粮食产量预测理论、方法与应用Ⅳ.粮食估产理论、模型及其应用

摘要

粮食估产的“通道-概率”理论：把属于最近通道的历年来的产量划分为5个气候年型通道，即丰产年、偏丰年、平产年、偏欠年、欠产年；计算产量出现在5个气候年型中的频率作为概率使用，估产年的初始估产值等于预测年各通道内平均产量与概率之积的和；估产值等于初始估产值与气候年型修正参数之积，专家根据当年气候条件和作物长势实时确定修正参数。预报单元为全国、省和县。应用结果表明：国家尺度上不需要修正，省和县级尺度需要气候年型参数修正；预测误差在3％以内；所述估产理论严谨、方法简单，参数少，参数来自原始数据本身和专家经验，易于推广使用。%The channel probability theory’of grain yield estimation is defined as follows:the historical production which belongs to recent channel is divided into 5 climatic-year types, i.e. high yield year, partial harvest year, normal year, partial less year, a bad crop year. The frequency(or probability)for the year appeared in 5 climatic types is calculated. The initial value of estimation yield is equal to the sum of the product of probability and the average yield within every channel for the forecasting year, and the value of estimation yield is equal to product of the initial estimation value and the climatic parameters. The parameters are real-time amended by experts according to the climat-ic conditions and crop growth. Prediction units are nation, province and county. The results of application showed that the parameters did not need to be modified on national scale, while the climatic parameters needed to be revised on province and county scales. The prediction error was within 3%. The theory would have good application effect and promising development due to its preciseness, simple method, and less pa-rameters which were from original data and expert experience.