【摘要】 基于流溪河模型构建了不同空间分辨率的新丰江水库入库洪水预报模型，采用粒子群算法优选模型参数进行洪水模拟，对比实测资料验证了模型性能。结果表明，建立的新丰江水库入库洪水预报流溪河模型模拟效果优良，符合新丰江水库入库洪水预报的精度要求，可用于新丰江水库入库洪水预报：90 m空间分辨率的模型模拟效果最好，200 m空间分辨率的模型也具有良好的模拟性能，500 m空间分辨率的模型模拟性能有明显降低。模型计算工作量随分辨率的提高呈指数增加，90 m空间分辨率的参数优选时间是200 m空间分辨率参数优选时间的7.04倍，而200 m空间分辨率的参数优选时间是500 m空间分辨率参数优选时间的3.53倍。更多还原

【Abstract】 This paper set up flood forecasting model of Xinfengjiang reservoir under different spatial revolution based on Liuxihe model, and simulated the observed hydrographs by optimizing model parameters with PSO algorithm, then verified model performance. The results show that the model performance is excellent with high simulation accuracy, and could be used for the real-time flood forecasting of Xinfengjiang reservoir. The model with 90 m spatial resolution has the best performance while the model performance with 500 m spatial resolution decreases obviously. The computation time needed for model run increases exponentially with the increase of spatial resolution, the time needed for running the 90 m spatial resolution model is 7.04 times as much as that needed for 200 m spatial resolution model, while the time needed for running the 200 m spatial resolution model is3.53 times as much as that needed for 500 m spatial resolution model.更多还原