Intercomparison of Global Positioning System-radio occultation and radiosonde vertical profiles of temperature and moisture over three mesoscale regions

To date, considerable promise has been seen in the limited amount of Global Positioning System-radio occultation (GPS-RO) data available for ingestion into numerical models, though many of the studies to date have only examined the suitability of this data for specific applications. As private industry seeks to expand the number of GPS-RO space-borne platforms, the greater availability of data suggests greater potential application of this data to shorter term and routine weather prediction/forecasting applications on both meso- and synoptic- scales. While these scales have been the focus of many observational system improvements over the past two decades, there are still data sparse areas across the globe that stand to benefit substantially (in terms of data coverage) from an increase in GPS-RO derived atmospheric profiles. The advent of greater GPS-RO sounding coverage, particularly in the free troposphere and lower stratosphere, implies that this issue needs to be revisited in a systematic, robust fashion.

According to anecdotal evidence as well as some published literature, the primary data-sparse regions that most often impact NOAA/National Weather Service forecasts (due to the impact the lack of data has on data assimilation) include the Eastern North Pacific Ocean (ENP), and what we define as the North Central Continental Zone (NCCZ; the states of ND, SD, MT along with the Canadian prairie provinces). Poor initialization in the Eastern North Pacific region can lead to poor forecasts nationwide on various time scales, while poor short- to medium-range forecasts for the eastern half of the country have been attributed to poor model initialization in the NCCZ. These issues motivate our study, which considers the impact of an expanded array of GPS-RO soundings on numerical forecasts, with a focus on these data-sparse regions.

In this paper, we present preliminary results from the initial phase of this study, namely an investigation into the relative quality of temperature and moisture profiles derived from the GPS-RO COSMIC datastream versus those available from conventional radiosondes over the ENP, NCCZ and the northeastern United States (in this case examining the utility of the COSMIC data for validation). In addition to an overall outline of our methodology, we present summary results for a multi-year period, individual years, seasons, months as well examining the impact of varying the “comparison radius” (an analogue to the scan radius used in simple Barnes and Cressman-type objective analysis schemes). The implications of the statistical results on the remaining aspects of the study will also be discussed.