The Prediction Skills of Atmospheric Rivers over the Western United States in three Subseasonal Hindcast Models

The atmospheric rivers (ARs) are recognized as long and narrow plumes with intense water vapor transport. ARs are important water resources over the western U.S., but ARs could also lead to damages such as flooding events. The assessment of prediction skills under the subseasonal time scale bridges the need between research and operation for the purpose of water management and damage control. In this study, the prediction skills are examined at the lead time of one to four weeks over the western U.S. during the boreal winter in three operational hindcast models: National Centers for Environmental Prediction (NCEP), European Center for Medium-Range Weather Forecasts (ECMWF), and Environment and Climate Change Canada (ECCC). The Modern-Era Retrospective Analysis for Research and Applications version 2 (MERRA-2) reanalysis dataset is selected as the reference to the three hindcast models. Compared to the MERRA-2 dataset, three hindcast models all show higher AR frequency at lag zero day over the western U.S. The mean of the ensemble members that detected ARs shows stronger integrated water vapor transport (IVT) and integrated water vapor (IWV) within the AR time steps over the western U.S. Both deterministic and probabilistic skill metrics will be calculated to quantify the prediction skill of IVT, IWV, and the wind and identify any forecast windows that have higher prediction skills.