摘要
传统的农业气象长期预报方法在选择预报因子时,往往忽略了因子间的独立性,降低了预报的实际效果。本文尝试应用协击方法对广州荔枝寒害进行长期预报。首先对1962-2005年的致灾因子与前期大气环流指数进行相关分析,初步入选预报因子;然后利用协击方法对入选因子进行独立性检验,并筛选独立性好的因子建立预报方程;最后对预报方程结果进行综合决策。通过对历史数据的回代检验,所建模型的模拟准确率达到86.4%。运用该预报模型对2007-2008年广州荔枝寒害进行试预报,结果与实际相符,表明应用此方法进行农业气象长期预报是有效的。
The independence of prediction factors is subjected to be ignored when the prediction factors are selected using a traditional long-term agrometeorological prediction method, thus reducing actual prediction effect. The purpose of this paper is to develop a Litchi cold damage prediction model for Guangzhou region using a different method named Xieji. Firstly, based on analyzing correlations between hazard factors and prophase atmospheric circu- lation factors for a period of 1962 -2005, preliminary predictors were selected, then independence of the selected factors were tested by Xieji method, and finally prediction equations using prediction factors with a good independence were established. The model was tested using historical data, and ac- curacy of model simulation was above 86. 4%. The model was then applied to Guangzhou for 2007 -2008 to predict Litchi chilling damage. The results of this application were close to real values, and it indicates that this method was valid for a long-term agrometeorological prediction.
出处
《中国农业气象》
CSCD
2008年第4期502-506,共5页
Chinese Journal of Agrometeorology
基金
国家科技支撑项目(2006BAD04B03)
国家科技支撑项目“东北平原中部(吉林)春玉米丰产高效技术集成研究与示范--玉米超高产环境条件研究”(2006BAD02A10)
关键词
农业气象长期预报
预报因子
独立性
协击
Long-term agrometeorological forecast
Predictor
Independence
Xieji
