Revisiting bare‐ground evaporation in a high‐resolution numerical weather prediction system

Abstract

As part of a project aimed at improving the Canadian High‐Resolution Deterministic Prediction System, the parameterization for bare‐ground evaporation is revisited and a new method, improving upon the soil‐resistance method used at the European Centre for Medium‐range Weather Forecasts (ECMWF), is proposed. Higher values of soil evaporation resistance are obtained, particularly in dry conditions. The proposed change, tested over the fall of 2021 and the spring of 2022, improves surface weather forecasts significantly in regions and seasons with large bare‐ground fractions. To confirm that the change also leads to better predictions of the latent heat flux, a point‐scale evaluation is performed at AmeriFlux station US‐Ne1 in Nebraska for three contrasted years in terms of precipitation and post‐harvest soil moisture (2015, 2016, and 2021). While significant improvements are observed, evaporation starts earlier in the day than predicted and evaporation following rainfall events is underestimated.