Item

Abstract

Rainfall-runoff models have been developed to predict runoff in response to rainfall. However, practical models do not adequately represent the rainfall-runoff processes and their applicability may decrease under specific conditions. Model parameter optimization is often required to obtain sufficient accuracy, but calibrated parameters do not always reflect watershed characteristics. Alternatively, models based on Richards’ equation are more precise and can directly use observed parameter values but are impractical due to their huge computational costs. In this study, we propose a vertical quasi-two-dimensional model for hillslope surface-subsurface flow (quasi-2D model), aiming at both appropriate representation of waterflow mechanism and reduction of computational cost. Subsurface flow is described by Richards’ equation, wherein the hydraulic gradient in the downward direction is approximated by the slope gradient. Surface flow is modeled using the kinematic wave model, which is computed separately from subsurface flow computation. This method can consider vertical soil moisture distribution and simplify the modeling of the runoff process. Rainfall-runoff simulations were conducted on a single slope using the quasi-2D model and compared with the results computed by a detailed model solving the two-dimensional Richards’ equation (2D model). For both subsurface and surface flows, the quasi-2D model reproduced the results of the 2D model well (NSE > 0.99). Furthermore, the computation time of the quasi-2D model was reduced to less than 1/10 of that of the 2D model. These results indicate that our quasi-two-dimensional modeling enables saturated-unsaturated flow simulations at lower computational costs, and is expected to be applied to basin-scale simulations.