An Active Microwave Limb Sounder for Profiling Water Vapor, Ozone, Temperature, Geopotential, Clouds, Isotopes and Stratospheric Winds

An Active Microwave Limb Sounder for Profiling Water Vapor, Ozone, Temperature, Geopotential, Clouds, Isotopes and Stratospheric Winds E. R. Kursinski 1,2, D. Feng 1, D. Flittner 1, G. Hajj 2, B. Herman 1, F. Romberg 2, S. Syndergaard 1, D.Ward 1, T. Yunck 2 1 Department of Atmospheric Sciences, University of Arizona, Tucson, AZ 85721, 2 Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, 91109 Abstract - We summarize our findings on the performance of a radio occultation system operating at cm and mm wavelengths selected to profile atmospheric water, ozone and other constituents such as water isotopes as well as temperature, the geopotential of atmospheric pressure surfaces and clouds. Furthermore winds in the upper stratosphere can be determined from the Doppler shift of the line center. Our analysis indicates that such a system will yield dramatically higher vertical resolution, precision and accuracy than present and planned passive radiometric systems. I. Introduction Our continual quest towards a deeper understanding of weather and climate and improved skill in predicting their future behavior depends critically on our knowledge of the present structure of the atmosphere and its variations. Radio occultations are well suited to characterizing processes and quantifying the vertical thermodynamic structure and constituents of the atmosphere and therefore also quite well suited for monitoring climate, detecting changes in climate, and improving weather prediction At present, the vertical information that can be derived from satellite radiometric measurements is limited to a few atmospheric layers. Even the highly touted AIRS produces only 6 pieces of information across the depth of the troposphere (roughly twice that of TOVS) yielding approximately 2 km vertical resolution for a constituent whose average scale height is 1.5 km. Radiosondes provide high vertical resolution but poor global, limited dynamic range and two samples per day. Here we discuss a particular design of the spacecraft radio occultation technique for characterizing the thermodynamic and compositional structure of the atmosphere. We refer to this design as BRIGHTOC, meaning Bi-static Radar Imaging of Geopotential, Humidity, Temperature, Ozone and Clouds. The 1 Kursinski et al. 2002

BRIGHTOC system described here would measure the phase and amplitude of several signals near the 22 and 183 GHz water lines and 195 GHz ozone line as they slice through the atmosphere during an occultation. From the measured phase and amplitude, we derive profiles of both the speed of propagation and the attenuation due to water absorption and in turn solve for the wet and dry density profiles directly from the occultation observations. BRIGHTOC provides very high signal-to-noise ratios (SNR) and vertical resolution as shown in Figure 1b. The result is a high vertical resolution, global, all-weather active limb-sounder yielding very precise and accurate moisture, temperature and geopotential profiles from the surface to the mesopause as well as liquid and ice clouds and profiles of other constituents such as ozone. With several orbiting platforms the diurnal cycle can be characterized as well including for the first time the portions of the atmosphere in and below the clouds. 40 Tangent Point b. LEO a r a α Earth a. GPS Minimum Ray Altitude (km) GPS 30 20 10 0 0.0 0.5 1.0 1.5 22 GHz 0.0 0.1 0.2 0.3 183 GHz 0.0 0.033 0.067 0.1 Fresnel Diameter (km) Fig. 1. a. Satellite to satellite occultation geometry. b. Diffraction limited occultation vertical resolution at 1.6, 22 and 183 GHz. (from Kursinski et al., 2002) Here we summarize the clear sky results of Kursinski et al. (2002) followed by an evaluation of the capabilities in cloudy conditions. We also describe briefly the potential for determining water isotope concentrations and upper stratospheric winds. Herman et al. (2003) describe the ozone retrieval capability elsewhere in this volume II. Inversion Theory: Retrieving Atmospheric structure from Amplitude A. Conversion of Occultation Absorption Profiles into Profiles of Absorption Coefficients Spacecraft radio occultations (such as those using GPS) have focused generally on deriving bending angle profiles from the changing Doppler shift during an 2 Kursinski et al. 2002

