为加强防灾减灾的预评估，建立了气候异常直接经济损失统计预测模型。模型由时间序列趋势项T（t）、气候异常影响项P（t）和随机波动项S（t）构成，在用统计资料进行模型参数估计的基础上，模拟预测了1989～1999年直接经济损失，模型整体拟合效果良好（R2=0.9847,F=150.45）.分别选取不同的SAAO/GDPPI比值，用模型对2000年气候异常直接经济损失进行预测，预测值均比1999年有所增加。模型预测结果显示，增加农业和减灾投入强度SAAO/GDPPI,可以大幅度减少气候异常直接经济损失。

In order to enhance the pre－ assessment of disaster reduction, a statistic predicting model is developed for simulating direct economic losses caused by climate abnormality. The model consists of time sequential term T(t), climate abnormality impact term P (t) and random variation term S (t). After parameter adjustment with statistic data, the model is run for the economic simulation losses from 1989 to 1999. The simulated results are quite consistent with the observed ones (R2: 0. 9847, F: 150.45). The economic losses in 2000 are predicted by. using the model with different SAAO/ GDPPI. In terms of the prediction results. the economic losses in 2000 will be higher than the ones in 1999. If the investment ratio of agriculture and disaster reduction, SAAO/ GDPPI, increases, the economic losses caused by climate abnormality would decrease by a wide. margin.