General response for all issues

Consolidated Comments from Gutman and Reale

Seth Gutman's General Comments:
Many of the issues and concerns raised by Seidel, Rosen, Peterson and Murrayare best discussed in the context of Kuo et al. 2004 which states that "Because of the optical geometry, the GPS radio occultation data have unique characteristics, which is very different from the point measurement of a radiosonde or an "area-average" measurement of a microwave sounder. A single GPS radio occultation measurement represents a weighted average over a "pencil-like" volume. The effective horizontal scale of the measurement along the ray is approximately several hundred km, while the cross-ray scale is quite small (~1 km)." To complicate matters further, the 'pencil of refractivity' that extends from the GPS satellite to the receiver in LEO is tilted in space, such that even two spacecraft in close temporal and spatial proximity can be sensing remarkably dissimilar air masses under certain circumstances.

Another point worth mentioning is based on the experience of trying to determine the absolute accuracy of water vapor observing systems at the DOE ARM SGP CART Site in North Central Oklahoma. Those of us who participated in the water vapor intensive observation periods between 1996 and 2000 encountered numerous problems or uncertainties related to the measurement or retrieval of water vapor from the observations. This experience accomplished three things: it allowed us to "iron out" problems in our data acquisition and processing systems; it allowed us to better understand some of the differences between various observing systems and data processing protocols; and it reconciled us to the fact that we have to live with uncertainty when it comes to observing the atmosphere. In the case of column-integrated precipitable water vapor, the uncertainty is at the single millimeter level as illustrated in Rivercomb et al 2003.

The final point I wish to raise before commenting on the issues raised is that refractivity in the neutral atmosphere at GPS/GNSS frequencies comes primarily from water vapor, temperature, and pressure. If you wish to retrieve a geophysical quantity such as temperature at a particular level in the atmosphere, you must have a priori knowledge of the pressure and moisture along the path of the occultation. Uncertainties or errors in these assumptions translate into retrieval errors plain and simple. Arguments about the lack of sensitivity of some retrieval errors to errors in the a priori's begs the question since we can never be certain of the validity of the assumption without independent verification. When it comes to GPS-RO, the observations are unique by any standard.

References
Kuo, Y. H., T. K. Wee, S. Sokolovskiy, C. Rocken, W. Schreiner, D. Hunt, 2004: Inversion and Error Estimation of GPS Radio Occultation Data, J. of the Meteorol. Society of Japan, 82, No. 1B pp.507-531.

Revercomb, H. E., D. D. Turner, D. C. Tobin, R. O. Knuteson, W. F. Feltz, J. Barnard, J. Bösenberg, S. Clough, D. Cook, R. Ferrare, J. Goldsmith, S. Gutman, R. Halthore, B. Lesht, J. Liljegren, H. Linné, J. Michalsky, V. Morris, W. Porch, S. Richardson, B. Schmid, M. Splitt, T. Van Hove, E. Westwater, and D. Whiteman, 2003. The Arm Program's Water Vapor Intensive Observation Periods. Bull. Amer. Meteorol. Soc., Volume 84, No. 2, 217-236.

Tony Reale's General Comments
As stated, the a-priori T and H20 are the critical issue for retrieval. These parameters must be included on all retrieval output files to segregate the impact of guess and retrieval on the final profile either as a "bulk" ensemble (global, regional...) quantity or individually for each profile. Comparing different retrieval approaches in most cases reduces to comparing individual a-priori approaches.

I agree with Kevin that these are good questions.

Are other satellite observing systems used for climate being submitting to this level of scrutiny? Jim Yoe indicated in one of his comments that he did not think other systems had demonstrated the levels of consistency as GPSRO already has. This evaluation process is very valuable and should not be limited to GPSRO, no matter what the level of "maturity" of the other observations.

I would also add that it is not clear based on the questions whether it has been appreciated that the independent variable in the RO measurements is height rather than pressure as it is for radiance measurements. Therefore RO provides some tight constraints on the geopotential height of pressure surfaces and all of the knowledge that accompanies that information such as balanced winds and the average temperature at lower altitudes. GPSRO allows one to see the troposphere expand with global warming and see the poleward migration of the jets which has been both predicted and is apparently being observed and GPSRO provides a very high SNR when observing this apparently anthropogenically-caused evolution as Stephen Leroy has pointed out.