UARS MLS DATA QUALITY DOCUMENT VERSION 5 DATA 21 JANUARY 2000 POINTS OF CONTACT: - General Dr. Nathaniel Livesey Information Jet Propulsion Laboratory and CH3CN Mail Stop 183-701 Pasadena CA, USA 91109 E-mail: livesey@mls.jpl.nasa.gov Phone: (818) 354-4214 - ClO Dr. Joe Waters (UARS MLS Principal Investigator) Jet Propulsion Laboratory Mail Stop 183-701 Pasadena CA, USA 91109 E-mail: joe@mls.jpl.nasa.gov Phone: (818) 354-3025 - H2O Dr. Hugh Pumphrey (stratosphere) Department of Meteorology The University of Edinburgh Edinburgh EH9 3JZ, Scotland E-mail: hcp@met.ed.ac.uk Phone: 0131-650-6026 (44-131-650-6026 if outside U.K.) - Upper. Trop. Dr. William Read Humidity(UTH) Jet Propulsion Laboratory Mail Stop 183-701 Pasadena CA, USA 91109 E-mail: bill@mls.jpl.nasa.gov Phone: (818) 354-6773 - HNO3 Dr. Michelle Santee Jet Propulsion Laboratory Mail Stop 183-701 Pasadena CA, USA 91109 E-mail: mls@mls.jpl.nasa.gov Phone: (818) 354-9424 - O3 Dr. Lucien Froidevaux Jet Propulsion Laboratory Mail Stop 183-701 Pasadena CA, USA 91109 E-mail: lucien@mls.jpl.nasa.gov Phone: (818) 354-8301 - Temperature Dr. Dong Wu and Geopot. Jet Propulsion Laboratory Height Mail Stop 183-701 Pasadena CA, USA 91109 E-mail: dwu@mls.jpl.nasa.gov Phone: (818) 353-1954 - WEB site (UARS MLS Homepage): http://mls.jpl.nasa.gov ****************************************************************************** TABLE OF CONTENTS INTRODUCTION SUMMARY INFORMATION ABOUT VERSION 5 UARS MLS DATA GENERAL INFORMATION ABOUT UARS MLS DATA PARAMETER # 1: TEMP (Temperature) CHANGES FROM VERSION 4 TO VERSION 5 AFFECTING TEMP DIFFERENCES BETWEEN AVERAGE VERSION 5 AND VERSION 4 TEMP ESTIMATED PRECISION AND ACCURACY OF TEMP KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN TEMP CAVEATS FOR TEMP PARAMETER # 2: O3_205 (Ozone from 205-GHz radiometer radiances) CHANGES FROM VERSION 4 TO VERSION 5 AFFECTING O3_205 DIFFERENCES BETWEEN AVERAGE VERSION 5 AND VERSION 4 O3_205 ESTIMATED PRECISION AND ACCURACY OF O3_205 KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN O3_205 CAVEATS FOR O3_205 PARAMETER # 3: O3_183 (Ozone from 183-GHz radiometer radiances) CHANGES FROM VERSION 4 TO VERSION 5 AFFECTING O3_183 DIFFERENCES BETWEEN AVERAGE VERSION 5 AND VERSION 4 O3_183 ESTIMATED PRECISION AND ACCURACY OF O3_183 KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN O3_183 CAVEATS FOR O3_183 PARAMETER # 4: H2O (Water Vapor) CHANGES FROM VERSION 4 TO VERSION 5 AFFECTING H2O DIFFERENCES BETWEEN AVERAGE VERSION 5 AND VERSION 4 H2O ESTIMATED PRECISION AND ACCURACY OF H2O KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN H2O CAVEATS FOR H2O PARAMETER # 5: UTH (Upper Tropospheric Humidity) DESCRIPTION CHANGES FROM VERSION 4.9 TO VERSION 5 AFFECTING UTH DIFFERENCES BETWEEN AVERAGE VERSION 5 AND VERSION 4.9 UTH ESTIMATED PRECISION AND ACCURACY OF UTH KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN UTH CAVEATS FOR UTH PARAMETER # 6: CLO (Chlorine Monoxide) CHANGES FROM VERSION 4 TO VERSION 5 AFFECTING CLO DIFFERENCES BETWEEN AVERAGE VERSION 5 AND VERSION 4 CLO ESTIMATED PRECISION AND ACCURACY OF CLO KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN CLO CAVEATS FOR CLO PARAMETER # 7: HNO3 (Nitric Acid) CHANGES FROM VERSION 4 TO VERSION 5 AFFECTING HNO3 DIFFERENCES BETWEEN AVERAGE VERSION 5 AND VERSION 4 HNO3 ESTIMATED PRECISION AND ACCURACY OF HNO3 KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN HNO3 CAVEATS FOR HNO3 PARAMETER # 8: CH3CN (Methyl Cyanide or Acetonitrile) [New Product] ESTIMATED PRECISION AND ACCURACY OF CH3CN KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN CH3CN CAVEATS FOR CH3CN PARAMETER # 9: GPH (Geopotential Height) [New Product] ESTIMATED PRECISION AND ACCURACY OF GPH KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN GPH CAVEATS FOR GPH ****************************************************************************** INTRODUCTION This document gives a brief description of the type and quality of Version 5 (V5) data from the Microwave Limb Sounder (MLS) aboard the Upper Atmosphere Research Satellite (UARS). It is an update to the previous document of this kind for MLS Version 4 (V4) data. Both documents are available from the MLS homepage (see above) and from the NASA Goddard Space Flight Center Distributed Access Archive Center (DAAC), which provides them to users requesting UARS MLS data. This document discusses data available in the Level 3 files for all parameters retrieved in Version 5 (see below), as well as Level 2 files for upper tropospheric humidity (UTH). The Level 3 data (produced from MLS software Version 5) are made publicly available from the Goddard DAAC after being transferred from Goddard's UARS Central Data Handling Facility (CDHF), where the data are initially stored. SUMMARY INFORMATION ABOUT VERSION 5 UARS MLS DATA There are 9 atmospheric parameters produced by MLS Version 5 processing and put in L3AT/L3AL data files (however, no L3AL files are produced for UTH - see below). These parameters are: (1) temperature (parameter name TEMP), (2) ozone from the 205-GHz band radiances (O3_205), (3) ozone from the 183-GHz band radiances (O3_183), (4) water vapor (H2O), (5) upper tropospheric humidity (UTH), (6) chlorine monoxide (CLO), (7) nitric acid (HNO3), (8) methyl cyanide or acetonitrile (CH3CN), and (9) geopotential height (GPH). The latter two are new products with little validation. Level 1 (radiance) data from Version 4 are used to produce the Level 2 and Level 3 data for Version 5. No Version 5 Level 1 files are produced. Also, no Standard Formatted Data Units (SFDU) documents are provided for the MLS Version 5 data. The Version 5 Level 3 files (L3AT, L3AL, L3PT, L3PL) have a format identical to the Version 4 files, so the Version 4 SFDU documents include all necessary information for the Version 5 Level 3 files. The Version 5 Level 2 file for UTH, which contains valuable diagnostic information for this product, is an ASCII file with a header providing all the information needed to understand its contents. The Version 5 `overall' Level 2 file (L2OUT) format has changed significantly; users wishing to read this file should contact the MLS team. The Level 3 H2O files include retrieved tropospheric H2O mixing ratios at 147, 215, 316, and 464 hPa; however, it should be emphasized that the tropospheric H2O and the stratospheric H2O in these files are from different radiometers, and that consistency between these two data sets at 100 hPa has not yet been achieved (100 hPa values are not considered reliable enough). The Level 3 UTH files give upper tropospheric humidity with respect to ice at 147, 215, 316, and 464 hPa - using temperatures from the National Centers for Environmental Prediction (NCEP). The Level 2 ASCII (daily) file is recommended for UTH data because it contains more information for screening bad data points; this file also gives the H2O mixing ratios, NCEP temperatures, and other diagnostic information useful for the MLS UTH data. This file's header contains information that should be sufficient for understanding the file's contents. Sulfur dioxide (SO2) was produced in Version 4, but is not a Version 5 product, due to the spectral similarity of its MLS signal to that of CH3CN. V5 CH3CN values are contaminated by Pinatubo SO2 from the start of the mission through about January 1992, and by the Lascar volcano SO2 for a local region over South America on 21 and 22 April, 1993. Temperature variances at gravity-wave scales have also been produced as a special product, but not as part of routine reprocessing. Contact Dong Wu for more information on these data. The primary changes in MLS Version 5 are: - retrieved values are produced on each of the standard UARS pressure surfaces, i.e. for six points per decade change in pressure, over most of the useful vertical range for each parameter (but retrievals in the upper mesosphere generally revert to a grid twice as coarse) - improved treatment of coupled temperature & tangent point retrievals - refinements in spectroscopy and estimated instrument sideband responses for some bands (O2, H2O, O3_183) - iterative retrievals in some bands More detailed parameter-specific information regarding changes from previous software versions, recommended pressure ranges, and uncertainties and caveats, is provided in later sections. In addition to noting the caveats for each parameter given later in this document, and the single-profile estimated uncertainties included in the data files, screening for bad data should follow the same procedures as in previous MLS data versions: - only use data from records for which the corresponding MMAF_STAT (in the appropriate L3PT/L3PL files) equals "G", "T", or "t" - only use data where the estimated precision (in the L3AT/L3AL files) has a positive value - only use data where the quality indicator (in the L3PT/L3PL files), e.g. "QUALITY_CLO", has value "4". * for CH3CN, use "QUALITY_CLO" as a quality flag * for HNO3, use "QUALITY_O3_205" as a quality flag * for GPH, use "QUALITY_TEMP" as a quality flag There may be additional screening needed for GPH, due to spacecraft attitude effects, but this has not been included in the V5 data set; users of this product should therefore use caution and keep in close communication with the MLS team. The 183-GHz radiometer failed in April 1993. The last good full day of data was April 15, 1993 (UARS Day 582), for ozone from the 183-GHz band and for stratospheric H2O. Therefore, ONLY FIELDS FOR O3_183 and H2O PRIOR TO APRIL 16 1993 SHOULD BE USED FOR STRATOSPHERIC STUDIES. NOTE, however, that the Version 5 Level 3 H2O files for days beyond this date are useful for the upper tropospheric portion of the data (pressures of 464 to 147 hPa). The MLS team expects to submit for publication various more detailed validation studies and other analyses of the Version 5 data. GENERAL INFORMATION ABOUT UARS MLS UARS MLS measurements are obtained from observations of millimeter- wavelength thermal emission (day and night) as the instrument field of view (FOV) is scanned vertically through the earth's atmospheric limb. The 63-GHz radiometer measures radiances from O2 emission (MLS band 1) to obtain tangent-point pressure and atmospheric temperature (as well as geopotential height). The 205-GHz radiometer provides chlorine monoxide (ClO) (bands 2 and 3), ozone (band 4), nitric acid (bands 2 and 4), upper tropospheric humidity (UTH) (band 2), methyl cyanide (CH3CN) (bands 2 and 3), and sulfur dioxide (bands 2 and 4; although it is not retrieved in Version 5). The 183-GHz radiometer provides water vapor (band 5) and ozone (band 6). The MLS standard operating mode consists of full vertical scans (approximately 0 to 90 km tangent heights) every 65.536 sec (an MLS major frame, MMAF), and the calibrated limb radiances from each band are used to retrieve atmospheric profiles; pressure is the independent variable. The MLS instrument calibration is described by Jarnot et al. [JGR 101, 9957-9982, 1996]. The UARS orbital motion (7 km/s) during limb scans smears the profile measurements over about 400 km in a direction perpendicular to the MLS line of sight; the profile is also "averaged" over a similar distance along the line of sight, because of radiative transfer "smearing." The antenna fields of view (half