GPS Radio Occultation (RO) Technique






The satellite radio occultation technique was created in the late 1960's with the flyby of Mariner IV around Mars.  Voyager sampled the outer planets except Pluto from 1977 to 1989.  By 1989, every planet in the solar system  had been profiled via RO except Earth and Pluto.  Pluto will be sampled by New Horizons.   

In 1987, researchers at JPL recognized the potential for using the Global Positioning System (GPS) satellites to make RO measurements of the Earth's atmosphere (Yunck et al., 1987).  I first became involved with GPS RO in 1990 following the Voyager Neptune encounter and worked to develop the NASA GPS Geoscience Instrument (GGI), chosen as a NASA EOS instrument in 1988.  Most of the first order technical issues of GPS occultations were worked out during GGI development in 1990-1992 with NASA funding.  However, JPL cost estimates for GGI were high and kept growing leading to GGI's cancelation in 1992.  I worked on a proposal with Yam Chu for a GPS RO receiver on the TIMED mission that was not selected but ushered in the possibility of relatively inexpensive GPS RO receivers in space.  The GPS RO idea was taken up by researchers at UCAR/NCAR who proposed the idea to NSF that resulted in the first GPS occultation mission, GPS/MET, (Ware et al., 1996) flown from April 1995 to February 1997. 

GPS RO has now been flown on a number of missions: GPS/MET, Oersted, CHAMP, SAC-C and now COSMIC.


GPS RO is a very simple technique where a GPS transmitter and GPS receiver on a low Earth orbiting (LEO) satellite are located on opposite sides of the planetary limb.  Orbital motion of one or both instruments produces the limb scanning geometry.  The technique provides a unique combination of high precision and vertical resolution at long wavelengths that penetrate through the atmosphere in virtually any conditions with very little sensitivity to clouds.  It provides the first capability to globally profile the gases (as opposed to particulates) of the atmosphere with vertical resolutions of approximately 200 m from orbit.  As a limb sounder, its along-track, horizontal resolution is approximately 300 km. 


Diurnal cycle:        The diurnal cycle is key to understanding the atmosphere, particularly moist convection.  Doing so requires global measurements of the 3D state of the atmosphere over the diurnal cycle.  Present measurements cannot provide such a capability except in focused regional field campaigns.  Geosynchronous satellites such as GOES satellites provide powerful IR imaging of the diurnal evolution of cloud tops but cannot routinely profile the entire atmospheric column.  With six orbiting receiving satellites like COSMIC, GPS RO will for the first time, provide the routine and global vertical profiling of the entire diurnal cycle. 

GPS RO accuracy estimates

Key to any remote sensing technique are its accuracy and resolution.  In the early EOS days, it was unclear how GPS RO compared with other Earth atmospheric remote sensing techniques such as the 2,000 channel Atmospheric Infra-Red Spectrometer (AIRS).  To address this question, I developed a systematic understanding and estimates of the GPS RO errors as part of my PhD research at Caltech (see Kursinski et al., 1995, 1997).  These estimates have been key in establishing the potential power and importance of the RO technique for climate characterization and monitoring and weather forecasting as well as a standard against which actual GPS RO results can be evaluated.

GPS RO Retrievals in the presence of Super-refraction

The GPS 19 and 24 cm wavelengths can penetrate virtually any conditions such that GPS RO provides an all-weather profiling of Earth's atmosphere.  We and Sergey Sokolovskiy at UCAR came to realize that a problem occurs frequently that limits the ability of GPS RO to probe the boundary layer (Kursinski et al., 2000; Sokolovskiy, 2003).  In particular, Sokolovskiy (2003) recognized that under certain conditions referred to as super-refraction, the measured bending angle profile is consistent with at least two different refractivity profiles, meaning there is a fundamental ambiguity in interpreting the GPS RO bending angle profiles.  (Super-refraction occurs when the vertical gradient of the index of refraction is so large that the radius of curvature of an occultation raypath becomes smaller than the radius of the Earth.)

To address this bending angle-refractivity ambiguity, we have developed a retrieval method that derives unique profiles of boundary layer refractivity from bending angle profiles when superrefraction occurs at the top of the planetary boundary layer (PBL) using a simple parameterization of the refractivity profile in the uppermost portion of the PBL and at least one additional constraint.  The method is described by Xie et al. (2006).  This approach will allow the GPS profiles to extend to the surface even in the tropical marine boundary layer. 

We are quantifying the accuracy of this approach including the effects of diffraction and extending it to separate temperature and water vapor in the boundary layer.

Fast forward operator for assimilation into NWP Forecasts and Analyses

The < 1 mm GPS measurement precision is very small in comparison with the ~1 km atmospheric delays in the lower troposphere suggesting refractivity precisions might approach 10-6.  However, the accuracy of the GPS RO refractivity profiles in the lowermost troposphere is only ~1%.  The limiting error is not measurement precision but rather horizontal variations in refractivity that cannot be represented in the standard Abel transform.  We developed a fast linear forward operator to use when assimilating GPS RO data into Numerical Weather Prediction (NWP) forecasts (Syndergaard et al., 2005).  By using this forward operator which far better represents the information in GPS profiles the GPS RO profiles can be weighted more heavily and therefore have a much larger effect on the weather forecasts.  The disagreement between the fast forward model and the GPS Abel result can be reduced by a factor of 5 in regions of severe weather.


GPS RO-Derived Temperatures for Climate

While the GPS RO temperature profiling capability is quite powerful, in the context of climate and climate change it suffers from 3 problems:

1.    Because the ionosphere affects GPS wavelengths significantly at stratospheric altitudes, incomplete calibration of the ionosphere leaves small residual signatures of diurnal, seasonal and solar cycle ionospheric variations in the GPS temperatures. 

2.    At least one piece of external information is required to initialize the hydrostatic integral to uniquely determine temperature.

3.    GPS tropospheric temperature profiles are limited to temperatures colder than approximately 240 K because of the water vapor contribution to refractivity at warmer temperatures (because of the strong temperature dependence of water vapor saturation vapor pressure via the Clausius-Clapeyron relation). 



Kursinski et al., 1995,

Kursinski et al.,1997

Syndergaard et al., 2005

Ware et al., 1996

Xie et al. (2006).

Yunck et al., 1987