Understanding the impact of model resolution on Atmospheric River representation over western North America

Abstract

Atmospheric rivers (ARs) are essential to many communities, providing a seasonal source of precipitation, however, ARs are also hazardous through flooding and landslides. The double-sided nature of ARs demands that they are well understood and are accurately modeled as we project their behavior under warming climate conditions. Projections of ARs are typically performed at low (2° or 1°) resolution, however, it remains unclear whether ARs can be resolved at such resolutions. To determine how model resolution influences the representation of ARs we analyze Preindustrial output from CESM 1.2 (Community Earth System Model) with spatial resolutions from 2° to 0.25°. We then focus on the western coast of North America as ARs are common there during winter months. We compare the output to ERA5 observation data to validate the accuracy of 2° and 0.25° simulations. Our study finds that higher resolutions (0.25°) represent ARs more accurately than 2° due to improved representation of precipitation systems over land when compared to observation data (ERA5). Looking at the IVT (integrated water vapor transport, a defining factor of ARs) anomaly between 2° and 0.25° also displays an increase around the west coast of North America with 0.25°. This increase is supported by the ERA5 data. Ultimately, this work enables future research to better understand and represent ARs and their impacts on western North America.