92nd American Meteorological Society Annual Meeting (January 22-26, 2012)

Tuesday, 24 January 2012: 3:45 PM
A Local Forward Operator for the Assimilation of GPS RO Observations
Room 340 and 341 (New Orleans Convention Center )
Lidia Cucurull, NOAA/NWS/NCEP, Suitland, MD; and R. J. Purser and J. Derber

GPS Radio-Occultation (RO) observations are an important component of the observing system. Their use in operational Numerical Weather Prediction (NWP) models has resulted in significant improvement in model skill along the whole vertical range of the atmosphere, in particular in the upper troposphere and lower stratosphere. The benefits of GPS RO are attributed to several unique characteristics: all-weather capability, very high vertical resolution (~ 100m), lack of biases, high accuracy temperature measurements (~ 0.2K), same accuracy over land as over ocean, and global and uniform coverage.

Most operational NWP centers started assimilating GPS RO observations a few years ago, after the launch in 2006 of the first GPS RO constellation, COSMIC, composed of six satellites. Some other GPS RO satellites have been incorporated into the assimilation system since then. For example, NCEP is currently assimilating GPS RO observations from 10 different satellites, which accounts for ~2,000 profiles/day.

Although GPS RO raw measurements are linked to an atomic clock measurement that requires no corrective calibration, several steps and assumptions are necessary in order to have products that can be easily used in a data assimilation system. Operational NWP centers use forward operators that ignore the non-horizontal gradients of refractivity (or density) in the atmosphere. These are the so-called “local” forward operators. More sophisticated operators that account for these horizontal gradients are available or under development, but they require higher computational resources and have not been optimized for use in an operational environment.

NCEP is currently using a refractivity (local) forward operator to assimilate GPS RO data in its global data assimilation system and has recently developed, tested and implemented a more accurate forward operator that assimilates observations of bending angle rather than refractivities. Bending angle is an earlier product than refractivity in the GPS RO processing chain, thus it is more desirable from a data assimilation standpoint. An earlier version of this operator was already available at NCEP, but this updated version has many improvements over the older code. This bending forward operator is under current pre-operational testing at NCEP.

During the talk, the design and characteristics of the NCEP's bending angle forward operator will be described and its performance against the refractivity code will be evaluated.

Supplementary URL: