早稻产量动态预报模型

利用早稻不同生育阶段气象因子的综合聚类指标选择气象相似年型，再根据相似年的产量变化确定分析年的产量气象影响指数，建立全国和区域早稻单产动态预报模型。通过对1995—2004年早稻产量进行动态预报，结果表明，4月30日、5月31日、6月30日和7月31日预报的早稻产量增减趋势的预报正确率均为100％，产量预报准确率为97．5％、96．6％、98．1％和96．9％，从而可实现早稻产量的连续、动态、定量化预报，促进有关业务服务的发展。

A Study on Dynamic Prediction Model of Early Rice Yield per Unit

Based on the synthesized cluster indicators of meteorological factors at different growing periods of early rice, the similar meteorological years were chosen to calculate meteorological impact index of yield for the prediction, and a dynamic prediction model of the early rice yield per unit was established. The early rice yield per unit of ten years was dynamically predicted from 1995 to 2004. The results showed that all of the predications of increasing or decreasing tendency were highly valid and the average accuracy of ten years was with 97.5%, 96.6%, 98.1% and 96.9% for April 30, May 31, June 30 and July 31 respectively. The dynamic prediction model realized a successive dynamic quantitative prediction of early rice yields.