基于信息扩散理论的干旱区农业旱灾风险分析——以甘肃省民勤县为例

 针对小区域历史灾情资料缺乏，获得的只是小样本数据，使用传统统计模型进行风险分析精度不高的特点，作者基于信息扩散原理，利用正态信息扩散技术将单一的样本观测值变为一模糊集的模糊数学方法，提出了农业旱灾风险定量分析模型，并以甘肃省民勤县为例，对干旱区农业旱灾风险进行了分析。计算结果表明，民勤的旱灾尽管年年发生，但农业受损程度并不大，受损程度主要集中在20%以内，受损程度在30%以上的可能性很小;但也不能忽视旱灾风险，受损10%以上旱灾发生概率在76%，结果符合实际。对信息扩散理论在现代风险分析技术中的应用提供了一个新途径。

Analysis on Agricultural Drought Risk based on Information Diffusing Principle-A Case Study of Minqin County in Arid Area

In general,the historical data recording for natural disasters in small region is not enough to be used for estimating the probability distribution in risk estimation,because the size of the observation sample is small.In this paper,the information diffusion method is used to change single sample observations into fuzzy sets,and a quantitatively analyzing model for agricultural drought risk is proposed.The model is applied in a case study in Minqin county of Gansu province.The results show that the drought frequency in study region is high,but the degree of agricultural loss is low.Most of the losses are below 20%,and the losses with degree more than 30% have little possibility.Even so we cannot ignore the drought risk,because the possibility of agriculture losses at more than 10% degree still reaches 76%.The simulating result is proved to be accordant with the natural phenomena.This research gives a new way of risk analysis using information diffusing principle.