occultation. (For a detailed description of the GPS occultation technique, resolution, and theoretical accuracy, see Kursinski et al., 1997). In such cases, the index of refraction, n, is taken to be real. However, n is in general complex, n c, because a medium affects both the speed and amplitude of signals via absorption as they pass through it. The information in n c is contained in the refractivity, N c, the non-unity portion of n c defined as N c = (n c -1)x10 6. N c has real (N ) and imaginary (N ) parts such that N c = N + in. Knowing profiles of both N and N provides the constraints needed to solve for profiles of moisture concentration, temperature and pressure. Ammonia concentrations in the outer planets [Lindal et al., 1981] and H 2 SO 4 concentrations in the atmosphere of Venus [Jenkins et al., 1994] have been inferred from spacecraft radio occultation absorption measurements. In the rest of the paper, for simplicity, we will write N = N as representing the real part of N c. Signal intensity, I, is reduced by absorption along the signal path as di = - I k dl where k (= 4πN /λ 0 x10-6 ) is the extinction coefficient, L is distance along the signal path and λ 0 is the signal wavelength in a vacuum. For each wavelength, the observed intensity, I, is related to the vacuum intensity, I 0 (the signal intensity with no atmosphere), as I = I 0 e τ where τ is the optical depth integrated along the signal path through the atmosphere. Therefore τ is determined from the intensity. τ = ln(i 0 /I) (1) While the measured optical depth is integrated along the occultation signal path, the desired quantity is the radial profile of the extinction coefficient, k. Given k and N as functions of r, the distance from the center of curvature (approximately the center of the Earth), we can derive a radial profile of atmospheric water. Under the assumption of local spherical symmetry, the optical depth and extinction coefficient are related by an abel transform pair, (2a) and (2b), analogous to the standard bending angle and index of refraction transform relation derived by Fjeldbo et al. [1971]. τ = kdl = r0 k n r dr 2 2 2 2 ( n r n r ) 0 0 1/ 2 1 da dτ k = 2π dr da a= a0 a0 da 2 2 ( a a ) 1/ 2 (2b) can be derived from (2a) via standard abel integral transform pair relations [Tricomi, 1985; Feng et al., 2001]. Note that the independent variable in (2b) is a, the asymptotic miss distance (see Fig. 1) defined as a = n r sin θ where θ is the angle between the ray path and radial direction. a is a constant for each ray path under the assumption of spherical symmetry and is derived from the atmospheric Doppler profile as described in Kursinski et al. [1997]. k is then derived as a function of r in (2b) using the fact that a 0 = r 0 n(r 0 ) where r 0 is the tangent radius of the raypath such that θ is π/2 and n(r 0 ) is derived from the bending angle profile via the standard abel equation. 0 (2) 3 Kursinski et al. 2002

B. Use of an off-line calibration tones to remove unwanted effects To remove noise and provide dynamic range, profiles of optical depth will be measured at several frequencies. Forming the ratio of the amplitudes of signals with similar frequencies to eliminate unwanted common noise and atmospheric effects. Therefore the optical depth used in (2b) will actually be the difference between the optical depths measured at 2 different frequencies, τ = τ τ = 12 1 2 ln where the subscripts, 1 and 2, refer to two frequencies, f 1 and f 2. The resulting extinction coefficient profile derived from (2b) will be k 1 k 2. It is important to note that the absolute signal amplitude is not relevant. Rather, the signatures of interest are the variations in amplitude during an occultation. The signal amplitudes will be normalized to the amplitude observed immediately before or after each occultation when the signal path is entirely above the atmosphere. The amplitude normalization of each and every occultation eliminates long-term drifts yielding a technique extremely well suited for observing long-term climate variations. I I 10 1 I I 2 20 C. Conversion of Absorption Coefficients and Refractivity into Temperature, Pressure, Water Vapor & Cloud Liquid In Earth s atmosphere, at frequencies less than 300 GHz, refractivity is related to temperature (T), total pressure (P t ) and partial pressure of water (e) as N = 77.6 (P t /T) + 3.73x10 5 (e / T 2 ) (3) In the upper troposphere and stratosphere, where there is no liquid water, the profiles of two observables, k 1 (r)-k 2 (r) and N(r), can be used to derive temperature (T), total pressure (P t ) and partial pressure of water vapor (e) by simultaneously solving 3 equations, the refractivity equation, (3), the absorption equation, (4), and the hydrostatic equation. k 1 (r) k 2 (r) = F(f 1, f 2, P t, e, T) (4) f 1 is positioned on the line to measure absorption and f 2 is positioned just offline to calibrate out unwanted effects. The absorption coefficient near the 22 and 183 GHz water lines is a strong function of e and a weaker function of P t and T, and therefore primarily constrains e. P t and T determine the line shape and the absorption due to O 2. Since the hydrostatic relation is a differential equation we need a boundary condition to initialize the hydrostatic integral such as temperature in the upper mesosphere. By measuring N and k at several frequencies to provide the dynamic range needed to sense water throughout the troposphere and middle atmosphere, the observations provide additional constraints. At altitude intervals where two 4 Kursinski et al. 2002

