Evaluation of the Forecast Skill of the North American Multi-Model Ensemble for Monthly and Seasonal Precipitation Forecasts over Central America

The North American Multi-Model Ensemble (NMME) precipitation forecast is one of the main tools to support policy-makers to make informed decisions in climate-related sectors across Central America. This study aims to quantify the precipitation forecast skill for both monthly and seasonal timescales using six NMME models and the multi-model ensemble mean. The CMORPH Climate Data Record satellite dataset is used to verify the models. In addition, we analyzed the teleconnections effects that El Niño Southern Oscillation (ENSO) might have in the rainfall representation on the NMME forecast. The evaluation is performed using both deterministic and probabilistic verification approaches. For the deterministic approach, we evaluated the accuracy of the ensemble mean using mean error, linear correlation, and root mean squared error maps. Meanwhile, for probabilistic evaluation, we measure the skill of the hindcasts using the Brier Score and the Ranked Probability Score. Evaluating the global forecast system provides a better understanding of predictability in the region, and it might help to improve future studies on sub-seasonal to seasonal forecasts that are essential for food security and water supplies.