GIS-BASED COUPLING OF GLEAMS AND REMM HYDROLOGY: II. FIELD TEST RESULTS

Abstract

Agricultural fields border riparian areas in many locations. These riparian areas provide important filtering capabilities for agricultural chemicals and sediment leaving upland agricultural areas. Several models have been developed for upland agricultural areas and for riparian zones. The need to integrate such models through a flexible structure which represents the field conditions is highly desirable. Since water is the primary mechanism for pollutant transport, the system must be capable of managing both surface and subsurface water movement. The hydrology components of the upland agriculturally oriented water quality model, GLEAMS, and the riparian ecosystem management model, REMM, were coupled through a GIS to allow evaluation of agricultural and forest management impacts on water movement for typical field situations in the southeastern Coastal Plain. Programs and subroutines were developed to allow delineation of field areas, identification and maintenance of data based on the field map, and incorporation of data to the input data format for each model. The model system was capable of representing the shallow groundwater levels from forest management practices of clear-cut, thinned, and mature forests in field comparisons with measured values from the Gibbs Research Farm near Tifton, Georgia, for 27 months of comparison data between 1992 and 1994. Simulated monthly average shallow groundwater levels were within 0.14, 0.08, and 0.25 m of measured groundwater levels for the clear-cut, thinned, and mature forest treatments, respectively. Correlation results for the same treatments indicated at least an r 2 of 0.78 for all simulated versus measured monthly shallow ground water results. The model system did not respond as well in representing field conditions on total monthly runoff. Average monthly differences in total simulated runoff were 0.73, 1.19, and 0.48 cm lower (highly significantly) than measured results for the clear-cut, thinned, and mature forest management treatments from the Gibbs Farm site, respectively. No correlation was indicated between simulated and measured total runoff. The model system did, however, indicate the runoff trends expected due to changes in forest management. As the number of trees and canopy increased, runoff decreased. The model system has the potential of providing a cost effective method of incorporating multiple model characteristics into management practice evaluations.