Skip to main content
ARS Home » Midwest Area » Columbia, Missouri » Cropping Systems and Water Quality Research » Research » Publications at this Location » Publication #224766

Title: Scaling up a model: Impact of soil map resolution and watershed discretization.

item Baffaut, Claire
item Sadler, Edward - John
item Ghidey, Fessehaie

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 3/27/2008
Publication Date: 9/3/2008
Citation: Baffaut, C., Sadler, E.J., Ghidey, F. 2008. Scaling up a model: impact of soil map resolution and watershed discretization. In: Proceedings of The National Sedimentation Laboratory: 50 Years of Soil and Water Research in a Changing Agricultural Environment, September 3-5, 2008, Oxford, Mississippi.

Interpretive Summary:

Technical Abstract: The modeling of a large river basin often starts with the modeling of a smaller watershed within that river basin. Reasons include the greater availability of flow and water quality data for smaller watersheds, a better knowledge of land management practices and the methodological step to start with a small, more homogeneous watershed to decrease uncertainties. When scaling up to a larger area, the selected detail of information included in the model cannot be maintained and “lumping” occurs. This analysis deals with two factors SWAT modelers face when scaling up a model: the number and size of the subbasins and the resolution of the soil map. SWAT was first calibrated and evaluated for Goodwater Creek Experimental Watershed (GCEW), a 70 km2 watershed located in north-central Missouri. Its performance and the parameter sets were compared when using SSURGO or STATSGO soil data sets on one hand, and 3 or 7 subbasins on the other hand. Then, the performance of SWAT with these parameter sets was evaluated for the Long Branch watershed, a 462 km2 (7 times larger) watershed that contains GCEW and has similar soils, land use, and cropping and management systems. The performance of the model in simulating stream flow and sediment yields using the STATSGO soil data was as good as that using the SSURGO soil data. The performance of the model in simulating stream flow from the Long Branch watershed was as good as that from GCEW. Results will include a comparison of the model’s goodness of fit for the different types of inputs for flow and sediment yields and between GCEW and Long Branch watersheds for flow. These results suggest that STASGO soil data and a coarser discretization of the river basin can be used for the modeling of Long Branch watershed and the Mark Twain Lake / Salt River Basin.