The initiative aims to develop an operational climate forecasting system on a monthly scale for Vale’s strategic areas. The company’s logistics, transportation, and production operations are directly impacted by weather conditions, especially in regions such as the Northern Corridor.
The project investigates statistical approaches, with an emphasis on machine learning techniques, which have shown promise in various meteorological applications. These techniques build predictive models from large volumes of historical data and identify patterns that can be used to forecast the future behavior of climate variables. The integration of dynamic and statistical models will also be explored as a hybrid alternative.