is not applicable to individual farmers and is too sophisticated to be used independently by citizen groups. However, stakeholders in the watershed are central to decision making and implementation. Therefore, the resource-based agencies should serve as facilitators for using WAMADSS with these stakeholders. Specifically, WAMADSS is well suited for facilitated sessions with citizen-based watershed alliances seeking a better understanding of their watershed.

A hypothetical management scenario: Assume farmers have an incentive to grow more corn next year based on high price expectations. This scenario assumes land currently planted in soybeans and sorghum will be planted in corn over time. Management practices, such as type of tillage and fertilizer and pesticide application rates are part of this change.

Scenario Implementation

How do the LUMP changes described in the hypothetical scenario above affect water quality and costs and returns in the watershed? A WAMADSS session is used to evaluate these impacts. Due to image space constraints not all WAMADSS menus are illustrated. Therefore, the following sections describe some of the major menus an end user may use when generating a scenario.

Begin the Session

The watershed management decision support system is activated by typing 'WAMADSS' at the command-line prompt in the subdirectory that contains the five required GIS layers and associated INFO relate files. The first menu that appears is the WAMADSS title menu (Figure 2). This menu gives the authors involved in creating WAMADSS, where it was created and the WAMADSS version number. The 'Acknowledgments' button generates a menu acknowledging those individuals and agencies that assisted in the development of WAMADSS. Select the 'Next Menu' button.

Figure 2. WAMADSS title menu  (click on the image to view full version)

The WAMADSS main menu prompts the decision maker to select from several options (Figure 1, above). Two environmental models are depicted in this menu: AGNPS and SWAT. WAMADSS is used as a template for watersheds in the Midwestern U.S. Therefore, if the user has not used WAMADSS in a particular watershed then the 'CARE and AGNPS Initialization' or 'SWAT and CARE Initialization' button is used to establish the baseline for the watershed. AMLs are used to generate the appropriate grid cell size (or subbasin) and input parameters required to establish a physical baseline. Select the 'CARE and AGNPS' button to proceed to the next menu.

The next menu prompts the user to select from a number of options (Figure 3). This menu is divided into four sections: viewing the watershed baseline layers; viewing or editing the CARE parameters for a given scenario; changing LUMPs; or viewing input and output information for the baseline and a given scenario if already generated. Select the 'landuse' checkbox to produce the land use layer for Goodwater Creek watershed. Use the utility bar at the bottom of the screen to zoom in on a particular area. Before proceeding to generate a scenario from the baseline it may be useful to look at the input and output information. Select the 'Input/Output' button to view the baseline or existing scenario information.

Figure 3. View watershed information and select module   (click on the image to view full version)

View Baseline Input and Output Information

The input/output screen is a powerful tool for viewing watershed characteristics and its corresponding output (Figure 4). The main feature of this screen is its capability to view and compare side-by-side watershed characteristics and economic and environmental results for two scenarios (or the baseline and a given scenario). Two identical menus, one on the bottom-left and the other on the bottom-right side of the screen, are used to make these comparisons. The user can compare up to six scenario combinations using these menus (i.e., compare the baseline to scenario 1, 2 or 3; compare scenario 1 to scenario 2 or 3; or compare scenario 2 to scenario 3). All queries made from the bottom-right menu are visualized on the right side of the screen while queries made from the other menu are visualized on the left side of the screen. Due to the limited number of images permitted, no side-by-side comparisons will be illustrated in this paper.

Figure 4. Input/Output screen  (click on the image to view full version)

The utility menu in-between the scenario menus enables the user to zoom in, zoom out, pan or select an extent for viewing. Select the 'Baseline' button on the bottom-left side of the screen to view input and output parameters for the baseline. These parameters are unique to the baseline and each scenario generated. Also, the land use layer and the associated management practices (type of tillage and fertilizer and pesticide use) are unique to each scenario. If scenario 1 was selected for viewing, then WAMADSS would reference all input and output parameters for that scenario and produce the layers that reflect LUMPs for that scenario.

After selecting the 'Baseline' button a menu pops up that prompts the user to select from four sections: viewing graphical AGNPS output; viewing spatial change between scenarios; viewing watershed characteristics; and viewing tabular CARE and AGNPS results.

