Assimilation of GPS Radio Occultation Data with WRF/DART

The atmospheric limb sounding technique making use of radio signals transmitted by the Global Positioning System (GPS) satellites has emerged as a powerful and relatively inexpensive approach for sounding the global atmosphere in all weather. Comparison of GPS radio occultation (RO) data with independent observations has shown that the GPS RO has the best data quality above the moist part of the atmosphere (approximately 8 km in tropics) to about 25 km. The assimilation of RO data in the upper troposphere and lower stratosphere has been shown to improve global analysis and prediction. Because GPS RO data are unbiased in these regions, the RO data have been used as anchors to improve the assimilation of satellite radiance data. However, the benefit of RO data assimilation in the tropical lower troposphere has not been fully demonstrated, as both the numerical models and observations are subject to considerable uncertainties. In this paper, we perform cycling assimilation of GPS RO soundings using the WRF model and its ensemble Kalman filter data assimilation (DART – Data Assimilation Research Testbed). We examine the sensitivity of bending angle data assimilation to observation error specification and data quality control procedures. We also compare the performance of bending angle and refractivity observation operators over the tropical lower troposphere, with a goal to optimize the strategy for the assimilation of GPS RO data for numerical weather prediction.