Zebra is a software package developed by the NSF funded Research Data Program within the Atmospheric Technology Division of the National Center for Atmospheric Research. Zebra is used for data integration, analysis, and display of a large variety of data sets.
The TOGA COARE Workshop Integrated Data Set (TCWIDS) is a collection of the diverse data sets that were presented at the TOGA COARE Workshop, all of which can be viewed and analyzed using Zebra software.
Section 2 of this document contains instructions for reading the data tapes.
Section 3 of this document describes each data set that is included in TCWIDS-Version 1.0. Acknowledgments and other important information about the data sets are included in this section.
Section 4 of this document briefly describes data sets that were not included in TCWIDS-Version 1.0, but may be included in a future release.
Section 5 contains information on obtaining Zebra software.
Tape 1 contains 64 tar files. The first tar file contains the tape index, named tcwids1.index, this document in postscript format, binaries for the Zebra program 'dsmanage' in all four supported platforms, and source code the the 'dsmanage' module DLoad.cc. File 2 contains data with start dates up to and including 01 Nov. 1992. File 3 contains data with a start date of 02 Nov. 1992, file 4 contains data with a start date of 03 Nov. 1992, etc., so file 62 contains data with a start date of 31 Dec. 1992. Tar files 63 and 64 contain FNOC and NMC model data, respectively.
Tape 2 contains 62 tar files. File 1 contains the tape index, named tcwids2.index, another copy of this document in postscript format, and copies of 'dsmanage' binaries and DLoad.cc source. Subsequent files contain daily data collected during January and February of 1993. So file 2 contains data with a start date of 01 Jan. 1993, file 3 contains data with a start date of 02 Jan. 1993,..., file 60 contains data with a start date of 28 Feb. 1993 and any data with a start date later than 28 Feb. 1993. FNOC and NMC model data is repeated on this tape in files 61 and 62.
Data may be extracted from the tapes using the standard `tar' utility or with the Zebra `dsmanage' program. You must have a running Zebra system to use `dsmanage' and file 1 must be extracted using tar before `dsmanage' can be used. Additionally, the NMC and FNOC model data cannot be accessed by `dsmanage' and tar must be used to extract this data.
A new widget will pop up that will allow you to select the data that you wish to extract from the tape. You will see: Select buttons, Platforms, Data present (what is currently on the local disk), and Data available (what is currently on the tape). When the widget is first brought up all the platforms are selected. The select buttons are in reverse video when they are on. To select the data to be moved onto the local disk, enter start and end times in the boxes at the top of the widget and click on Apply to selected platforms. This will select all the data for each selected platform that falls between these times. The number of files and size of the data will appear in the upper right corner.
Modifications to the selected data can be made. Entire platforms may be un-selected by clicking on the `select' button to the left of the platform name and then clicking on `Apply to selected platforms'. Individual files can be selected by clicking on the platform name. Finally, large data sets such as satellite or model data can be selected at specified intervals by clicking on the `only select every' button. After selecting the data to be moved to the local disk click on `Load the data'. A widget will pop up that gives the status of the load. These last two widgets go away when the load is complete. You can pop up the `Load Data' widget and look at the Data Available column to verify that the load was successful and `Quit' from the disk manager window when finished.
The field Data Directory is both the name of the platform and the name of the subdirectory where the data files are stored. A Source is given for each data set. Format describes the format in which Zebra stores the data. Most data sets have been converted to this form by a Zebra ingest module written specifically for that purpose.
Additional processing information is given in this section, telling what the original format of the data was, and how it was converted to Zebra format. Other information about the data, and acknowledgments also appear in this section.
If you have questions or comments about TCWIDS or Zebra contact: rdp-support@atd.ucar.edu
The GMS Honiara dataset was collected by the NCAR Research Data Program during the COARE intensive field phase. Data arrived in Honiara via MCIDAS and was converted from their area file format to Zebra's raster file format. Counts were calibrated to IR temperature values during this process.
The GMS Honiara dataset was collected by the NCAR Research Data Program during the COARE intensive field phase. Data arrived in Honiara via MCIDAS and was converted from their area file format to Zebra's raster file format. Counts were calibrated to IR temperature values during this process.
This GMS dataset was produced by Pierre Flament and R. Bernstein and published on CDROM. Their "makebin" program was used to convert the PostScript files to calibrated binary; the result was incorporated into the Zebra raster file format with no change in the pixel values.
