摘要
依据陕西29个苹果基地县1971~2007年4月极端最低气温资料,将陕西果区苹果花期冻害年型划分为严重冻害（NTd≤-2℃≥10）,中等冻害（NTd≤-2℃〈10＆NTd≤0℃≥15）,轻度冻害（NTd≤-2℃〈10＆NTd≤0℃〈15＆NTd≤2℃≥19）,和正常年型（NTd≤-2℃〈10＆NTd≤0℃〈15＆NTd≤2℃〈19）四个状态。在逐年状态分型的基础上,运用马尔科夫链转移概率计算模式计算了各种状态下的转移概率矩阵,建立了陕西苹果花期冻害年型的马尔科夫链预测模型,进行了预测实例分析和回代检验,准确率达83%,并提出了使用中的注意事项。该模型预测效果较好,可为防灾减灾赢得时间,提供依据。
By using extreme temperature in April from 1971～2007 in 29 apple-producing counties,annual type division index of frost damage in apple blossom stage in Shaanxi was proposed： serious frost damage（NTd≤-2℃≥10）,moderate frost damage（NTd≤-2℃10 NTd≤0℃≥15）,light frost damage（NTd≤-2℃10 NTd≤0℃15 NTd≤2℃≥19）,and normal frost damage（NTd≤-2℃10 NTd≤0℃15 NTd≤2℃19）.We calculated transition probability matrix under various conditions by using Markov Chain transition probability calculations based on the state classification year by year,and then the Markov Chain Model for Prediction of frost damage in apple blossom stage in Shaanxi was established.We conducted the case study and return test of annual type prediction,and the results showed better effect,with accuracy rate of 83%,and what should be paid attention to in the use of the model was suggested.The model has better results of prediction,and it can gain time for the disaster prevention and mitigation and provide basis for it.
出处
《干旱地区农业研究》
CSCD
北大核心
2011年第2期272-275,280,共5页
Agricultural Research in the Arid Areas
基金
陕西省气象局2009年度研究型业务重点科研项目"政策性农业保险气象服务研究与开发"(2009Z-3)
关键词
苹果
花期冻害
年型
预测
马尔科夫链
apple
frost damage
annual type
prediction
Markov Chain
