基于随机森林的内江市降水量中长期预测

Medium-to Long-Term Forecast of Precipitation of Neijiang City Based on Random Forest

科学地对区域降水进行中长期预测对地方工农业生产调整及其转型发展具有指导意义.本文采用随机森林对1971~2013年内江市降水量进行建模分析,并预测未来10年内江市降水情况,得出以下结论:(1)随机森林模型预测精度高、稳定性好;(2)研究发现,内江市未来十年年际降水量变化较小,2014年和2019年降水相对较多,而2018年降水较少,其中2020年夏季降水偏少,局部地区可能会遭遇干旱,应该做好相应的防汛抗旱措施,完善防灾抗灾机制,确保工农业生产正常进行.

Scientific medium-to long-term forecast of regional precipitation has a guiding significance in promoting the adjustment and transformation of local industrial and agricultural production. This thesis applies random forest to make modeling analysis for precipitation of Neijiang City from 1971 to 2013 and forecast precipitation of the city in the next 10 years. The following conclusions are made:( 1) Model forecast of random forest has higher accuracy and better stability;( 2) According to relevant research,inter-annual precipitation of Neijiang City in the next 10 years will have smaller changes,precipitation is relatively more in 2014 and 2019,while there will be less precipitation in 2018. In the summer of 2020,there will be relatively less precipitation,thus some areas may suffer drought. Therefore,relevant departments should take corresponding flood-control and drought-resistant measures and perfect the mechanism of controlling and resisting disasters so as to ensure normal industrial and agricultural production.