different pairs of frequencies each provide independent estimates of the extinction coefficients and therefore e, P t, and T, the overlapping constraints provide the information needed to determine the hydrostatic boundary condition. For each additional constituent at least one additional frequency tone is needed. For example, to solve simultaneously for cloud liquid water, C l, and ozone along the path requires at least two additional tones. It may require more because of the calibration tone needed for each of the new constituents. If the new tone frequencies are sufficiently close, a calibration tone placed between the 183 GHz water and 195 GHz ozone lines could satisfy the calibration needs of both lines. Given an occultation profile of observations, we then combine the refractivity and absorption observations, with the set of equations like (4) and solve simultaneously for each constituent, the bulk density, temperature and pressure at each altitude. III. Clear Sky Results We now discuss the accuracies of retrieved water, temperature and geopotential for clear sky conditions based on covariance results. We have a vector of observations, y, from an occultation profile from which we want to derive the atmospheric state vector, x, consisting of the atmospheric variables of interest: water vapor, temperature and surface pressure. Assuming a linear set of equations relate y and x, the statistically optimal weighted least squares solution for x is x T 1 1 T [ K S K ] K S y = (5) y Y where K represents the gradient of y with respect to x and S y is the observation error covariance. The error covariance, S x, of x is T 1 [ K S ] 1 Sx = y K (6) altitude (km) 80 60 40 While (1) and (4) are somewhat nonlinear, the fractional observational errors are quite small and (6) provides a representative estimate of the error in x resulting from errors in y 183.31/179.0 183.30/179.0 lo-band: 8.0, 13.0, 17.5, 20.0, 22.21, 32.0 + hi-band: 179.0, 182.2, 183.0, 183.2, 183.3, 183.31 + high altitude pressure boundary condition [Rodgers, 1990]. Fig. 2 shows that a combined 22 and 183 GHz occultation system can profile 20 0 lo-band: 8.0, 13.0, 17.5, 20.0, 22.21, 32.0 + hi-band: 179.0, 182.2, 183.0, 183.2, 183.3, 183.31 without the high altitude pressure boundary condition lo-band only: 8.0, 13.0, 17.5, 20.0, 22.21, 32.0 0.01 0.1 Fractional RMS water vapor error 5 Fig. 2. Fractional water vapor error for tropical, clear sky conditions [from Kursinski et al., 2002]. Kursinski et al. 2002

water vapor to ~1-2% or better precision from near the surface to 70 km altitude in clear conditions. The frequencies of the occultation tones are given in the figure. The Fig. 2 results are for 250 m vertical resolution in the lower half of the troposphere, 500 m resolution in the upper troposphere and lower stratosphere and 1 km resolution in the stratosphere. Fig. 3 shows an example of very dry, high latitude winter conditions where the 22 GHz line information is limited to the lowest 4-5 km of the atmosphere (dashed-dotted line in Fig. 3: Water vapor error under Fig. 3). Sampling by tones at 179 (solid line), high latitude winter conditions. 176 (dotted line) and 165 and 176 GHz (dashed line) dramatically improve the moisture characterization above 3.5 km altitude. Kursinski et al. (2002) showed BRIGHTOC will profile temperature to a few tenths of a Kelvin precision up to ~80 km and the geopotential height of pressure surfaces to ~10 m up to ~60 km. Combining the moisture and temperature information will yield relative humidity profile precisions of a few percent. Significantly better accuracies will be achieved when profiles are averaged to characterize climatological behavior. Fig. 4. Liquid water profile for cloud near 4 km under tropical conditions IV. Results in the Presence of Clouds A. Tropical liquid water cloud at 4 km Fig. 4 shows a simple liquid water cloud at 4 km altitude for tropical conditions. Figures 5a, 5b and 5c show the errors in retrieved temperature, water vapor and cloud liquid water. In Figures 5a and 5b, the small differences between the clear sky retrieval (dotted line) and the full with-cloud retrieval (solid line) shows that the presence of the cloud causes little degradation in the retrieved 6 Kursinski et al. 2002

