UARS MLS DATA QUALITY DOCUMENT For questions on this document, please contact Lucien Froidevaux Mail Stop 183-701 Jet Propulsion Laboratory 4800 Oak Grove Dr. Pasadena, CA 91109 USA email: lucien@mls.jpl.nasa.gov telephone: 818-354-8301 ******************************************************************* Document History -------------------------------------------------------- 19 Oct 1993 Initial Release. 02 Mar 1994 Modified to include mention of programming errors in certain L3AL ClO files, and document history added. 04 Nov 1996 Modified to make Lucien Froidevaux point of contact (instead of Evan Fishbein) --------------------------------------------------------- ******************************************************************* VERSION NO.: VERSION 3 FILES PARAMETER #1: TEMPERATURE UARS Standard Single Profile Estimated Estimated Index Pressure 1-sigma Precision (1) Accuracy (2) 20 0.46 hPa 3 K 7 K 18 1 hPa 2.5 K 5 K 16 2.2 hPa 1.5 K 5 K 14 4.6 hPa 1.5 K 5 K 12 10 hPa 1.5 K 4 K 10 22 hPa 1.5 K 4 K Notes: 1) The precisions are estimated from RMS differences between near colocated measurements on adjacent orbits (1.5 hours separated) at the orbit turn-around points. 2) The accuracies are obtained from RMS differences between the MLS profiles and the NMC daily analyses interpolated onto the MLS orbit tracks. The numbers should be interpreted as preliminary upper bounds on the 1-sigma accuracy. RESOLUTION: The retrieved vertical resolution is two UARS pressure surfaces (delta log10(P) = 0.33, or about 6 km). The retrieved temperature profile is represented as a piecewise-linear function with breakpoints at alternate (even-numbered) UARS pressure surfaces (e.g. 10, 4.6, 2.2, 1 hPa). The temperatures on the even-numbered surfaces (level 3AT files only) are the retrieved breakpoint values, while those on the odd-numbered surfaces (e.g 6.8, 3.2, 1.5 hPa) are averages of the temperatures on adjacent even-numbered surfaces. The level 3AL profiles have an additional linear interpolation with respect to latitude to generate an evenly spaced latitude grid. SYSTEMATIC EFFECTS: Orbit and Yaw-period dependencies are observed in the temperature field. Zonal-mean differences and zonal RMS differences between the MLS temperature field and the NMC analysis show biases which depend on: whether the data is obtained during the ascending or descending sides of the UARS orbit, whether MLS is looking north or south, and when during a yaw period the measurements are taken. These systematics are 1-3 K in the stratosphere, but can be as much as 10 K in the lower mesosphere. Comparisons of zonal-mean cross sections before and after yaw maneuvers suggest that the north-looking cross sections may be shifted northward 1-3 degrees in latitude relative to the south-looking cross sections. MLS temperatures are biased 1-2 K lower than NMC temperatures between 22 hPa and 1 hPa. Lapse rates in the mesosphere are too negative due to decreasing sensitivity. A notch, usually negative, is observed in many MLS temperature profiles at 0.22 hPa. CAVEATS The retrieval is based on the sequential estimation algorithm with an a priori estimate containing both the NMC daily analysis (when available) and a month-dependent, latitude-dependent climatology developed by the UARS science team. Although the profiles extend from 1000 hPa to 0.0001 hPa, useful information is provided by MLS only between 22 hPa and 0.46 hPa. Above 0.22 hPa the profiles relax slowly to the climatology. Below 22 hPa, the profiles are linearly interpolated from NMC daily analyses (or climatology when necessary) onto the even-numbered surfaces. TEMPERATURES OUTSIDE THE RANGE, 22 hPa TO 0.4 hPa SHOULD NOT BE USED WITHOUT THE ENDORSEMENT OF THE MLS TEAM. Currently a linearized forward