NOAA -11, 12 satellite images (AVHRR SST) were collected at Townsville, Australia (Emery) and reprocessed by Gary Wick, William Emery, Phil Mislinski, and Keith Cherkauer at the University of Colorado - Boulder, CO USA. These data are currently being reprocessed using an "improved" cloud filter and new composites may be available in the future.
An abstract of a presentation by G. Wick, Science Groups: Oceans and Air-Sea Fluxes gives more information about this data set.
Full resolution AVHRR data of the TOGA COARE domain was collected from a ground station in Townsville, Australia throughout the experiment and processed into maps of sea surface temperature (SST) at the University of Colorado. One-km resolution HRPT data from NOAA-11 and 12 was collected between October 1, 1992 and March 30, 1993 at a ground station operated by the Northeast Australian Satellite Imagery System (NASIS) at James Cook University in Townsville, Australia. Data was obtained from both morning and evening passes of NOAA-12 and nighttime passes of NOAA-11. The ground station in Australia provided coverage of about 2/3 of the COARE domain defined as from 10 N to 10 S and 140 to 180 E. Only the extreme northeast portion of the domain lacks satellite coverage. Additional data was obtained from sites in Guam and Kwajalein in an attempt to fill in the rest of the domain but limited data and data format problems have limited the contribution of this data.
The sea surface temperature dataset was obtained from Gary Wick of the University of Colorado. Each grid was converted from two-byte integer to floating-point format and stored in netCDF.
Platforms ssmi_s2, ssmi_s4, and ssmi_s5 contain antenna temperatures only. Platforms ssmi_s2_tb, ssmi_s4_tb, and ssmi_s5_tb contain brightness temperatures and rainfall rates. The s2 refers to the satellite `craft id' 2 aka F-8, s4 refers to the satellite `craft id' 4 aka F-10, and s5 refers to the satellite `craft id' F-11. Only satellite scans within the general TOGA COARE region were included in the files.
For each satellite, temperatures are calculated for each of the 7 channels of the microwave imager:
`Low' frequencies, A scans only, 64 cells per scan:
-----
ta19v Antenna temp., 19 GHz vertical polarization
ta19h ...and so on...
ta22v
ta37v
ta37h
`High' frequencies, A and B scans, 128 cells per scan:
-----
ta85v
ta85h
Likewise the low-frequency brightness temperatures fields are named tb19v, tb19h, tb22v, tb37v, tb37h and the high-frequency brightness temperatures: tb85v, tb85h.
Rainfall rates, field `rr', are derived using Liu and Curry 1992: JGR,97,9959-9974.
if TB37V - TB37H < 38 and TB19H - TB85H + 88 > 0
RR=5.5e-3*(TB19H-TB85H+88)**1.6,
else
RR = 0
RR is the rainrate in mm/hr, and the TB variables are the brightness temperatures in Kelvin for the 85GHz, 37GHz and 19GHz channels. H and V indicate the Horizontal and Vertical polarizations. The algorithm only works over the ocean surface, but no distinction of surface type is made when calculating RR. Consequently, an outlandish value for rainrate usually indicates either one of the temperatures is bad or that the cell is located over land. Since the calculation of rainrate depends upon low-frequency temperatures, the rainrate is only calculated on the 64-cell low-frequency grid of the A scans, yielding a resolution of about 25 km by 25 km.
The SSMI ingestor works by compiling consecutive scans into one set, verifying that the scans are close to each other in time. The lat/lon location of each cell of each scan is converted to coordinates on a km grid whose origin is 9.42 degrees south latitude and 159.97 degrees east longitude (the origin of the coare CAP plots). A grid is designed according to the range of the scan cells and the Resolution parameter (12.5 km). Since the high and low frequencies have different coverage, two different mappings are generated for mapping scan cells coordinates into kilometer grid points. A grid point receives the closest cell within a specific radius. (The Radius parameter used for this dataset is 50 km.) For each of the high- and low- mappings, the cell temperature values are filled into the grid. The mapping does not include any interpretation or analysis. Grid points which do not fall into the range of the mapping are filled with zero. The scale and offset for compressing the temperatures (K) into bytes are 1 and 100, respectively. So fill values will uncompress temperatures of 100 K. As long as the center and step of a raster plot do not include 100 K, the raster plots will give a good representation of the satellite scans (though the black fill areas in the rectangular images will still cover any plots beneath the image). Because of the radius limit, a series of bad scans will appear broken up.
