青海牧区雪灾月尺度精细化直接预测方法研究

Study on Monthly-scale Fine Direct Prediction Method of Snow Disaster in Qinghai Pastoral Areas

通过统计1961～2008年青海省牧区出现的雪灾情况,将出现雪灾的地区分为6个区,对应每个区域,将前期的降水、气温、降水日数以及同期的降水、气温作为因子,在保证信度≥0.1的基础上,确定相关系数较大的因子,运用线性回归方程建立预测模型,对青海省牧区分区域、分月份的积雪日数和积雪量进行预测,并依据"青海省雪灾标准(DB63/T372-2001)"进行相应级别的月尺度雪灾预警。通过检验,模型预测效果比较理想,是一种值得研究并不断完善的精细化直接预测方法。

By the statistics of the snow disasters occurred in pastoral areas of Qinghai Province during 1961-2008,regions where snow disasters occurred are divided into six sub-regions,for each sub-region,precipitation,air temperature and rainfall days during prophase period and precipitation,air temperature during current period as factors,based on ensuring the reliability greater than 0.1,factors of bigger correlation coefficient are determined.Using linear regression equation prediction model,space-partitioned and monthly-variational snow days as well as snow quantity in pastoral areas of Qinghai Province are predicted.According to ＂snowstorm standard（DB63/T372-2001）in Qinghai＂,early warming can be made of corresponding levels of the monthly scale snowstorm.By testing,the model has a relatively ideal prediction effect,and it is a fine direct method which is worthy of study and continuous improvement.