The Magazine for
Esri Software Users
Dynamic Hydrological Modeling Using ArcView GIS
by Joep C. Luijten, GIS Specialist and Systems Modeler
Editor's Note: The International Center for Tropical Agriculture (CIAT) in Cali, Colombia, is a nonprofit, nongovernment research organization dedicated to alleviating poverty and preserving natural resources in developing countries. CIAT has been using GIS to map poverty and identify links between poverty and land use. This article describes a spatial analysis tool CIAT built for performing hydrological modeling.
Water shortages and the degradation of water supplies threaten the food security and health of people in many parts of the world. This is particularly true in developing countries that are experiencing rapid population growth but have limited means to manage water resources. Steeply sloping regions of some Latin American countries where small-scale farming is the predominant production activity and source of food are examples of these threatened areas. CIAT cooperates with community-level organizations to improve water resources management in hillside areas. Currently, research efforts are focused on the Tascalapa watershed, which covers 12,160 hectares in central Honduras and the 3,250-hectacre Cabuyal watershed in southwest Colombia.
In both these watersheds, streams are the primary source of water to satisfy domestic, industrial, and agricultural needs. Decreasing water availability caused by the lack of local government regulation, inadequate resource management strategies, and nearly unregulated use of stream water has caused concern. Local stakeholders are generally not very aware of the interdependencies between land and water resources, the impact of using these resources, or any upstream-downstream connections.
Quantitative information on these landscapes is essential for improving local-level decision making and guiding local rural development and water management. CIAT developed the Spatial Water Budget Model (SWBM), a spatial analysis tool that builds on the hydrological modeling capabilities in ArcView Spatial Analyst to supply this information.
Using GIS to Hydrological Models
Over the past decade, numerous interfaces have been developed between ArcView GIS and ArcInfo and various hydrological models such as Simulator for Water Resources in Rural Basins (SWRRB), Environmental Policy Integrated Climate (EPIC), Groundwater Loading Effects of Agricultural Management Systems (GLEAMS), TR20, HEC-1, and HEC-2. In each case, GIS is used for preprocessing and postprocessing of data. ArcView Spatial Analyst provides excellent capabilities for computing and visualizing the static hydrological properties of landscapes such as flow direction, flow accumulation, and flow length for data that does not incorporate a time factor. SWBM, a truly dynamic simulation model entirely written in Avenue, demonstrates that ArcView Spatial Analyst can be effectively used for complex spatial-temporal modeling.
Spatial Water Budget Model
SWBM is a continuous, distributed-parameter, watershed-scale model that simulates water supply and demand on a daily basis. SWBM measures water as it flows through the watershed and can simulate the effects of water use and flow control by dams. Land unit water balance, water flow to streams, stream water flow balance, water storage in dams, and water use from streams and dams are all processes that can be simulated on a daily basis by SWBM.
These daily calculations of soil-water balance for each land unit include the relationships-expressed mathematically-of canopy interception, evapotranspiration, surface runoff, and infiltration and drainage out of the root zone. These are factors typically incorporated in established hydrological models. Next, the flow accumulation function is applied to calculate surface runoff and lateral flow toward streams using a two-compartment distributed delay function. A constant volumetric flow rate of threshold of 1,300 cubic meters per day (m3d-1), which is equivalent to 15 liters per second (L/s), was applied to the accumulated flow grid to delineate streams. This value gave the best match between streams delineated using digital elevation model (DEM) data and existing mapped streams in the Cabuyal watershed. Water that flows into a stream is routed to the watershed outlet on the same day. Finally, the flow accumulation function is applied again to calculate the downstream changes in stream flow rates caused by dams and streams at multiple locations throughout the watershed.
Daily dynamics were implemented in Avenue by using a For-Each loop that performs the above calculations on grid data on a day-by-day basis for any number of days. Daily river flow grids are saved to a permanent location on the disk for postprocessing. Daily flow rates at specific user-defined locations are also saved in a dBASE file. If desired, the state of all other components of the land unit and stream water balance can be saved as a grid, or an average value for the watershed can be calculated and saved in a table.
