Machine Learning for Climate Precipitation Prediction Modeling over South America

Many natural disasters in South America are linked to meteorological phenomena. Therefore, forecasting and monitoring climatic events fundamental issues for society various sectors of the economy. In last decades, machine learning models have been developed tackle different society, but there is still a gap applications applied physics. Here, evaluated precipitation prediction over America. Currently, numerical weather unable precisely reproduce patterns due many factors such as lack region-specific parametrizations data availability. The results compared general circulation atmospheric model currently used operationally National Institute Space Research (INPE: Instituto Nacional de Pesquisas Espaciais), Brazil. Machine able produce predictions with errors under 2 mm most continent comparison satellite-observed climate seasons, also outperform INPE’s some regions (e.g., reduction from 8 central winter). Another advantage computational performance models, running faster much lower computer resources than based on differential equations operational centers. it important consider forecasts centers way improve forecast quality reduce computation costs.