Climate Change Prediction on Middle East Using RNN Model and Hybrid LSTM&GRU Model

The issue of climate change is regarded as one of the most significant global concerns of our day due to its unparalleled impact on Earth's natural system. The main effects of climate change are increasing frequency of extreme weather events, fast windfalls and storms, altered rainfall patterns, and dramatic temperature variations. One of the areas most susceptible to climate change is the Middle East. In addition to other detrimental effects of climate change, it is subject to regular heat waves, a shortage of food and water, and other challenges. This area is warming at a rate that is almost twice as fast as the rest of the world, which puts it at danger of significant impacts and increases the risks to food and water security. Similarly, the Middle East's agricultural output is heavily dependent on rainfall since the region is extremely vulnerable to variations in rainfall and temperature brought on by climate change, which has a detrimental impact on the region's economy. Weather forecasts are crucial for providing early notice of how the weather will affect several facets of human life. The objective of meteorologists has always been to forecast meteorological conditions accurately and promptly. Research and applications based on deep learning in a range of domains, including computer vision, speech recognition, translation, and time series prediction, have made it clear how vital deep learning is and how to invest in and use it in many fields. Additionally, it performed incredibly well when it came to removing spatial and temporal aspects from the data. As a result, deep learning-based weather prediction (DLWP) will function well in addition to conventional techniques. Two deep learning models were applied in this work to forecast Middle Eastern climate change. Recurrent neural networks (RNNs) make up the first model, while a mix of gated recurrent units (GRU) and long- and short-term memory (LSTM) makes up the second. There were six climate parameters taken into consideration: humidity, wind speed, sun radiation, maximum and minimum temperatures, and rainfall rate. Nine Middle Eastern capitals were the subject of the investigation. Using statistics from 1981 to 2022, they are: Riyadh (Saudi Arabia), Ankara (Turkey), Baghdad (Iraq), Beirut (Lebanon), Cairo (Egypt), Doha (Qatar), Abu Dhabi (UAE), Sanaa (Yemen), Tehran (Iran). NASA's Modern Era-Retrospective analysis for Research and Applications (MERRA-2) provided the data that was used. To assess how well these deep learning algorithms performed, they were compared using statistical criteria such as mean absolute error (MAE), root mean square error (RMSE), and coefficient of determination (R2).