model is used to fit radiances in a one pass retrieval through the data. In the winter at high latitudes, the atmosphere may differ from the linearization point by more than 20K, especially when wave activity is enhanced. During these periods, systematic errors from nonlinearities may be of the order of 5-10K. WAVE AMPLITUDES MAY BE MISREPRESENTED DURING PERIODS OF LARGE WAVE ACTIVITY. The `quality' field in the level 3A files is the retrieval's estimated uncertainty, includes random and systematic components, and is obtained by propagating precisions of the radiance measurements, estimates of constrained parameter uncertainties, forward model inaccuracies, and some calibration uncertainties through the retrieval software. The quality should be interpreted as a lower bound on the accuracy. At the conclusion of the retrieval, the estimated uncertainty is compared with the a priori uncertainty. When the ratio is greater than 0.5, or the temperature is more than 25% climatology, the quality is set negative to flag the dependence of the retrieved temperature on the a priori. Profiles with all qualities negative were usually not retrieved and are the a priori; including these profiles in scientific analyses is not recommended. The level 3 parameter files (described in the "Standard Formatted Data Units, MLS Level 3TP Parameter File" and "Standard Formatted Data Units, MLS Level 3LP Parameter File" documents) contain the diagnostic flag `MMAF_STAT'. FOR OPTIMAL USE OF THE MLS DATA, THE PARAMETER FILES SHOULD BE EXAMINED, AND PROFILES WITH MMAF_STAT NOT SET TO 'G', 'T', OR 't' SHOULD BE DISREGARDED. PARAMETER:#2: O3_205 (Ozone retrieved from 205 GHz radiometer radiances) UARS Standard Single Profile Estimated Estimated Index Pressure 1-sigma Precision (1) Accuracy (2) 20 0.46 hPa 0.5 ppmv 10% 18 1 hPa 0.3 ppmv 7% 16 2.2 hPa 0.3 ppmv 7% 14 4.6 hPa 0.3 ppmv 5% 12 10 hPa 0.2 ppmv 5% 10 22 hPa 0.2 ppmv 7% 8 46 hPa 0.2 ppmv 30% 6 100 hPa 0.5 ppmv >50% Notes: 1) The estimated precisions are based on observed variability in latitude bands where meteorological variability is small, hence the true precisions may be somewhat better than these estimates. These numbers are 1-sigma (rms) precisions and are consistent with theoretical estimates obtained by propagating the radiance measurement precisions through a sequential estimation retrieval algorithm. 2) The accuracy estimates are based on statistical comparisons with other data sets (including SAGE II, ozonesonde and balloon correlative data) and are preliminary. RESOLUTION: The retrieved vertical resolution is two UARS pressure surfaces (delta log10(P) = 0.33, or about 6 km). The retrieved ozone volume mixing ratio profile is represented as a piecewise-linear function with breakpoints at alternate (even-numbered) UARS pressure surfaces (e.g. 100, 46, 22, 10 hPa). The mixing ratios on the even-numbered surfaces (level 3AT files only) are the retrieved breakpoint values, while those on the odd- numbered surfaces (e.g 68, 32, 15 hPa) are averages of the mixing ratios on adjacent even-numbered surfaces. The level 3AL profiles have an additional linear interpolation with respect to latitude to generate an evenly spaced latitude grid. SYSTEMATIC EFFECTS: The 46 hPa ozone mixing ratios usually have a negative bias (in comparison with other data sets) of approximately a few tenths of a ppmv, whereas the 100 hPa ozone mixing ratios generally have a positive bias of approximately 0.1 to 0.2 ppmv; these biases account for most of the accuracy figures at 46 and 100 hPa. A small (approximately 5%) positive bias relative to SAGE II measurements exists in the middle and upper stratosphere. Similar biases are seen in other (but not all) comparisons near the ozone volume mixing