The parameters used to ingest this dataset, taken directly from the tcl script:
set Format level1b ;# "rss", "level1b", or "1b"
set Resolution 12.5 ;# kilometers
set MinScans 32 ;# min number of scans to make an image
set DisableLimits true ;# take every scan in the file
set Radius 50.0 ;# kilometers
Origin -9.42 159.97 ;# use the same origin for all images
The 85 GHz channels on F-8 are frequently bad and so raster plots will show up as a solid color. Even though all of the images are created with 12.5 km resolution, only the 85 GHz channels are actually sampled at that resolution. The low-frequency cells are about 25 km apart.
Configuration notes:
There is a module file, `ssmi', in the modules directory, and a generic ssmi
plot component, `tc_ssmi', in the pd directory. The module file only
contains a field menu, and the plot description is similar to p_raster except
for setting the middle menu to the field menu in the `ssmi' module file. The
IconBar includes three submenus under the satellite icon menu, one for each
satellite. Each submenu selects one of the fields for one of the three
satellites. Icons for each of the satellite names are in the `icons'
directory.
The defaults pd (pd/defaults.pd) contains some generic center-step defaults for each of the fields. These are not tailored to each specific frequency or polarization, consequently some of the plots may show up with gaps from out-of-range values.
A simple display configuration is in dconfig/ssmi.dc. Enter `display ssmi' at the dm> prompt to see it.
The source for the ingestor can be found in zeb/src/ingest/ssmiplus in the Zebra source distribution. This is the README file from that directory. The directory also contains an example of a script used to process the data files, called ssmi.tcl.
Our thanks go to Michael Goodman for making this data set available.
The NASA flight-level dataset was obtained from the University of Washington and came to us in ascii format. The ascii files included both INS and GPS time and location fields. GPS fields were used when converting these files to netCDF. In certain fields the units were converted; knots -> m/s, feet -> meters, feet/min -> m/s, etc. to be more consistent with other data sets.
Our thanks go to Pat Hrubiak for making this data set available.
The NASA flight-level dataset was obtained from the University of Washington and came to us in ascii format. The ascii files included both INS and GPS time and location fields. GPS fields were used when converting these files to netCDF. In certain fields the units were converted; knots -> m/s, feet -> meters, feet/min -> m/s, etc. to be more consistent with other data sets.
Our thanks go to Pat Hrubiak for making this data set available.
Electra flight-level data was produced by the NCAR Research Aviation Facility in GENPRO format. A subset of fields was selected from this dataset and converted over to netCDF.
Electra flight-level data was produced by the NCAR Research Aviation Facility in GENPRO format. A subset of fields was selected from this dataset and converted over to netCDF.
These P3 flight data were produced in the field by NOAA MRD. These images have been converted to the Zebra netCDF file format.
Our thanks go to John Daugherty for making this data set available.
The C-130 flight tracks came to us through the University of Washington.
These are the first-release version of the processed "slow" (1
second/sample) data set from flights by FIAMS' Cessna research aircraft
during the TOGA COARE experiment.
PLEASE NOTE that intercomparisons of the cessna with other platforms is not
yet complete, and consequently this data is subject to alteration as further
results come to hand from those continuing efforts. It was felt, however,
that this data set was in a suitably fit condition for release at this stage,
as a service to the general TOGA COARE community.
A more final version of this data set is expected to be released after the intercomparison efforts are fully complete, sometime in the middle of 1995.
A detailed description of the full data set, including a "road map" of the flights and runs contained therein, information about the aircraft, its systems and instrumentation, and comments on calibrations, processing steps, failures and the form of the data files, can be found in the following document:
"Description of Instrumentation and Flights Conducted with FIAMS' Cessna 340A Aircraft EOS During the International Experiment TOGA COARE", Technical Report No. 10, FIAMS, August 1993.
