Application of Ensemble Sensitivity during the AR-Recon 2019 Experiment

Atmospheric Rivers (AR) are the source of significant precipitation along the West Coast of the United States. These features originate over the open Pacific Ocean, meaning that these features are not captured by many in situ observations. As a consequence, modeling systems could be characterized by significant initial condition uncertainty associated with these features, which in turn could translate into higher forecast variability. Despite this, there have been relatively few studies that have quantitatively documented how uncertainty in specific aspects of the AR (e.g., wind, moisture content) and surrounding synoptic features (e.g., position of upper troughs) could impact the subsequent precipitation forecast. This study evaluates the sensitivity of California precipitation forecasts for cases during the AR-Recon experiment during February 2019. Here, sensitivity is computed using the ensemble-based sensitivity method applied to real-time ECMWF ensemble forecasts. This methodology includes a dynamic forecast metric definition scheme that highlights regions of large precipitation variability. Preliminary results indicate that the precipitation forecast sensitivity is primarily associated with the position of shortwave troughs that moved over the top of the prominent Eastern Pacific ridge during this period. The evolution of these features subsequently modulated various aspects of the AR and in turn impacted where the precipitation was maximized.