Spatio-Temporal Deep Neural Modelling for Climate Anomaly Detection using CNN-LSTM Networks

When it comes to studying environmental problems, it is growing increasingly vital to discover climatic anomalies and measure temperature changes, particularly in areas like Ethiopia, where the impacts of climate change are more damaging. In this research work, a composite CNN-LSTM algorithm was developed based on deep learning to identify and interpret the temperature fluctuation in Ethiopia. This investigation evaluated the 26 years of average temperature from 1995 to 2020. CNN layers were employed in this study to analyze transitory swings, whilst LSTM was used to assess climatic trends and the requisite tendencies. Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE) are three typical error metrics used to quantify the performance of the models. Thus, the model could recognize the existence of temperature anomalies and discovered a huge regional difference in temperature patterns. Combining publicly accessible data with a sort of deep architecture, such as CNN-LSTM, as shown in the present study, might also yield surprising insights into the dynamics of local climatic behavior. Conducting successful environmental research and monitoring is also related to preparation, which is the topic of attention.