基于Stacking机器学习模型的南海北部海温预报

Sea temperature forecast in the northern South China Sea base on Stacking machine learning model

基于Stacking(ET-ET)的机器学习算法,利用美国国家环境预报中心再分析数据和MGDSST海温融合数据,建立了一套高效的海温长期预报方法,并在南海北部海域开展了1 a的表层海温长期预报实验。结果表明:基于Stacking(ET-ET)机器学习模型的表层海温长期预报的均方根误差降至0.52℃,平均绝对百分比误差降至1.58%,明显优于基于支持向量机、人工神经网络和长短期记忆模型的预报结果。

In this paper, an efficient long-term SST forecast method is established based on Stacking(ET-ET)machine learning algorithm using reanalysis data of National Centers for Environmental Prediction and Mergid satellite and in situ data Global Daily sea surface temperature(SST) fusion data, and long-term SST forecast experiment is carried out in the northern South China Sea for one year. The results show that the root mean square error of long-term SST forecast based on Stacking(ET-ET) machine learning model is reduced to 0.52 ℃, and the mean absolute percentage error is reduced to 1.58%, which is significantly better than the forecast results based on the support vector machine, artificial neural network and long short-term memory model.