Identification of the Effect of Geographical Coordinates on the Accuracy Prediction of Total Rainfall Using Multi-layer Perceptron Neural Network

Rainfall prediction is an interesting topic in the meteorology or hydrology field since it is directly related to agriculture, management of water resources in hydrologic basins, and water scarcity. The aim of this study is predict the total rainfall in Lebanon using monthly precipitation in Mediterranean coastal Multi-Layer Perceptron Neural Network (MLPNN) model. For this aim, three scenarios with various input variables are proposed. Scenario 1 and 2 are developed using geographical coordinates (latitude, longitude, and altitude) and climate conditions (average temperature, relative humidity, wind speed and solar radiation), respectively. Scenario 3 is developed by adding geographical coordinates to the climate condition data to see the impact of geographical coordinates on the accuracy of the prediction of total rainfall. The results indicated that the Scenario 3 has decreased the RMSE and MAE by 10%.and 13%, respectively. Consequently, the Scenario 3 can be recommended for modeling the complexity of interactions for rainfall-climate conditions-geographical coordinates and predicting rainfall.