Construction of draught estimation model with combination of improved TOPMODEL and PDSI

Abstract: The frequent occurrence of drought for the semiarid region in spring period in recent years seriously influences the grain production. Distributed hydrological model and draught estimation model for semiarid region are not well developed, and they are not physically-based and universal. In this study, grid hydrological model and draught estimation model for semiarid region were established based on the data of typical northern semiarid region of Songnen Plain. The runoff process was estimated by saturated and unsaturated zone flow model, which was established based on LBM method. For the flow confluence process, flow confluence numerical model was constructed by solving the slope and river flow equation with LBM method，and furthermore, LBMGTOPMODEL was built by means of loose coupling for each module in the grid areas. From the respective of hydrologic cycle, combining with hydrological model LBMGTOPMODEL and Palmer Drought Index (PDSI), draught estimation model for semiarid region was formed by simulating the temporal and spatial variation of parameters in the water circulation process of Taoer River, which is located in a typical semiarid region. The results showed that the average absolute value of the relative error of the 10-field flood simulation runoff was 4.45%, the average absolute value of the relative error of the flood peak flow was 5%, and the mean value of the deterministic coefficient was 0.76. It could be seen that the established LBMGTOPMODEL hydrological model could simulate the process of rainfall and runoff with a high accuracy. Meteorological hydrological constants such as meteorology, water supplement, runoff, water loss and climate characteristics showed a certain spatial distribution state. Among the meteorological hydrological constants, the spatial distribution of climatic characteristic constants was 0-7.23, which showed the most obvious difference. The average values of meteorological hydrological constants changed with the variation of selected computational time period. When the established draught estimation model was used to assess the draught condition of investigated area, the model estimation results were in accordance with the actual conditions of Taoer River basin. When the draught conditions estimated by the established draught estimation model were compared with the draught conditions evaluated by rainfall deviation index of single factor drought index and draught conditions of the actual situation, it could be concluded that PDSI index evaluated by the distributed hydrological model could well reflect the effects of various factors in the hydrological process on the drought and the temporal and spatial evolution process of drought. PDSI index evaluated by the distributed hydrological model could provide guidance for the draught condition assessment.