power beamwidths) are 3.0, 3.6, and 9.5 km for the 205-GHz, 183-GHz, and 63-GHz measurements, respectively. The measurement latitudinal coverage is from 34 degrees on one side of the equator to 80 degrees on the other. UARS performs a yaw maneuver at roughly 36-day intervals (every "UARS month" or "yaw period"), when MLS high-latitude coverage switches between north and south. Within each UARS month the UARS orbit plane precesses slowly with respect to the Earth-Sun line. The orbit precession causes the measurements at a given latitude to sweep through essentially all local solar times during the course of a UARS month, becoming 20 min earlier each day. The MLS retrievals for Version 5 use the standard optimal estimation technique. The retrievals occur on a profile by profile basis with no `memory' between profiles. Temperature and tangent pressure are retrieved first from the 63 GHz measurements. This is followed by an initial retrieval of upper tropospheric humidity. Stratospheric water vapor and 183-GHz ozone are then retrieved from the 183-GHz radiometer (for days when that radiometer was operating). The remaining species (205-GHz ozone, ClO, HNO3 and CH3CN) are then retrieved from the 205-GHz band. Finally, the estimate of upper tropospheric humidity is refined. At each stage the results from the previous stage (e.g. Temperature) are used as constrained forward model parameters. The retrieved errors for these are propagated through the successive retrieval calculations. The error propagation is rigorous for pressures larger than or equal to 21.5 hPa; i.e., it includes correlated errors in forward model radiances, with zero correlation used at lower pressures to achieve faster calculations. After June 15, 1997 (UARS day 2104), only the 205 GHz radiometer is operational, since the 63 GHz radiometer was switched off to lower the MLS power requirements. Data from this period are retrieved in a slightly different manner, with temperature not produced and tangent pressure deduced from the 205-GHz ozone line. As stated below, details of possible biases introduced by this change in instrument operational mode remain to be fully investigated and documented. People wishing to use data beyond this day should consult with the MLS science team. Retrieved profile values are produced for each UARS pressure surface (delta log10(P) = 1/6, or slightly less than 3 km) over most of the vertical range. At higher altitudes (pressures less than 0.1 hPa), the resolution degrades to 2 UARS surfaces, with the L3AT/L3AL data for the `odd' surfaces containing interpolations. The profiles are represented as piecewise-linear functions with breakpoints at each pressure surface (e.g. 100, 68, 46 hPa). The retrieved values should be interpreted as the breakpoints of the piecewise-linear profile that best fits the MLS radiances. They are *not* point values at each surface. The Level 3AL profiles are linearly interpolated (in space) from the L3AT data to generate a sample evenly spaced in latitude (with fewer profiles than the L3AT data). The `quality' field in the Level 3A files is the retrieval's estimated uncertainty and includes mostly random components. It is obtained by propagating precisions of the radiance measurements, estimates of constrained parameter uncertainties, and forward model inaccuracies (and some calibration uncertainties) through the retrieval software. Note, however, that the quality fields used for TEMP and UTH are based on an estimate of the forward model accuracy, since the random uncertainties (from radiance noise) are significantly less than our ability to accurately compute the forward model for these parameters. At the conclusion of the retrieval, the estimated uncertainty is compared with the assumed a priori uncertainty, and if the ratio is greater than 0.5, the quality is set negative, indicating that the retrieved value is influenced more than 25% by the a priori. The Version 5 data generally exhibit less scatter than would be expected from the `quality' field. This is due to the influence of the a priori data and its vertical smoothing on the retrieved products. Additional quality indicators exist in the MLS Level 3 parameter files (one file/day); these files are described in the "Standard Formatted Data Units, MLS Level 3TP Parameter File" and "Standard Formatted Data Units, MLS Level 3LP Parameter File" Version 4 documents, also available on the MLS homepage. The diagnostic flag MMAF_STAT should be used to screen poor data (see more discussion below) on an individual major frame basis; possible values of this (single character) flag are described in the Table below. Also, other diagnostic parameters in the MLS Level 3 parameter files are related to radiance quality and radiance fits in each band (flags have names corresponding to the primary parameter being retrieved in each band (QUALITY_TEMP, QUALITY_CLO, QUALITY_O3_205, QUALITY_H2O, and QUALITY_O3_183 for bands 1, 2, 4, 5, and 6, respectively). These flags have a value less than 4 under (very infrequent) abnormal conditions (at least among major frames with MMAF_STAT equal to 'G', 't', or 'T'). Table describing the major frame quality indicator MMAF_STAT _____________________________________________________________________ |_____________________________________________________________________| | | | MMAF_STAT MMAF_STAT | | Value Meaning | |_____________________________________________________________________| | | | 'G' Good Data | |_____________________________________________________________________| | | | 't' Temperatures missing from NCEP data at heights below 22 hPa | |_____________________________________________________________________| | | | 'T' Temperatures missing from NCEP data at heights below 100 hPa| |_____________________________________________________________________| | | | 'M' Missing too many tangent points from scan | |_____________________________________________________________________| | | | 'P' Pointing anomaly | |_____________________________________________________________________| | | | 'S' Scan mode anomaly (i.e., not normal full scan range) | |_____________________________________________________________________| | | | 'B' Bad Data | |_____________________________________________________________________| |_____________________________________________________________________| Occasional interference effects (induced by the switching mirror stepper motor at low spacecraft battery voltage) can perturb the radiances and retrieved parameter values. This problem occurred largely between September 1992 and June 1993, typically just before sunrise (at the satellite location) for a few minutes. Some diagnostics are sensitive to this effect (quality fields in the Level 3 parameter files show a degradation for ClO and O3_205), but this is not reflected in the error bars (quality values) given in the Level 3A files. After 2.3 years in orbit (in late December 1993), the antenna-scanning mechanism began to exhibit signs of wear. March 1994 through May 1994, and July 1994, were periods of testing and significantly reduced data gathering; these months have from one third of the days with bad data to almost all bad days (days with no profiles retrieved). August and September of 1994 contain mostly good data, but the months of October 1994 through January 1995 again have very few days of useful atmospheric profile data. Reverse scanning and other modifications to the operations (including short periods of "mechanism rest" every orbit) have been imple- mented since February 1, 1995. Very little limb data gathering (typically only a few days per month) occurred from February through July 1995, which was a period during which instrument power sharing began for UARS (in May 1995) because of poor solar array performance. Since June 1995, MLS has been in a mode of operation characterized by off periods for power savings and on periods during which typically 2 days of full (reverse) scans are obtained followed by one day of limb tracking at altitudes near 18 km. The limb tracking days do not lead to standard catalogued profile data files, although there is some information on the atmosphere from those days. From this time period (mid-1995) until mid-June 1997, the number of good full scan days was limited to as little as a few per month up to about half the month, depending on power allocation and scan problems. After June 15, 1997 (UARS Day 2104), MLS was operated with only the 205-GHz radiometer, in order to conserve power, and temperature is not retrieved (but vertical pointing information is obtained from the 205-GHz ozone linewidth sensitivity to pressure). The days from June 15, 1997 on are therefore subject to possible biases if compared to prior days or years. Based on our analyses at the time of writing for this document, we believe that these biases are small. Problems with the MLS antenna scan intensified greatly after June 29, 1998 (UARS Day 2483), and a much more limited number of good profiles per day was obtained after this date (typically from a few percent to 20 percent of possible maximum coverage), thereby significantly degrading the availability and scientific value of MLS data. This limited data availability continued to worsen until July 27, 1999 (UARS Day 2876), the last attempt in 1999 at obtaining MLS data. The MLS instrument was "mothballed" in the fall of 1999. It will be powered up for occasional tests, and for a few correlative data studies (in particular, towards the goal of supporting the EOS CHEM mission, if possible). Calendars describing the various instrument modes (especially useful for time periods since March 1994) are available on the MLS Web site (http://mls.jpl.nasa.gov); however, access of the diagnostic parameters mentioned above should generally allow users to appropriately screen for bad quality data and is needed for proper use of the MLS data. PARAMETER #1: TEMP (Temperature) MLS measurement of temperature is described in detail in the Version 3 validation paper by Fishbein et al. [JGR 101, 9938-10016, 1996], and users of MLS temperature data are advised to study that paper. CHANGES FROM VERSION 4 TO VERSION 5 AFFECTING TEMP MLS V5 temperature is retrieved on every UARS surface while the V4 retrieval is only on every other (even) UARS surface with interpolated values placed on the odd surfaces. Reliable temperatures in V5 are at pressure surfaces of 32-0.46 hPa. V5 software also retrieves temperature in the mesosphere (0.32-0.01 hPa), where values are mainly climatological in V4. The retrieval of the mesospheric temperature is made by incorporating the saturated radiances of channels 7-9 in the 63-GHz radiometer. The mesospheric temperature is still a research product and further quantitative validation needs to be carried out. Users should consult the MLS science team regarding the use of these measurements. V5 corrects an interpolation error for the 1000-68 hPa temperatures which should relax to NCEP values. This error has affected the temperatures in MLS files prior to V5 and can be 1-5 K at odd UARS surfaces. The a priori temperature uncertainty in the V5 retrieval is set to 10 K at 46-3.2 hPa and linearly increased to 46 K from 3.2 hPa to 0.0001 hPa, whereas in V4 the values are 20 