3560 IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 46, NO. 11, NOVEMBER 2008 CloudSat’s Cloud Profiling Radar After Two Years in Orbit: Performance, Calibration, and Processing Simone Tanelli, Stephen L. Durden, Senior Member, IEEE, Eastwood Im, Fellow, IEEE, Kyung S. Pak, Dale G. Reinke, Philip Partain, John M. Haynes, and Roger T. Marchand Abstract—The Cloud Profiling Radar, the sole science instru- ment of the CloudSat Mission, is a 94-GHz nadir-looking radar that measures the power backscattered by hydrometeors (clouds and precipitation) as a function of distance from the radar. This instrument has been acquiring global time series of vertical cloud structures since June 2, 2006. In this paper, an overview of the radar performance and status, to date, is provided together with a description of the basic data products and the surface clutter Fig. 1. CPR’s first image, May 20, 2006. Vertical section of frontal system in rejection algorithm introduced for the Release 04 data product. the North Sea. The green band is the ocean surface in the adaptive-height CPR science data window (in color in the electronic version). Index Terms—A-train, clouds, CloudSat, radar. water distribution, and as input to numerical weather-prediction I. INTRODUCTION models. In order to take full advantage of observations by other types NEW satellite mission, the CloudSat Mission [1], was of spaceborne atmospheric remote-sensing instruments, the jointly developed by the National Aeronautics and Space A CloudSat spacecraft flies in formation as part of the Afternoon Administration (NASA), the Jet Propulsion Laboratory (JPL), Constellation of satellites (also referred to as “A-Train,” from the Canadian Space Agency, Colorado State University, and the Billy Strayhorn’s old jazz tune). In particular, CloudSat flies U.S. Air Force. CloudSat’s payload, the Cloud Profiling Radar in close formation with CALIPSO [3], which carries a lidar (CPR) [2], is the first spaceborne 94-GHz (W-band) radar and system, so that their respective beams cover the same vertical is contributing vertical cloud profiles over the globe. CloudSat column within about 15 s. was launched on April 28, 2006; CPR instrument operations began on June 2, 2006. Since that time, CPR has been ac- quiring the first-ever continuous global time series of vertical II. CPR SYSTEM cloud structures and vertical profiles of cloud liquid water and CPR is a short-pulse profiling radar; it measures the power ice content with a 485-m vertical resolution and a 1.4-km backscattered by atmospheric targets (hydrometeors, ice crys- antenna 3-dB footprint. Fig. 1 shows the vertical structure tals, and cloud droplets) and any other target intercepted by of a squall line in the North Atlantic observed on May 20, the antenna beam (e.g., the Earth’s surface, aircraft, etc.) as 2006, immediately after the activation of CPR for a brief a function of time. Time is converted into range based on checkout test. CPR data are providing valuable information r = c0t/2, where c0 is the speed of light in vacuum. In a normal for studies of cloud physics, radiation budget, atmospheric data-acquisition mode, CPR points at nadir (0.16◦ off-geodetic nadir since August 15, 2006); therefore, ranging resolves tar- gets in altitude. The narrow antenna beam and the motion of Manuscript received February 29, 2008; revised May 19, 2008. Current version published October 30, 2008. This work was supported by the NASA the spacecraft resolve targets in the along-track direction. All Earth Science Pathfinder Program (ESSP). CPR products are therefore vertical slices of the troposphere, S. Tanelli, S. L. Durden, and K. S. Pak are with the Radar Science and Engi- as shown in Fig. 1. Table I shows some of the main parameters neering Section, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109 USA (e-mail: [email protected]). of interest, including the definition of symbols used throughout E. Im is with the Earth Science Technology Research and Advanced Con- this paper. The values included in Table I are provided at a cepts Office, Jet Propulsion Laboratory, California Institute of Technology, higher resolution than the approximate values used throughout Pasadena, CA 91109 USA. D. G. Reinke is with the Cooperative Institute for Research in the the text. Atmosphere, Colorado State University, Fort Collins, CO 80523-1375 USA. The A-train sun-synchronous orbit covers the latitude ranges P. Partain is with the METSAT Division, Science and Technology Corpora- ◦ ◦ tion, Fort Collins, CO 80521 USA, and also with the CloudSat Data Processing between 82.5 S and 82.5 N, with a repeat cycle of 16 days Center, Cooperative Institute for Research in the Atmosphere, Colorado State (or 233 orbits). CloudSat reaches its maximum altitude above University, Fort Collins, CO 80532-1375 USA. the Earth’s geoid when over Antarctica (about 732 km above J. M. Haynes is with the Department of Atmospheric Science, Colorado State University, Fort Collins, CO 80523-1371 USA. sea level) and its lowest altitude when at the equator (about R. T. Marchand is with the Joint Institute for the Study of the Atmosphere 705 km). CloudSat’s reference ellipsoid is the WGS84 for and Ocean, University of Washington, Seattle, WA 98195 USA. all official products. One vertical profile is acquired for each Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. integration time interval TI =0.16 s, corresponding to a sub- Digital Object Identifier 10.1109/TGRS.2008.2002030 satellite motion of 1.09 km (±10 m, depending on latitude). 