temperature and water vapor results. The cloud liquid water errors at the altitude of the cloud are ~0.005 g/m 3 or 5% at the level of the cloud. The cloud liquid water errors increase near the surface reaching a maximum of ~0.04 g/m 3. Fig. 5. a. Temperature errors. b. Water vapor errors. c. Liquid water errors B. Marine stratus case Marine stratus clouds such as those lying off the coast of Southern California are an important and variable contributor to Earth s albedo. Because of their remote location, routine characterization of these clouds can only be done via remote sensing, posing a serious remote sensing challenge because of their close proximity to the surface. Because of the horizontal homogeneity of stratus decks, the absence of topography over the oceans, in combination with the high vertical resolution and ability to penetrate clouds of microwave occultations, these observations will characterize marine stratus clouds vertical structure far better than passive nadir viewing observations. Figures 6a, 6b and 6c show the temperature, water vapor and liquid water profiles of representative marine stratus cloud deck based on dropsonde measurements [Stevens et al., 2002]. The liquid water profile is a thick marine stratus deck with a peak liquid water density of 0.8 g/m 3 near 800 m altitude. Fig. 6. a. Marine stratus case temperature profile. b. water vapor profile. c. liquid water profile Figures 7a, 7b and 7c show the errors in retrieved temperature, water vapor and liquid water respectively for the Fig. 6 case. The dotted line in each of the Fig. 7 panels shows the error when we solve for liquid water when no cloud is present. This establishes a noise floor to the retrievals when clouds may be present. Kursinski et al. [2002] showed that utilizing a low frequency tone to measure absorption by the 60 GHz O 2 feature significantly reduces errors in the lowermost troposphere under warm and wet conditions. The solid line in each of the Fig. 7 7 Kursinski et al. 2002

panels represents the error when the cloud is present and the amplitude of the 8 GHz frequency tone has been used to constrain the abundance of molecular oxygen and therefore the atmospheric temperature. The dashed line in the panels shows the temperature error when the 8 GHz tone is not used. Fig. 7. a. Temperature errors, b. Water vapor errors, c. Liquid water errors As shown in Fig. 7a, the temperature error is less than 1 K above 2 km altitude in all cases. Below 1.5 km altitude the temperature error depends on the conditions. When no cloud is present, the temperature error is 2 K or less. When a cloud is present but there is no O 2 absorption constraint, the near-surface temperature error is 2.5 to 3.5 K rising to more than 4 K right at the surface. With the 8 GHz O 2 absorption constraint, the error within 1.5 km of the surface is 2 to 2.5 K rising to ~3 K right at the surface. This amounts to a ~0.5 K degradation relative to the no-cloud case. In Fig. 7b the fractional water vapor error using the 8 GHz O 2 constraint is 4 to 5% in the lowest km. The clear sky errors near the surface are 1 to 2%. With no O 2 absorption constraint errors are 6 to 8% within 1 km of the surface. Fig. 7c shows the errors in the retrieved liquid water profile. Errors when solving for liquid water when none is present (dotted line) are 0.01 g/m 3 near 1 km altitude and increase to 0.02 g/m 3 at the surface. Above 2 km altitude errors are 0.002 g/m 3 or less. The errors when the cloud is present and the 8 GHz O 2 constraint is used (solid line) reach a peak of 0.04 g/m 3 at the cloud peak amounting to a fractional error of 5%. The errors in the well mixed layer beneath the cloud are ~0.015 to 0.02 g/m 3. Above the cloud, the errors are essentially those when no cloud is present. The liquid water errors without the O 2 constraint are somewhat worse reaching a maximum of 0.05 g/m 3 near cloud top and 0.04 g/m 3 below the cloud. The Fig. 7 results reveal the potential for BRIGHTOC to characterize the marine boundary layer to high accuracy and vertical resolution in clear conditions and when marine stratus clouds are present. C. The impact of ice clouds The occultation tones near 183 GHz will be sensitive to scattering by ice particles [e.g. Evans and Stephens, 1995]. However, contamination by ice above the freezing level should not be a problem because the scattering effects are quite broadband such that the on-line/off-line ratioing described in Section II.B will 8 Kursinski et al. 2002

