Evaluating the Impacts of Assimilating Airborne Radio Occultation (ARO) Observations on Precipitation Forecasts during an 11-Flight Sequence of ARs

Airborne Global Navigation Satellite System (GNSS) radio occultation (ARO) observations are collected during Atmospheric River Reconnaissance flights, increasing the sampling of AR’s during the flights beyond that of traditional dropsondes. In past studies, the assimilation of ARO refractivity profiles has been investigated, but the assimilation of ARO bending angle profiles is yet to be explored. In this study, we investigate the impacts of assimilating ARO bending angle observations on precipitation forecasts over the northeast Pacific region against estimates of rainfall from satellite data and the U.S. STAGE-IV precipitation product. For this study we use the Model for Prediction Across Scales – Atmosphere (MPAS-A) coupled with the Joint Effort for Data assimilation Integration (JEDI). For the ARO bending angles simulation, we use the modified two-dimensional bending angle observations operator of the EUMETSAT ROMSAF Radio Occultation Processing Package (ROPP) interfaced with the JEDI-Unified Forward Operator (UFO). This study focuses on the 2022-2023 AR Recon campaign, in which an unprecedented sequence of ARO data collection was possible between 6 to 17 January 2023 with the National Oceanic and Atmospheric Administration (NOAA) G-IV aircraft. Two numerical experiments are performed assimilating all available global conventional data for this time period along with Atmospheric Motion Vector (AMV) winds, spaceborne GNSSRO bending angles and clear-sky radiances from AMSUA and MHS with variational bias correction. In addition, ARO bending angle observations are assimilated for comparisons. Preliminary results indicate that modifications in the humidity and temperature analyses bring substantial improvements concentrated on day-1 and day-2 forecasts of temperature, water vapor mixing ratio and zonal and meridional wind fields when compared against the GFS analyses. These improvements are also observed in the analysis of the precipitation forecast, suggesting that the 24- and 48-h forecasts provide a better representation of the precipitation when compared to STAGE-IV, especially for higher thresholds. These results give indications that by using targeted observations and sampling important features of the ARs 48 hours ahead of their landfall in the west coast of the US, we can reduce the uncertainties in the initial conditions and consequently improve the precipitation forecasts.