|ZANG Shuaihong,LI Zhijia,HUANG Yingchun,LI Qiaoling.Calculation method and application of river network regression coefficient Cs based on self-similarity of river network structure. J. Lake Sci.2019,31(3):788-800. DOI:10.18307/2019.0317
|关键词: 新安江模型 河网消退系数 参数规律 汇流计算
|Calculation method and application of river network regression coefficient Cs based on self-similarity of river network structure
ZANG Shuaihong1, LI Zhijia1,2, HUANG Yingchun1, LI Qiaoling1
1.College of Hydrology and Water Resources, Hohai University, Nanjing 210098, P.R.China;2.National Cooperative Innovation Center for Water Safety & Hydro-Science of Hohai University, Nanjing 210098, P.R.China
|The river network parameter Cs of Xin'anjiang model has significant influence on the simulation of flood peak, but it is difficult to estimate and transfer directly for the data-limited regions. Therefore, the determination of Cs is a difficult problem and has to be solved urgently for the application of hydrological models in ungauged basins. Based on the flow calculation process of the self-similar river network structure, this study presents an estimation method of Cs based on the river chain storage equation. For 11 selected catchments, including humid, semi-humid and semi-arid areas, the Cs values were calculated and statistically analyzed. Compared with the model based optimized method, the difference of the Cs values are extremely small and the model performances are similar. Which indicates that the Cs estimation method has certain applicability. Chenhe and Tunxi catchments are used to investigate the effect of the sub-catchment properties on the estimation of Cs. Results show that with the increasing of catchment size and the number of river chains, the Cs values increase. Meanwhile, the closer the sub-catchment is to the outlet, the higher the Cs value becomes, which indicates that the Cs value of each sub-catchment changes after the watershed blocked. Hence it must cause the error of the confluence calculation with the same Cs value. In addition, the Cs value for the sub-catchment is normally smaller than the that for the whole catchment. Because the whole catchment Cs value represents all storage function of the whole catchment, and when catchment is divided into blocks, each sub-catchment uses its own Cs for storage, and the river below the outlet of the sub-catchment is calculated by the Muskingum algorithm, and Cs value for the whole catchment is larger than that for the sub-catchment. In order to further improve the application effect of this method in the data-limited regions, the confluence calculation process of the Xin'anjiang model has been improved in this study. The confluence calculation module of the Xin'anjiang model is modified so that the confluence is calculated in each sub-catchment separately with different Cs values. The simulation results of the Xin'anjiang model and the modified model in study catchments show that both of two models could obtain reasonable forecasting results. In contrast, the modified model shows more advantages, as it considers the spatial variability of catchment and establishes relationship between model parameters and surface conditions to strengthen the physical mechanism and parameter independence of the model, which can obviously enhance the flood forecasting accuracy in ungauged basins.
|Key words: Xin'anjiang model river network regression coefficient parameter regionalization confluence calculation