remove most of the scattering effect. Cloud ice can be sensed by examining the attenuation profiles of the unratioed individual tones. 2 or more tones spaced widely in frequency will yield information on ice amount and other attributes such as particle size distribution. Use of polarization may help sort out ice particle properties and orientations. The combined 22 and 183 GHz sensitivities will shed insight into the combination of liquid and ice particles known to exist between 0 and 40 o C, a poorly understood and characterized regime of behavior. V. Characterizing Water Isotopes The ratios of atmospheric and precipitation HDO to H 2 O and H 2 18 O to H 2 O has proven to be diagnostic of the sources, sinks and evolutionary history of atmospheric moisture. Measurements of isotopic ratios from space would place very important and unique constraints on the history of air parcels, improving our understanding of the hydrological cycle and providing critically needed constraints for evaluating and improving models. Fig. 8. Precision of individual water isotope profiles derived via occultation at 1 km vertical resolution at the equator and 60 o S in Altitude (km) 25 20 15 10 5 H2_18O Equator HDO (255 GHz) Equator H2-18O 35S HDO (255 GHz) 35S 0 0.00% 5.00% 10.00% 15.00% 20.00% 25.00% 30.00% Precision(%) Measuring isotopic concentrations is challenging because the isotopic concentrations are so small. H 2 18 O /H 2 O ratios are ~0.2%, and HDO/H 2 O ratios are an order of magnitude smaller near 0.03%. Atmospheric HDO varies by roughly a factor of 7 more than H 2 18 O making the absolute variations of HDO and H 2 18 O comparable. H 2 18 O can be characterized via a strong line near 203 GHz. While HDO has a number of strong lines starting at 80 GHz, the line at 255 GHz appears to have the best combination of line strength and frequency separation from strong lines of other constituents. Fig. 8 shows the precisions of retrieved H 2 18 O and HDO achievable with a one second voltage SNR of 2000 and 1 km vertical resolution for equatorial and high latitude winter conditions. Precision in the upper troposphere and stratosphere is limited by the isotopic line strength, shape and concentrations. At lower altitudes, the precision decreases rapidly where the continuum absorption by the dominant isotope of H 2 O begins masking the absorption by the minor isotopes. In the tropics, the best precisions of ~5% are achieved near 10 km altitude where isotopic concentrations are 9 Kursinski et al. 2002

relatively large while the H 2 O concentrations are sufficiently small to not mask the attenuation signature of the minor isotopes. Because of the cold air and low H 2 O concentrations at 60 o S in June, the characterization of isotopic concentrations extends approximately 5 km deeper into the troposphere. The precisions in Fig. 8 represent the random error in individual profiles and are adequate for measuring the factor of 2 variations in HDO observed across the vertical extent of the troposphere [e.g. Kuang et al. 2002]. Significantly higher accuracies can be achieved with averaging. VI. Deriving Stratospheric Winds In the upper stratosphere, the line width of water and other constituent lines near 200 GHz is sufficiently narrow that the shift of the line center due to the component of wind along the occultation path can be accurately determined. In this way the atmospheric winds can be sensed directly with high vertical resolution. Because at each altitude we must solve for both the constituent density and the line center shift, an additional tone is required. An example of the achievable accuracies is given in Kursinski et al. (2003). VII. Other Sources of Error A. Horizontal errors The errors we have described here have ignored those due to horizontal variations in atmospheric structure. Syndergaard et al. (2003) evaluated the accuracy of occultation retrievals near weather fronts and found worst-case retrieval errors of ~5% when the retrieved occultation quantities were treated as horizontally weighted averages. In most cases the errors were significantly less and treating the results properly as horizontal averages generally reduced the larger errors by a factor of 5 relative to comparisons with point values. Therefore, the 0.5 to 3% moisture errors due to finite measurement SNR and residual diffraction effects described here are quite relevant even under the rather severe horizontal conditions associated with fronts. B. On-orbit spectroscopic calibration Achieving accuracies consistent with the ~1% precisions which BRIGHTOC is capable of achieving are likely to be limited by our spectroscopic knowledge of the absorption spectra of the atmosphere of Earth and Mars. To address this issue, the BRIGHTOC tones will be tunable such that we can oversample the line shape 10 Kursinski et al. 2002

