短期降雨预测的随机微分模型  

Short-term rainfall prediction of stochastic differential models

作　　者：陈森发[1] 张文红[1] 张建坤[1] 于栋华[1] 

机构地区：[1]东南大学系统工程研究所,江苏南京210096

出　　处：《系统工程学报》2004年第3期239-243,共5页Journal of Systems Engineering

摘　　要：在Nakakita等的确定型短期降雨预测模型的基础上,利用随机微分方程理论,给出了预测降雨强度的随机模型.通过随机分析理论,求出了随机模型的解,同时研究了求解的近似数值方法和利用Fourier变换转化该随机降雨模型方程为常微分方程组的可行性.最后把文章提出的随机波动模型和Nakakita提出的确定型短期预测模型运用到Nakakita1996年给出的降雨事件中进行预测和比较,效果良好.In this paper based on the model of Nakakita, etc. we advance a random model to predict short-term rainfall. In contrast to the original deterministic model, the new one using stochastic differential equations (SDEs) can describe the whole rainfall process better, especially under the situation with the uncertain, dynamic and variable properties. By theories of stochastic analysis, time factor has been introduced into the new model. After solving the SDEs, with Fourier transform, we also present an approximate algorithm and a method to transform this model's equations into ODEs(ordinary differential equations). At the end, the method is applied to a rainfall prediction example proposed by Nakakita in 1996. Results show that for the extended lead time of two and four hours, the prediction of the rainfall distributions is improved.

关 键 词：随机微分方程 傅里叶变换 降雨强度 

分 类 号：S42[农业科学—植物保护]