A two-stage partitioning monthly model and assessment of its performance on runoff modeling

Abstract Monthly hydrological models are very useful tools for modeling of catchment hydrology. In this study, a two-stage partitioning monthly water balance model (TSPM) with three parameters is developed based on a generalized proportionality hypothesis. A total of four model scenarios are conducted to investigate the effects of calculation procedures of the soil water content on model performances by incorporating the hyperbolic tangent function used in the two-parameter monthly water balance model (TWBM) by Xiong and Guo (1999) . The developed monthly models are then tested in 205 Catchment Attributes and Meteorology for Large-sample Studies (CAMELS) basins across the United States. Application results demonstrate that the three-parameter model TSPM1 performs well and has better performance than the modified two-parameter model TSPM4 using the equations of soil moisture and runoff from the TWBM. Five monthly models with different structures and model parameter numbers are compared to the two developed TSPM models. Results show that the TSPM1 and TSPM4 models have a median performance among these models. They show an excellent performance in high flow, but a relatively weak performance in low flow, especially for the TSPM4 model. The relationship between the model performances and catchment characteristics is explored and results demonstrate that the climate and vegetation conditions of catchment (e.g., aridity index and forest fraction) exert very significant effects on model performances, including entire goodness-of-fit, low and high flows in runoff simulations.