The spatial change section is used to visualize where land use, fertilizer application rates, pesticide application rates, tillage practices, etc. are changed in the watershed. For example, when the user clicks on a spatial change button for land use, an image of the watershed pops ups highlighting in red only those fields being changed. The watershed in Figure 4 depicts the baseline land use. Therefore, viewing change is not applicable. Once the proposed scenario is developed in this session the LUMP changes can be visualized.

The 'Watershed Characteristics' menu is used to examine input parameters and generate statistics for soils and land use. Specifically, the soils table illustrates the percent contribution of a given soil type to the watershed. The land use table illustrates the percent contribution of a given land use to the watershed for the baseline. These percentages are updated for each scenario. To view the field-level information click on a field using across-hair. A pop-up window appears describing characteristics of that field: land use, total acres, and the acres, yield goal, and the receipts from production by soil type. The spatial variability in yield within a field can be used by farmers to determine the recommended nitrogen and phosphorus application rates.

After viewing watershed characteristics for the baseline or a given scenario, exit the Input/Output screen menu to generate a scenario.

Create a Scenario

The menu illustrated in Figure 3 pops up again. This time the menu is used to generate the proposed scenario (Figure 5). Select the'LUMP' button to modify the LUMPs based on the scenario which assumes that land currently in soybeans and sorghum will be planted in corn. Management practices, such as type of tillage and fertilizer and pesticide application rates, are reflected in this change.

Figure 5. Generate a scenario to modify LUMPs (click on the image to view full version)

A menu pops up and advises the user to generate a scenario (1, 2, or 3) based on either the baseline or one of the three scenarios (12 possible permutations in all). Select the 'Scenario 1' button which assumes that scenario 1 is developed from the baseline. A mirror image of all baseline CARE and AGNPS (or SWAT) INFO input files are copied as scenario 1 INFO input files. The purpose of copying these files is to ensure that no LUMP modifications are made to the baseline. Rather, all permutations are made to the copied file. Select the 'Apply' button in Figure 5 to initiate the copying process and initiate the next menu.

Generate Yield Goals and Price Information

A menu appears that details yield goals for each crop by soil type and associated price information for that crop (Figure 6). The price received for each crop is used to determine the gross receipts from production at the watershed scale. The yield estimates are based on the soil productivity index (PI) contained in published NRCS soil surveys. This rating system provides an index for ranking soils based upon their suitability to produce crops. Yields are not sensitive to tillage type. This is consistent with findings by Williams et. al. (1989) that there is no statistical difference among yields for different tillages.

Figure 6. View/modify yield goal and price information (click on the image to view full version)

The purpose of illustrating subfield variability is to capture the importance of soil type in determining yield goal, which in turn is used to decide the recommended application rates for nitrogen and phosphorus. The yield goal also serves as the foundation for determining the receipts from production and the input costs for nitrogen and phosphorus based on those recommended application rates.

Nitrogen and phosphorus application rates are based on the yield goals calculated by WAMADSS for each crop in each soil type. These rates are a significant improvement over the preset fertilizer levels established by the models because they are sensitive to crop yields which, in turn, are sensitive to soil type. Instead of the user manually entering the N and P rates, WAMADSS generates these parameters across the whole watershed. The nitrogen and phosphorus recommendations are based on equations adopted from the Soils Test Interpretations and Recommendations Handbook and the FERTREC software program developed by the University of Missouri. These recommendations also serve as the foundation for determining input costs for nitrogen and phosphorus and serve as input parameters for AGNPS. The handbook is used to assist individuals in deriving suggested fertilizer treatments based on soil test levels. The equations are imbedded into WAMADSS using AML.

Generation of WAMADSS Input Parameters

Subsequent menus prompt the end user to modify LUMPs by selecting individual fields or by changing LUMPs across the entire watershed by land use type (i.e., change all pasture fields in the watershed to soybeans). Menus guide the user through a series of prompts in order to update all the CARE, AGNPS or SWAT parameters needed to reflect the new land use activity or management practice selected. For example, the following input parameters are updated for the selected fields: the curve number, Manning's roughness coefficient, cropping factor, surface condition constant, and chemical oxygen demand factor. AMLs are used to relate between the updated land use layer and its relational INFO file to extract the appropriate values for these parameters. Specifically, land use codes for scenario 1 are modified in the land-soils layer (this layer is created to generate unique polygons that subdivide fields by soil type in order to take into account variability within the field). AMLs are also used to set flags for where land use activities, tillage practices, fertilizer applications or pesticide applications were changed from the baseline to scenario 1.