If you do not already have a copy of this report, please contact A.G. Williams and J.M. Hacker at:
Flinders Institute for Atmospheric and Marine Sciences
(FIAMS)
School of Earth Sciences, Flinders University
G.P.O. Box 2100, Adelaide 5001, Australia
The channels used are:
temp - air temperature (Celsius)
altrad - radar altitude (m)
altpres - pressure altitude (m) [ref: 1013.25hPa contour]
lat - latitude (decimal degrees)
lon - longitude (decimal degrees)
static - air pressure (hPa)
tas - true air speed (m/s)
rho - air density (kg/m^3)
uair - east/west wind component (m/s) [+ve east]
vair - north/south wind component (m/s) [+ve north]
wair - vertical wind component (m/s) [+ve up]
theta - potential temperature (Celsius)
sphum - specific humidity (g/kg)
tempsf - radiometric surface temperature (Celsius)
Sin - incoming short-wave radiation flux (W/m^2)
Sout - outgoing short-wave radiation flux (W/m^2)
Lin - incoming long-wave radiation flux (W/m^2)
Lout - outgoing long-wave radiation flux (W/m^2)
RNet - net radiation flux (W/m^2)
The "tempsf" channel (radiometric surface temperature) has NOT been corrected for effects of reflected incoming long-wave radiation. An algorithm is currently under development to correct for this effect, so that a corrected version of "tempsf" may become available in the future.
Although this data set is essentially public domain, FIAMS requests that all users keep them informed as to the type of study they are engaged in, so that they can keep you informed regarding latest revisions (adjustments, bugs found and so on) to the data set. They also request that reference be made to the origin of this data set in any subsequently published papers.
During processing the pressure altitude was used as the altitude field.
Our thanks go to Alistair Williams for making this data set available.
The P3 composite flight-level radar images were produced in the field by NOAA MRD. These images have been converted to the Zebra raster file format. These images are in the process of being updated and new composites will be available from NOAA MRD sometime in the future.
Our thanks go to John Daugherty for making this data set available.
The MIT and TOGA radars were processed into the Zebra raster image format by NASA/TRMM using the Zebra rasterizer subsystem using source data from Paul Kucera. Only the 0.8, 1.3 and 1.8 degree tilts were selected out of the source data. There are still some problems with the Cruise 1 data included in this data set so these should be used with caution. Updated Cruise 1 data should be available in a future release.
Our thanks go to Paul Kucera for making this data set available.
The MIT and TOGA radars were processed into the Zebra raster image format by NASA/TRMM using the Zebra rasterizer subsystem using source data from Paul Kucera. Only the 0.8, 1.3 and 1.8 degree tilts were selected out of the source data.
Our thanks go to Paul Kucera for making this data set available.
The ISS omega sounding and surface datasets were produced in netCDF format by the NCAR Surface and Sounding Systems Facility and are displayed directly by Zebra.
Our thanks go to Erik Miller for making this data set available.
These data are available from NOAA/PMEL via anonymous FTP and come in a variant of netCDF substantially different from that used by Zebra. We would like to thank the TAO project office at PMEL, Dr. Michael J. McPhaden, Director and Nancy Soreide for making these data available. These data may be subject to further editing and re-released by PMEL.
Two separate TAO datasets are part of this release. The platform called simply "tao" consists of 24-hour averaged data at the sea surface. The platform "tao_hr" contains hourly data; winds data are a six-minute average centered on the hour, while the temperature and humidity data are an average of six, ten-minute samples over the previous hour. Missing data are given the fill value 1.0e35 for all variables. The time words indicate the center of the averaging interval.
If you use TAO data in publications, please acknowledge the TOGA-TAO Project Office, Dr. Michael J. McPhaden, Director. Also, the TAO project office would appreciate receiving a preprint and/or reprint of those publications utilizing TOGA-TAO data for inclusion in the TAO bibliography. These publications should be sent to:
The TOGA-TAO Project Office
NOAA/Pacific Marine Environmental Laboratory
7600 Sand Point Way NE
Seattle, WA 98115
The ascii file WHOIbuoy1b.dat was obtained via anonymous ftp from Woods Hole and converted to netCDF.
Our thanks go to Robert Weller for making these data available. The following is a reprint of the readme file that accompanies WHOI data.
FILE WHOIbuoy.dat - ASCII (from togaflux5b; vawrpost26.epic)
Release 1.0 August 3, 1994
Release 1.0b August 8, 1994
This data was collected with support from the U.S. National Science Foundation, grant OCE91-10559. If you use any of these data in a publication, please acknowledge Dr.Robert Weller and the Woods Hole Oceanographic Institution when using the data.
The data is from the Woods Hole Oceanographic Institution surface mooring deployed in the center of the IFA during the COARE IOP. (156 E, 1 45 S).
