Modeling spatio-temporal distribution of soil moisture by deep learning-based cellular automata model

A novel macroscopic cellular automata (MCA) model is used by combining DBN and MLP to predict the SMC over an irrigated corn field in the Zhangye oasis, Northwest China and shows that the DBN-MCA model performs better than the MLP-M CA model, and provides a powerful tool for predicting SMC in highly non-linear forms.