Remote Sensing of the Stratosphere: Recent Publications

The database of MIPAS level-2 products retrieved with the 2-dimensional GMTR algorithm (MIPAS2D) is made available at MIPAS Bologna Facility to promote the use of MIPAS products and stimulate scientific collaboration.

MIPAS2D result files can be downloaded via scp from MBF.
The credentials can be obtained sending an e-mail to: bm.dinelli (at) isac.cnr.it or to: massimo.carlotti (at) unibo.it.

The currently released MIPAS2D version 1.0 consists of the retrieval of observations of MIPAS main targets (pressure, temperature, H2O, O3, HNO3, CH4, N2O and NO2) from the entire Full Resolution (FR, July 2002 to March 2004) mission, and its new configuration with the Optimized Resolution (OR, January 2005 to date). Version 1.0 has been obtained analysing MIPAS level-1b data produced with different versions of the ESA Instrument Processing Facility (IPF): version 4.61 and 4.62 for the FR mission, and version 4.65 (up to 14 July 2006), and version 4.67 (afterwards) for the OR mission. Version 2.1 has been obtained analysing MIPAS level-1b data produced with IPF version 5.02. The database is continuously updated with new measurements. MIPAS2D results are calculated in the range 6 to 68 km over a common fixed vertical grid, at steps of 3 km up to 42 km and at the altitudes of 47, 52, 60 and 68 km (MIPAS original nominal altitudes). A thorough desciption of the database can be found in the AMT paper "The MIPAS2D database of MIPAS/ENVISAT measurements retrieved with a multi-target 2-dimensional tomographic approach"

The MIPAS2D database is available with two different horizontal-grid definitions:

NOMINAL GRID (NOM) - Each vertical profile is aligned at the average latitude of the tangent points of MIPAS individual limb-scans. The horizontal retrieval grid nodes are thus placed where the maximum information is expected, although MIPAS observations are performed at variable latitude intervals.
FIXED GRID (GRD) - The horizontal retrieval grid is fixed for all the analyzed orbits: vertical profiles are aligned at latitudes equispaced by 5° over the whole latitude range -90° to 90°. At negligible costs of precision of the retrieval, this dataset is useful to minimize the need of interpolation for the end user (averaging processes or intercomparison with other observational/model datasets).

NOM - The NOMINAL observation mode (that is different for the FR and OR measurements)
UTL - The UTLS-1 observation mode (for OR measurements only)
MAT - The Middle Atmosphere observation mode (for OR measurements only)

The list of level-1b files processed for the production of the database can be found here, where the tag on the right column indicated the MIPAS observation mode (for example NOM-FR means: nominal mode, full resolution).

The geophysical parameters pressure, temperature, H2O and O3 were determined simultaneously with a Multi-Target Retrieval (MTR); the other targets were determined in cascade on retrievals following the MTR, each new step exploiting the atmospheric fields previously determined.

All datasets are obtained using optimal estimation without regularization, and high cloud index filtering. Geolocation information, random errors and chi-squares of the retrievals are stored together with the individual information gain (that represents the gain of information obtained in the retrieval when compared to the a-priori assumption).

The data are stored into directories whose name is of the general form TYPE_OBSMOD_RETGRID_VERSION. Currently, TYPE "FR" corresponds to the Full Resolution mission and "OR" corresponds to the Optimised Resolution mission, OBSMOD is the label for the analysed observation mode, RETGRID is the label of the corresponding horizontal retrieval grid and VERSION is a progressive number that labels the improved version of the database.

GMTR level-2 products reported in the database have passed a basic quality check and a filtering process based on the chitest value of their fitting procedure (chi-square < 7.0, only on MTR retrieval). The filtered results are stored on the server as compressed .tgz files, one for each orbit. In the FR dataset, out of 6794 analyzed level-1b files, 5725 have generated acceptable NOM result files and 5798 GRD files. Given that the full advantage of the 2-dimensional retrieval is due to exploiting the information of adjacent limb-scans, orbit chunks having less than 20 minutes of observations were rejected (a full orbit is about 100 minutes). Because of the use of the optimal estimation, results should be always used together with their individual information gain (a value of 0 implies that the final results are based only on the apriori information). Negative values of the individual information gain are used to flag data to be rejected. Other further details can be made available on request.

