关键词: 主成分分析, 主成分分析

Abstract: In order to examine the technical methods of flowering-year forecast of freeze injury of fruit trees, so as to improve high quality meteorological services for policymakers and the production line of our farmers. Based on the data of mean temperature, average minimum temperature and extreme minimum temperature in April from 1961 to 2010 in Guide County, Qinghai Province, the author analyzed the trends of mean temperature, average minimum temperature and extreme minimum temperature in April in the past 50 years. At the same time, by using the data of extreme minimum temperature in April, annual type division index of frost damage in pear florescence in Guide County was proposed: serious frost damage (Td≤-4.0℃), moderate frost damage (Td≤-2.0℃), light frost damage (Td≤0.0℃), and normal frost damage (Td≤2.0℃). On the basis of the yearly state classification, the author calculated transition probability matrix under various condition by using Markov Chain transition probability Model, then the Markov Chain Prediction Model for florescence frost damage in Guide County was established, and the prediction showed that serious frost damage would be a significant probability from 2011 to 2015. According to the test, the prediction of 2011 and 2013 was consistent with the actual situation.