Testing the Transfer of Hydrologic Model Parameters Across Scales: Modeling the Emory River, Daddy’s Creek, and Crooked Fork Watersheds.

Modeling the hydrologic response of small watersheds is often hindered by lack of data. We cannot calibrate robust models without ample and sufficiently accurate precipitation and discharge measurements. Limited data prevent the calibration process from estimating parameter values that adequately depict the physical characteristics of the watershed. But physical characteristics change gradually over space and hydrologic processes behave in predictable ways across scales. These factors suggest that the modeling of data-limited small watersheds could be improved by using the physical characteristics of a large, data-rich watershed to estimate the characteristics of a small watershed.

This thesis presents such an application. I estimate the characteristics of the large Emory River watershed by calibrating the parameters of a VenSWM hydrologic model. VenSWM is based on the algorithms of the Stanford Watershed Model. I test how well the calibrated parameters estimate the characteristics of two subwatersheds, as determined by accuracy in predicting stream discharge in the subwatersheds, by applying them to models of the Daddy’s Creek and Crooked Fork subwatersheds. I compare the adequacy of these estimates with the best possible characterization of the properties of the subwatersheds, which I obtain by calibrating hydrologic models specifically for the subwatersheds.

I find that the Emory parameters provide very accurate predictions for the Daddy’s Creek watershed, which has comparable physical characteristics to the larger watershed. The Emory parameters also provide reasonably accurate predictions for the Crooked Fork watershed, which has somewhat distinct physical characteristics from the larger watershed.