A GIS-Based Watershed Quality Model for Identifying Land-Based Pollution Problems

Christopher J. O'Bara, Paul P. Piszczek, Yvette R. Clark, Dennis B. George, Jason A. Weeks and Jack D. Wall
Tennessee Technological University, Center for the Management, Utilization and Protection of Water Resources, Cookeville, Tennessee 38505 USA

The recent focus on watershed-based approaches to land and water resource issues has prompted the need to develop assessment tools. In many watersheds, water resources serve municipal, industrial, recreational, and ecological purposes; therefore community decisions makers must consider all user interests and not to upset the holistic balance of water resource uses. To address the need for resource assessment tools, a watershed quality integrity model (WQI) was developed using universally accepted model components and readily available data. The WQI model is a GIS-based model that links three watershed-process models, the Agricultural NonPoint Source model (AGNPS), the Modular Three-Dimensional Finite-Difference Groundwater Flow model (MODFLOW), and the RCHRES module of the Hydrological Simulation Program-Fortran model (HSPF). Input geographic data included soils, hydrography, land use, land cover, topography, and locations of discharge and recharge points. Watersheds are divided into hydrologically connected subwatersheds and each subwatershed is further subdivided into cells of 0.9 ha. The integration of these models and the GIS coverages provides comprehensive analysis of watershed quality as depicted by the watershed quality index. The watershed quality index is a unitless score derived by simulated output of 12 chemical and physical parameters. Simulations can be made by single-storm events with rainfall of between 13 mm to 125 mm or a time-series event of a one month period. The resulting watershed quality index provides the user an understanding of the resulting quality of a receiving stream following land use and/or watershed management practices. Additional, a costing module allows the user an understanding of the estimated costs of a proposed action.

The WQI model was applied to a Richland Creek watershed of the large Elk River basin in south-central Tennessee, USA. The Elk River basin drains 17,929 km² of primarily farm and forest lands. Water resource development includes a large hydro-electric impoundment, a cooling water impoundment, and several small water supply impoundments. The Richland Creek watershed drains 1254 km² primarily of karst geology, and gently rolling or steep-sided hills. Farmland account for 60% of land cover, with the remaining 37% in forest and 3% urban. Richland Creek is the main water source for a human population of 25,000. One non-point source remediation approach modeled was the reforestation of several subwatersheds. This practice increased the watershed quality index in the treated subwatershed as well as connected subwatersheds by as much as 35%. The cost of this approach would be $22,280,000 (USA). A second alternative approach simulated was the conversion of buffer zones adjacent to the streams. This approach resulted in improvements approaching 30% with a projected costs of only $460,700 (USA). Point-source techniques were also applied by upgrading the sewage treatment plant (STP) discharging directly into Richland Creek. Improvements to the STP by upgrading to tertiary treatment only improved by WQI by less 10% at a cost of $15,000,000 (USA).