中文摘要：

采用关中地区39 a的月平均降水量数据,计算了该地区不同时间尺度的标准化降水指数（SPI）值,运用ARIMA模型对SPI序列进行分析建模,并进行12步预测。结果表明,ARIMA模型对所有时间尺度的SPI 3、6、9序列可进行精度在10%以下的1步预测,对SPI 12、24序列可进行平均精度在10%以下的9步预测,说明ARI-MA模型较适合SPI 3、6、9序列的短期预测,适合SPI 12、24序列的长期预测。

英文摘要：

The time series of SPI at different temporal scales are calculated from monthly precipitation data, which was collected from 36 weather stations in Guanzhong Plain and Weibei Tablelands. ARIMA models are developed to forecast and simulate SPI series. The results show that ARIMA models can predict SPI 3,6,9 series with 1-month lead-time at the precision of under 10 %, and can predict the SPI 12,24 series with 9-month lead-time at the average precision of under 10 %. Forecasting precision increases with the increase of temporal scales.