and spectrum and solve for the absorption line parameters while in orbit. With such a capability BRIGHTOC can become the absolute standard against which other constituent observations are calibrated. VIII. Conclusions An orbiting occultation system using frequencies near the 22 and 183 GHz water lines and the 195 GHz ozone line, will measure moisture, temperature, geopotential of pressure surfaces and cloud profiles with an unprecedented combination of precision, accuracy, vertical resolution and dynamic range in clear and cloudy air. We have shown that we can profile the water isotopes that provide critical constraints and insight into how moisture in the troposphere and stratosphere are controlled. Temperature accuracies of individual profiles will be sub-kelvin from ~1 km to 70 km altitude. Accuracies of geopotential heights of pressure will be 10 to 20 m from the surface to 60 km altitude. The errors described here are random such that climatological averages derived from this data should be significantly more accurate. As we have shown, these levels of performance will be degraded only slightly (generally by less than a factor of 2) in the presence of clouds. The degradation is primarily in the lowermost troposphere and is significantly less at higher altitudes. The technique can accurately profile liquid water clouds and may be able to similarly profile ice clouds. These profiles will be horizontally averaged liquid water owing to the limb viewing geometry. The technique is therefore best suited for characterizing stratiform-type clouds. The along-track resolution is comparable to the 200 to 300 km of the GPS occultation observations but the shorter 22 and 183 GHz wavelengths improve the diffraction-limited vertical resolution to 100 to 300 m. Figures 12, 13 and 14 compare the resolution and precision of BRIGHTOC and other present and planned satellite sensors. The figures indicate that the combined dynamic range, accuracy, vertical resolution and ability to penetrate and characterize clouds of BRIGHTOC extend significantly beyond the capabilities of other space-borne atmospheric sensors. In characterizing and understanding changes in climate, we must have observations decoupled as much as possible from models to provide the independent information needed to determine how the climate system works. While the need for decoupled observations seems obvious, in practice it is difficult to achieve. For instance, the passive radiometric radiances that dominate the operational satellite observations are non-uniquely related to the atmospheric properties such as temperature, pressure and constituent densities. As a result deriving these atmospheric properties from spaceborne radiances relies heavily on additional constraints from modeled and climatological estimates of these quantities. In contrast, the radio occultation technique provides estimates of these quantities directly with no assumptions or constraints except local spherical symmetry. As such, unique retrievals are generated completely independent of 11 Kursinski et al. 2002

models. Therefore the BRIGHTOC observations are significantly better at characterizing the true signatures of underlying climatic processes and assessing the accuracy of climate models. A constellation of such sensors would provide an all-weather, global remote sensing capability including full sampling of the diurnal cycle for process studies related to water, and climate research and weather prediction in general. 10 SSM/I Fig. 12. Comparison of BRIGHTOC performance with other existing and planned satellite sensors. Vertical resolution (km) 1 GIFTS AIRS HSB TOVS MLS GPS Occ BRIGHTOC 0.1 2 20 200 600 Horizontal resolution (km) Water vapor mixing ratio (ppmv) 10 100 1000 10000 AMOS MLS 3 km vert. res. MLS 2 km vert. res. AIRS (clear air, 2 km vert. res. ) MLS 1 km vert. res. GPS IASI (mid-lat winter Collard&Healy) AIRS (microwave only) 1 10 100 Fractional precision (%) Fig. 13. Fraction water vapor precision versus water vapor mixing ratio and vertical resolution of BRIGHTOC, GPS occultation, MLS and AIRS observations. Solid lines are 1 km or better vertical resolution. Dashed lines are 2-3 km vertical resolution. Thicker lines indicate higher vertical resolution. Fig. 14. Comparison of vertical resolution versus precision of individual MLS and BRIGHTOC water vapor profiles near the tropical tropopause. For fine scale layering observed by in-situ observations see Newell et al. (1999). Vertical Resolution (km) 3 2.5 2 1.5 1 0.5 0 Resolution & precision needed to remotely observe water and ozone structures observed by in-situ instruments AMOS at 16 km MLS at 16 km 1 10 100 Precision (%) 12 Kursinski et al. 2002

