洪水预报误差分布的极大熵法

首先根据实际洪水预报误差出现在有限区域的特点,应用极大熵原理,建立了洪水总量预报误差分布的极大熵模型;通过几个不同流域的计算,得出随着降雨量的增大,产流预报误差趋于一个稳定值的结论. 同时将该模型计算的分布与正态分布进行了比较,结果表明用极大熵法求得的误差概率分布能更好地描述洪水总量预报误差的分布特性,可以根据实际降雨量的大小确定不同的最大不确定性的误差分布,为分析不同量级洪水预报的风险提供依据.

Maximum entropy method used in error distribution of flood forecast

Based on the characteristics of error distribution of flood forecast presented in the limited area, the error distribution of flood volume forecast is established using the principles of maximum entropy, and the maximum entropy model of error distribution of flood volume forecast is put forward. The corresponding solving method is also given in details. The computation results show that the runoff forecast error is always inclined to a constant value with the rainfall volume increasing. And the constant values of several different basins are obtained using this method. After that, the comparison analysis between the obtained model and the normal distribution is done in details. The results show that the probability distribution of error obtained by the method of maximum entropy can describe the characteristics of error distribution of flood volume forecast. The method can be used to determine the different maximum likelihood probability distribution of error according to the different real-time precipitation. At the same time, it can also be used to provide the basic reference for risk analysis on different levels of flood forecast.