Numerical Weather Forecast Biases by Region and Climate Zones

Understanding the sets of weather conditions and regions where numerical weather forecast models are typically more or less reliable is helpful to forecast users and model developers. In this study, we examine two operational weather forecast models: the US Navy’s Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPS) and NOAA’s Global Forecasting System (GFS). We match forecasts with hourly weather station data to evaluate predictions at a 48 hour lead time over regions including North America, Europe, and Asia. We examine both the winter and summer seasons, focusing on conditions at 7 AM and 3 PM local time which represent estimated daily minimum and maximum temperatures in the diurnal cycle. Model biases will be evaluated regionally and by Koeppen Geiger climate zones to determine to what degree biases are consistent across similar environmental conditions across the globe.