Evolution of a parsimonious rainfall–runoff model using soil moisture proxies

Abstract

Pre-storm soil moisture conditions play a crucial role in the analysis of the physical processes involved in conceptual rainfall–runoff modeling and for demonstration of the soil–water–atmosphere interaction. Runoff quantification is of common interest to most hydrologists who study rainfall–runoff modeling. The scientific community prefers hydrological models with a smaller number of parameter requirements and superior performance, because of the increased water resource applications. In order to circumvent the impacts of sudden undesirable jumps (e.g., in the conventional CN model) in runoff estimation, we investigated the integral effects of soil moisture proxies before and after rainfall occurrence so as to reduce the structural inconsistencies of the previous models. Accordingly, these sudden jumps were avoided by incorporating a new expression that varied from storm to storm and with prior rainfall. Similarly, the utilization of additional information to calibrate new parameters reduced the applicability to ungauged watersheds. By persisting with the principle of simplicity and avoiding over-parameterization, a new model was developed and validated using measured data of 1804 storm rainfall–runoff events from 39 South Korean watersheds. The significant results exhibited by the suggested one-parameter model in comparison to those of the other contenders were hydrologically justified using three statistical metrics and scatter plots.