Please direct any question about this data set to Jennifer Logan, jal@io.harvard.edu, 617-495-4582. Send an e-mail to jal@io.harvard.edu if you download any of these data. You will be informed of any updates. Please cite the following papers as the source of this dataset. J. A. Logan, An analysis of ozonesonde data for the troposphere: Recommendations for testing 3-D models, and development of a gridded climatology for tropospheric ozone, J. Geophys. Res., 104, 16,115-16,149, 1999. J. A. Logan, An analysis of ozonesonde data for the lower stratosphere: Recommendations for testing models, J. Geophys. Res., 104, 16,151-16,170, 1999. The original data were obtained from the World Ozone and Ultraviolet Data Center, Downsview Ontario (http://www.tor.ec.gc.ca/woudc), Sam Oltmans (NOAA/CMDL), Volker Kirchhoff (INPE, Brazil), Greg Bodeker (Lauder, New Zealand), the NASA/GTE data archive, and the microfiche supplement to Chatfield and Harrison [J. Geophys. Res, 82, 5969-5976, 1977]. I selected 1980-1993 as the base period of analysis, but not all sonde stations operated during this period. Earlier data were included to increase the geographical coverage, especially in the tropics, and both earlier and later data were included to increase the number of soundings. A postscript file (number.sondes.ps) summarizes the number of soundings each month at each station at 500 mbar. Note that there are often fewer soundings by 10 hPa. To be included in the analysis, each sounding had to meet the following ranges for the correction factors: ECC, KC (Japanese) sondes, 0.8-1.2 Brewer Mast, Indian sondes, 0.9-1.35. For a discussion of the correction factors see Logan [1985, 1994]. (Note that the header for the ozone files includes the correction factors. The range may appear as 0.8-1.35 if the station changed from BM to ECC sondes; the analysis code used the correct CF range for the period of each type of sonde.) Ozone concentrations for each sounding that met the correction factor criteria were interpolated to give values at 22 pressure levels from 1000 to 10 mbar. Tables were compiled for each station containing the monthly mean, the standard deviation, and the number of soundings that went into the mean. pub/sonde/profiles contains these results for each ozonesonde station on pressure levels. Results relative to the tropopause height may be found in the directory pub/sonde/tropp There is one file for each station, containing the average concentrations (nbar), the number of values that went into the mean at that level, and standard deviations (nbar) for each of 22 levels. The 22 pressure levels in mbar are: 1000 900 800 700 600 500 400 300 250 200 150 125 100 90 80 70 60 50 40 30 20 10 Ozone is given in units of nbars, as partial pressures, the original format in which it was obtained. ****To convert to ppb, multiply by (1000/[Pressure level (in mbar)]).**** DO NOT FORGET TO CHANGE UNITS IF YOU WANT PPB OZONE. This has been the most common error with previous users of these files. The data are arranged according to numerical codes designated by WMO. Stations that did not have a WMO code were given a three digit code starting with 500 or 800. NOTE that tropospheric values are omitted for 3 stations. For Lindenberg (80-93) and Legionowo (80-93) tropospheric levels below 300 mbar are omitted, since the OSE sondes (used until 92-93) systematically overpredicted tropospheric ozone. To prevent anyone thinking these values were valid, I simply omitted them. For Payerne (156), there are anomalously high values for tropospheric ozone in the early 1990s (see Logan (1994). For this reason I recommend use of the file for 80-89 for the troposphere. The file for 80-93 omits values below 300 mbar. Further documentation is given in the postscript files described below. =================== sonde.table.ps =================== This postscript file is a table that lists the WMO station code, station name, latitude, longitude, altitude, sonde type, number of acceptable soundings, and dates used for each ozonesonde station. ============= station.codes ============= This ascii file contains the WMO station code, latitude, longitude, and station name for each data file in this directory. ==================== sondeXXX.aa.bb.cc.dd ==================== The ascii files named sondeXXX.aa.bb.cc.dd contain ozonesonde data for WMO station XXX for the dates aa/bb (month/year) to cc/dd. These files contain 4 header lines, followed by 3 fields of data, each with a one line header. The first line of the file gives the WMO station code, the station name, the limits of acceptable correction factors, and the dates within which data were used. The second line gives the column headings, and lines 3 and 4 are blank. The first field of data contains monthly mean ozone (in nbar) for 22 pressure levels. The second field of data gives the number of soundings that went into each monthly mean. The third data field gives the standard deviation of each monthly mean. +++++++++++++++++++++ read_sonde.tar.Z +++++++++++++++++++++ This contains IDL routines to read the sonde data files. It was written by Martin Schultz (mgs@io.harvard.edu), and any questions about the routine should be directed towards Martin. I do not use IDL. --------------------------------------------------------------------- Jennifer A. Logan Phone: 617-495-4582 Harvard University Fax: 617-495-9837 108 Pierce Hall 29 Oxford Street e-mail: jal@io.harvard.edu Cambridge MA 02138 U.S.A. ----------------------------------------------------------------------