Prediction of Monthly Rainfall in Plateau Area Based on Convolutional Neural Network

Jinfeng Xiao ； Shixing Han

《World Scientific Research Journal》 8卷7期 (2022/07) Pp. 349-353

https://doi.org/10.6911/WSRJ.202207_8(7).0049

摘要

More accurate rainfall prediction plays a vital role in agricultural development and water resources management. Because the influencing factors of rainfall are many and complex, nonlinear and unstable, this paper proposes a convolution neural network (CNN) monthly rainfall prediction method based on deep learning after the traditional time series analysis method. Taking the rainfall data of Lhasa from 2009 to 2021 as the research object, taking the rainfall of five months as the step, the rainfall of the next month is predicted, and the optimal method is tried to be found by adjusting the parameters for many times. In order to test whether the prediction results have a certain improvement in accuracy, the prediction results are compared by establishing differential integrated moving average autoregressive model (ARIMA). Compared with the regression model, the root mean square (RMSE) error and absolute error (MAE) of the convolution neural network rainfall prediction model are reduced by 1.32 and 1.61 respectively. The results show that the convolution neural network rainfall prediction model has relatively high rainfall prediction accuracy in the plateau area, can extract more rainfall variation characteristics, and has advantages in convergence.

關鍵字

Convolutional neural network ； Autoregressive ； Rainfall ； Prediction

參考文獻

Ebtehaj I, Bonakdari H, Zeynoddin M, et al. Evaluation of preprocessing techniques for improving the accuracy of stochastic rainfall forecast models[J]. International Journal of Environmental Science and Technology, 2020, 17(1): 505-524.

Chen L, Li Y, Cheng Z. An overview of research and forecasting on rainfall associated with landfalling tropical cyclones[J]. Advances in Atmospheric Sciences, 2010, 27(5): 967-976.

Sokol Z, Szturc J, Orellana-Alvear J, et al. The role of weather radar in rainfall estimation and its application in meteorological and hydrological modelling—A review[J]. Remote Sensing, 2021, 13(3): 351.

Sun Xiaoting, Ren ganghong, Du Kun, Feng Yan, Zhou Ming Monthly rainfall prediction based on grey correlation method [J] Journal of irrigation and drainage,2019,38(01):90-95.

Ji Yali, Jia Keli, Li Changyou, et al Application of LS-SVM and RBF neural network model in rainfall prediction [J] Water science and Engineering Technology, 2012 (2): 1-4.