K at all levels. DIFFERENCES BETWEEN AVERAGE VERSION 5 AND VERSION 4 TEMP The V5 temperature is generally warmer (by 1-3 K) than the V4 in the stratosphere but cooler (by ~1 K) at the stratopause. These changes reduce the sharpness of the stratopause from that seen in the V4 temperature. Typical V5-V4 differences are listed in the table below, showing some variations with latitude and season. As in V4, the V5 temperature is warmer than NCEP temperature in the stratosphere. Differences between Version 5 and Version 4 Temperatures ----------------------------------------------------- | UARS |Pressure| V5-V4 Differences (1) | | standard| | (K) | | index | (hPa) | global | tropical | polar winter | |-----------------------------------------------------| | 20 | 0.46 | 0.0 | -0.6 | 2.0 | | 19 | 0.68 | 1.2 | 0.2 | -0.1 | | 18 | 1.0 | - 0.7 | -0.4 | -2.1 | | 17 | 1.5 | 1.9 | 1.8 | 1.3 | | 16 | 2.2 | 3.0 | 2.2 | 3.2 | | 15 | 3.2 | 1.8 | 1.2 | 1.7 | | 14 | 4.6 | 1.7 | 1.7 | 0.6 | | 13 | 6.8 | 1.2 | 0.7 | 1.3 | | 12 | 10 | 2.8 | 2.3 | 2.8 | | 11 | 15 | 1.9 | 1.8 | 1.7 | | 10 | 22 | 2.8 | 1.3 | 3.1 | | 9 | 32 | 1.5 | 1.9 | 0.5 | ----------------------------------------------------- Note: The differences are based upon the first full year of MLS data. ESTIMATED PRECISION AND ACCURACY OF TEMP ---------------------------------------------------------------------- | | | Typical | Typical | | | |UARS | P | Vertical | Single profile| Accuracy | V5-ISAMS | |index| |Resolution (1)| precision | |Difference (2)| | | (hPa) | (km) | (K) | (K) | (K) | |----------------------------------------------------------------------| | 30 | 0.010 | 20. | TBD | TBD | 9.7 | | 29 | 0.015 | 20. | TBD | TBD | 8.8 | | 28 | 0.022 | 25. | TBD | TBD | 7.1 | | 27 | 0.032 | 25. | TBD | TBD | 6.5 | | 26 | 0.046 | 30. | TBD | TBD | 5.5 | | 25 | 0.068 | 20. | TBD | TBD | 5.8 | | 24 | 0.10 | 15. | TBD | TBD | 3.4 | | 23 | 0.15 | 10. | TBD | TBD | 0.3 | | 22 | 0.22 | 10. | TBD | TBD | -0.4 | | 21 | 0.32 | 10. | TBD | TBD | 1.1 | ---------------------------------------------------------------------- ------------------------------------------------------------------------- | | | | Typical | | | | |UARS | | Vertical |Single Profile| |Estimated | V5-NCEP | |Index| P |Resolution(1)| Precision (3)| Ratio(4)|Accuracy(5)| Diff. (6)| | |(hPa)| (km) | (K) | | (K) | (K) | |-------------------------------------------------------------------------| | 20 | 0.46| 5.0 | 3.3 | .7 | 5 | -1.5 | | 19 | 0.68| 7.0 | 2.1 | .5 | 5 | 5.2 | | 18 | 1.0 | 6.5 | 1.8 | .5 | 5 | 8.6 | | 17 | 1.5 | 7.0 | 1.7 | .5 | 5 | 6.6 | | 16 | 2.2 | 6.5 | 1.5 | .5 | 4 | 4.3 | | 15 | 3.2 | 6.5 | 1.5 | .5 | 4 | 1.4 | | 14 | 4.6 | 6.0 | 1.4 | .5 | 5 | 0.9 | | 13 | 6.8 | 6.5 | 1.4 | .5 | 4 | 0.1 | | 12 |10. | 6.5 | 1.3 | .4 | 4 | 1.1 | | 11 |15. | 6.0 | 1.2 | .4 | 4 | 1.1 | | 10 |22. | 7.0 | 0.8 | .3 | 4 | 1.6 | | 9 |32. | 6.5 | 0.9 | .3 | 6 | 2.0 | ------------------------------------------------------------------------- Notes: (1) Vertical resolution given here is the full width at half-maximum of the rows of the averaging kernel matrix computed for the nominal MLS operational scan and radiometer noise. (2) V5-ISAMS differences are based on the data in UARS month 9110 (4 Dec 1991 to 14 Jan 1992) when both instruments have a similar sampling coverage. (3) Precisions (1-sigma) are estimates obtained by computing the minimum monthly rms variability of individual measurements in the 5S to 5N latitude band (during Oct. 1991 to Sep. 1992). Uncertainties quoted in the Level 3 files are generally intended to be estimates of precision. They are typically larger than the precision estimates described in this table because of the influence of the a priori data and its vertical smoothing on the retrieved products. (4) These ratios are the typical precisions divided by typical uncertainties in the Level 3 files. The values given in the Level 3 files account for variations in the uncertainty that might occur from profile to profile for various reasons (e.g., missing channels or tangent point scan positions would increase the uncertainties); however, as mentioned above, they overestimate the actual precision of the measurements. The estimated uncertainties given in the Level 3 files should be multiplied by these ratios to obtain the best estimate of the precision. (5) Accuracy is estimated from the error analysis described in the temperature validation paper [JGR 101, 9983-10,016, 1996] that includes instrument and spectroscopic uncertainties. (6) The V5-NCEP differences are for global averages from the first full year of MLS data. KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN TEMP There is a ~0.5 K systematic artifact in the V5 temperature at 32-0.46 hPa that is synchronized with the UARS orbit and yaw cycles. This artifact is reduced by nearly 50% from that in V4. CAVEATS FOR TEMP The temperature retrieval uses a priori values based upon NCEP daily analyses and a monthly climatology [Fleming et al., NASA Tech. Memo., 100697, 1988]. Below 1 hPa, the a priori profile is mainly the NCEP temperature, while above that it is mostly monthly climatology. The MLS temperature retrieval relaxes to the a priori profile as MLS loses sensitivity to temperature. FOR THE TEMPERATURE MEASUREMENTS AT 32-0.46 hPa, THE PARAMETER FILES SHOULD BE EXAMINED AND ONLY PROFILES WITH MMAF_STAT='G', 't', or 'T', AND QUALITY_TEMP = 4 SHOULD BE USED FOR SCIENTIFIC STUDIES. Moreover, data with negative estimated uncertainties (quality field in Level 3 data files) should not be used (although such values are rare in Version 5 data, for the vertical range recommended above). The MLS 63-GHz radiometer was turned off on June 15, 1997 (UARS Day 2104) to conserve spacecraft power. The temperature after that day should not be used. PARAMETER #2: O3_205 (Ozone retrieved from 205-GHz radiometer radiances) For useful information on previous data versions, the user is referred to the MLS O3 validation paper by Froidevaux et al. [JGR 101, 10,017- 10,060, 1996], along with the articles by Cunnold et al. [JGR 101, 10,061-10,075, 1996, and JGR 101, 10,335-10,350, 1996]. Version 4 data quality has been discussed in the WMO Ozone Report No. 43 ["Assessment of Trends in the Vertical Distribution of Ozone", May 1998, edited by N. Harris, R. Hudson, and C. Phillips]; see also Cunnold et al. ["Uncertainties in upper stratospheric ozone trends from 1979 to 1996", JGR Atmospheres, in press, 2000]. CHANGES FROM VERSION 4 TO VERSION 5 AFFECTING O3_205 The main change is the use of a finer retrieval grid in the stratosphere; this grid is on every UARS pressure surface for pressures larger than 0.1 hPa, and every other UARS surface for pressures less than 0.1 hPa. The use of a larger a priori uncertainty value leads to better retrieval sensitivity, particularly in the mesosphere (where the a priori uncertainties were significantly lower in Version 4). The recommended vertical range now extends from 100 hPa up to 0.1 hPa (see also caveats in a later sub-section). In Version 5, the algorithm for setting the quality flag "QUALITY_O3_205" (which helps to screen bad data) was modified because of changes in the chi-square statistic describing the fit to the radiances. The chi-square statistic for this band is now less correlated with anomalies in the retrieved O3_205 values than it was in Version 4. More records in Version 5 have been assigned quality flags less than "4", indicating bad fits to the radiances (and/or bad radiances), than were flagged bad in Version 4. Except for unusual months, this parameter, together with the other criteria for selecting "good" O3_205 retrievals (see the "Caveats" section below), discards about 2% of the Version 5 data, compared to about 1% in Version 4. Thus some individual profiles that passed the recommended quality control measures in Version 4 may be screened out using the same procedures with the Version 5 data, even though some of these retrieved profiles do not *appear* obviously bad. DIFFERENCES BETWEEN AVERAGE VERSION 5 AND VERSION 4 O3_205 The following table shows differences between MLS Version 5 and Version 4 average profiles. Because in Version 4 the retrievals were performed only on the even UARS surfaces and were reliable only at certain pressure surfaces, only the differences on these surfaces are tabulated. Separate comparisons are made for different conditions, as follows: (1) global: ~400,000 profiles from all latitudes for the first 10 full UARS months (2) tropical: ~60,000 profiles from 10S-10N latitude for the first 10 full UARS months (3) midlatitude: ~25,000 profiles from 35-45N and 35-45S latitude for the first 10 full UARS months Differences between Version 5 and Version 4 O3_205 --------------------------------------------- | P | global | tropical | mid- | | | | | latitude | | (hPa) |(ppmv) (%) | (ppmv) (%) | (ppmv) (%) | |-------------------------------------------- | | 0.46 | 0.0 0 | -0.05 -3 | 0.0 0 | | 1 | 0.2 7 | 0.3 10 | 0.2 6 | | 2.2 |-0.1 -2 | -0.1 -2 | -0.1 -2 | | 4.6 |-0.1 -1 | 0.0 0 | -0.1 -2 | | 10 |-0.1 -1 | -0.3 -3 | 0.0 0 | | 22 |-0.1 -1 | -0.2 -3 | -0.1 -1 | | 46 | 0.6 37 | 1.1 304 | 0.5 20 | | 100 |-0.5 -53 | -0.9 -82 | -0.4 -41 | --------------------------------------------- Note: differences are V5-V4; percentages are relative changes from V4. Version 5 data exhibit an overall decrease from V4 of 1 to 3 percent between 10 and 2 hPa, with a 5 to 10 percent increase at 1 hPa. The lower stratosphere shows the largest differences, particularly in the tropics; V5 values are systematically larger than V4 values at 46 hPa (by about 0.5 to 1 ppmv) and smaller at 100 hPa (by about 0.5 to 1 ppmv). Differences in the polar regions are generally smaller than the mid- latitude differences shown above (the decrease from V4 to V5 at 100 hPa is often only 10 to 20%, although it is about 40% during the southern winter of 1992). Except at 100 hPa, where the noise (precision) is better than the V4 value, V5 data are generally noisier than V4 data (typical precision is 0.3 ppmv rather than 0.2 ppmv, see Table below); note, however, that the retrieval grid is twice as fine as in Version 4. ESTIMATED PRECISION AND ACCURACY OF O3_205 -------------------------------------------------------------------------- | | | Typical | Typical | Ratio | | |UARS | | Vertical | Single Profile| of precision | Estimated | |Index|Pressure| Resol. (1)| Precision (2) | to uncertainty| Accuracy (4)| | | (hPa) | (km) | (ppmv) (%) |in L3 files (3)|(ppmv) (%) | |--------------------------------------------------------------------------| | 24 | 0.1 | 8 | 0.5 | See (5)| 0.6 | TBD | TBD | | 23 | 0.15 | 11 | 0.4 | See (5)| 0.5 | TBD | TBD | | 22 | 0.22 | 6 | 0.4 | See (5)| 0.6 | TBD | TBD | | 21 | 0.32 | 8 | 0.35 | 25 | 0.6 | TBD | TBD | | 20 | 0.46 | 5 | 0.35 | 20 | 0.6 | 0.3 | 15 | | 19 | 0.68 | 4 | 0.3 | 12 | 0.6 | 0.3 | 12 | | 18 | 1.0 | 5 | 0.3 | 10 | 0.6 | 0.3 | 10 | | 17 | 1.5 | 5 | 0.3 | 7 | 0.7 | 0.3 | 7 | | 16 | 2.2 | 4 | 0.3 | 5 | 0.7 | 0.3 | 5 | | 15 | 3.2 | 4 | 0.3 | 4 | 0.8 | 0.3 | 4 | | 14 | 4.6 | 4 | 0.3 | 4 | 0.8 | 0.3 | 4 | | 13 | 6.8 | 4 | 0.3 | 4 | 0.8 | 0.3 | 4 | | 12 | 10 | 3.5 | 0.3 | 4 | 0.8 | 0.3 | 4 | | 11 | 15 | 3.5 | 0.3 | 4 | 0.9 | 0.3 | 4 | | 10 | 22 | 3.5 | 0.3 | 5 | 0.9 | 0.4 | 5 | | 9 | 32 | 3.5 | 