IX. References Evans, K. F. and G. L. Stephens, 1995, Microwave radiative transfer through clouds composed of realistically shaped ice crystals, Part I: Single scattering properties, J. Atmos. Sci., 52, 2041-2057. Feng, D. D., S. Syndergaard, B. M. Herman, E. R. Kursinski, T. P. Yunck, and F. W. Romberg, 2001, Deriving atmospheric water vapor and ozone profiles from active microwave occultation measurements", in Sensors, Systems, and Next-Generation Satellites IV, edited by H. Fujisada, J. B. Lurie, A. Ropertz, and K. Weber, Vol. 4169 of SPIE Proceedings Series, pp. 299-308. Fjeldbo, G., A. J. Kliore, and V. R. Eshleman, 1971, The neutral atmosphere of Venus as studied with the Mariner V radio occultation experiments, Astron. J., 76, 123-140. Herman, B., D. Feng, D. Flittner, E. R. Kursinski, S. Syndergaard, D. Ward, 2003, An overview of the University of Arizona ATOMS project, OPAC-1 proceedings, September 2002, Graz, Austria, Springer-Verlag.. Jenkins, J. M., P.G. Steffes, D. P. Hinson, J. D. Twicken, and G. L.Tyler, 1994, Radio occultation studies of the Venus atmosphere with the Magellan spacecraft. 2. Results from the October 1991 experiments, Icarus, 110, 79-94. Kuang, Z., G. Toon, P. Wennberg, and Y. Yung, 2002, Evidence for the convective control of stratospheric humidity from HDO measurements in the tropical tropopause layer, submitted to Geophys. Rev. Lett.. Kursinski, E. R., G. A. Hajj, J. T. Schofield, R. P. Linfield and K. R. Hardy, Observing Earth s atmosphere with radio occultation measurements using the Global Positioning System, J. Geophys. Res., 102, 23429-23465, 1997. Kursinski, E. R., G. A. Hajj, S. S. Leroy and B. Herman, 2000a, The Radio Occultation Technique, Terrestrial, Atmospheric and Oceanic Sciences, 11, 53-114. Kursinski, E. R., S. B. Healy and L. J. Romans, 2000b, Initial Results of Combining GPS Occultations with ECMWF Global Analyses within a 1Dvar Framework, Earth, Planets and Space, 52, 885-892. Kursinski, E. R. and G. A. Hajj, 2001, A comparison of water vapor derived from GPS occultations and global weather analyses, J. Geophys. Res., 106, 1113-1138. Kursinski et al., 2002, A microwave occultation observing system optimized to characterize atmospheric water, temperature and geopotential via absorption, J. Oceanic and Atmos Tech., Dec. Kursinski, E. R., G. A. Hajj and L. J. Romans, 2002, The zonal variability of of water vapor derived from GPS occultations and ECMWF global weather analyses, submitted to J. Geophys. Res.. Kursinski, E. R., et al., 2003, The Mars Atmospheric Constellation Observatory (MACO) Concept, OPAC-1 Book, Springer-Verlag. Lindal, G. F., et al., 1981, The atmosphere of Jupiter: An analysis of the Voyager radio Occultation measurements, J. Geophys. Res., 86, 8721-8727. Newell, R. E., V. Thouret, J. Y. N. Cho, P. Stoller, A. Marenco and H. G. Smit, Ubiquity of quasihorizontal layers in the troposphere, Nature, 398, 316-319, 1999. Poli, P., J. Joiner and E. R. Kursinski, 2002, 1DVar analysis of temperature and humidity using GPS radio occultation refractivity data, in press J. Geophys. Res. Stevens, B., et al., 2002: Dynamics and Chemistry of Marine Stratocumulus -- DYCOMS-II, BAMS, submitted February 7, 2002. Syndergaard, S, D. Flittner, E. R. Kursinski, and B. Herman, 2003, The accuracy of occultation retrievals near weather fronts, OPAC-1 proceedings, September 2002, Graz, Austria, Springer- Verlag.. Tricomi, F. G., Integral Equations, Dover, Mineola, N. Y., 1985. 13 Kursinski et al. 2002