We gratefully acknowledge Clayton Paulson, (OSU,USA), Chris Fairall (NOAA/WPL, USA) and Mike McPhaden (NOAA/PMEL, USA) for sharing their observations.
Release 1.0b August 8, 1994
Notes:
The wind speed is now the measured wind speed rather than the relative as
given in the previous release.
The stmp is now the measured temperature at .45m depth. The Tskin is the
sea surface skin temperature and is now an additional output.
The TOGA COARE Bulk Flux Code vers.2.0.b is used. The output of Tskin is
corrected in this version to include the cool skin which was previously not
included.
Release 1.0 August 3, 1994
Notes:
WARNING this is a preliminary processed file and should be used with caution.
Please forward an email message to S. Anderson or R. Weller when you have
picked up the data. We will place you on our mailing list and notify you
when the file is updated.
The file is uniformly spaced in time.
The surface fluxes are estimated using the TOGA COARE Bulk Flux Code Ver.2.0.
(C. Fairall, E. Bradley, D. Rogers, J. Edson, G. Young (1994)). The warm
layer and cool skin is from C. Fairall, E. Bradley, J. Godfrey, G. Wick,
J. Edson, and G. Young (1994).
The rain data comes from a combination of observations from various surface
observing platform. These platforms include the WHOI surface buoy, R/V
Wecoma, R/V Moana Wave and the ATLAS buoy at 2deg South. The wet bulb
temperature is used as an estimate of the rain temperature. The rain data has
been block averaged over 24 hours.
A four day gap 12/9/92 00:33:45 to 12/13/92 05:11:45 was filled with hourly
data from Mike McPhaden's (PMEL) nearby ATLAS buoy (156 E, 2 S) that was
regridded to our sample interval (wind, rh, sst, atmp). The previous four
days of SW and BP were duplicated and patched into the gap as a temporary
solution for these missing variables.
Some gaps were present and the time series was linearly interpolated in time
to fill the gaps.
Wind velocities are relative the to sea surface and calculated by subtracting
the surface currents from the observed winds. Wind vectors are given as
oceanographic convention (direction towards).
The sea surface temperature is estimated using the COARE Bulk Flux algorithm
which include a warm layer extrapolation of the measured sea surface
temperature (at 0.45m) to the surface and a cool skin correction.
Magnetic Variation correction has been applied = +7.67 degrees. Averaging
Interval 60 minutes.
Latitude = -1.755540 Longitude = 155.994507
Output file processing depth bins
start rec = 1 0.000000m
end rec = 1 0.000000m
Output file processing records
start rec 10/21/1992 19:33:45
end rec 03/04/1993 00:03:45
The following are positive for ocean cooling.
qlat W/m**2 Latent heat flux
qsen W/m**2 Sensible heat flux
qlw W/m**2 Net Longwave radiation
The shortwave radiation is positive for ocean heating.
qsw W/n**2 Net Shortwave radiation
taue N/m**2 East component of wind stress
taun N/m**2 North component of wind stress
rain mm/hr rainfall rate
Train degC rainfall temperature (wet bulb temp)
atmp, degC, air temperature (adjusted for solar heating error)
time, UTC, Time at beginning of the averaging interval
wnde, m/s, relative wind east component
wndn, m/s, relative wind north component
hrh, percent, relative humidity
bpr, mb, barometric pressure
stmp, degC, sea surface temperature (measured)
Tskin, degC, sea surface skin temperature (from Ver.2 code)
Measurement heights above the sea surface:
wind cup 3.54m
vane 3.26m
hrh 2.74m
atmp 2.78m
bpr 3.00m
stmp 0.45m
swadj2 3.54m
lwadj 3.54m
Contact Dr. Robert Weller or Dr. Steven Anderson if you have questions.
Dr. Robert Weller
Woods Hole Oceanographic Institution
Woods Hole, MA 02543
(508)457-2000 x2508
rweller@whoi.edu R.WELLER omnet
Dr. Steven Anderson
Woods Hole Oceanographic Institution
Woods Hole, MA 02543
(508)457-2000 x2876
sanderson@whoi.edu S.ANDERSON omnet
ASCII ship position files are available via anonymous ftp from the COARE office. They have been converted to netCDF with no modification.
These data were given to us at the TOGA COARE workshop by Peter Coppin, and converted to netCDF.
This data set was provided by Chris Fairall and converted from an ascii file with no modifications or thresholding.
