The estimation of rainfall and precipitation variation during 2011 convective system using an artificial neural network over Tawau, Sabah

Abstract

For succeed the early warning system development programme for space activity plan at Tawau, Sabah Malaysia, we studied the variation of rainfall and precipitation over convective system activity. We use five variables data such as the surface meteorological data (Pressure, Temperature and Relative Humidity), rainfall data, and precipitation. The surface meteorological data are taken from weather underground website, the rainfall data we obtain from rain gauge sensor Department of Irrigation and Drainage (DID) Malaysia, and also the precipitation data was obtained from NASA Satellite (Tropical Rainfall Measuring Mission). In this study, we attempt to estimate two variable targets (rainfall and precipitation) using Artificial Neural Network (ANN). The estimation was constructed as an alternative method to develop rainfall and precipitation model using surface meteorological data during convective system activity on summer monsoon over Tawau area. In this study, we processed surface meteorological, rainfall and precipitation data from 1 June 2011 to 31 August 2011. The Multilayer Perceptron (MLP) architecture and Levenberg Marquardt (LM) algorithms have been deployed in this study. We obtain the good result of estimation with one output target. The Root Mean Square Error (RMSE) value of rainfall and precipitation were 4.3540e-004 and 2.7834e-004 respectively and also Variance Accounted For (VAF) of rainfall and precipitation were 94.344% and 98.923% respectively during the training process. Testing results showed an error of 5.656% and 1.077% for precipitation and rainfall, respectively. However, we found the weakness of estimation model using three parameter inputs and two parameter output targets with the same structure. It means that for future work we suggest using 3 years data over convective system activity during summer monsoon is considered in the output target (rainfall and precipitation).