Rainfall Estimation From GMS Imagery DataUsing Neural Network

Rainfall estimation at the mesoscale level generally involves with a huge amount of data which demands a time consuming process, as well as, a big computing memory of the computer. This problem can be solved by the use of Neural Network. In this study, Neural Network is used for forecasting the rainfall by using remote-sensing image cloud data, which can express many characteristics of clouds such as topography, temperature, surrounding atmosphere etc. Neural Network can store data of image variation of clouds between the connections of the network. By utilizing these stored data, the rainfall fields are estimated. From the result of the investigation of cloud image data and rain intensity, it is clear that the data correspond to May to July are distributed above the 30% albedo line and below the -10°C line, while the data correspond to August to October are distributed above the 30% albedo line and below the -20°C line. Therefore the estimation can be conducted using these thresholds.