ASCII ship position files are available via anonymous ftp from the COARE office. They have been converted to netCDF with no modification.
Our thanks go to Robert Pinkel for making this data set available.
These CTD data were collected during the JAPACS (Japanese Pacific Climate Study) COARE cruise in February 1993 by the R/V Natsushima with a Seabird SBE 9/17 CTD (Seabird Electronics, Seattle).
CTD observations were performed every 6 hours while on station (156E 0N) and every 1 degree of latitude along the ship's track. The sensors were calibrated before and after this cruise by SBE. The drift of the sensors was small. Although data processing was performed to reject noise, the surface data (0-5 meters) contain noisy salinity data.
Our thanks go to Kentaro Ando for making this data set available.
Seventeen casts were made 14-24 February 1992 with a SEACAT PROFILER SBE19.
Our thanks go to Gavin Fisher, Peter Isaac and Robert Knobben for making this data set available.
If you use these data in any publications please acknowledge Marie-Helene Radenac, groupe FLUPAC, ORSTOM Noumea.
Our thanks go to Dr. Radenac for making these data available.
Our thanks go to Roger Lukas at the University of Hawaii for making these data available.
CTD data from the R/V Le Noroit were gathered along 18 repetitive sections at 156E, from 5N-5S, during December 7, 1992- February 24, 1993. CTD observations were made with a SeaBird SBE-9 probe, down to at least 1000 dbar, and processed to a 2-dbar vertical resolution. Calibration is based on Laboratory calibration of sensors (SeaBird Inc. at Seattle) and rosette samples.
Our thanks go to Thierry Delcroix and Joel Picaut for making this data set available.
CTD casts were taken with a Seabird CTD SBE 9/11. We have used CTD data collected on three legs. Leg 1 ran from Nov. 8 to Dec. 8 1992, from Pohnpei to Guam and CTD casts were taken daily, usually at approximately 1000 local (2300 Z). Leg 3 ran from Jan. 22 to Feb. 22 1993, from Guam to Guam. Casts were taken daily, usually at approximately 1300 local (0100 Z).
Our thanks go to Lisa Bogar for making Legs 1 and 3 available. Dr. Mike Gregg, from the Applied Physics Laboratory at the University of Washington, was the principal investigator for Legs 1 and 3, and the work was funded by ONR.
Our thanks go to Dr. James Moum of Oregon State University for making Leg 2 available. Leg 2 ran from Dec. 12, 1992 to Jan. 16, 1993, from Guam to Guam.
The following is an excerpt from the file mwlegs13_ctd.readme:
Both our raw Sea-Bird CTD files and "semi-processed" CTD data are
available. We have no plans currently to process the data any further. The
"semi-processed" files were used to quickly check the calibration
of our microstructure profiler temperature and conductivity data.
The CTD files need further processing before any analysis takes place.
We [University of Wash.] would suggest the following procedure:
1. Smooth the pressure with a filter such as a 1 second Bartlett filter.
2. Find the appropriatfe number of scans to advance the conductivity in order
to minimize salinity spiking. You can do this by determining the frequency
at which the phase spectrum of the temperature and conductivity drop to -180
deg. Use phase spectra which use relatively homogeneous water.
3. Correct for the differences in probe response time constants. Try
examining the conductivity filtered with doublepole and single recursive
filters and pick the filter with the most stable density profile.
4. Correct for the conductivity cell thermal mass error. We have done this
in the past by empirically adjusting the values of the Lueck-Picklo model
(JOAT, 1990) to close the T-S trace difference of an up-down cast.
5. Edit pressure loops. We have followed Giles & McDougall
(DSR,1986,33(9), 1253-1274) by passing the though the data once to find the
mean pressure values of the reversals and then passing the data through again
to cut the file at the pressure the first time it is reached.
6. Grid the data. We have used a least-squares fit to achieve a 20-cm grid
in the past.
Our thanks go to Ming Feng at the University of Hawaii for making this data set available. Please contact Ming Feng (mfeng@iniki.soest.hawaii.edu) if you use these data in a publication.
The NMC model output was obtained in GRADS format from Mike Fiorino of Lawrence Livermore National Laboratory. The GRADS data is read directly by Zebra and was not modified.
There are nine fields; height, u wind, v wind, temperature, relative humidity, surface pressure, precipitable water, u wind at surface, and v wind at surface.
