Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Hydrology and Remote Sensing Laboratory » Research » Publications at this Location » Publication #301337

Title: Simultaneous state and forcing data correction for improved rainfall-runoff modeling

item CHEN, F - Science Systems, Inc
item Crow, Wade
item RYU, D - University Of Melbourne

Submitted to: Journal of Hydrometeorology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/15/2014
Publication Date: 10/1/2014
Citation: Chen, F., Crow, W.T., Ryu, D. 2014. Simultaneous state and forcing data correction for improved rainfall-runoff modeling. Journal of Hydrometeorology. 15(5):1832-1848. DOI: 10.1175/JHM-D-14-0002.1.

Interpretive Summary: The estimation of stream flow using rainfall-runoff models (and observed precipitation) is a valuable tool for water resource monitoring in agricultural watersheds. Recent research has focused on the potential for improving such estimates via the use of remotely-sensed surface soil moisture retrievals. This research presents the first real data demonstration of a new approach for integrating remotely-sensed surface soil moisture estimations into a rainfall-runoff model. The approach is based on utilizing remotely-sensed surface soil moisture retrievals to simultaneously update both pre-storm soil moisture values (required to determine the infiltration capacity of the landscape) and within-storm precipitation totals. Preliminary results support the premise that this approach will eventually yield improved stream flow predictions for water resource (and water quality) management applications in agricultural watersheds.

Technical Abstract: Uncertainties in precipitation input and pre-storm soil moisture conditions are important sources of error in stream flow predictions. All rainfall accumulation estimates - particularly those derived via remote sensing - are prone to error. Using a synthetic twin experiment, Crow and Ryu (2009, ‘CR09’) demonstrate that error in both antecedent soil moisture conditions and rainfall input forcing can be simultaneously filtered by assimilating remotely-sensed surface soil moisture retrievals. This opens up the possibility of applying satellite soil moisture estimates to address each of these key sources of error in hydrologic model predictions. Here, in an attempt to extend the CR09 synthetic analysis into a real data environment, two satellite-based surface soil moisture products - based on passive and active microwave remote sensing respectively - are assimilated to enhance the stream flow forecasts. A bias correction scheme is implemented to remove bias in background forecasts caused by synthetic perturbations in the ensemble filtering routines, and a triple collocation-based technique is adopted to derive both rescaled observations and error variances for data assimilation. Results are largely in agreement with the earlier synthetic analysis, that is, corrected satellite rainfall alone is able to improve stream flow prediction, especially for the relatively high-flow periods. In contrast, pre-storm soil moisture correction is more efficient in improving the baseflow component. When rainfall and state corrections are combined, RMSE of both the high- and low-flow components can be reduced by ~40% and ~30%, respectively. However, an unresolved issue is that soil moisture data assimilation also leads to under-prediction of very intense precipitation/high-flow events.