Convection embedded in an atmospheric river: exploring precipitation sensitivity to convective parameterizations

Atmospheric rivers (ARs) are major drivers of orographic precipitation along themountainous west coasts and often involve embedded convection delivering extreme localbursts of rainfall associated with flash floods. Understanding the convection within ARs iscritical for assessing flood risks and improving precipitation forecasts and numerical weathermodels. This study evaluates the sensitivity of the 2019 Valentine’s Day Atmospheric Riverprecipitation in California to Grell-Freitas and Tiedtke convective schemes using the WeatherResearch and Forecasting (WRF) model. Ensemble simulations at a 9 km horizontalresolution- a scale within the gray zone where some convection is resolved butparametrization remains necessary - are compared with seven observational datasets,including ground and satellite-based products. Over the ocean, WRF successfully captureskey synoptic features and the overall distribution of total precipitation due to the dominanceof stratiform precipitation. Convective precipitation, however, is simulated differently: Grell Freitas produces it within the AR, whereas Tiedtke simulates it west of the AR. Given thechallenges in diagnosing AR-embedded convection at 9 km, an alternative method based onoutgoing longwave radiation and precipitation rate thresholds is tested. The alternativeconvective precipitation aligns with observations, occurring within the AR over the ocean,coastal mountains, and Sierra Nevada. Using the alternative convective method, the overallpattern of convective precipitation is generally consistent between the Grell-Freitas andTiedtke schemes. This agreement contrasts with the patterns of parameterized convectiveprecipitation. These results are relevant for global models or regional large ensemblesapproaching gray zone resolution, guiding diagnostics of embedded convection in ARs.