ratio profile maximum. Integrated column ozone using data at 100 hPa and above tend to be biased high compared to estimated columns from other data sets (TOMS or ozonesondes); these biases may be caused by inadequate vertical resolution. A cyclic O3_205 fluctuation synchronized to the UARS yaw period (roughly every 36 days) is present primarily in the tropics at the lowest altitudes, and affects column ozone by approximately 5 to 10 percent. The zonal-mean O3_205 trend shows small `jumps' coincident with some satellite yaw maneuvers. CAVEATS: O3_205 has received greater scrutiny than O3_183 and is currently the recommended ozone for stratospheric studies. It has consistently agreed with other data sets better than O3_183, and the 205 GHz radiances show better radiance residual closure than the 183 GHz radiances. However, O3_183 is better for mesospheric studies (up to 0.05 hPa) where the 205 GHz radiances lose sensitivity. The two fields are in good agreement in the upper stratosphere where both 183 and 205 GHz radiances have similar sensitivities. The retrieval is based on the sequential estimation algorithm with an a priori based on a month-dependent latitude-dependent climatology developed by the UARS science team. While the O3_205 profiles have 37 grid points extending from 464 hPa to 4.6e-4 hPa, the values at pressures outside the range 100 hPa to 0.46 hPa are mostly climatological. O3_205 at 100 hPa currently exhibits some large biases and therefore has limitations for use in scientific studies. MIXING RATIOS AT PRESSURES LARGER THAN 100 hPa OR SMALLER THAN 0.46 hPa SHOULD NOT BE USED FOR SCIENTIFIC STUDIES, AND VALUES AT 100 hPa SHOULD NOT BE USED WITHOUT CONSULTING THE MLS TEAM. The `quality' field in the level 3A files is the retrieval's estimated uncertainty, includes random and systematic components, and is obtained by propagating precisions of the radiance measurements, estimates of constrained parameter uncertainties and forward model inaccuracies through the retrieval software. The quality should be interpreted as a lower bound on the accuracy. At the conclusion of the retrieval, the estimated uncertainty is compared with the a priori uncertainty. When the ratio is greater than 0.5, or the estimated mixing ratio is more than 25% climatology, the quality is set negative to flag the dependence of the retrieved mixing ratio on the a priori (note that the 100 hPa and 0.2 hPa levels are generally flagged with negative quality). Profiles with all qualities negative were usually not retrieved and are climatology; including these profiles in scientific analyses is not recommended. The level 3 parameter files (described in the "Standard Formatted Data Units, MLS Level 3TP Parameter File" and "Standard Formatted Data Units, MLS Level 3LP Parameter File" documents) contain the diagnostic flag `MMAF_STAT'. FOR OPTIMAL USE OF THE MLS DATA, THE PARAMETER FILES SHOULD BE EXAMINED AND PROFILES WITH MMAF_STAT NOT SET TO 'G', 'T', OR 't' SHOULD BE DISREGARDED. PARAMETER #3: O3_183 (Ozone retrieved from 183 GHz radiometer radiances) UARS Standard Single Profile Estimated Estimated Index Pressure 1-sigma Precision (1) Accuracy (2) 26 0.05 hPa 0.4 ppmv 24 0.1 hPa 0.3 ppmv 22 0.22 hPa 0.2 ppmv 20 0.46 hPa 0.2 ppmv 18 1 hPa 0.2 ppmv 16 2 hPa 0.3 ppmv 14 5 hPa 0.3 ppmv 12 10 hPa 0.2 ppmv 10 22 hPa 0.2 ppmv 8 46 hPa 0.2 ppmv Notes: 1) The estimated precisions are based on observed variability in latitude bands where meteorological variability is small, hence the true precisions may be somewhat better than these estimates. These numbers are 1-sigma (rms) precisions and are consistent with theoretical estimates obtained by propagating the radiance measurement precisions through a sequential estimation retrieval algorithm. 