ARIMA模型在新疆喀什地区中长期降水量预测中的应用研究

为研究新疆喀什地区降水量中长期预测问题,结合ARIMA模型对新疆喀什地区的降水量进行预测,利用喀什地区两个子流域1950-2015年实测年降水数据分析其模型的适用性和预测精度。结果表明:ARIMA模型可较好的模拟喀什地区的年降水量,具有较好的预测精度,模拟降水量和实测降水量之间的误差相对值低于20%,绝对误差也可控制在15 mm以内;从空间角度分析,降水总体分布从东向西逐步递减,预测结果和实际降水空间分布状况较为吻合,可达到乙级精度标准。研究成果对于喀什地区中降水量长期预测具有较好的参考价值。

Application of ARIMA model in prediction of medium and long-term precipitation in Kashi area,Xinjiang

In order to study the medium and long-term prediction of precipitation in Kashgar, Xinjiang, the ARIMA model was used to predict the precipitation in Kashgar, Xinjiang. The applicability and prediction accuracy of the model were analyzed using the measured annual precipitation data from 1950 to 2015 in the two sub-basins of Kashgar. The results show that the ARIMA model can better simulate the annual precipitation in Kashgar region and has better prediction accuracy. The relative error between simulated precipitation and measured precipitation is less than 20%, and the absolute error can be controlled within 15 mm. From the perspective of space, the overall distribution of precipitation gradually decreases from east to west, and the predicted results are in good agreement with the actual spatial distribution of precipitation, which can reach the B-level accuracy standard. The research results have a good reference value for the long-term prediction of precipitation in Kashgar.