The YYYYMMDD-YYYYMMDD-ESACCI-L3C_AEROSOL-AEX-GOMOS_ENVISAT-AERGOM_5days-fv3.00.nc files contain the binned stratospheric aerosols properties derived from the AerGom algorithm using GOMOS v601 data. - AOD and extinction are now given at 355, 440, 470, 550 and 750 nm - PSCs have been flagged based on PSC I(NAT) formation temperature - Angstrom coefficient for extinction and AOD [400 - 800]nm - size related parameters are now provided, see details below CHANGES ------------ *02.07.2016* Version 3.00 is the first effective GOMOS-AerGOM CDR including size information. It provides GOMOS products on a grid of 5 degree latitude x 60 degree longitude x 1 km altitude x 5 days time interval. Event selection criteria have been modified, and are currently: Star magnitude < 3 Solar zenith angle > 100 degree The extinction and AOD are provided at 355, 440, 470, 550 and 750 nm. The extinction is calculated from the retrieved spectral parameterization. There is no particular change in the extinction retrieval by AerGOM. Besides the total extinction, extinction time series are now also provided at the same wavelengths for the sulfate aerosol fraction and for polar stratospheric clouds. Size related parameters are the total particle number density (in cm-3), the mean radius (in micron) and the variance (in micron2) of the particle size distribution, the surface area density (in micron2 .cm-3), the volume density (in micron3.cm-3), the effective radius (in micron), the single scattering albedo and the asymmetry parameter. All these parameters are computed from the retrieved sectional particle size distribution (20 size bins centered on size values from 0.02 to 3 microns) using Mie theory. All size-related quantities are provided for the sulfate aerosol fraction and for the polar stratospheric clouds, over an altitude grid similar to the one used for the extinction profiles, but limited to 35 km as maximum height. All products are provided with their uncertainty computed as the average of all values between percentiles 10 and 90; if less than 10 events are available, all the values are used for the averaging. As previously, a number of diagnostic files are provided: the number of events used for the computation of the extinction (including the number of polar stratospheric cloud occurences) and the median tropopause height. This first version of the particle size distribution should be considered an experimental product. This version has been provided to the user on August 2nd, 2016, and uploaded on ICARE on October 3rd, 2016. *30.05.2016* Version 2.21 is the very first version including size information derived from GOMOS-AerGOM extinctions. This version presents errors and is replaced by version 3.00 . *28.05.2015* Version 2.19 is very similar to version 2.15, but will contain more data due to a relaxation of the selection parameters for the profile regarding the illumination conditions: for v2.15, only profiles in dark limb conditions were processed, whereas in v2.19, some profiles will be in twilight/straylight conditions. To be more precise, only profiles where the mean SZA of the measurements was larger than 105 deg. were selected. All other selection criteria remained the same as for v2.15. The effect will mainly be seen at high latitudes, where the coverage was sparse before. Keep in mind that that these newly added profiles may exhibit strange behavior due to the presence of straylight. Therefore, v2.19 should be considered an experimental product. *23.04.2015* Version 2.15 is an experimental product that uses different binning parameters (5° latitude x 60° longitude x 1km altitude x 5 days) than those previously used (2.5° latitude x 10° longitude x 1km altitude x 1 month) and which could be useful for model studies. This version contains data produced by the same AerGom algorithm as v2.1, and the entire available GOMOS time series has been processed, spanning the years 2002 - 2012. Note that since the processing was done on a yearly basis, it may be that the last file of a given year does not contain 5 days worth of data. *23.04.2015* Files version 2.14 contain data produced by the same AerGom algorithm as v2.1 but were produced in a separate processing. The entire available GOMOS time series has been processed, spanning the years 2002 - 2012. The file, variable and dimension names in the netCDF files have been