Prediction of hatay meteorology radar images with artificial neural network using python programming language

In association with global warming, unusual weather events, heavy rains and floods occur in many parts of the world. When the global climate crisis is evaluated together with urbanization and rapid population growth, it is seen that making consistent and timely weather forecasts, which have the quality of early warning, have vital importance. In this thesis, meteorology radar data, which is a remote sensing device used frequently in early warnings and short-term strong precipitation forecasts, has been processed and used. PPI images taken from Hatay meteorological radar are trained with codes written in Python programming language with ConvLSTM model, an Artificial Neural Network (ANN) algorithm. With the developed model, the last 6 radar images were trained and the next image, which has not yet been recorded, was estimated. The image time interval selected for the dataset is 30 minutes. Model training accuracy is in the range of 80-84%. The similarity ratios of the model prediction image and the actual image were measured by Root Mean Square Error (RMSE), Peak Signal to Noise Ratio (PSNR), and Structural Similarity Index (SSIM). According to the results obtained from the applied model, the predictive values with the best performance were calculated as RMSE 0.000013, PSNR 97.795 and SSIM 0.998. With the ANN model used, it has been shown that sequential radar images can be trained with various time intervals and the image prediction that will occur after the specified time interval can be made. In order to increase model accuracy, the data set should be increased and diversified Key Words: Python, ConvLSTM, Remote Sensing, Meteorology, Data Science