Acreage base, yield, and recommended nitrogen and phosphorus application rates for the 12 cropping systems are automatically generated by WAMADSS and CARE input files are generated that reflect those changes. Similarly, the input parameters for AGNPS or SWAT are generated and an input file is generated that reflects the updated LUMPs in the watershed.

It is the 'where' on the landscape that dictates the yield goal. Since crop yields are based on soil type and the budgets are based on crop type, the yields are aggregated from the crop/soil type delineation to the crop-specific delineation. These aggregated yields are exported to CARE. A weighted average of the cost and returns items in each budget, by the number of acres in each crop, is used to derive the net returns for the entire watershed. These watershed-level net returns can be used by resource-based agencies to better assess the economic impacts of changing LUMPs. By understanding these impacts, agencies can more efficiently allocate limited resources to those areas in the watershed that need attention.

After all calculations are processed for this loop the user has the option of initiating another loop to generate further changes to the existing scenario or proceeding to run WAMADSS. The end user clicks on the 'Submit' button (not illustrated) to execute the models and import the model results into WAMADSS for viewing purposes.

Visualizing Economic and Environmental Impacts

WAMADSS offers enhanced viewing capabilities over the stand-alone CARE, AGNPS and SWAT because it not only displays the output in a more organized fashion, it enables the user to visualize and compare watershed characteristics and the economic and environmental consequences of scenarios (or the baseline and a scenario).

The intent of this paper is not to illustrate all possible menu and submenu options and their corresponding information. Nor is the intent to interpret the economic and enviromental results based on generating the scenario. Rather, this section highlights the 'Spatial Output' and 'Tabular Output' sections in Figure 4.

The 'Input/Output' screen permits the user to view results in both graphical and tabular formats. The 'Spatial Output' section provides AGNPS output in graphical format and the 'AGNPS Tabular Results' section affords output in tabular format (SWAT output menus are not depicted in this figure). WAMADSS permits the user to view side-by-side any combination of graphical and tabular results for different scenarios (i.e., graphical soil loss results for the baseline and scenario 1) or for a given scenario (i.e., viewing side-by-side graphical and tabular results for the baseline).

With respect to the 'CARE Tabular Results', three levels of spatial detail are presented for the costs and returns items: (1) budgets by tillage practice, (2) budgets by crop type, and (3) a total cost and returns budget for the watershed. These levels, presented on a per-acre and total acreage basis, are illustrated in Figure 7.

Figure 7. CARE tabular results

The purpose of this research is to demonstrate how the linkages between LUMPs and the landscape affect nonpoint source pollution and cost and returns. The economic and environmental output generated in WAMADSS are discussed in this section because they are interrelated. The management decisions made regarding the crops planted, tillage practices used, and where on the landscape these activities take place, have an impact on the economic output generated by CARE. However, these decisions also impact the environmental output generated by AGNPS or SWAT. Specifically, the soil erosion rates, sediment delivery and nutrient/pesticide concentrations are influenced by these management decisions. For example, corn planted in conventional tillage on an a steeper slope in a less productive soil type will generate different environmental output than corn planted in no till on a flatter slope in a highly productive soil type.

Conversely, inherent landscape features such as slope and soil type have an impact on the environmental output generated by AGNPS or SWAT. However, these landscape features also impact the economic output generated by CARE. Specifically, the yield goals and nitrogen and phosphorus application rates depend on the spatial orientation of soils on the landscape. These physical characteristics impact net returns in terms of receipts from production and the input costs of applying nitrogen and phosphorus and other inputs. For example, corn planted on a less productive soil type will generate a lower yield goal than corn on a highly productive soil type, resulting in a lower gross receipt from production.

Summary, Conclusions and Future Research

This study involves the integration of knowledge and information from several disciplines into a functional computer-based interactive decision support system (IDSS). The research emphasizes the development of a software tool to aid decision makers in watershed management rather than evaluating results based on the tool and its policy implications. Specifically, A Watershed Management Decision support System (WAMADSS) is developed to provide decision makers with an interdisciplinary tool for evaluating the economic and environmental impacts of changing land use/management practices (LUMPs) in a timely fashion.

CARE, AGNPS and SWAT are seamlessly integrated into WAMADSS to evaluate LUMPs for Goodwater Creek watershed. The CARE program is a crop budget generator and the AGNPS model simulates sediment, runoff, and nutrient transport from agricultural watersheds. ARC/INFO GIS, its programming language - AML, and C programs are used to bind the external programs together in an IDSS framework.

