Modeling Precipitation Forecasting with Deep Learning

Accurate precipitation forecasting can capture climate change and may provide information to predict flooding and other environmental changes in a timely manner. Precipitation is one the key factors that is used to describe climate change and extreme weather events. Precise forecasting of precipitation is required by policymakers to develop strategies for extreme weather events such as heatwaves, heavy rainfalls, droughts, wildfires, and flooding. The potential impact of high precipitation and temperature in the Malaysian context could be heavy rainfall, increased flooding, more drought, and rising sea levels. Accurate simulations of these factors are important to develop precautionary measures to mitigate the risk of extreme events and adapt to climate change. In this work, we use the Long Short-Term Memory (LSTM) networks to simulate precipitation for 14 metrological stations in Malaysia. The LSTM has widely been used for time-series data analysis due to their ability to capture long-term dependencies and make accurate predictions. We use Root Mean Squared Error (RMSE), Mean Absolute Error, and Nash-Sutcliffe Efficiency ratio to estimate the model’s performance. The results show that the model performs better for the shorter day-ahead prediction.