Twelve hour accumulated precipitation during the 0 - 12 MRF forecast.
Twelve hour accumulated precipitation during the 12 to 24 MRF forecast.
Sea level pressure was calculated hypsometrically from the 1000 mb geopotential heights and temperature.
There are three fields; chi (m**2/sec), u and v.
The FNOC model output was obtained in GRADS format from Mike Fiorino of Lawrence Livermore National Laboratory. The GRADS data is read directly by Zebra and was not modified.
There are four fields of data; total cloud, low cloud, mid cloud, and high cloud.
There are two fields; sst and ocean mixed layer depth.
There are two fields; total precipitation and conv. precipitation.
There are two fields; total precipitation and conv. precipitation.
One field containing the sea level pressure analyses.
There are six fields; slp, u component, v component, temperature, geopotential height, and q (kg water / kg air * 1000).
There are two fields, u wind and v wind.
There are three fields; chi (m**2/sec), u and v.
There are two fields; u wind component and v wind component.
Potential users of this data should read the TOGA COARE Integrated Sounding System Data Report - Volume I prepared by Erik Miller, National Center for Atmospheric Research, Boulder, CO 80307. That document contains a full description of the profiler system, data collection and data analysis procedures.
Profiler data products
Profilers were operated for TOGA COARE at six ISS sites. Four on land at Kavieng, Kapingamarangi, Manus and Nauru, and two at sea on board the R/V Kexue #1 and Shiyan #3. In addition there was a profiler aboard the R/V Moana Wave. Profiler data products have been prepared for all these sites throughout the IOP. For Manus additional data has been prepared for the 4 months preceding and following the IOP. The same additional periods were prepared for Nauru but most of the usable data came in June 93. In addition data sets obtained with the National Oceanic and Atmospheric Administration's 50 MHz profilers operating in a high height coverage mode at Biak and Christmas Island have been processed in the same manner from July 92 through June 93.
Half Hour Wind Profiler Data
The files came in ASCII format and were converted to netCDF. The ASCII files include a quality number Q1 which serves two different purposes. Normally, like all other quality flags, it is zero. However, if all the data in this record came from quick look data files, then it is incremented by 1. On shipboard systems it is incremented by 2 if the latitude and longitude were interpolated over an interval such that there might be an error in the least significant digit or it is incremented by 8 in circumstances where a larger error is possible. During reprocessing values with a quality flag Q1 >= 7 were set to the missing data value.
Throughout the table 9999 is used to flag missing data. When the data comes from post processing, missing data will be associated with a count of zero items in the average. However, when quick look data are used, some columns can be missing data while others contain valid data on the same line.
Half hour RASS data
Processing of the RASS data was similar to that of Profiler Data. The ASCII files also contained a quality flag, and values where Q1 >= 7 were set to the missing data value.
Other volumes in that set, Volumes X through XVIII, give further details pertinent to individual sites and show daily plots of the available data.
Our thanks go to Tony Riddle for making these data sets available.
See Volume X, Biak for more information about this 50 M Hz profiler.
See Volume XI, Christmas Island for more information about this 50 M Hz profiler.
See Volume XIII, Kapingamarangi for more information.
See Volume XII, Kavieng for more information.
See Volume XIV, R/V Kexue #1 for more information.
See Volume XV, Manus for more information.
See Volume XVI, R/V Moana Wave for more information.
See Volume XVII, Nauru for more information.
See Volume XVIII, R/V Shiyan #3 for more information.
The ISS sounding data was gathered, quality-checked, and reformatted into CLASS format by Erik Miller. The result was then converted to netCDF.
The dropsonde data was gathered, quality checked, and reformatted into CLASS format by Erik Miller. The result was then converted into netCDF.
The ERS-1 "value added" wind dataset was obtained from Tim Liu.
The winds data were converted to netCDF grid format through the following
process:
- Data from all consecutive passes over the COARE area were accumulated;
- A Barnes analysis was performed to convert the data into a regular grid;
- The grids were stored in netCDF files.
The non-iss soundings were not available at this time.
Zebra is available in both source and binary form. Binaries for Zebra are
provided for four combinations of architecture and operating system:
Sun SPARC architecture and SunOS 4.1.x
Sun SPARC architecture and Solaris 2.3 (SunOS 5.3)
HP 700 series architecture and HP-UX 9.0
SGI Indy architecture and IRIX 5.2
ris