Calibration of a Watershed Model for Metropolitan Atlanta

The Metropolitan North Georgia Water Planning District was created in the 2001 Georgia legislative session to provide a mechanism for regional coordination on water supply, wastewater treatment, and stormwater management. The District's first mandate included the development of a Watershed Management Plan (WMP) to provide for compliance with water quality standards that are associated with TMDLs, while allowing for continued sustainable growth in the region. To accomplish coordinated management efforts across jurisdictional boundaries, a modeling approach consistent through out the District was needed. Additionally, the model needed to reflect both NPS and point source influences on up-stream loadings to downstream water quality. Therefore, the Better Assessment Science Integrating Point and Nonpoint Sources (BASINS 3.0) tool (USEPA, 2001) and the Hydrological Simulation Program – FORTRAN (HSPF) (Bicknell et al., 2001) were used for this study.Model calibration and validation of current conditions, along with the ability to perform future conditions analyses, were needed for both hydrology and water quality. Water quality constituents modeled in the study included TSS, DO, ammonia, nitrate-nitrite, orthophosphorous, BOD, water temperature, fecal coliform, and metals. An additional challenge in this study was quantifying and reporting results at a detailed level (259 12-digit HUCs) for a large area (over 6640 mi2). Reporting at this level was crucial to the process of prioritizing watersheds for implementation strategies such that maximum benefit is achieved with limited financial resources.A key element of the watershed model set up and calibration effort was selection of calibration sites with subwatersheds that were representative of the study area. Factors such as availability of observed streamflow data, diversity of subwatershed size and land use, and representation of the major river basins in the area were considered. The calibration watersheds were spread throughout the study area, allowing the project team to analyze variability of hydrologic response in the study area. To accurately represent this variability, significant effort was spent developing a robust database of model inputs. Along with the data layers provided in BASINS, local data were collected and compiled for meteorological, land cover, hydrology, water quality, point source, and water withdrawal data requirements.As part of the model development process, many components of the BASINS 3.0 system were used, namely WinHSPF and WDMUtil for pre-processing and GenScn for post-processing. Additionally, an expert system for HSPF hydrology calibration (HSPEXP) (Lumb, 1994) was used. Along with these interactive tools, a suite of scripts was developed for running batch processes across the entire study area. These scripts were used to perform such tasks as meteorological data compilation, HSPF input (UCI) file development, model production runs, and summarizing results.