partly modified to conform with the CCI and CF conventions. Here is an overview of the changes, presented as 'New_Name' ('old name'): - file naming: '200212-ESACCI-L3C_AEROSOL-AEX-GOMOS_ENVISAT-AERGOM_MONTHLY-fv2.14.nc' ('200803000000-ESACCI-L3C_AEROSOL-AEX-GOMOS_ENVISAT-AERGOM_MONTHLY-fv2.03.nc') - dimensions: 'altitude' ('alt'); 'latitude' ('lat'); 'longitude' ('lon'). - 'S_AOD550' ('aod550'); 'S_AOD550_uncertainty' ('aod_uncertainty550'); S_AOD550_Ndata ('aod_Ndata550') - 'AEX550' ('Aerosol_extinction550'); 'AEX550_uncertainty' ('Aerosol_extinction_uncertainty550'); 'AEX550_Ndata' ('Aerosol_extinction_Ndata550') - 'ANG400-800-AOD' ('Angstrom_AOD_400_800'); 'ANG400-800-AOD_uncertainty' ('Angstrom_AOD_400_800_uncertainty'); - 'ANG400-800-AEX' ('Angstrom_EXT_400_800'); 'ANG400-800-AEX_uncertainty' ('Angstrom_EXT_400_800_uncertainty'); The new global attribute 'inputFileList' has also been added to the netCDF files, containing a string that identifies Level 2 files used to calculate the values in each Level 3 file. *11.06.2013* Files version 2.1 contain data produced by the AerGom algorithm. New variables have also been added to the netCDF files including "Number of Observations" for the extinction and the AOD, as well as the PSC occurrence frequency. *29.10.2012* Files version 2.02 fixes some problems with the naming convention and some netCDF global attributes ('project' is no more absent and the institution has been updated). *01.07.2012* Files version 2.01 now follow the guidelines concerning the naming conventions, as currently agreed upon by the Aerosol_CCI consortium. *01.05.2012* The new version (2.0) of the dataset reflects the use of the long-awaited GOMOS data v6.01 as an input for the AERGOM retrieval. This permits the retrieval of the AOD and extinction at 550nm (compared to 500nm before). Because the new data allows a retrieval of the extinction/AOD at various wavelengths, this information is used do derive the Angstrom coefficient, both for the extinction and the AOD. The Angstrom coefficient is computed using a least-square fit on a log/log plot of the extinction and wavelength between 400 and 756nm, inclusively. As before, the PSCs have been filtered based on PSCI(NAT) formation temperature. The AOD is calculated by integrating the extinction between tropopause height+2km to 50km. The tropopause height calculation has been improved, using the dynamical tropopause at mid- and high-latitudes and thermal tropopause elsewhere. A bug has also been fixed in the thermal tropopause calculation, which will probably improve the validation with CALIPSO. Another change to the data is the error calculation of the AOD and extinction, which is now simply computed as the (geophysical) standard deviation of the P25-P75 data throughout the month. This will be a better reflection of the state of our knowledge of the data, rather than the error of the monthly mean. The data vertical grid is now finer, with a 0.5km increment. There is no filtering of cirrus clouds, but they should not contribute to the AOD either assuming that the tropopause height is accurate. *29.08.2011* In addition to the data already present in the v1.0 of the product, two additional variables have been added to the file: - Cirrus_fraction - PSC_fraction In the case of GOMOS, the cloud coverage should be understood at the number of cirrus clouds/PSC clouds detection divided by the total amount of observations, in a given bin during the month. These variables are to be used with care since they are not very precise, nor have they been verified. They are provided as indication only. > The PSC detection is currently not based on any GOMOS measurements but on the comparison of PSC > I(NAT) formation temperature and ECMWF data. > The cirrus detection is based on a purely empirical and arbitrary approach, probably not very > accurate: if the extinction in the lower portion of the atmosphere (0-20km) reaches values equal or > larger than 25E-3 km-1, the profile is assumed to contain a cirrus clouds. *06.06.2011* The files have been updated so that the array dimensions follow the guidelines of Aerosol_CCI, namely that the arrays are in [lat, lon] format, and not the contrary. The 3-dimensional variables have also been changed so that their dimensions now have the format [lat,lon,alt]. Some errors in the data representation shave been fixed.