2) The accuracy estimates for O3_183 should be similar to, but somewhat larger than those for O3_205 based on comparisons of the two products. RESOLUTION: The retrieved vertical resolution is two UARS pressure surfaces (delta log10(P) = 0.33, or about 6 km). The retrieved ozone volume mixing ratio profile is represented as a piecewise-linear function with breakpoints at alternate (even-numbered) UARS pressure surfaces (e.g. 10, 4.6, 2.2, 1 hPa). The mixing ratios on the even-numbered surfaces (level 3AT files only) are the retrieved breakpoint values, while those on the odd-numbered surfaces (e.g 6.8, 3.2, 1.5 hPa) are averages of the mixing ratios on adjacent even-numbered surfaces. The level 3AL profiles have an additional linear interpolation with respect to latitude to generate an evenly spaced latitude grid. SYSTEMATIC EFFECTS: O3_183 may be biased high by 5 to 10 percent (based on what is known for O3_205 and limited comparisons between the O3_205 and O3_183 fields). CAVEATS: The O3_183 data have not been scrutinized as critically as the O3_205 data because of efforts to resolve observed post-launch biases. O3_205 has consistently agreed with other data sets better than O3_183, and the 205 GHz radiances show better radiance residual closure than the 183 GHz radiances. Although reasonable consistency exists with the O3_205 data, detailed studies of possible systematic effects have not been completed -- THIS SHOULD BE KEPT IN MIND WHEN ATTEMPTING TO ANALYZE O3_183 DATA. O3_183 is better than O3_205 for mesospheric studies (up to 0.05 hPa) where the 205 GHz radiances lose sensitivity. The two fields are in good agreement in the upper stratosphere where both 183 and 205 GHz radiances have similar sensitivities. The retrieval is based on the sequential estimation algorithm with an a priori based on a month-dependent latitude-dependent climatology developed by the UARS science team. While the O3_183 profiles have 37 grid points, extending from 464 hPa up to 4.6e-4 hPa, the MIXING RATIOS AT PRESSURES OUTSIDE THE RANGE 46 hPa TO 0.046 hPa ARE MOSTLY CLIMATOLOGICAL AND SHOULD NOT BE USED FOR SCIENTIFIC STUDIES. The `quality' field in the level 3A files is the retrieval's estimated uncertainty, includes random and systematic components, and is obtained by propagating precisions of the radiance measurements, estimates of constrained parameter uncertainties and forward model inaccuracies through the retrieval software. The quality should be interpreted as a lower bound on the accuracy. At the conclusion of the retrieval, the estimated uncertainty is compared with the a priori uncertainty. When the ratio is greater than 0.5, or the estimated mixing ratio is more than 25% climatology, the quality is set negative to flag the dependence of the retrieved mixing ratio on the a priori. Profiles with all qualities negative were usually not retrieved and are climatology; including these profiles in scientific analyses is not recommended. The level 3 parameter files (described in the "Standard Formatted Data Units, MLS Level 3TP Parameter File" and "Standard Formatted Data Units, MLS level 3LP Parameter File" documents) contain the diagnostic flag ``MMAF_STAT''. FOR OPTIMAL USE OF THE MLS DATA,THE PARAMETER FILES SHOULD BE EXAMINED AND PROFILES WITH MMAF_STAT NOT SET TO 'G', 'T', OR 't' SHOULD BE DISREGARDED. PARAMETER #4: H2O UARS Standard Single Profile Estimated Estimated Index Pressure 1-sigma Precision (1) Accuracy (2) 22 0.22 hPa 0.4 ppmv 15 - 30 % 20 0.46 hPa 0.4 ppmv 15 - 30 % 18 1 hPa 0.3 ppmv 15 - 30 % 16 2.2 hPa 0.2 ppmv 15 - 30 % 14 4.6 hPa 0.2 ppmv 15 - 30 % 12 10 hPa 0.1 ppmv 15 - 20 % 10 22 hPa 0.2 ppmv 15 - 20 % 8 46 hPa 0.2 ppmv 15 - 20 % Notes: 1) The estimated precisions are based on observed variability in latitude bands where meteorological variability is small, hence the true precisions may be somewhat better than these estimates. 