A Monthly Rainfall Forecasting from Sea Surface Temperature Spatial Pattern

Abstract

The ocean surface temperatures or sea surface temperatures have a significant influence on local and global weather. The change in sea surface temperatures will lead to the change in rainfall patterns. In this paper, the long-term rainfall forecasting is developed for planning and decision making in water resource management. The similarity of sea surface temperature images pattern that was applied to analyze and develop the monthly rainfall forecasting model will be proposed. In this work, the convolutional neural network and autoencoder techniques are applied to retrieve the similar sea surface temperature images in database store. The accuracy values of the monthly rainfall forecasting model which is the long-term forecasting were evaluated as well. The average value of the model accuracies was around 82.514%.