New flood protection planning based on the use of distributed rainfall runoff simulation system

本研究通过基于物理的分布式降雨径流模型和随机降雨生成模型，寻求一种新的方法来建立新的洪水灾害规划。为达到研究目的，对分布式降雨径流模型所需的雷达数据尺度进行了分析，建立了随机降雨生成模型，并利用平移模型进行了降雨预测误差分析。本文的研究成果主要有：1)建立了随机降雨生成模型，建立了基于随机级联理论的随机降雨生成模型。该方法成功地考虑了地形对降雨分布的影响。2)分析了平移模型预报降水的误差结构，并将误差结构建模为对数正态随机场模型。3)分析了分布式降雨径流模型所需雷达数据的空间尺度，明确了所需雷达降雨尺度与流域尺度的关系。4)分析了分布式降雨径流模型所需输入数据的空间尺度，发现模型参数的几何空间信息对径流模拟不太敏感。

This research seeks a new methodology to establish a new flood disaster planning by using a physically based distributed rainfall runoff model and a stochastic rainfall generation model. To achieve the research purposes, radar data scale analysis required by a distributed rainfall runoff model, development of a stochastic rainfall generation model, and a rainfall prediction error analysis with the use of a translation model have been conducted. The followings are summery obtained by the research:1) Development of a stochastic rainfall generation modelA stochastic rainfall generation model by using random cascade theory has been developed. In the method, topographic effect on rainfall distribution is successfully incorporated.2) An error structure of rainfall prediction by a translation model was analyzed and the error structure has been modeled as a lognormal random field model.3) Spatial scale of radar data required by a distributed rainfall runoff model was analyzed and a relationship between required radar rainfall scale and catchment scale was clarified.4) Spatial scale of input data required by a distributed rainfall runoff model was analyzed and it was revealed that geometrical spatial information of model parameters were less sensitive for runoff simulation.