Multi-Model Ensemble Prediction of Summer Precipitation in China Based on Machine Learning Algorithms

The development of machine learning (ML) provides new means and methods for accurate climate analysis prediction. This study focuses on summer precipitation prediction using ML algorithms. Based BCC CSM1.1, ECMWF SEAS5, NCEP CFSv2, JMA CPS2 model data, we conducted a multi-model ensemble (MME) experiment three tree-based algorithms: the decision tree (DT), random forest (RF), adaptive boosting (AB) On this basis, explored applicability algorithms seasonal in China, as well impact different hyperparameters accuracy. Then, MME predictions based optimal were constructed regions China. results showed that all had an maximum depth less than 2, which that, current amount could only predict positive or negative anomalies, extreme was hard to predict. importance each ML-based quantitatively evaluated. CFSv2 higher eastern part Finally, China predicted tested from 2019 2021. According results, method provided more main rainband mean ACC 0.3, improvement 0.09 over weighted average 0.21 2019–2021, exhibiting significant other methods. shows have great potential improving short-term