Improving flood forecasting capability of physically based distributedhydrological models by parameter optimization

摘要

Physically based distributed hydrological models (hereafter referred to asPBDHMs) divide the terrain of the whole catchment into a number of gridcells at fine resolution and assimilate different terrain data andprecipitation to different cells. They are regarded to have the potential toimprove the catchment hydrological process simulation and predictioncapability. In the early stage, physically based distributed hydrologicalmodels are assumed to derive model parameters from the terrain propertiesdirectly, so there is no need to calibrate model parameters. However,unfortunately the uncertainties associated with this model derivation are veryhigh, which impacted their application in flood forecasting, so parameteroptimization may also be necessary. There are two main purposes for thisstudy: the first is to propose a parameter optimization method for physicallybased distributed hydrological models in catchment flood forecasting by usingparticle swarm optimization (PSO) algorithm and to test its competence and to improve its performances; thesecond is to explore the possibility of improving physically baseddistributed hydrological model capability in catchment flood forecasting byparameter optimization. In this paper, based on the scalar concept, a generalframework for parameter optimization of the PBDHMs for catchment floodforecasting is first proposed that could be used for all PBDHMs. Then, withthe Liuxihe model as the study model, which is a physically based distributedhydrological model proposed for catchment flood forecasting, the improvedPSO algorithm is developed for the parameteroptimization of the Liuxihe model in catchment flood forecasting. Theimprovements include adoption of the linearly decreasing inertia weight strategyto change the inertia weight and the arccosine function strategy to adjustthe acceleration coefficients. This method has been tested in two catchmentsin southern China with different sizes, and the results show that theimproved PSO algorithm could be used for the Liuxihe model parameter optimization effectively and could improve the model capability largely in catchment flood forecasting, thus proving that parameter optimization is necessary to improve the flood forecasting capability of physically based distributed hydrological models. It also has been found that the appropriate particle number and the maximum evolution number of PSO algorithm used for the Liuxihe model catchment flood forecasting are 20 and 30 respectively.