A Climatology of Forecast Sensitivity for Landfalling Atmospheric Rivers along the US West Coast

Atmospheric Rivers (AR) are the source of significant precipitation along the West Coast of the United States. These features form over the open Pacific Ocean and previous research has shown that ARs and their predecessor features are often characterized by a lack of observations, including with satellites. Consequently, these features can have significant initial condition uncertainty, which in turn can result in high subsequent forecast uncertainty once these features make landfall. For the past few winters, ensemble-based sensitivity has been utilized during AR Recon to identify regions where uncertainty in the model state has the largest impact on the subsequent AR and precipitation forecasts over CONUS, which in turn is used to direct observation assets. Each of these cases are evaluated independently; however, there has not been a climatological analysis of forecast sensitivity over many cases.

This study evaluates the climatology of forecast sensitivity by utilizing a large set of landfalling AR forecasts at multiple points along the US West Coast. This climatology is obtained by applying ensemble-based sensitivity to 2016-2024 ECMWF ensemble forecasts obtained from the TIGGE database and employing a uniform process for defining metrics for all cases. Preliminary results indicate that the highest forecast sensitivity for the full suite of cases is along the upper-tropospheric midlatitude waveguide, with the maximum shifting westward toward Asia with increasing forecast lead time. Near the surface, there is a region of higher sensitivity to integrated vapor transport and equivalent potential temperature that appears to originate north of Hawaii and translates east with time. Forecast sensitivities from individual cases will be clustered by weather regimes to identify whether the climatologically sensitive region varies by the synoptic setting.