0.3 | 8 | 0.9 | TBD | TBD | | 8 | 46 | 3.5 | 0.25 | 10 | 0.7 | TBD | TBD | | 7 | 68 | 4 | 0.25 | 20 | 0.6 | TBD | TBD | | 6 | 100 | 4 | 0.4 | >50 | 0.7 | TBD | TBD | -------------------------------------------------------------------------- Notes: (1) Vertical resolution given here is the full width at half-maximum of the rows of the averaging kernel matrix computed for the nominal MLS operational scan and radiometer noise. (2) Precisions are estimates obtained by computing the minimum monthly rms variability of individual measurements in the 5S to 5N latitude band (from each of the first 10 full UARS months of the MLS mission; i.e., Oct. 1991 to Sep. 1992). Uncertainties quoted in the Level 3 files are generally intended to be estimates of precision. They are typically larger than the precision estimates quoted above because of the influence of the a priori data and its vertical smoothing on the retrieved products. (3) These ratios are the typical precisions divided by typical uncertainties in the Level 3 files. The values given in the Level 3 files account for variations in the uncertainty that might occur from profile to profile for various reasons (e.g., missing channels or tangent point scan positions would increase the uncertainties); however, as mentioned above, they overestimate the actual precision of the measurements. The estimated uncertainties given in the Level 3 files should be multiplied by these ratios to obtain the best estimate of the precision. (4) Accuracies are preliminary estimates based on various error sources and statistical comparisons with other data sets, SAGE II data in particular; they should be viewed as "1-sigma" estimates. Accuracy values will be finalized later, particularly for the mesosphere and lower stratosphere, but we expect the values to be similar to or better than the values published for previous data versions. (5) At pressures lower than about 0.2 hPa, day/night differences in ozone mixing ratios become significant enough that absolute (ppmv) precision or accuracy becomes the most convenient quantity to use. KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN O3_205 Biases between Version 4 data and other data sets in the lower stratosphere have been significantly reduced, based on comparisons of Version 5 MLS and SAGE II data, in particular. The changes in the upper stratosphere reduce the already small biases that existed in that region between SAGE II and MLS Version 4 O3_205 data. Furthermore, significant biases that existed between MLS V4 values for O3_205 and SAGE II ozone values at 46 and 100 hPa (especially in the tropics) have been largely removed in Version 5 MLS data. The 205 GHz and 183 GHz ozone retrievals from MLS give results which generally track each other well, with some differences in the lower stratosphere (especially at low latitudes) probably mostly caused by problems with the O3_183 retrievals (see caveats in the O3_183 section of this document). Slight variations (a few percent or less) still exist, in connection with the 36-day yaw cycle of the satellite, but this effect seems to be smaller than in Version 4 data. More detailed analyses with MLS V5 O3_205 data are needed to draw quantitative conclusions about any remaining systematic effects in these data. CAVEATS FOR O3_205 The retrieval for ozone uses an a priori estimate based on a month- dependent latitude-dependent climatology developed by the UARS and MLS science teams. While the O3_205 profiles extend from 464 hPa to 4.6e-4 hPa, the values at pressures larger than 100 hPa are mostly climatological and do not contain useful independent information from MLS. O3_205 values at 100 hPa, although apparently improved over V4 data, are still to be used with caution, and with attention to the error bars, especially for single profiles. Some information exists at pressures less than the top pressure tabulated above (0.1 hPa), but the vertical resolution is poorer and the influence of the a priori increases; users should consult with the MLS science team if they are interested in using some values at these lower pressures (where averaging significant amounts of data would generally be required). O3_205 is the recommended ozone for stratospheric studies, primarily because of the better calibration and accuracy. However, the precision of O3_183 is generally somewhat better than the O3_205 precision, and O3_183 is the recommended choice for studies of the mesosphere, where the 205-GHz radiances lose sensitivity faster than the 183 GHz-radiances. Unfortunately, the O3_183 data set is limited to the time period from late September 1991 to April 15, 1993 (as mentioned in the summary information earlier on). THE PARAMETER FILES SHOULD BE EXAMINED AND ONLY PROFILES WITH MMAF_STAT='G', 't', or 'T' AND QUALITY_O3_205=4 SHOULD BE USED FOR SCIENTIFIC STUDIES. Profiles with MMAF_STAT set to 'T' (which occurs quite infrequently) may have slight biases at pressures larger than or equal to 46 hPa. Moreover, data with negative estimated uncertainties (quality field in Level 3 data files) should not be used (although such values are rare in Version 5 data, for the vertical range recommended above). PARAMETER #3: O3_183 (Ozone retrieved from 183-GHz radiometer radiances) For useful information on previous data versions, the user is referred to the MLS O3 validation paper by Froidevaux et al. [JGR 101, 10,017- 10,060, 1996], and the mesospheric O3 validation paper by Ricaud et al. [JGR 101, 10,077-10,089]. CHANGES FROM VERSION 4 TO VERSION 5 AFFECTING O3_183 A finer retrieval grid is used in the stratosphere; this grid is on every UARS pressure surface for pressures larger than 0.1 hPa, and every other UARS surface for pressures less than 0.1 hPa. New values were deduced for the sideband ratios and spectral parameters. This was done by retrieving the parameters along with zonal means of O3_183, from zonal mean radiances. Details are in Pumphrey and Buhler [JQSRT, vol.64/4, pp.421-437 (1999)]. An iterative retrieval is used as the problem is somewhat nonlinear. Version 5 makes use of a full forward model for the 5 channels furthest from line center. Radiance measurements with tangent heights below 100 hPa are not used, nor are those with opacities greater than 1.0 (mainly for reasons of computer efficiency). Channel 76 (a far-wing channel for this ozone band) is not used at all, since it does not appear to be stable and reliable enough. DIFFERENCES BETWEEN AVERAGE VERSION 5 AND VERSION 4 O3_183 The following table shows differences between MLS Version 5 and Version 4 average profiles. Because in Version 4 the retrievals were performed only on the even UARS surfaces and were reliable only at certain pressure surfaces, only the differences on these surfaces are tabulated. Separate comparisons are made for different conditions, as follows: (1) global: ~400,000 profiles from all latitudes for the first 10 full UARS months (2) tropical: ~60,000 profiles from 10S-10N latitude for the first 10 full UARS months (3) midlatitude: ~25,000 profiles from 35-45N and 35-45S latitude for the first 10 full UARS months Differences between Version 5 and Version 4 O3_183 --------------------------------------------- | P | global | tropical | mid- | | | | | latitude | | (hPa) |(ppmv) (%) | (ppmv) (%) | (ppmv) (%) | |--------------------------------------------- | 0.046 | 0.01 2 |-0.03 -5 | -0.05 -7 | | 0.1 | 0.04 4 | 0.03 3 | 0.05 5 | | 0.22 | 0.2 18 | 0.2 18 | 0.2 20 | | 0.46 | 0.2 9 | 0.1 8 | 0.15 8 | | 1 | 0.4 12 | 0.4 14 | 0.4 12 | | 2.2 | 0.5 9 | 0.5 9 | 0.5 9 | | 4.6 | 0.0 0 | 0.0 0 | -0.1 -1 | | 10 | 0.5 6 | 0.5 5 | 0.5 7 | | 22 |-0.5 -8 | -0.4 -6 | -0.6 -10 | | 46 | 0.5 23 | 0.8 97 | 0.4 15 | --------------------------------------------- Note: differences are V5-V4; percentages are relative changes from V4. Version 5 data exhibit an overall increase from Version 4 of about 5 to 10 percent (and occasionally as much as 20 percent) between 10 and 0.1 hPa. Version 5 values are larger than Version 4 values (by 0.4 to 0.8 ppmv) at 46 hPa, especially in the tropics, but Version 5 data at 22 hPa show a 5 to 10% decrease versus Version 4. Differences in the polar regions are generally comparable to midlatitude differences at and above 10 hPa (where a few to 10% increase occurs in Version 5, versus Version 4); in the polar stratosphere, Version 5 values are typically slightly smaller than Version 4 values (by up to about 10%). Version 5 data are generally noisier than Version 4 data (typical precision is 0.2 to 0.3 ppmv rather than 0.1 to 0.2 ppmv in the stratosphere, see the Table below); note, however, that the retrieval grid is twice as fine as in Version 4. ESTIMATED PRECISION AND ACCURACY OF O3_183 --------------------------------------------------------------------------- | | | | Typical | Ratio | | |UARS | | Vertical | Single Profile| of precision | Estimated | |Index|Pressure| Resol. (1)| Precision (2) | to uncertainty| Accuracy (4) | | | (hPa) | (km) | (ppmv) (%) |in L3 files (3)|(ppmv) (%) | |---------------------------------------------------------------------------| | 32 | 0.005 | 14 | 0.3 | See (5)| 0.3 | TBD | TBD | | 31 | 0.007 | 13 | 0.3 | See (5)| 0.3 | TBD | TBD | | 30 | 0.01 | 12 | 0.3 | See (5)| 0.4 | TBD | TBD | | 29 | 0.015 | 9 | 0.25 | See (5)| 0.4 | TBD | TBD | | 28 | 0.022 | 7 | 0.25 | See (5)| 0.4 | TBD | TBD | | 27 | 0.032 | 7 | 0.2 | See (5)| 0.4 | TBD | TBD | | 26 | 0.046 | 6 | 0.2 | See (5)| 0.5 | <0.5 | See (5)| | 25 | 0.068 | 6 | 0.15 | See (5)| 0.4 | <0.4 | See (5)| | 24 | 0.1 | 6 | 0.15 | See (5)| 0.4 | <0.3 | See (5)| | 23 | 0.15 | 8 | 0.15 | See (5)| 0.3 | <0.3 | See (5)| | 22 | 0.22 | 5 | 0.15 | See (5)| 0.3 | <0.2 | See (5)| | 21 | 0.32 | 7 | 0.15 | 10 | 0.3 | <0.2 | <10 | | 20 | 0.46 | 3.5 | 0.15 | 8 | 0.4 | <0.2 | <10 | | 19 | 0.68 | 3.5 | 0.2 | 8 | 0.7 | <0.3 | <15 | | 18 | 1.0 | 4 | 0.2 | 6 | 0.7 | <0.4 | <15 | | 17 | 1.5 | 3.5 | 0.2 | 5 | 0.7 | <0.6 | <15 | | 16 | 2.2 | 3.5 | 0.25 | 4 | 0.9 | <0.8 | <15 | | 15 | 3.2 | 3.5 | 0.25 | 4 | 0.8 | <0.8 | <10 | | 14 | 4.6 | 3 | 0.3 | 4 | 1 | <0.9 | <10 | | 13 | 6.8 | 3 | 0.3 | 3 | 1 | <0.9 | <10 | | 12 | 10 | 3 | 0.3 | 3 | 0.9 | <0.9 | <10 | | 11 | 15 | 3 | 0.3 | 4 | 1 | <0.8 | <10 | | 10 | 22 | 3 | 0.3 | 5 | 1 | <0.8 | <20 | | 9 | 32 | 3 | 0.25 | 6 | 1 | <0.8 | <30 | | 8 | 46 | 3.5 | 0.2 | 8 | 0.9 | <0.8 | <40 | -------------------------------------------------------------------------- Notes: (1) Vertical resolution given here is the full width at half-maximum of the rows of the averaging kernel matrix computed for the nominal MLS operational scan and radiometer noise. (2) Precisions are estimates obtained by computing the minimum monthly rms variability of individual measurements in the 5S to 5N latitude band (from each of the first 10 full UARS months of the MLS mission; i.e., Oct. 1991 to Sep. 1992). Uncertainties quoted in the Level 3 files are generally intended to be estimates of precision. They are typically larger than the precision estimates quoted above because of the influence of the a priori data and its vertical smoothing on the retrieved products. (3) These ratios are the typical precisions divided by typical uncertainties in the Level 3 files. The values given in the Level 3 files account for variations in the uncertainty