WAMADSS is not only seamlessly integrated via menu interfaces but it exploits the strengths of its respective models, the spatial relational data base, and the GIS programming language, and generates results that would otherwise not be possible as stand-alone, disjointed tools. The whole is greater than the sum of its parts.

The focus of this research was on developing WAMADSS. Emphasis now needs to be placed on rigorously applying and building on the tool to evaluate economic and environmental impacts of a number of alternative management scenarios. For example, the economic and environmental impacts of establishing filter strips on cropped fields along stream corridors can be evaluated. Filter strips or vegetative buffers along stream corridors and riparian areas can reduce peak flows, stabilize banks, and trap sediments and nutrients.

This exercise leads us to a current deficiency of watershed-level studies - addressing the farm-level decision making process within a watershed context. Property rights were not addressed in this research because impacts addressed society's point of view rather than the private stakeholder's. Although WAMADSS can be used by stakeholders in a facilitated group session to inform them about watershed conditions, the decisions about what LUMPs to adopt reside at the farm level. This is not to say that WAMADSS may not impact their decision making process in terms of educating them about the collective interrelatedness of their actions.

Emphasis must be placed on finding a common ground between watershed-level and farm-level assessments. This is an ambitious endeavor given the paucity of financial information stakeholders are willing to make public. However, from a landscape perspective, the layers in the GIS can readily be adapted to focusing on farm-level analyses. What may emerge is a three-tiered approach to managing agroecosystems within a watershed. For example, a watershed-level assessment can be used to identify those areas within the watershed that need attention. Farm- and field-level assessments can be employed to evaluate the best strategy for economically improving water quality or soil erosion.

Tools based on good science are not necessarily used. These tools must be couched in an IDSS that is tailored to users' needs. Specifically, users must be involved in the IDSS development process to ensure that the tool is adopted. Therefore, emphasis must be placed on refining WAMADSS by allowing target end users to provide feedback on how to improve the tool.

Economic analyses of nonpoint source pollution is traditionally conducted in a non-interactive fashion. The time and data required to characterize a watershed precludes the possibility of generating 'what-if' scenarios in a timely fashion. Therefore, economic studies that have incorporated H/WQ models have focused on objectives that only addressed specific aspects of nonpoint source pollution. Researchers have been restricted to this 'from the ground up' approach because no tools existed to serve as a springboard for conducting more rigorous economic inquires. WAMADSS serves as this springboard.

SUJALA WATERSHED PROJECT:ECONOMICS OF GROUNDWATER RECHARGE FOR SUSTAINABLE WATERSHED DEVELOPMENT
 

Introduction

Initial and premature failures of irrigation wells are a predicament to farmers in hard rock areas due to cumulative well interference induced by drought situation. While demand side policies promote rapid extraction of groundwater, thereby exacerbating the predicament of well failure, supply side policies like watershed development programmes help dampen negative externalities. With the primary survey data from farmers of Basavapura watershed in Karnataka, India, this study proves that watershed development programmes have potential to alleviate the effect of drought by increasing groundwater recharge. This has contributed to increased physical and economic access to groundwater for farmers through increased pumping at reduced costs of extraction. Watershed development programmes are currently absorbing huge funds out of state and central schemes. Over the years, the focus of the programme has changed substantially. It began from a technically dominated programme and culminated into peoples' participatory schemes. However, throughout the travelogue, there are sporadic instances of single impact focused watershed development programmes. One of the important impact parameters visualised and utilised under the watershed development programme is the groundwater recharge. Therefore there is a need to design and implement programme specifically focusing on the role of watershed programme in augmenting groundwater resources. In hard rock areas, the life of irrigation wells and their groundwater yield is gradually declining due to factors singularly or in combination inter alia, interference of irrigation wells due to violation of isolation distance among wells, overdraft of groundwater without regard to recharge. In Karnataka, given the distribution of holdings, obscurity in property rights and frequently occurring droughts, interference among wells is a negative externality.