2) These accuracies are first-order estimates based on comparisons of MLS H2O data with other UARS and/or correlative measurements. RESOLUTION: The retrieved vertical resolution is two UARS pressure surfaces (delta log10(P) = 0.33, or about 6 km). The retrieved water vapor volume mixing ratio profile is represented as a piecewise-linear function with breakpoints at alternate (even-numbered) UARS pressure surfaces (e.g. 10, 4.6, 2.2, 1 hPa). The water vapor mixing ratios on the even-numbered surfaces (level 3AT files only) are the retrieved breakpoint values, while those on the odd-numbered surfaces (e.g 6.8, 3.2, 1.5 hPa) are averages of the mixing ratios on adjacent even-numbered surfaces. The level 3AL profiles have an additional linear interpolation with respect to latitude to generate an evenly spaced latitude grid. SYSTEMATIC EFFECTS: At high latitudes, especially in winter, the estimated uncertainty (see below) can increase to 1.8 ppmv at 46 hPa, and to 1 ppmv at 22 hPa, whereas the low latitude uncertainties are around 0.9 ppmv and 0.8 ppmv. At these pressures, the a priori uncertainty is 2 ppmv, so that for these estimated uncertainties, the estimated mixing ratios are more than 80% or 25% climatology. Detailed studies suggest that this is due to a combination of the atmosphere being optically thick and very low temperatures (especially during winter), resulting in a loss of information content. A 'notch' of high H2O values appears in the lower mesosphere (in the region near 0.1 hPa) which is thought to be unreal and needs further study. Comparisons with other UARS data and/or correlative data suggest that MLS H2O may be ~10-30% too high in the range 46 hPa to 0.2 hPa. CAVEATS: The `quality' field in the level 3A files is the retrieval's estimated uncertainty, includes random and systematic contributions, and is obtained by propagating precisions of the radiance measurements, estimates of constrained parameter uncertainties, and forward model inaccuracies through the retrieval software. The quality should be interpreted as a lower bound on the accuracy. The useful vertical range for MLS H2O is 46 hPa to 0.2 hPa. At 46 hPa there is a loss of information at high latitudes. H2O VALUES AT 46 hPa SHOULD NOT BE USED IN SCIENTIFIC STUDIES WITHOUT THE ENDORSEMENT OF THE MLS TEAM. The retrieval is based on the sequential estimation algorithm and uses an a priori estimate based on a month-dependent latitude-dependent climatology developed by the UARS science team. At the conclusion of the retrieval, the estimated uncertainty is compared with the a priori uncertainty. When the ratio is greater than 0.5, or the mixing ratio is more than 25% climatology, the quality is set negative to flag the dependence of the retrieved mixing ratio on the a priori. ONLY DATA HAVING POSITIVE QUALITY SHOULD BE USED FOR SCIENTIFIC PURPOSES. The level 3 parameter files (described in the "Standard Formatted Data Units, MLS Level 3TP Parameter File" and "Standard Formatted Data Units, MLS Level 3LP Parameter File" documents) contain the diagnostic flag ``MMAF_STAT''. FOR OPTIMAL USE OF THE MLS DATA, THE PARAMETER FILES SHOULD BE EXAMINED AND PROFILES WITH MMAF_STAT NOT SET TO 'G' SHOULD BE DISREGARDED. PARAMETER #5: ClO UARS Standard Single Profile Estimated Estimated Index Pressure 1-sigma Precision (1) Accuracy (2) 20 0.46 hPa 1.6 ppbv 0.15 ppbv and 30% 18 1 hPa 1.3 ppbv 0.1 ppbv and 20% 16 2.2 hPa 0.8 ppbv 0.1 ppbv and 15% 14 4.6 hPa 0.5 ppbv 0.1 ppbv and 15% 12 10 hPa 0.4 ppbv 0.15 ppbv and 15% 10 22 hPa 0.4 ppbv 0.20 ppbv and 15% 8 46 hPa 0.5 ppbv 0.50 ppbv and 15% 6 100 hPa 1.5 ppbv 0.70 ppbv and 40% Notes: 1) The estimated precisions given here are typical values obtained by propagating the radiance precisions through the retrieval algorithm and are consistent with the observed variance in situations where ClO is below the instrument noise level. The quality field in the level 3A files give the estimated precisions on individual profiles. Precision can be improved by averaging together individual profiles. 