Remark: GMTR operates simultaneously on measurements acquired during a complete orbit. The analysed orbit originates and closes at the North Pole where the geographical distance between the air masses observed by the first and the last limb-scans is minimum. Since the first limb-scan of the level-1b files occurs far from North Pole, level-1b files of two consecutive orbits are joined by GMTR to produce a full orbit of measurements to be processed.

the year YYYY, month MM and day DD of the first orbit,
the number of the two consecutive orbits used in the retrieval NUM1 + NUM2. If the level-1b of the subsequent orbit does not exist, the retrieval is performed only on the first orbit and _singl is used in place of +NUM2.

Untarring the .tgz file with the command (tar -zxvf YYYYMMDD_OUT_GMTR_NUM1+NUM2.tgz) creates a directory OUT_GMTR_NUM1+NUM2 containing:

X_results.dat: file of the results for target X reporting latitude (LAT), longitude (LONG), Orbital Coordinate (O.C.), altitude (ALT), the target quantity (P, T or VMR), the Estimated Standard Deviation (ESD), date and time (Time) and the information content (Inf). For convenience of the user, in the VMR result files pressure and temperature values are also reported in additional columns after the "Time" column. The fortran routine used to write these files can be found in write_results.f,
X_chitest.dat: chitest evolution at each iteration in the retrieval procedure.

In order to handle MIPAS2D datasets, additional programs are provided in the directory TOOLS. Inside, the README file reports the instructions to complile and use this additional software.

extact_data: the program is interactive and is given just as an example on how to extract data from the database. It enables to extract data from the database on a given latitude, longitude and time box.
extract_alldb: the program is interactive and enables to extract data for all observation modes on the GRD horizontal retrieval grid

Systematic errors (in %) for a global average atmosphere (glw) can be found in: PTHO_000.err_1cm for pressure, temperature, H2O and O3, ch4_000.err, hno3_000.err, no2_000.err, n2o_000.err for the other target quantities. In these files, the total systematic error and the single components that contribute to this error are reported for each retrieval altitude. Pressure, temperature, H2O and O3 errors have been computed for the 2D Multi Target Retrieval System. CH4, HNO3, NO2 and N2O errors were generated for a 1D retrieval system, so they include the systematic error due to the assumption of horizontal homogeneity of the atmosphere (GRA) that should be disregarded in this 2D retrieval. The total error for each altitude is the quadratic summation of the ESD (reported in the X_results.dat files) and the systematic error components. For a detailed explanation of the error-file formats use the link MWMAKE where the Error Analysis File is described.

The list of microwindows used for the retrieval of each target in the FR and OR measurements can be foud here . An estimate of the total systematic error is also reported.

Plots showing average and median values along with the Estimated Standard Deviations (ESDs) computed for four different periods of years 2005-2010 (only 2008 for ESDs) are also available.

JFM=January, February, March; AMJ=April, May, June; JAS=July, August, September; OND=October, November, December; DJF=December, January, February; MAM=March, April, May; JJA=June, July, August; SON=September, October, November

NP= North Pole; MIDN=Mid latitude North; EQUN=Equatorial North; EQUS=Equatorial South; MIDS=Mid latitude South; SP=South Pole

In addition to the MIPAS2D data, a separate directory AK is used to store examples of the 2-Dimensional Averaging Kernels of MIPAS2D database. Due to the size of the AK files the AK of the individual retrievals are not stored into the database. However they can be made available to the users upon request. The AK directory contains examples of the Averaging Kernels for the GRD products of MIPAS2D database and a program to compute the vertical and horizontal resolution from any AK matrix produced by the GMTR analysis system.

The program for reading the AK files and compute the vertical and horizontal resolution is stored in the directory PROG. The averaging kernels files are organized into 4 directories:

Additional files (i.e. columns, spectra etc.) can be made available upon request

Database distribution policy: the database is released for academic and non profit purposes. Invitation to co-authorship on scientific papers should be considered.