Impacts of Dropsonde Observations on the Predictability of Two Landfalling Atmospheric River Events in February 2016

Landfalling Atmopsheric Rivers (ARs) can bring large environmental and societal impacts. In particular, they can produce between 30-50% of the precipitation for the Western United States (U.S.) and are cause of the major flooding events. Accurate forecasts of a landfalling AR can improve water management decisions and reduce the flooding risks. Sparse observations over the central and eastern Pacific have limited the improvement of forecast skills for the Western U.S. due to the poor upstream initial conditions. While the numerical weather prediction reaps the benefits of satellite data over the oceans, those data are known to poorly represent the low-level circulation and the vertical structure of water vapor in ARs. For example, the landfall position error is on the order of +/- 400 km at 3 days lead time in NCEP GFS model. In February 2016, two C-130 aircraft carried out reconnaissance missions targeted AR condition between Hawaii and the U.S. West Coast to support improved 1-3 day prediction of AR impacts. Each aircraft released dropsondes that measured the detailed vertical profiles of wind, water vapor, temperature and pressure below the aircraft. In this study, the impact on the forecast accuracy of landfalling ARs by assimilating these dropsonde data is evaluated.

Two AR storms are chosen as case studies, utilizing a total of 182 dropsonde records taken during February 13th-16th 2016. Four experiments are conducted using the Weather Research and Forecasting (WRF) model with the Data Assimilation Research Testbed (DART) system, including CONTROL run with no data assimilated, the DROP run with the dropsonde data assimilated, REGULAR run with the conventional data (including satellite data) assimilated, and ALLDATA run with both conventional data and dropsonde data assimilated. Comparisons between DROP and CONTROL show that the 1-3 day in-land precipitation forecast error reduces by 5-25% for the two storms when assimilating dropsonde observations in the regional WRF model. With the dropsonde data assimilated, the WRF model can better simulate the small features of the integrated water vapor transport and the finer structures of the precipitation and its coverage area. The value of dropsonde observations in addition to the conventional observations is assessed by comparing the outputs from REGULAR and ALLDATA runs. This work will also investigate the impacts of employing different error-covariance matrix on the ensemble generations for AR events used for DART ensemble-based data assimilation process.