2) The estimated accuracies are a root sum square of a bias error (accuracy given in ppbv) plus a scaling error (the product of the fractional accuracy and the estimated mixing ratio). These accuracies do not include the random noise which, for a single profile, is the estimated precision. RESOLUTION: The retrieved vertical resolution is two UARS pressure surfaces (delta log10(P) = 0.33, or about 6 km). The retrieved ClO volume mixing ratio profile is represented as a piecewise-linear function with breakpoints at alternate (even-numbered) UARS pressure surfaces (e.g. 100, 46, 22, 10 hPa). The ClO mixing ratos on the even-numbered surfaces (level 3AT files only) are the retrieved breakpoint values, while those on the odd-numbered surfaces (e.g 68, 32, 15 hPa) are averages of the mixing ratios on adjacent even-numbered surfaces. The level 3AL profiles have an additional linear interpolation with respect to latitude to generate an evenly spaced latitude grid. SYSTEMATIC EFFECTS: HNO3 and N2O affect the lower stratospheric ClO retrievals, and the algorithms producing the version 3 level 3A files assume climatological values for these species. This leads to systematic errors of up to (approximately) 0.5 ppbv in the polar vortices and up to (approximately) 0.2 ppbv in other situations. These systematic errors can be reduced by taking appropriate day - night differences. CAVEATS: The `quality' field in the level 3A files is the 1-sigma precision estimated by the retrieval software. The retrieval is based on the sequential estimation algorithm and uses an a priori estimate based on results from the Lawerence-Livermore 2-D model. At the conclusion of the retrieval, the estimated precision is compared with the a priori uncertainty, and if the ratio is greater than 0.5, (the mixing ratio is more than 25% climatology), the quality is set negative. ONLY DATA HAVING POSITIVE QUALITY SHOULD BE USED FOR SCIENTIFIC PURPOSES. Additional quality indicators in the level 3 parameter files should also be examined, and ONLY DATA HAVING MMAF_STAT=G AND QUALITY_CLO=4 SHOULD BE USED FOR SCIENTIFIC PURPOSES. Further details on quality indicators available in L3 parameter files are given in the "Standard Formatted Data Units, MLS Level 3TP Parameter File" and "Standard Formatted Data Units, MLS Level 3LP Parameter File" documents. THESE PARAMETERS SHOULD BE EXAMINED FOR OPTIMUM USE OF THE MLS DATA. ************************************************************************ A programming error occurred in producing some MLS Version 3 L3AL ClO files for data having dates before 15 April 1993 (UARS day 0582). This caused an isotopic correction to be applied twice, and ClO values in the affected files are too large by a factor of 1.32. Users of L3AL ClO files for this period should contact the MLS team for information on which files are affected. L3AT files and published results are not affected in any way by this error. *********************************************************************** PARAMETER #6: SO2 UARS Standard Single Profile Estimated Estimated Index Pressure 1-sigma Precision (1) Accuracy (2) 16 2.2 hPa 3.3 ppbv 3 ppbv and 15% 14 4.6 hPa 3.0 ppbv 3 ppbv and 15% 12 10 hPa 3.0 ppbv 3 ppbv and 15% 10 22 hPa 4.5 ppbv 3 ppbv and 15% 8 46 hPa 6.7 ppbv 3 ppbv and 15% Notes: 1) The estimated precisions given here are typical values obtained by propagating the radiance precisions through the retrieval algorithm. The quality field in the level 3A files should be examined to obtain the precisions on any specific profile. 