WORK PLAN TIME SCHEDULE

A. Field Investigation

Activity

Months

Approach and Methodology Outlining Various Steps for Performing the Study

Introduction
The rainfed areas support 40 percent of India's population, contribute to 9 percent of GDP, 36 percent of India's agricultural exports, 44 percent of food production, 91 percent of coarse cereals, 80 percent of oil seeds and 65 percent of cotton. These areas are fraught with problems of erosion, land degradation and loss of productivity, which have serious equity implications as they affect the very subsistence of dry land farmers, where more than 80 percent of precipitation is lost through runoff. Karnataka has the largest arid zone area after Rajasthan. About 63 per cent of the geographical area and 71 per cent of the net sown area in Karnataka are in dry agro climatic zones. With about 23 per cent of the area under irrigation, 62 per cent of agricultural production is from dry lands, which includes entire output of pulses. Karnataka's agriculture is a gamble with monsoon and frequently faces partial or severe droughts. Low productivity in dry lands and over exploitation of groundwater for irrigation have emerged as major problems. Thus, the watershed development program is implemented for in situ moisture conservation to improve productivity of dry land through judicious use of soil and water, agronomic practices, dry land horticulture, forestry, and other soil and water conservation measures. These result in restoration of ecological balance and improve the socio economic conditions of dry land farmers.

In this study we hypothesize that benefits inter alia, in situ moisture conservation, in situ conservation of predator population and other useful biodiversity, ecological benefits such as enhanced surface water, groundwater, control of soil and water erosion, aesthetics, air and water quality and other environmental and ecological benefits due to watershed development program are apparent benefits for farmers. This has a positive impact on farmers and others in the watershed who suffer economic losses inter alia due to drought, soil erosion, low soil productivity, biotic and abiotic stress, ambient environment, premature and initial failure of irrigation wells, who all stand to gain from the overall watershed development program impact including the recharged groundwater. This further improves their resilience when compared with non-watershed program areas.

Need for Economic impact assessment of watershed

Economic impact assessment of watershed inter alia includes impact on (i) rain fed farming in the watershed, in management and utilization of soil moisture under rainfed farming, discerning individual components of watershed management; (ii) upstream and downstream areas covering forestry, agriculture, horticulture and environment; (iii) groundwater recharge in turn reflected in contribution from irrigation. In these, direct and indirect, use and non-use, tangible and intangible benefits are to be valued from watershed development program. In this process of estimating the total economic value of watershed development, positive externalities due to watershed development program inter alia water infiltration, flood control, erosion control, bio diversity and species-sustaining services not valued by the market forces, get the due weighting

Since watershed development programs are public funded, the social benefits and costs need to be considered. Thus, the need for valuation of external benefits of watershed contributes for public policies to protect natural habitats.

In recent years both central and state governments have drawn up programmes on watershed development with internal and external assistance. Given the complexity of activities in the watershed development programme and their linkages, economic evaluation, discerning tangible and intangible benefits is essential to justify investment of scarce financial resources. This will add for better formulation, modification and implementation of WDPs with appropriate institutions for sustainable management of watershed. With this background, this study will be undertaken in Sujala watershed to assess the economic impact on agriculture productivity, land use and cover, groundwater recharge, watershed ecosystem and sustainability of watershed development program.

Groundwater recharge

Groundwater development in watershed development program is inevitable since the recharged aquifers need to be utilized for farm development. In the context of increased the probability of premature and initial failures of irrigation wells thus raising the implicit and explicit costs of extraction of groundwater, the contribution of Sujala watershed development program in this regard if significant, will have equity implications on small and marginal farmers. Studies on well failure (being a well that fails or goes dry, loses its yield, or requires deepening, because of interactive effects of pumping in neighboring wells or new wells coming in, but not because of low rainfall or technical deficiency in drilling, construction, pump size) rates being dampened by watershed development program (Chandrakanth, Bisrat and Bhat, Economic and Political Weekly pp 1164-1170, 2004) are reported. As the demand for groundwater is increasing, well drilling is increasing (at a compound growth rate of 10 percent per year in some parts of Karnataka). Since there are no regulations on well drilling and groundwater use, wells are drilled without regard to the degree of recharge and the recommended isolation distance (of 600 feet between two open wells and 850 feet between two borewells). This is resulting in cumulative well interference in general in all areas where intensity of groundwater use in relation to recharge is higher. In the watershed development program, reduction in the predicament of cumulative well interference is a crucial contribution.
Considering the relatively poor access to water resources in dry lands in Karnataka, the average cost of watershed treatment per hectare (Rs. 4000) and the incremental yield increase of 50 percent, the cost of providing major irrigation (Rs. 1 lakh) and the incremental yield increase of 400 percent, the cost is in the ratio of 25:1 while the returns are in the ratio of 8:1, between irrigation and watershed treatment (KNR Sastry, Development of natural resources: A solution to environmental problems, in Anil Agarwal (ed) The Challenge of the Balance, Center for Science and Environment, New Delhi, 1994, p.125). This provides the economic rationale for watershed development program. In addition, undertaking watershed development program in dry lands where more than 50 percent of our population ekes out their living, itself addresses equity concerns.

