[2]See for example: K. R. Hardy et. al., Atmospheric Profiles From Active Space-Based Radio Measurements, Conference on Satellite Meteorology and Oceanography, January 5-10, 1992, Atlanta, GA. Published by the American Meteorological Society, Boston, Mass.

[3]Report of the Small Climate Satellites Workshop, Committee on Earth and Environmental Sciences (CEES), February 12-14, 1992, Ft. Belvoir, VA.

[4](IPCC, 1990, 1992)

[5]G. A. Hajj, E. R. Kursinski, Analysis of Errors in the Vertical Temperature Profiles Recovered From GPS Occultation Observations, Trans. Amer. Geophys. Union (EOS), 72,372, 1991.

[6]W. G. Melbourne et. al., GPS Geoscience Instrument for EOS and Space Station, JPL proposal to NASA AO OSSA-1-88, July 15, 1988.

[7]R. P. Malla, et. al., Geocenter Location and variations in Earth Orientation using GPS, submitted to JGR- Solid Earth, 1992.

[8]Fjeldbo et al. 1971; Newman et al. 1984.

[9]Fjeldbo and Eshleman 1968; Lindal et al. 1979.

[10]Lindal et al. 1981, 1985, 1987, 1990.

[11]Kliore et al. 1975; Lindal et al. 1983; Tyler et al. 1989.

[12]C. Rocken and C. Meertens, Monitoring Selective Availability Dither Frequencies and Their Effect on GPS Data, Bulletin Geodesique, 71, 162-169, 1991.

[13]K. R. Hardy et. al., Atmospheric Profiles From Active Space-Based Radio Measurements, Conference on Satellite Meteorology and Oceanography, January 5-10, 1992, Atlanta, GA. Published by the American Meteorological Society, Boston, Mass.

[14]According to a 1991 study conducted by G. A. Hajj and E. R. Kursinski (JPL), a simple linear combination correction is accurate to 1 mm for night-time occultations below 80 km. The error increases above 80 km. Because of high electron density variability in the day-time ionosphere, the simple linear combination correction is only good to 1-5 cm below 80 km.

[15]J. J. Spilker, Jr., GPS Signal Structure and Performance Characteristics, Journal of Navigation, 25, No. 2, pp 121-146, Summer 1978.

[16]The ionospheric correction described is the standard dual frequency correction. For ground observations, the effect of higher order terms, caused primarily by the Earth's magnetic field, may be removed in the manner described by Brunner and Gu (Manuscripta Geodetica, 16, 205, 1991). We will investigate the potential of this refinement.

[17]Total atmospheric delay is a function of two factors: ray bending due to refraction gradients, and reduced propagation velocity in the atmosphere.

[18] M. E. Gorbonov and S. V. Sokolovskiy, Space Refractivity Tomography of the Atmosphere: Modeling of Direct and Inverse Problems, Max-Planck Institute fore Meteorology Report, Hamburg, 1994

[19]It may be possible to overcome this limitation if a constellation of LEO satellites is used.

[20]See also, M. Bevis, S. Businger, T. A. Herring, C. Rocken, R. A. Anthes, and R. H. Ware, GPS Meteorology: Remote Sensing of Atmospheric Water Vapor using the Global Positioning System, Journal of Geophysical Research, VOL. 97, NO. D14, 15,787-15,801, October 20, 1992.

[21]The International GPS Service for Geodynamics

[22]For the Phase I Proof of Concept, A/S is expected to be on most, but not all of the time. Thus, we expect to receive some A/S free data in Phase I. For an operational system, it is expected that Y code receivers would be used (or equivalent). Thus, errors due to phase noise in the operational context are expected to be similar to P code phase noise (1 mm).