2) The estimated accuracies are a root sum square of a bias accuracy (accuracy given in ppbv) plus a scaling error (the product of the fractional accuracy and the estimated mixing ratio). These accuracies do not include the random noise which, for a single profile, is the estimated precision. RESOLUTION: The retrieved vertical resolution is two UARS pressure surfaces (delta log10(P) = 0.33, or about 6 km). The retrieved SO2 volume mixing ratio profile is represented as a piecewise-linear function with breakpoints at alternate (even-numbered) UARS pressure surfaces (e.g. 100, 46, 22, 10 hPa). The mixing ratios on the even-numbered surfaces (level 3AT files only) are the retrieved breakpoint values, while those on the odd-numbered surfaces (e.g 68, 32, 15 hPa) are averages of the mixing ratios on adjacent even-numbered surfaces. The level 3AL profiles have an additional linear interpolation with respect to latitude to generate an evenly spaced latitude grid. SYSTEMATIC EFFECTS: Inadequately calculated spectral baselines caused by contaminant species (e.g. HNO3) or instrument characterization inaccuracy will cause SO2 retrieval errors as high as +/- 3ppbv. These errors imply that MLS can measure SO2 only when the mixing ratio greatly exceeds atmospheric equilibrium values (approximately 0.1 ppbv), such as after large volcanic eruptions. An improved estimate of the transient SO2 mixing ratio can be obtained by subtracting an estimated background SO2 mixing ratio from the estimated transient mixing ratio. The background can be estimated from long term measurements following the volcanic event or using estimated mixing ratios from locations outside the volcanic cloud. The 15% accuracy is predominantly caused by retrieval numerics and assumes the ClO accuracy above. CLO and SO2 are retrieved concurrently in the same spectral band. CAVEATS: SO2 HAS NOT BEEN VALIDATED BY COINCIDENT PROFILE COMPARISONS WITH OTHER EXPERIMENTS because these data do not exist. However, comparison of derived quantities, such as initial loading and decay rate following the Pinatubo eruption, and MLS measurements extrapolated to times when other data exist, agree to within 10-20% with TOMS, SBUV and aircraft measurements. The quality field appearing in the level 3A files is the 1-sigma precision estimated by the retrieval software. The retrieval is based on the sequential estimation algorithm and uses an a priori estimate of 0.1 ppbv at all levels. At the conclusion of the retrieval, the estimated precision is compared with the a priori uncertainty, and if the ratio is greater than 0.5, (the mixing ratio is more than 25% climatology), the quality is set negative. ONLY DATA HAVING POSITIVE QUALITY SHOULD BE USED FOR SCIENTIFIC PURPOSES. The profiles contained in the level 3A files extend from 464 hPa to 2.2 hPa. ONLY VALUES AT PRESSURES BETWEEN 46 hPa TO 2.2 hPa SHOULD BE USED FOR SCIENTIFIC STUDIES. Analysis of SO2 from small volcanoes is limited by the coarse spatial resolution of the MLS measurements (except near the tangent track orbital turn-around latitudes). Additional quality indicators in the level 3 parameter files should also be examined, and ONLY DATA HAVING MMAF_STAT=G AND QUALITY_CLO=4 SHOULD BE USED FOR SCIENTIFIC PURPOSES. Further details on quality indicators available in L3 parameter files are given in the "Standard Formatted Data Units, MLS Level 3TP Parameter File" and "Standard Formatted Data Units, MLS Level 3LP Parameter File" documents. THESE PARAMETERS SHOULD BE EXAMINED FOR OPTIMUM USE OF THE MLS DATA.