that might occur from profile to profile for various reasons (e.g., missing channels or tangent point scan positions would increase the uncertainties); however, as mentioned above, they overestimate the actual precision of the measurements. The estimated uncertainties given in the Level 3 files should be multiplied by these ratios to obtain the best estimate of the precision. (4) Accuracies are preliminary estimates based mainly on previous data versions and will be finalized later, particularly for the mesosphere and lower stratosphere, but we expect these to be similar to or better than the accuracy values published for previous data versions. There is good agreement between O3_205 values and SAGE II and other data sets (often within 5% or less); this, coupled with the knowledge that global differences (over a year) between the O3_183 and O3_205 values (at 46 to 0.1 hPa) average about 7% or less, leads to a preliminary accuracy estimate for Version 5 O3_183 data of less than 10 to 15% for most of this altitude region (but see section below on systematic effects). (5) At pressures lower than about 0.2 hPa, day/night differences in ozone mixing ratios become significant enough that absolute (ppmv) precision or accuracy becomes the most convenient quantity to use. Even at the top of the vertical range shown above, a precision of 20% is possible, for nighttime measurements. KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN O3_183 The 205 GHz and 183 GHz ozone retrievals from MLS give results that generally track each other well, with differences probably mostly caused by problems with the O3_183 retrievals. Average profiles for these two retrieved products show the existence of systematic notches in the O3_183 profiles (values at 22 and 68 hPa tend to dip lower than the O3_205 values), particularly at high latitudes. Also, the O3_183 values exhibit a systematic positive bias of several tenths of a ppbv (or more) in the tropics at 100 hPa, with respect to the O3_205 values, along with tropical values at 68 hPa that are often biased in the opposite sense (and/or negative on average) and show unrealistic variations that seem tied to the 36-day yaw cycle of the satellite. We therefore do not recommend general use of the O3_183 data set at pressures larger than 46 hPa, even though the tracking between the 2 MLS ozone data sets at mid- to high latitudes appears to be very good, even at 46 to 100 hPa. On average, for global results over the first year of UARS MLS data, the O3_183 values are about 2 to 8% larger than the O3_205 values, except for the negative O3_183 notch at 22 hPa (where the O3_183 values are lower than the O3_205 values by about 4%); also, O3_183 data are 2 to 7% smaller than the O3_205 values (on average) for pressures between 0.2 and 0.1 hPa. CAVEATS FOR O3_183 The retrieval for ozone uses an a priori estimate based on a month- dependent latitude-dependent climatology developed by the UARS and MLS science teams. While the O3_183 profiles extend from 464 hPa to 4.6e-4 hPa, the values at pressures larger than 100 hPa are mostly climatological and do not contain useful independent information from MLS. As mentioned above, we do not recommend general use of the O3_183 data set at pressures larger than 46 hPa, because of systematic effects (especially in the tropics). Some information exists at pressures less than the top pressure tabulated above (0.005 hPa), but the vertical resolution is poorer and the influence of the a priori increases; users should consult with the MLS science team if they are interested in using some values at these lower pressures (where averaging significant amounts of data would generally be required). O3_205 is the recommended ozone for stratospheric studies, primarily because of the better calibration and accuracy. However, the precision of O3_183 is generally somewhat better than the O3_205 precision, and O3_183 is the recommended choice for studies of the mesosphere, where the 205-GHz radiances lose sensitivity faster than the 183-GHz radiances. Unfortunately, the O3_183 data set is limited to the time period from late September 1991 to April 15, 1993 (as mentioned in the summary information earlier on). THE PARAMETER FILES SHOULD BE EXAMINED AND ONLY PROFILES WITH MMAF_STAT='G', 't', or 'T' AND QUALITY_O3_183=4 SHOULD BE USED FOR SCIENTIFIC STUDIES. Profiles with MMAF_STAT set to 'T' (which occurs quite infrequently) may have slight biases at pressures larger than or equal to 46 hPa. Moreover, data with negative estimated uncertainties (quality field in Level 3 data files) should not be used (although such values are rare in Version 5 data, for the vertical range recommended above). PARAMETER #4: H2O (Water Vapor) For useful information on previous data versions, the user is referred to the following documents: Version 3 is described in Lahoz et al. [JGR 101, 10,129-10,149, 1996]. Version 4, and an interim prototype tagged as Version 104, are described by Pumphrey [JGR vol.104, No.D8, pp.9399-9412, 1999]. CHANGES FROM VERSION 4 TO VERSION 5 AFFECTING H2O A finer retrieval grid is used in the stratosphere; this grid is on every UARS pressure surface for pressures larger than 0.1 hPa, and every other UARS surface for pressures less than 0.1 hPa. New values were deduced for the sideband ratios and several spectral parameters. This was done by retrieving the parameters along with zonal means of H2O, from zonal mean radiances. Details are in Pumphrey and Buhler [JQSRT, vol.64/4 pp.421-437 (1999)]. An iterative retrieval was used as the problem is somewhat nonlinear. Radiance measurements with tangent heights below 100 hPa were not used, nor were those with opacities greater than 1.0. As a result of these restrictions and the a priori covariance matrix used, the lowest level at which the retrieved product contains useful information is 68 hPa. One of the channels (channel 69) was not used at all. For reasons we still do not understand, this channel appears to be subject to large systematic errors. The Level 3 H2O files include retrieved tropospheric H2O mixing ratios at 147, 215, 316, and 464 hPa, based on the UTH (upper tropospheric humidity) retrievals, converted to mixing ratio units. However, it should be emphasized that the tropospheric H2O and the stratospheric H2O in these files are from different radiometers, and that consistency between these two data sets at 100 hPa has not yet been achieved (100 hPa values are not considered reliable enough). For pressures larger than 464 hPa, a constant relative humidity equal to the value at 464 hPa is assumed and converted to ppmv units for the Level 3 H2O files. DIFFERENCES BETWEEN AVERAGE VERSION 5 AND VERSION 4 H2O The following table shows differences between MLS Version 5 and Version 4 average profiles, based on global comparisons for ~400,000 profiles for the first 10 full UARS months. Because in Version 4 the retrievals are performed only on the even UARS surfaces and are reliable only at certain pressure surfaces, only the differences on these surfaces are tabulated. Differences between Version 5 and Version 4 H2O --------------------------------- | | | | | UARS | Pressure | Global Diff. | | Index | (hPa) | (ppmv) (%) | |---------------------------------| | 22 | 0.22 | -1.2 -18 | | 20 | 0.46 | 0.2 4 | | 18 | 1 | -0.1 -1 | | 16 | 2.2 | -0.35 -6 | | 14 | 4.6 | -0.2 -4 | | 12 | 10 | -0.05 -1 | | 10 | 22 | 0.3 8 | | 8 | 46 | 0.3 6 | --------------------------------- Note: differences are V5-V4; percentages are relative changes from V4. The differences above are representative of differences at midlatitudes and in the tropics. In the dehydrated Antarctic polar vortex at 46 hPa, V5 is between 1 and 1.5 ppmv wetter than V4. This difference is probably due to the different vertical resolutions of these versions interacting with the large vertical gradient in H2O between 22 and 46 hPa. It is hard to validate any of the MLS data in this region as there are almost no correlative data. ESTIMATED PRECISION AND ACCURACY OF H2O ---------------------------------------------------------------------- | | | | | Ratio | | | | | Typical | Typical | of precision | | | UARS| | Vertical |Single Profile|to uncertainty | Estimated | |Index|Pressure|Resol. (1)|Precision (2) |in L3 files (3)| Accuracy (4)| | | (hPa) | (km) | (ppmv) (%) | | (ppmv) (%) | |----------------------------------------------------------------------| | 30 | 0.01 | 8 | 0.3 | 16 | 0.5 | 1.0 | 51 | | 29 | 0.015 | 7 | 0.3 | 9 | 0.4 | 1.0 | 35 | | 28 | 0.022 | 6 | 0.3 | 7 | 0.4 | 1.0 | 26 | | 27 | 0.032 | 6 | 0.15 | 3 | 0.3 | 1.0 | 22 | | 26 | 0.046 | 6 | 0.2 | 4 | 0.4 | 1.0 | 19 | | 25 | 0.068 | 6 | 0.15 | 3 | 0.3 | 1.0 | 18 | | 24 | 0.1 | 6 | 0.25 | 4 | 0.4 | 1.0 | 18 | | 23 | 0.15 | 7 | 0.2 | 3 | 0.2 | 0.9 | 16 | | 22 | 0.22 | 5 | 0.2 | 4 | 0.4 | 0.9 | 15 | | 21 | 0.32 | 7 | 0.2 | 3 | 0.3 | 0.8 | 14 | | 20 | 0.46 | 3.5 | 0.15 | 3 | 0.4 | 0.8 | 13 | | 19 | 0.68 | 3.5 | 0.2 | 3 | 0.4 | 0.8 | 13 | | 18 | 1 | 4 | 0.2 | 3 | 0.5 | 0.7 | 12 | | 17 | 1.5 | 3.5 | 0.2 | 3 | 0.6 | 0.7 | 11 | | 16 | 2.2 | 3.5 | 0.15 | 3 | 0.5 | 0.6 | 12 | | 15 | 3.2 | 3 | 0.15 | 3 | 0.5 | 0.6 | 11 | | 14 | 4.6 | 3 | 0.15 | 3 | 0.5 | 0.5 | 11 | | 13 | 6.8 | 3 | 0.1 | 3 | 0.4 | 0.5 | 11 | | 12 | 10 | 3 | 0.1 | 3 | 0.4 | 0.5 | 10 | | 11 | 15 | 3 | 0.1 | 3 | 0.4 | 0.5 | 12 | | 10 | 22 | 3.5 | 0.15 | 3 | 0.5 | 0.5 | 12 | | 9 | 32 | 3.5 | 0.15 | 3 | 0.6 | 0.5 | 11 | | 8 | 46 | 3.5 | 0.15 | 4 | 0.5 | 0.5 | 12 | | 7 | 68 | 5 | 0.25 | 6 | 0.4 | 0.8 | 19 | ---------------------------------------------------------------------- Notes: (1) Vertical resolution given here is the full width at half-maximum of the rows of the averaging kernel matrix computed for the nominal MLS operational scan and radiometer noise. (2) Precision estimates are made by calculating the standard deviation in the 5S-5N latitude bin, where the natural variability is small. The results shown were calculated using four separate groups of 5 days in the middle of yaw periods during the first year of operations. The groups of 5 days were processed separately and the results shown are the minima of the results from the four groups. The groups of 5 days were in the middle of yaw periods to reduce the effect of yaw-cycle dependence on the estimated precision. Uncertainties quoted in the Level 3 files are generally intended to be estimates of precision. They are typically about two or three times as large as the precision estimates quoted above because of the influence of the a priori data and its vertical smoothing on the retrieved products. The quoted uncertainties are still rather lower than the accuracy. (3) These ratios are the typical precisions divided by typical uncertainties in the Level 3 files. The values given in the Level 3 files account for variations in the uncertainty that might occur from profile to profile for various reasons (e.g., missing channels or tangent point scan positions would increase the uncertainties); however, as mentioned above, they overestimate the actual precision of the measurements. The estimated uncertainties given in the Level 3 files should be multiplied by these ratios to obtain the best estimate of the precision. (4) Accuracy estimates are typical values for the rms difference between MLS and other