Resource information
Deriving a priori gridded parameters is an important step in the development and deployment of an operational distributed hydrologic model. Accurate a priori parameters can reduce the manual calibration effort and/or speed up the automatic calibration process, reduce calibration uncertainty, and provide valuable information at ungauged locations. Underpinned by reasonable parameter data sets, distributed hydrologic modeling can help improve water resource and flood and flash flood forecasting capabilities. Initial efforts at the National Weather Service Office of Hydrologic Development (NWS OHD) to derive a priori gridded Sacramento Soil Moisture Accounting (SAC-SMA) model parameters for the conterminous United States (CONUS) were based on a relatively coarse resolution soils property database, the State Soil Geographic Database (STATSGO) (Soil Survey Staff, 2011) and on the assumption of uniform land use and land cover. In an effort to improve the parameters, subsequent work was performed to fully incorporate spatially variable land cover information into the parameter derivation process. Following that, finer-scale soils data (the county-level Soil Survey Geographic Database (SSURGO) (Soil Survey Staff, 2011a,b), together with the use of variable land cover data, were used to derive a third set of CONUS, a priori gridded parameters. It is anticipated that the second and third parameter sets, which incorporate more physical data, will be more realistic and consistent. Here, we evaluate whether this is actually the case by intercomparing these three sets of a priori parameters along with their associated hydrologic simulations which were generated by applying the National Weather Service Hydrology Laboratory’s Research Distributed Hydrologic Model (HL-RDHM) (Koren et al., 2004) in a continuous fashion with an hourly time step. This model adopts a well-tested conceptual water balance model, SAC-SMA, applied on a regular spatial grid, and links to physically-based kinematic hillslope and channel routing models. Discharge and soil moisture simulated using the different set of parameters are presented to show how the parameters affect the results and under what conditions one set of parameters works better than another. In total, 63 basins ranging in size from 30km² to 5224km² were selected for this study. Sixteen of them were used to study the effects of different a priori parameters on simulated flow. Simulated hourly flow time series from three cases were compared to hourly observed data to compute statistics. Although the overall statistics are similar for the three different sets of parameters, improvements in simulated flow are observed for small basins when SSURGO-based parameters are used. Fifty-seven basins covering different climate regimes were used to analyze differences in the modeled soil moisture. Results again showed that the use of SSURGO-based parameters generate better soil moisture results when compared to STATSGO-based results, especially for the upper soil layer of smaller basins and wet basins.
