中文摘要：

利用1959—2008年云南省125个气象站的逐日资料,采用模糊信息分配方法和三分段法进行了雨、雾淞（冰冻）天气发生的气象条件研究;基于模糊信息分配和超越极限概率法进行了云南冰冻灾害气候风险评估和风险区划研究。结果表明：模糊信息分配方法用于概率密度估计,可减少分析、评估误差,且不需要假定数列总体概率分布函数,就能得出较为接近实际概率密度分布的分析和评估结果;雨、雾淞（冰冻）出现当日最低气温≤1℃的气候概率达99.8%,最低气温≤1℃是出现雨、雾淞（冰冻）天气的阈值;当温度条件达到时,微量以上降水、日照时数≤1 h、日平均相对湿度大于80%是出现雨、雾淞（冰冻）的阈值。云南省大部分地区（96个站,占全省总站数的77%）没有出现冰冻灾害的风险,无风险的区域分布在云南省的中、西部和南部地区;冰冻灾害风险较大的是滇东北和滇东地区,其中有12个站累积概率值在50%～100%之间,有年年出现或2年1遇的气候风险,风险较大,特别是镇雄、鲁甸2个县的概率值为100%,有年年出现冰冻灾害的气候风险,风险很大。

英文摘要：

With daily data of 125 meteorological stations in Yunnan Province during 1959-2008,meteorological conditions for the occurrence of rain and rime（frost） were studied using the fuzzy information distribution method and three-segments method.Risk assessment and zoning of frost disaster in Yunnan Province were studied using the fuzzy information distribution method and the exceedance limit probability method.Results show that using fuzzy information distribution method to estimate the probability density could reduce errors in analysis and evaluation.Also,analytical and evaluation results that are close to actual probability density distribution can be obtained without assuming the overall probability distribution function of series.The climatic probability of the low temperature≤1℃ is as high as 99.8% when rain or rime（frost） appears,and that the low temperature≤1℃ is the threshold for the rain and rime（frost） to appear.When the temperature condition is met,the threshold of rain and rime（frost）＇s appearance denotes the no-less-than-the-minimum precipitation,≤1h sunshine and daily average≥80% relative humidity.There is no risk of frost in most part of Yunnan Province（96 stations,77% of all）;the risk-free areas are located in the middle,west and south of Yunnan Province,whereas the risk of frost disaster is higher in northeast and east of Yunnan,among which the cumulative probability of 12 stations is between 50%~100%.The risk is higher in some areas where climatic risk appears once in one or two years,especially in Zhenxiong County and Ludian County,where the probability is 100%.The risk is the highest where climatic risk appears in every year.