SWBM was created using ArcView GIS 3.2 with the ArcView Spatial Analyst and Dialog Designer extensions. The ArcView Spatial Analyst extension was used for the hydrological computations and Dialog Designer for developing a user-friendly GUI. This interface can be used for inputting data, specifying simulation output options, controlling and displaying progress of a simulation run, postprocessing, and displaying simulation results.
Base Data for the Cabuyal Watershed
Some of the input data that SWBM uses in grid format includes DEM, slope, land use, SCS curve numbers [a set a standard empirical curves used to estimate runoff published by the Natural Resources Conservation Service, formerly known as Soil Conservation Service], hydraulic conductivity, drained upper limit, field saturation data, and the water contents at wilting point. The DEM data, created by digitizing contour maps, was used to determine the direction of surface flow and the location of streams. Soil data was based on field measurements and existing Pescador Soil Series data. A land use grid with 10 classes was created through a maximum likelihood classification of a Landsat 4 Thematic Mapper image from August 9, 1989.
All input grids were in ArcInfo GRID format and were clipped to the same spatial extent. These grids were resampled at 100-meter resolution, the maximum resolution that adequately represented the steep topography of this watershed. The coordinate system used was a Transverse Mercator projection with the Bogota Observatory (BOO) datum, International 1909 ellipsoid, and Western Colombian origin (4 35' 56.57" N W 77 4' 51.3").
Canopy interception parameters, rooting depth, and crop evapotranspiration coefficients had to be specified for each land use class in order to characterize the hydrological properties of the landscape. These parameters were entered using dialog boxes and saved in tables. SWBM automatically combines these parameters with land use grids to create parameter grids. Daily rainfall, solar radiation, and minimum and maximum temperatures were measured at a farm in the middle of the watershed from January 1994 through December 1997 and stored in an ASCII file.
Water Use and Dams
Water is removed from streams or dams to meet domestic, industrial, and agricultural needs. The locations where water is extracted are identified on maps showing the streams. Water use rates may vary from day to day. For every location, daily water extraction rates and their applicable period(s) are specified in dialog boxes and saved in a table.
In a similar fashion, dam locations are selected from stream maps. Dams may be located anywhere along a stream and used to store water for use at a later time and/or to regulate flow rates. The storage capacity and operational characteristics of each dam are entered in a dialog box. The rate of water flow out of a dam may be fixed or can be calculated automatically if the dam storage should be kept at an equilibrium level. The user has the option of specifying minimum required river flow rates and maximum allowed river flow rates. The dam operation may vary from day to day, or it can be kept constant for as long as a year.
Some Model Results
SWBM was used to analyze water availability and water use under three scenarios-corporate farming, ecological watershed, and business as usual-plausible development pathways for the Cabuyal watershed in 2025. Simulated stream water availability and the percentage of water that was extracted were different for each scenario and varied greatly over space and time. Simulation results showed that the locations and length of streams with flow rates of at least 15 L/s vary over time. There are relatively few streams in the dry period, which lasts from June to September, because smaller streams dry up. This means that people have to walk farther to get water.
In the corporate farming scenario, up to 61 percent of the available water would be extracted if there were no dams. This high percentage raises concerns about the effects on aquatic and riparian ecology, concentrations of potential contaminants, and maintenance of water reserves for especially low rainfall years. SWBM was used to determine the locations and storage capacities of dams needed to sustain minimum flow rates during dry seasons and to provide access to sufficient water year round within a reasonable distance of households and irrigated fields.
SWBM is currently being used to explore the long-term variability in water accessibility and to conduct a water scarcity risk analysis for the Tascalapa watershed. For these analyses, the SWBM model incorporated 100 years of stochastically generated weather. Simulated flow rates at every stream segment are statistically analyzed to identify areas that may have a critically high water demand/water supply ratio during at least part of the year. SWBM will be demonstrated and results of the Cabuyal and Tascalapa studies presented at local stakeholder workshops later this year.
For more information, contact
About the Author
Joep C. Luijten is GIS specialist and systems modeler with the Agricultural and Biological Department, University of Florida, Gainesville, Florida, and the International Center for Tropical Agriculture, Cali, Colombia.