Research papersEvent-based hydrological modeling for detecting dominant hydrological process and suitable model strategy for semi-arid catchments

- This paper explores suitable models for flood processes in semi-arid catchments.
- Suitable models need at least one nonlinear component to simulate surface runoff.
- Two nonlinear components should be arranged in parallel, rather than in series.
- Parallel nonlinear components should be combined by multiplication.
- The dominant hydrological process is the infiltration excess surface runoff.

To simulate the hydrological processes in semi-arid areas properly is still challenging. This study assesses the impact of different modeling strategies on simulating flood processes in semi-arid catchments. Four classic hydrological models, TOPMODEL, XINANJIANG (XAJ), SAC-SMA and TANK, were selected and applied to three semi-arid catchments in North China. Based on analysis and comparison of the simulation results of these classic models, four new flexible models were constructed and used to further investigate the suitability of various modeling strategies for semi-arid environments. Numerical experiments were also designed to examine the performances of the models. The results show that in semi-arid catchments a suitable model needs to include at least one nonlinear component to simulate the main process of surface runoff generation. If there are more than two nonlinear components in the hydrological model, they should be arranged in parallel, rather than in series. In addition, the results show that the parallel nonlinear components should be combined by multiplication rather than addition. Moreover, this study reveals that the key hydrological process over semi-arid catchments is the infiltration excess surface runoff, a non-linear component.