Assimilation of GPS Radio Occultation Refractivity Data from CHAMP and SAC-C Missions over High Southern Latitudes with MM5 4DVAR

In this study, the GPS radio occultation (RO) data from the Challenging Minisatellite Payload (CHAMP) and Satellite de Aplicaciones Cientificas-C (SAC-C) missions are assimilated. An updated version of the fifth-generation Pennsylvania State University–National Center for Atmospheric Research Mesoscale Model (MM5) four-dimensional variational data assimilation system (4DVAR) is used to assess the impact of the GPS RO data on analyses and short-range forecasts over the Antarctic. The study was performed during the period of intense cyclonic activity in the Ross Sea, 9–19 December 2001. On average 66 GPS RO soundings were assimilated daily. For the assimilation over a single 12-h period, the impact of GPS RO data was only marginally positive or near neutral, and it varied markedly from one 12-h period to another. The large case-to-case variation was attributed to the low number of GPS RO soundings and a strong dependency of forecast impact on the location of the soundings relative to the rapidly developing cyclone. Despite the moderate general impact, noticeable reduction of temperature error in the upper troposphere and lower stratosphere was found, which demonstrates the value of GPS RO data in better characterizing the tropopause. Significant error reduction was also noted in geopotential height and wind fields in the stratosphere. Those improvements indicate that early detection of the upper-level precursors for storm development is a potential benefit of GPS RO data. When the assimilation period was extended to 48 h, a considerable positive impact of GPS RO data was found. All parameters that were investigated (i.e., temperature, pressure, and specific humidity) showed the positive impact throughout the entire model atmosphere for forecasts extending up to 5 days. The impact increased in proportion to the length of the assimilation period. Although the differences in the analyses as a result of GPS RO assimilation were relatively small initially, the subtle change and subsequent nonlinear growth led to noticeable forecast improvements at longer ranges. Consequently, the positive impact of GPS RO data was more evident in longer-range (e.g., greater than 2 days) forecasts. A correlation coefficient is introduced to quantify the linear relationship between the analysis errors without GPS RO assimilation and the analysis increments induced by GPS RO assimilation. This measure shows that the growth of GPS RO–induced modifications over time is related to the prominent error reduction observed in GPS RO experiments. The measure may also be useful for understanding how cycling analysis accumulates the positive impact of GPS RO data for an extended period of assimilation.