0196-2892/$25.00 © 2008 IEEE Authorized licensed use limited to: IEEE Xplore. Downloaded on December 5, 2008 at 12:29 from IEEE Xplore. Restrictions apply. TANELLI et al.: CloudSat’s CLOUD PROFILING RADAR AFTER TWO YEARS IN ORBIT 3561 TABLE I pulselength c0τ/2), and La is the two-way atmospheric loss. CLOUDSAT AND CPR PARAMETERS CloudSat’s L1B product includes the calibrated power input to the receiver (Pr = Prec/Grec) sampled every 240 m, the radar constant C, and the orbit-averaged estimate of the transmit power Pt. Reflectivity is then converted (in L2B-GEOPROF) into the equivalent (attenuated) reflectivity factor according to the standard relation λ41018 Ze = η 5 2 (2) π |Kw| 6 3 where Ze has units of mm /m and is often expressed in ∗ | |2 decibels as dBZe =10 log10(Ze). KW is set to 0.75 (representative for water at 10 ◦C at the W-band) [9], [10]. CloudSat’s L2B-GEOPROF product also includes vertical pro- files of La estimated from the European Center for Medium- Range Weather Forecasts’ (ECMWF) auxiliary products. The second quantity of interest generated within the L1B 0 The actual integration of pulses lasts for 96.8% of TI (±0.1%, product is the normalized radar cross section (σ ) of the Earth’s depending on latitude). The resulting horizontal resolution after surface, expressed in decibels. integration and including latitudinal dependence is between 1.3 3 2 2 and 1.4 km cross track and between 1.7 and 1.8 km along track 0 Prec(4π) r La 1 CΔ Precr La σ = 2 2 = (3) (defined at the 6-dB point of the two-way resulting weighting Ptλ GrecG Ω cos(θ) cos(θ) Pt function). This estimate for the along-track resolution replaces the oft-quoted value of 2.5 km, derived early in the program where θ is the incidence angle; at nadir, the cosine term can from purely geometrical considerations. be neglected. This quantity, labeled “Sigma-zero” in the L1B CPR acquires 125 samples per profile, one every ∼240 m. product, is also affected by atmospheric loss. A more detailed discussion on this product is provided later. Considering that the spacecraft altitude hs varies with the position along the orbit, the radar timing parameters (includ- CloudSat’s data-processing plan included a one-year re- ing PRF and M) are adjusted based on a lookup table to processing, which has been implemented as planned in 2007. keep the troposphere inside the 30-km science data window. The resulting Release 04 (R04) of the data products includes In particular, the parameters are adjusted so that the Earth’s improvements discussed in the following sections and in the surface return is observed in the data window’s lower portion Appendix. and that a 5–10-km cloud-free region (of the stratosphere) is included in the upper portion. The former provides a useful IV. CPR HIGH-LEVEL DESIGN AND reference for retrieval algorithms (see, e.g., [4] and [5]), and the PRELAUNCH CALIBRATION latter allows for accurate noise-floor estimation used to perform Clouds are weak scatterers of microwave radiation, partic- noise subtraction, and achieve the required minimum detectable ularly in contrast to the reflection of the underlying Earth’s reflectivity (or sensitivity). surface. The overriding requirement on CPR was to achieve a minimum detectable cloud reflectivity factor (Ze) of −28 dBZ III. CPR MEASUREMENTS at the beginning of life (BOL) and −26 dBZ at the end of life (EOL, 22 months after launch). By comparison, the re- The discussion on CPR data products presented in this pa- flectivity for rain is typically 10–50 dBZ; the Tropical Rainfall per is limited to the L1B-CPR [6] and portions of the L2B- Measuring Mission (TRMM) precipitation radar (PR) [11] has GEOPROF [7] data products. The quantities of interest are a sensitivity of around +17 dBZ. The CPR required range the calibrated and geolocated parameters used as input to all resolution was 500 m. subsequent atmospheric radar retrieval algorithms. In order to achieve such requirements, CPR design and For volume scattering, the quantity of interest is the radar implementation included a high-efficiency antenna assembly cross section per unit volume or the radar reflectivity η.From (AA), a low-loss quasi-optical transmission line (QOTL), high- the radar equation [8], η is calculated as power amplifiers [HPA, each one composed of a high-voltage 3 2 2 power supply and an extended interaction klystron (EIK)], and a Prec(4π) r La Precr La η = = C (1) low-noise-figure receiver.