instruments. Comparisons were made against HALOE, ATMOS, WVMS and NOAA Frostpoint hygrometers. These estimates should be regarded as pessimistic as they contain the errors of those other instruments. KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN H2O MLS Version 5 H2O has generally smaller biases against other measurements than does Version 4. In particular, the large errors at 0.1 hPa and the large ascending/descending differences associated with them have been eliminated in Version 5. Various other systematic errors remain, however. At 68 hPa (and to some extent at 46 hPa) there is a large systematic error associated with the UARS yaw cycle. The retrieved values at the ends of a yaw month are up to 0.8 ppmv lower than in the middle of the month. The retrieved values at certain levels appear to be consistently higher or lower than the surrounding levels. These spikes are not large, but they are persistent and tend to affect averaged profiles as well. The most persistent such spike is at 1.4 hPa, where values seem high by about 0.2 ppmv. CAVEATS FOR H2O For studies involving the lower stratosphere, between 100 and 46 hPa, we recommend the interim "Version 104" data set described by Pumphrey [JGR vol.104, No.D8, pp.9399-9412, 1999]. This is available from the BADC (British Atmospheric Data Centre) at http://www.badc.rl.ac.uk/ or direct from the scientist responsible at http://www.met.ed.ac.uk/~hcp/watvap.shtml. These data have smoother profiles than V5 and lack some of the systematic effects mentioned above. They have better information content at 68 hPa and some usable information at 100 hPa. However, they have a consistent dry bias in the stratosphere with respect to V5 and to several other data sets. Some information exists at pressures less than the top pressure tabulated above, but the vertical resolution is poorer and the influence of the a priori increases; users should consult with the MLS science team if they are interested in using some values at these lower pressures. THE PARAMETER FILES SHOULD BE EXAMINED AND ONLY PROFILES WITH MMAF_STAT='G', 't', or 'T' AND QUALITY_O3_205=4 SHOULD BE USED FOR SCIENTIFIC STUDIES. Profiles with MMAF_STAT set to 'T' (which occurs quite infrequently) may have slight biases at pressures larger than or equal to 46 hPa. Moreover, data with negative estimated uncertainties (quality field in Level 3 data files) should not be used (although such values are rare in Version 5 data, for the vertical range recommended above). PARAMETER #5: UTH (Upper Tropospheric Humidity) DESCRIPTION MLS Upper Tropospheric Humidity (UTH) is reported at pressure surfaces of 464, 316, 215 and 147 hPa. The values given on these surfaces should be interpreted as the breakpoints of a piecewise linear profile that best fits the measured MLS radiances. UTH is archived in L3AT files in units of percent relative humidity with respect to ice (%RHi). Also available are Level 2 L2UTH files, an ASCII file containing diagnostic information and UTH profiles in %RHi and parts per million volume (ppmv) mixing ratio units. There are no UTH L3AL files produced for v5. The UTH measurement is described in W. G. Read et al. "Upper Tropospheric Water Vapor from UARS MLS", Bull. Amer. Meteor. Soc. Vol. 76, pp. 2381- 2389, (1995). The onion peeling retrieval described in that paper has been updated with an optimal estimation retrieval for the Version 5 data set. The a priori UTH profile is 50 +/- 150 %RHi with a 4 km (0.25 in log pressure) correlation length. The primary spectroscopy data required for the UTH retrieval are the N2/O2 (dry) and water vapor (wet) continuum functions. These were determined from MLS radiometric data coincident with Vaisala UTH profile soundings and driest atmospheric conditions. The UTH quality is primarily limited by the accuracy with which the dry and wet continuum parameters are determined. Contributions from nitrous oxide, ozone, and nitric acid (minor species correction) are included to improve accuracy. Temperatures used in the retrieval are from NCEP reanalyses. The UTH validation is described in more detail in a manuscript currently being prepared. CHANGES FROM VERSION 4.9 TO VERSION 5 AFFECTING UTH The differences between UTH Versions 4.9 (or 490) and 5 are listed below: The dry and wet continuum functions are determined differently in Version 5 and Version 4.9. Version 4.9 estimated the dry continuum by fitting an appropriate pressure-squared function to the smallest radiance profiles (after making a small correction for minor species) for a few selected days and assuming no water vapor contribution. Then the wet continuum was determined using the largest signal profiles from the same days, assuming atmospheric conditions of 100 %RHi and the dry emission and the minor species correction. Ancillary data such as limb tangent pressure, temperature, and the minor species contributions came from Version 4 retrievals. Version 5 determines the wet and dry continuum functions simultaneously from a larger sample of selected profiles with improved algorithms. Small signal radiance profiles assuming 0 %RHi were combined with radiance profiles coincident (1 degree longitude by 1 degree latitude by three hours) with Vaisala radiosonde UTH measurements to establish the dry and wet continuum functions. Because of this approach the average Version 5 UTH between 300-150 hPa agrees within 5% with the Vaisala sondes. The starting UARS pressure index for a Version 5 L3AT UTH file is 2 (or 464 hPa). It is 0 (1000 hPa) for Version 4.9 (values reported for levels 0 and 1 are equal to the %RHi at 464 hPa). Version 5 L3AT and L2UTH files have negative uncertainties whenever the estimated error is greater than 75 %RHi or the received radiance is contaminated by cloud scattering. Version 4.9 L3AT files have negative uncertainties only when a retrieval is not performed, which is usually caused by an insufficient number of tropospheric radiances. The Version 5 L2UTH MMAF time tag is in msec, whereas in Version 4.9 it is in UT hours. The UTH L3AT MMAF time is 32768 msec later in Version 5 than in Version 4.9. The Version 5 H2O L3AT files have UTH data stored in levels 0-5 in mixing ratio units (parts per volume). UTH in levels 0-2 is the 464 hPa %RHi converted into mixing ratio; these values will vary vertically because the conversion is very sensitive to temperature. There are no UTH data in Version 4 H2O L3AT files. DIFFERENCES BETWEEN AVERAGE VERSION 5 AND VERSION 4.9 UTH The table below shows differences between Version 5 and Version 4.9 average profiles for 1992. The comparisons are averages of all data records where Version 4.9 has a positive uncertainty and Version 5 has an estimated uncertainty less than the a priori uncertainty. The comparisons are separated into the following regions: 1) global: all coincident data. 2) low-latitude: 30S-30N. 3) mid-latitude: 60S-30S and 30N-60N. 4) high-latitude: 90S-60S and 60N-90N. Differences between Version 5 and Version 4.9 UTH (1) ----------------------------------------------------------------- | P | global | low-latitude | mid-latitude | high-latitude| |(hPa)|(%RHi) (%) |(%RHi) (%) |(%RHi) (%) |(%RHi) (%) | |-----------------------------------------------------------------| | 147 | -9 -19 | -13 -15 | -7 -54 | -4 -36 | | 215 | -4 -11 | -4 -9 | -4 -14 | -5 -38 | | 316 | -5 -14 | -1 -4 | -7 -16 | -8 -16 | | 464 | -3 -10 | 2 9 | -4 -13 | -14 -27 | ----------------------------------------------------------------- Note: Difference is V5-V4.9 and percent difference is 100*(V5-V4.9)/V4.9 Version 5 data are generally drier than Version 4.9 because of the different approaches for determining the wet and dry continuum functions. ESTIMATED PRECISION AND ACCURACY OF UTH ----------------------------------------------------------------------- |UARS| Press | typical | Latitude Ranges | |ind | | vertical |-----------------------------------------------| | | |resolution| global | low-lat. | mid-lat. | high-lat. | | | | (1) | Prec. Acc.| Prec. Acc.| Prec. Acc.| Prec. Acc.| | | | | (2) (3) | (2) (3) | (2) (3) | (2) (3) | | | (hPa) | km | %RHi %RHi | %RHi %RHi | %RHi %Rhi | %RHi %RHi | |-----------------------------------------------------------------------| | 5 | 147 | 3 | 24 26 | 37 39 | 11 11 | 30 27 | | 4 | 215 | 3 | 11 25 | 17 23 | 5 24 | 21 33 | | 3 | 316 | 3 | 9 25 | 7 23 | 10 24 | 14 38 | | 2 | 464 | 3-6 | 20 52 | 22 64 | 26 50 | 21 45 | ----------------------------------------------------------------------- Notes: (1) Vertical resolution values given here are the full width at half- maximum (FWHM) of the rows of the averaging kernel matrix computed for the nominal MLS scan and radiance uncertainty. The FWHM at 464 hPa broadens with increasing moisture above 30 %RHi. (2) Precision is the root sum square of radiance noise (0.1 K), tropospheric limb tangent pressure precision (150 m), temperature uncertainties (2 K), and minor species corrections (0.4 ppmv O3, 1.5 ppbv HNO3, and 15 ppbv N2O) projected onto the UTH profile. (3) Accuracy is based on propagating the mean radiance fit residual to the wet and dry continuum functions into the individually retrieved UTH profiles. The UTH uncertainties given in the L3AT files are the estimated *accuracy*, as described in note (3). This is different from other MLS constituent retrievals where the uncertainties given in the data files are *precision*. The accuracy of the forward model can affect the retrieved UTH in a non- linear way. Since the accuracy is much worse than the precision from radiance noise, it was decided to use radiance modeling accuracy instead of radiance precision for uncertainties given in the data files. The average quality of the radiometric fit for the derived continuum functions and atmospheric state is considered a reasonable accuracy estimate. KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN UTH The most significant limitations to UTH accuracy are due to uncertainties in the continuum emission. The water vapor continuum function for Version 5 used Vaisala humicap sondes as the UTH reference. Therefore any biases associated with the sonde will be present in MLS Version 5 UTH but not in Version 4.9; the driest regions of the globe are probably slightly wetter than indicated by MLS. UTH at 464 and 316 hPa can be severely underestimated (by as much as 100%) by MLS when the mean tropospheric moisture above 500 hPa is greater than 50 %RHi. This is associated with a nonunique solution producing unusually dry UTH retrievals at 316 and 464 hPa when the upper two levels are moist. The problem is not flagged in the data files. Thick clouds can scatter the radiation received by MLS. The Version 5 forward model does not account for scattering in the atmosphere and therefore is unsuitable for these situations. Scattering usually causes very high %RHi values to be retrieved in the upper two levels (> 200 %RHi) and low values at 316 and 464 hPa. The Version 5 algorithm sets the uncertainty negative when scattering is detected. CAVEATS FOR UTH The 147 and 215 hPa levels contain the best quality data, which are thought to be reliable whenever the uncertainty is positive; data at these levels with negative reported uncertainties should not be used. All values greater than 125 %RHi (chosen to include