To support the TOGA/COARE project, Langley is developing cloud and radiation datasets from the Japanese Geostationary Meteorological Satellite (GMS) visible (VIS, 0.65 micrometer) and infrared (IR, 11 micrometer) radiances taken every hour over the tropical western Pacific. The GMS pixel data are taken at a nominal resolution of 5 km and analyzed on three different grids to accommodate a variety of potential needs. The synoptic scale domain consists of a 2.5 x 2.5 deg latitude-longitude grid extending from 12.5 S to 10N and from 180 to 137.5 E. The mesoscale domain has a 1 deg grid superimposed on the area between 8S to 6N and 162E to 148E. Finally, the IFA domain is divided into a 0.5 deg covering the area from 6S to 4N and from 160E to 150E.
Because of problems in the collection of GMS data during the TOGA/COARE period, it was necessary to combine data from several sources to achieve the best possible coverage and hourly sampling. The bulk of the data was provided by Don Wylie at the University of Wisconsin (see Hinton). Some data were acquired by NASA Langley during the January-February period in Townsville. Other gaps were filled using ISCCP B1 data that originated from the Japanese Meteorological Service and are archived by NOAA-NESDIS. Despite all efforts to develop a complete hourly dataset, some significant gaps remain especially during mid-January when the Australian data reception center suffered a multiday failure. Other problems, such as incomplete pictures and bad scan lines were also encountered more frequently than expected. We attempted to remove or minimize the impacts of such problems, but some remain. The following GMS data is available from our analysis by month - Nov'92, Dec'92, Jan'93 and Feb'93.
The GMS data are analyzed to produce cloud amount, altitude, IR temperature, optical depth, thickness, and VIS reflectance for the total scene as well as at three levels: low (z < 2 km), middle (2 < z < 6 km), and high (z > 6 km). Clear-sky IR temperature and VIS albedo are also determined. For optically thin clouds, the cloud temperature and altitude are adjusted as a function of the optical depth. Both the observed and corrected values are given for daytime data. At night, no VIS information is available. Therefore, no corrections are applied at night. This lack of optical depth data is especially problematic in the tropics where thin cirrus clouds are abundant. Thus, the corrected cloud heights are the most reliable for estimating the true cloud altitude. Beware of the nocturnal cloud altitudes and temperatures. The broadband shortwave (SW, 0.2 - 5.0 micrometer) albedos and longwave (LW, 5 - 50 micrometer) fluxes at the top of the atmosphere are also included in this dataset.
The techniques described by Minnis et al. (1987, F16) were used to determine the clear-sky parameters at each hour. These values were used in the methodologies of Minnis et al. (1993,L35 L36) and Minnis et al. (1995,F43) to produce the cloud products for the 3 grids. Nighttime cloud amounts were determined using a simple IR threshold method. The GMS VIS and IR radiances were calibrated against the NOAA-11 AVHRR channels 1 and 4, respectively, by matching nearly simultaneous GMS and AVHRR views taken at nearly the same viewing and azimuth angles. The AVHRR channel 1 was corrected for drift as in Minnis et al. (1995). The conversions of the VIS albedos to SW albedos and IR temperatures to LW fluxes were accomplished using the methods described by Minnis et al. (1995) using a set of VIS-SW and IR-LW coefficients based on the correlation of GMS and Earth Radiation Budget Satellite (ERBS) data taken over the synoptic domain during December 1986 and January 1987. It is assumed that the derived relationships are valid for the TOGA/COARE period from November 1992 through February 1993.
Many of the caveats and cautions given in Minnis et al. (1995) apply to the products developed for the TOGA/COARE area. A copy of the Minnis et al. (1995) report can be obtained electronically at http://techreports.larc.nasa.gov/ltrs/95/NASA-95-rp1366.refer.html or a hard copy can be mailed to you from Patrick Minnis at p.minnis@larc.nasa.gov. . The publication contains a number of color figures. Any questions concerning the content and quality of the data should be directed to Patrick Minnis. Questions regarding the format should be sent to Rabi Palikonda. If you wish to be kept informed about availability of data and changes in processing methods, please add your name to the mailing list for E-mail updates.
The datasets will be made available through the Internet as the processing is finished. At this time, prelimilary data from Nov'92, Dec'92, Jan'93 & Feb'93 are available. This document will be updated when new datasets become available. The TOGA/COARE dataset will be formally documented when all of the GMS analyses are complete. If you plan to use the data in a formal document, it is requested that you contact Patrick Minnis before proceeding until the data are documented.
Additional analyses of the GMS data are being performed at a higher resolution as part of our efforts to support the ARM program's Tropical Western Pacific database. We are also processing AVHRR data taken over the TOGA/COARE area as part of our participation in FIRE. These data will complement the GMS data because higher resolution products will be made available with additional cloud parameters including cloud particle size and phase and nocturnal optical depths (thin clouds only) and corrected heights. Notice of these products will be posted on the relevant home pages.
The hourly GMS narrowband radiance have been recalibrated during TOGA-COARE and converted to broadband fluxes. The TOA fluxes (longwave and shortwave reflected albedo) agree with the ERBS WFOV SF 10 degree data available on the NASA-langley DACC http://eosdis.larc.nasa.gov. The narrowband to broadband relationship was derived from Dec86 GMS and ERBS scanner 2.5 degree data.The results are discussed in this COARE98 abstract.The file naming convention used is toga..<grid_res>.yyddd_bb.nc.gz ,where grid_res is the grid resolution (eg: 2_5,1_0,0_5), yy is the year and ddd is the julian day.
NOTE:The units for "total broadband longwave flux at TOA (Total_BB_LW_Flux_TOA)" should be Watts per square meter and not Kelvin in these netcdf files.
| GRIDRES/MONTH | 92'NOV | 92'DEC | 93'JAN |
|---|---|---|---|
| 0.5 DEG | X | X | X |
| 1.0 DEG | X | X | X |
| 2.5 DEG | X | X | X |
Download PDF Viewer 