OBJECTIVES / ISSUES ADDRESSED IN THIS STUDY

In this proposal, the economic impact on agriculture productivity, land use and cover, groundwater recharge, watershed ecosystem, sustenance of watershed technologies/ practices and the synergistic relationship among watershed treatments, locations, community institution building will be assessed with the following objectives / issues:

1. Economic impact assessment of Sujala watershed development program on rainfed land, irrigated land, livestock, common property resources and other natural resource endowments in the watershed and outside watershed
2. Assessment of equity in distribution of benefits of Sujala watershed development program on marginal and small farmers, women, tribals and other disadvantaged sections of the society in up stream and downstream areas of the watershed
3. Estimation of economic benefits due to synergistic roles of surface water bodies, other in-situ conservation efforts and the community organization in augmenting water resources for irrigation
4. Examination of potentialities for application of ecological economics and institutions such as correlative rights, legal structures, local user groups and negotiations between interest groups in sustainability of watershed development program

REVIEW OF LITERATURE PERTINENT WORK DONE AT THE UNIVERSITY/ KARNATAKA/ INDIA/ ABROAD

Lokesh (Economic Impact Assessment of Watershed Development Programe: A study of Kallambella Watershed, Karnataka, unpublished Ph.D. theses, University of Agricultural Sciences, 2004) estimated the total economic value of Kallambella watershed development program (WDP) employing "With and Without Project" framework using field data for 2000-01. The data covered socio-economic features, investment on various components of watershed, economics of rainfed and groundwater irrigated crops and other related information for 2000-01. Discounted cash flow techniques and hedonic models were used in data analysis. This study has endeavored to measure inter alia, contribution of rainfed land in WDP.

Due to WDP, the value of rainfed dry land increased by 17 per cent (Rs. 5371 per acre) over non-watershed project area. The contribution from rainfed field crops, agri-horticulture, agro-forestry and silvi-pasture is 46 percent, while contribution from groundwater irrigation is 49 percent of the total economic value of Rs. 20375 per acre. Intangible benefits formed one per cent and livestock three per cent.

Net return per acre of dry land field crops (ragi and groundnut) in watershed area was Rs. 1437, which was Rs.719 in non-watershed project (NWP) area. Due to recharge of groundwater, gross irrigated area per irrigation well increased by 53 per cent (by 2.4 acres).

The perennial crops in agri-horticulture and silvi-pasture contributed almost equally (around 45 percent) to groundwater irrigation in this watershed. While groundwater irrigation is investment intensive, perennial crops in agri-silvi system are not. Hence higher budgetary allocation in watershed be made for this sector to augment farm incomes. This also complements environmental goals through carbon sequestration.

Chandrakanth, Bisrat and Bhat (Economic and Political Weekly, pp 1164-1170, 2004) indicated that initial and premature failure of irrigation wells are a predicament to farmers in hard rock areas due to cumulative well interference induced by drought situation. While demand side policies promote rapid extraction of groundwater, exacerbating the predicament of well failure, supply side policies like watershed development program are dampening the negative externalities. With the primary survey data from farmers of Basavapura watershed in Karnataka, India, this study proves that Watershed Development Program has potential to dampen the effect of drought by increasing the groundwater recharge. This has contributed to increased physical and economic access to groundwater for farmers in the upstream and downstream of watershed, through increased pumping at reduced costs of extraction.

IMPACT INDICATORS

1. Improved cropping intensity.
2. Equity: Class, gender and number of farmers located in upstream, middle stream and downstream who have received different benefits inter alia physical and economic access to (surface and ground) water resource.

3. Additional farm and non-farm income.
4. Extent of cultivation of waste land.
5. Income from agri-silvi, horti-silvi, silvi-pastoral crops.
6. Income from animal husbandry, dairy activities.
7. Availability of fuel wood from public and common lands.
8. Grazing of livestock on public and common lands.
9. Permanent improvements on the farm (like well construction, fencing, farm mechanization).

10. Improvement in farm assets and non farm assets.
11. Improvement in technology of agriculture.
12. Physical and economic access to different markets for better prices for farm products.