a 25 %RHi uncertainty) indicate that the measurement was probably taken in the presence of ice. Values greater than 200 %RHi are probably indicative of thick cirrus associated with a convective system. The Level 2 L2UTH files are recommended over the L3AT UTH files for studies using the 316 and 464 hPa levels because they contain the single-layer %RHi, which is a useful data quality indicator for those levels. Only use 316 or 464 hPa UTH data if they are not more than 5 %RHi drier than the single-layer average and the magnitude (regardless of sign) of the uncertainty is less than 110 %RHi. This rejects most erroneous 464 and 316 hPa UTH data where the upper levels are wet or cloudy, situations that give the retrieval algorithm most difficulty. A dry bias is expected for 464 and 316 hPa levels when the atmosphere is wetter than 50%, but qualitative information about the atmospheric morphology should still be useful in these situations. The measurement footprint is approximately 100 km x 100 km x 3 km (horizontal x horizontal x vertical). UTH variability can be greater than the expected precision within the measurement footprint. Therefore comparing an individual MLS UTH profile with a profile from another measurement technique can show worse agreement than the precision. The use of single isolated MLS UTH profiles requires some care. The interpretation of the data is most reliable if several profiles can be combined together (i.e., averaged or gridded to produce maps), when small scale variability within the MLS measurement footprint averages out. This should be borne in mind in scientific investigations. The lower limit on UTH sensitivity is 8-10 ppmv. These data should not be used to study phenomena that vary less than this, for example dynamics of stratospheric water vapor. This limitation prevents a simple joining of the MLS stratospheric water vapor product to the tropospheric product. PARAMETER #6: CLO (chlorine monoxide) This document supplements the MLS ClO validation paper by Waters et al. [JGR 101, 10,091-10,127, 1996], which should be studied by any user of these data. Certain terms used here are defined in the JGR publication. As in previous versions, the retrieved data should be interpreted as values at the breakpoints of a piecewise-linear representation of the profile that best fits the radiances measured by MLS. As in Version 4, HNO3 retrievals are done in V5 which reduce (or eliminate) the HNO3-caused bias in enhanced polar lower stratospheric ClO present in Version 3 (V3) as described in the JGR ClO validation paper. Also, as in V4, a linear forward model was used in the V5 ClO retrievals; thus the ~10% scaling uncertainty due to "closure" described in the ClO validation paper is not expected to be significantly reduced in V5, although the V5 retrievals on every surface may have reduced it. Results of further investigations on this issue will be reported in the MLS V5 data validation publication. Also, as in V4, the 'scaling error' due to inadvertently using an outdated ClO line strength file in Version 3 is corrected in V5. As before, uncertainty values given in the data files are made negative if the "a priori" contributes more than 25% to the retrieved value. Also, as in previous versions, there are no constraints in the retrieval algorithms that force retrieved ClO values to be positive - as doing so would cause undesirable bias artifacts in averages of individual retrieved profiles. The individual retrieved values will often be negative due to instrument noise. CHANGES FROM VERSION 4 TO VERSION 5 AFFECTING CLO The major changes in ClO from Version 4 (V4) to Version 5 (V5) are due to: (1) retrievals on each UARS surface in V5, compared to V4 where "odd" UARS surfaces (68, 32, 15 hPa, ..) have values linearly interpolated between the retrievals done on the "even" surfaces (46, 22, 10 hPa, ...), and (2) retrievals of CH3CN in V5. Weak HNO3 lines within the ClO band are also used in the V5 MLS HNO3 retrievals, which was not done in previous versions (the Version 4 HNO3 retrieval used only the stronger HNO3 lines occurring just outside the spectral range covered by the O3_205 spectrometer). However, this is not thought to have had a significant effect on ClO. Although V5 ClO retrievals are done on each UARS surface (between 100 and 0.46 hPa), the vertical resolution of V5 is approximately the same as V4 (about 5 km, see table later in this section). The V5 ClO resolution is coarser than the separation of individual UARS surfaces, mainly because the V5 algorithms include vertical smoothing to provide a stable retrieval. Although the vertical resolution of V5 ClO is approximately the same as V4 ClO, the additional free parameters in V5 allow better definition of the vertical profile. The CH3CN retrievals in V5 allow a better fit of the measured radiances in MLS spectral bands 2 and 3, including a fit of some residual "negative curvature" in the spectra that previously led to unrealistic negative values in averaged nighttime ClO between ~22 and ~4.6 hPa. The V5 data at 100 hPa are more stable and have more realistic values than in previous versions. We believe the Version 5 ClO data at 100 hPa are acceptable for use in scientific studies but, as with all MLS data, their uncertainties must be appreciated by the user. ClO data in the files at pressures outside the 1-100 hPa range should not be used. The algorithm for setting the ClO "quality indicator", QUALITY_CLO, to value 4 (indicating good radiances and retrievals) has been changed due to changes in the V5 chi-square statistic describing the fit to the radiances. The chi-square statistic for band 2 is not as highly correlated with "spikes" in retrieved ClO as it is in V4. Furthermore, primarily because of how uncertainties in retrieved tangent pressure and other parameters are propagated in the V5 algorithms, the resulting chi-square values are much larger than unity for good fits to the radiances. The algorithm for setting QUALITY_ClO to "4" was determined empirically by choosing criteria which eliminate most (possibly all, based on examination of the first full year of data) unreasonable "spikes" in the data while not throwing out an excessive amount of good data. The resulting V5 criterion is QUALITY_CLO=4 if CHISQ_MMAF(2)<20 and CHISQ_MMAF(3)<20 and CHISQ_MMAF(4)<80, where CHISQ_MMAF(n) is the V5 MMAF (major frame) chi-square for MLS band n. (This criterion "throws out" ~1-3% of data records with ClO profiles that do not *appear* obviously bad, based on inspection of 3 months of data.) The criteria for selecting good ClO retrievals (QUALITY_CLO=4, MMAF_STAT=G/T/t, positive uncertainties in the data files) "throw out" ~3% of all V5 data for the first year of the mission (excluding UARS months 9108 and 9205, which had abnormal operations), compared to ~1% for V4. DIFFERENCES BETWEEN AVERAGE VERSION 5 AND VERSION 4 ClO Following sections compare V5 and V4 ClO for (A) low and middle latitude ClO, (B) Antarctic vortex enhanced ClO, and (C) Arctic vortex enhanced ClO. Data used in all comparisons were selected by MMAF_STAT="G", "T" or "t", QUALITY_CLO=4, and positive estimated uncertainty. All values are in parts per billion (ppbv) of ClO. A. Low and middle latitude ClO The following table compares V5 and V4 ClO for averages of all data taken from 21 September 1991 to 20 September 1992 between 45S and 45N latitudes. The major change is ~0.1-0.2 ppbv increase over V4 values between 46 and 4.6 hPa, due to retrieval of CH3CN in V5. This change removes the negative values that are present in previous versions of ClO average nighttime data at these altitudes. However, the V5 nighttime values of ~0.1 ppbv at 68 to 22 hPa are unrealistically large. The daytime values are, correspondingly, thought to be typically too large by ~0.1 ppbv at 68 to 22 hPa. Day-night differences in low and middle latitude data have changed by less than 0.03 ppbv at all altitudes. ------------------------------------------------------------------------- | (A) low-mid latitude ClO: 45S-45N, 21 Sep 91 - 20 Sep 92. | | "Day" is solar zenith angle (sza) < 90 deg, and is the average | | of 80456 individual profiles for V5 and 89188 for V4. | | "Night" is sza > 95 deg and local solar time 0-6 am, and is the | | average of 95324 individual profiles for V5 and 104949 for V4. | |------------------------------------------------------------------------| | hPa | ------ day -------- | ----- night ------- | --- day-night --- | | | V4 V5 V5-V4 | V4 V5 V5-V4 | V4 V5 V5-V4 | | 1.0 | 0.14 0.15 0.01 | 0.13 0.12 -0.01 | 0.01 0.03 0.02 | | 1.5 | 0.30 0.25 -0.05 | 0.18 0.15 -0.03 | 0.12 0.10 -0.02 | | 2.2 | 0.45 0.36 -0.09 | 0.24 0.16 -0.08 | 0.21 0.20 -0.01 | | 3.2 | 0.39 0.43 0.04 | 0.08 0.12 0.04 | 0.31 0.31 0.00 | | 4.6 | 0.33 0.40 0.07 | -0.07 0.03 0.10 | 0.40 0.37 -0.03 | | 6.8 | 0.23 0.31 0.08 | 0.09 -0.02 0.07 | 0.32 0.33 0.01 | | 10 | 0.14 0.27 0.13 | -0.11 0.02 0.13 | 0.25 0.25 0.00 | | 15 | 0.06 0.23 0.17 | -0.13 0.04 0.17 | 0.19 0.19 0.00 | | 22 | -0.02 0.21 0.23 | -0.15 0.07 0.22 | 0.13 0.14 0.01 | | 32 | -0.02 0.18 0.20 | -0.10 0.10 0.20 | 0.08 0.08 0.00 | | 46 | -0.02 0.11 0.13 | -0.06 0.09 0.15 | 0.04 0.02 -0.02 | | 68 | 0.02 0.07 0.05 | 0.01 0.07 0.06 | 0.01 0.00 -0.01 | | 100 | 0.06 -0.01 -0.07 | 0.08 0.02 -0.06 | -0.02 -0.03 -0.01 | -------------------------------------------------------------------------- B. Antarctic vortex enhanced ClO The following two tables compare V5 and V4 ClO values in the Antarctic vortex for (B1) mid-August 1992 and (B2) mid-September 1992. The mid-August 1992 V5 Antarctic daytime peak value of 2.31 ppbv at 22 hPa agrees to within 0.04 ppbv with that of V4, and V5 has 0.3 ppbv more ClO at 32 hPa and 0.4 ppbv more at 100 hPa. Other mid-August 1992 daytime changes, and the nighttime changes, are generally less than 0.2 ppbv and within the noise of the average values. The altitude of the profile minimum, separating upper and lower stratospheric ClO, is lower in V5 (at 6.8 hPa) than in V4 (at 4.6 hPa). The upper stratospheric profile, especially in daytime, is smoother than in V4 but differences are within the noise. The nighttime values in V5 are unrealistically negative by ~0.15 ppbv at 4.6 and 6.8 hPa, noticeably above the expected noise of ~0.04 ppbv in the average. V5 in this regard is, however, an improvement over V4, which is negative by 0.33 ppbv at 4.6 hPa. The major difference between V5 and V4 for mid-September 1992 Antarctic enhanced ClO is that V5 has more daytime ClO at 100 hPa (0.84 ppbv for V5, compared to -0.04 ppbv for V4). The altitude of the profile minimum, separating upper and lower stratospheric ClO, is lower in V5 (at 15 hPa) than in V4 (at 10 hPa), and this altitude has lowered since mid-August as it has in V4. The peak altitude of the enhanced lower stratospheric ClO has also lowered since mid-August in both V4 and V5. Nighttime average values are negative at 10 and 15 hPa, but by only 0.04 ppbv which is the approximate noise level for these averages. More data will be examined to see whether the above differences are present in other winters, and results will be described in the MLS V5 data validation paper to be prepared. Values in the "(~eq.)" column of the tables are nighttime ClO values that could be expected due to thermal decomposition of ClOOCl. These are rough estimates based on (1) equilibrium calculations by Dr. Ross Salawitch, (2) the NCEP temperatures for that time and region, and (3) the measured amount of daytime ClO in that region to provide an estimate of total activated chlorine. The MLS nighttime V5 ClO values agree to within 0.05 ppbv of what could be expected, except for 100 hPa in mid-August 1992 where the difference is ~0.2 ppbv. --------------------------------------------------------------------------- | (B1) Antarctic vortex ClO: 15-17 Aug 1992, 70S to 80S, -120E to 90E. | | This is where daytime ClO is enhanced at 15-100 hPa, and NCEP temps | | are generally 180-190K at 22-100 hPa, and 185-195K at 15 hPa. "Day" is | | sza <87 deg, and is average of 25 individual profiles for V5, 26 for V4.