Lots of water projects are built in river basins to mitigate water resources shortage and develop regional economy.This changes the underlying surface conditions of the basin, accordingly impacting evaporation, infiltration,runoff generation and flow concentration, etc, resulting in that the conventional flood forecast model cannot represent the law of runoff generation and flow concentration.Through the decision tree technology, in the paper, the law of flood trapped or discharged by the upstream reservoirs is explored.By establishment of the inference model for the flood detention ratio of separate flood, the flood forecast is modified to improve the precision of the flood forecast .%为缓解水资源缺乏以及发展区域经济，人们在流域内修建了大量的水利工程，改变了流域的下垫面条件，从而影响了蒸发、入渗、产流、汇流等特性，导致传统的洪水预报模型难以反映流域产汇流规律。文章文通过决策树技术，挖掘上游水库群拦蓄或泄放洪水的规律，建立场次洪水拦洪比推理模型，对洪水预报予以修正，有效提高洪水预报精度。
展开▼