| | "Night" is 100 deg < sza < 110 deg, and is average of 95 individual | | profiles for V5 and V4. The "(~eq.)" column gives nighttime values | | that could be expected (see text preceding this table). | |-------------------------------------------------------------------------| | hPa | ------ day ------- | -------- night --------- | --- day-night --- | | | V4 V5 V5-V4 | V4 V5 (~eq.) V5-V4 | V4 V5 V5-V4 | | 1.0 | -0.10 0.15 0.25 | 0.26 0.27 0.01 | -0.36 -0.12 0.24 | | 1.5 | 0.28 0.31 0.03 | 0.21 0.29 0.08 | 0.07 0.02 -0.05 | | 2.2 | 0.65 0.42 -0.23 | 0.16 0.20 0.04 | 0.49 0.22 -0.27 | | 3.2 | 0.33 0.49 0.16 | -0.08 0.03 0.11 | 0.41 0.46 0.05 | | 4.6 | 0.01 0.29 0.28 | -0.33 -0.15 0.18 | 0.34 0.44 0.10 | | 6.8 | 0.25 0.05 -0.20 | -0.11 -0.17 -0.06 | 0.36 0.22 -0.14 | | 10 | 0.48 0.34 -0.14 | 0.10 -0.03 -0.13 | 0.38 0.37 -0.01 | | 15 | 1.41 1.32 -0.09 | 0.17 0.16 (0.15) -0.01 | 1.24 1.16 -0.08 | | 22 | 2.35 2.31 -0.04 | 0.24 0.21 (0.15) -0.03 | 2.11 2.10 -0.01 | | 32 | 1.93 2.25 0.32 | 0.10 0.15 (0.10) 0.05 | 1.83 2.09 0.26 | | 46 | 1.50 1.37 -0.13 | -0.04 0.04 (0.05) 0.08 | 1.54 1.33 -0.21 | | 68 | 0.79 0.76 -0.03 | 0.11 0.11 (0.10) 0.00 | 0.68 0.65 -0.03 | | 100 | 0.09 0.53 0.44 | 0.25 0.23 (0.05) -0.02 | -0.16 0.30 0.46 | --------------------------------------------------------------------------- --------------------------------------------------------------------------- | (B2) Antarctic vortex ClO: 17-19 Sep 1992, 75S to 80S, 0E to 360E. | | This is where ClO is enhanced at 32-100 hPa and NCEP temperatures are | | ~188-198K at 100 hPa, ~190-197K at 68 hPa, ~190-203K at 46 hPa, and | | ~190-210K at 32 hPa. "Day" is sza < 90 deg, and average of 151 indiv. | | profiles for V5 and 155 for V4. "Night" is 95 deg 105 deg, but all data at 95-105 deg sza | | are in early morning before sunrise and should well represent "night". | | The "(~eq.)" column gives nighttime values that could be expected (see | | text preceding this table). | |-------------------------------------------------------------------------| | hPa | ------ day ------- | -------- night --------- | --- day-night --- | | | V4 V5 V5-V4 | V4 V5 (~eq.) V5-V4 | V4 V5 V5-V4 | | 1.0 | 0.33 0.20 -0.13 | 0.17 0.25 0.08 | 0.16 -0.05 -0.21 | | 1.5 | 0.39 0.40 0.01 | 0.29 0.32 0.03 | 0.10 0.08 -0.02 | | 2.2 | 0.45 0.51 0.06 | 0.40 0.38 -0.02 | 0.05 0.13 0.08 | | 3.2 | 0.45 0.41 -0.04 | 0.27 0.31 0.04 | 0.18 0.10 -0.08 | | 4.6 | 0.45 0.41 -0.04 | 0.13 0.22 0.09 | 0.32 0.19 -0.13 | | 6.8 | 0.25 0.33 0.08 | 0.06 0.03 -0.03 | 0.19 0.30 0.11 | | 10 | 0.05 0.15 0.10 | -0.02 -0.04 -0.02 | 0.07 0.19 0.12 | | 15 | 0.25 0.04 -0.21 | 0.01 -0.04 -0.05 | 0.24 0.08 -0 16 | | 22 | 0.45 0.41 -0.04 | 0.04 0.00 -0.04 | 0.41 0.41 0.00 | | 32 | 1.23 1.54 0.31 | 0.12 0.21 (0.25) 0.09 | 1.11 1.33 0.22 | | 46 | 2.00 1.47 -0.53 | 0.21 0.23 (0.20) 0.02 | 1.79 1.24 -0.55 | | 68 | 0.98 0.99 0.01 | 0.24 0.14 (0.10) -0.10 | 0.74 0.85 0.11 | | 100 | -0.04 0.84 0.88 | 0.26 0.06 (0.05) -0.20 | -.30 0.78 1.08 | --------------------------------------------------------------------------- C. Arctic vortex enhanced ClO The following table compares V5 and V4 ClO values in the Arctic vortex on (C1) 8-10 January 1992 and (C2) 29-31 January 1996, when and where lower stratospheric ClO was enhanced in the Arctic (Waters et al., Nature 362, pp. 597-602; Santee et al., GRL 23, pp. 3207-3210, 1996). Daytime ClO mixing ratios at the profile peak are in very close agreement (within 0.01 ppbv) between V5 and V4, but the altitude of the profile peak is higher in V5 (at 32 hPa) than in V4 (at 46 hPa). The very large unrealistic negative value in the Jan 1996 V4 data at 100 hPa (-1.05 ppbv, which is generally representative of the individual profiles that went into the average, and not the result of a single very bad profile) is not present in V5, which has ~0.5 ppbv 100 hPa daytime ClO in both years. The daytime profile in the upper stratosphere is smoother in V5, although the differences here are within the noise on the average. V5 and V4 nighttime ClO values for both years agree to within the noise of the averages shown here except at 10 hPa in January 1996 where V4 is more unrealistically negative (-0.22 ppbv) than V5 (-0.05 ppbv). The nighttime values in V5 (and V4) agree to within ~0.1 ppbv with what could be expected due to thermal decomposition of ClOOCl. --------------------------------------------------------------------------- | (C1) Arctic vortex ClO: 8-10 Jan 1992, 60-80N, -30E to 60E. | | This is where ClO is enhanced at 22-68 hPa and NCEP temperatures are | | ~198-203K at 100 hPa, ~195-200K at 68 hPa, and generally < 185-195K at | | 46, 32 and 22 hPa. "Day" is sza <87 deg, and average of 13 individual | | profiles for V5, 14 for V4. "Night" is sza>100 deg, and average of 116 | | individual profiles for V5 and 118 for V4, with data taken at 2-9 am | | local time. The "(~eq.)" column gives nighttime values that could be | | expected; see text before Table B1 earlier in this section. | |-------------------------------------------------------------------------| | hPa | ------ day ------- | -------- night --------- | --- day-night --- | | | V4 V5 V5-V4 | V4 V5 (~eq.) V5-V4 | V4 V5 V5-V4 | | 1.0 | 0.70 0.34 -0.36 | 0.27 0.23 -0.04 | 0.43 0.11 -0.32 | | 1.5 | 0.30 0.24 -0.06 | 0.22 0.22 0.00 | 0.08 0.02 -0.06 | | 2.2 | -0.09 0.16 0.25 | 0.16 0.19 0.03 | -0.25 -0.03 0.22 | | 3.2 | 0.33 0.43 0.10 | 0.11 0.11 0.00 | 0.22 0.32 0.10 | | 4.6 | 0.75 0.45 -0.30 | 0.06 0.07 0.01 | 0.69 0.38 -0.31 | | 6.8 | 0.30 0.33 0.03 | -0.01 0.01 0.02 | 0.31 0.32 0.01 | | 10 | -0.15 0.15 0.30 | -0.07 -0.02 0.05 | -0.08 0.17 0.25 | | 15 | 0.62 0.22 -0.40 | 0.03 0.00 -0.03 | 0.59 0.22 -0 37 | | 22 | 1.39 1.22 -0.17 | 0.13 0.08 (0.10) -0.05 | 1.26 1.14 -0.12 | | 32 | 1.61 1.83 0.22 | 0.10 0.16 (0.15) 0.06 | 1.51 1.67 0.16 | | 46 | 1.84 1.42 -0.42 | 0.08 0.10 (0.10) 0.02 | 1.76 1.32 -0.44 | | 68 | 0.78 0.89 0.11 | 0.12 0.15 (0.10) 0.03 | 0.66 0.74 0.08 | | 100 | -0.27 0.44 0.71 | 0.16 0.19 (0.10) 0.03 | -0.43 0.25 0.68 | --------------------------------------------------------------------------- --------------------------------------------------------------------------- | (C2) Arctic vortex ClO: 29-31 Jan 1996, 60-90N, 45E to 105E. | | This is where ClO is enhanced at 22-100 hPa, and NCEP temperatures | | are ~200-203 at 100 hPa, ~195-205K at 68, 46 and 22hPa, and ~200-210 K | | at 22 hPa. "Day" is sza < 90 deg and average of 30 individual profiles | | for V5 and V4. "Night" is 110 deg 5 sigma) "spikes" in the retrieved HNO3 mixing ratios; occasional anomalous retrievals must be eliminated by inspection on an individual basis. PARAMETER #8: CH3CN (Methyl Cyanide, or Acetonitrile) CH3CN is a new product in Version 5, and detailed validation of this product is yet to be performed. People wishing to use these data are advised to consult the MLS science team. The data are thought not to be reliable at or below 100 hPa, or above 1 hPa. ESTIMATED PRECISION AND ACCURACY OF CH3CN -------------------------------------------------------------------------- | | | | | Ratio | | | | | Typical | Typical | of precision | | |UARS | | Vertical |Single Profile| to uncertainty| Estimated | |Index| P |Resolution (1)| Precision (2)|in L3 files (3)| Accuracy (4) | | |(hPa)| (km) | (pptv) | | | |--------------------------------------------------------------------------| | 18 | 1.0 | 8 | 90 | 0.9 |10 pptv and 20% | | 17 | 1.5 | 8 | 60 | 0.7 |10 pptv and 20% | | 16 | 2.2 | 7 | 60 | 0.8 |10 pptv and 20% | | 15 | 3.2 | 4 | 50 | 0.7 |10 pptv and 20% | | 14 | 4.6 | 6 | 50 | 0.7 |10 pptv and 20% | | 13 | 6.8 | 5 | 40 | 0.7 |10 pptv and 20% | | 12 | 10 | 4 | 40 | 0.7 |10 pptv and 20% | | 11 | 15 | 4 | 30 | 0.7 |10 pptv and 20% | | 10 | 22 | 4 | 30 | 0.7 |10 pptv and 20% | | 9 | 32 | 4 | 30 | 0.7 |10 pptv and 20% | | 8 | 46 | 4 | 30 | 0.7 |10 pptv and 20% | | 7 | 68 | 4 | 30 | 0.7 |10 pptv and 20% | | 6 | 100 | 4 | 60 | 0.9 | TBD, see (5) | -------------------------------------------------------------------------- Notes: (1) Vertical resolution given here is the full width at half-maximum of the rows of the averaging kernel matrix computed for the nominal MLS operational scan and radiometer noise. (2) Precisions are estimates obtained by computing the minimum monthly rms variability of individual measurements in the 5S to 5N latitude band (from each of the first 10 full UARS months of the MLS mission; i.e., Oct. 1991 to Sep. 1992). Uncertainties quoted in Level 3 files are typically 10 to 40 percent larger, because of the influence of the a priori data and its vertical smoothing on the retrieved products. (3) These ratios are the typical precisions divided by typical uncertainties in the Level 3 files. The values given in the Level 3 files account for variations in the uncertainty that might occur from profile to profile for various reasons (e.g., missing channels or tangent point scan positions would increase the uncertainties); however, as mentioned above, they overestimate the actual precision of the measurements. The estimated uncertainties given in the Level 3 files should be multiplied by these ratios to obtain the best estimate of the precision. (4) This column gives the "bias" (i.e., additive) uncertainty in pptv, and the "scaling" (i.e., multiplicative) uncertainty in %. Values given here are believed to represent 60-70% confidence levels (roughly "1-sigma"). These estimates were obtained by analogy with ClO, accounting for the relative line strengths of CH3CN and ClO and assuming 10% uncertainty in the CH3CN pressure-broadened linewidth parameter. The overall estimate of accuracy is the root sum square of the bias uncertainty and the scaling uncertainty (the product of the retrieved mixing ratio value with the percentage given here). The overall uncertainty for a datum is the root sum square of the accuracy and the precision. (5) The data at 100 hPa may be contaminated by contributions from H2(18)O. The magnitude of this contamination has yet to be completely determined, and CH3CN data at 100 hPa are not yet considered sufficiently understood for use in scientific studies. KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN CH3CN A detailed study of the systematic effects in this data set remains to be undertaken. CH3CN and SO2 have similar spectral signatures in the MLS 205-GHz radiometer, so it is not possible to accurately retrieve them both simultaneously operationally. As SO2 is not retrieved in Version 5, the algorithm ascribes the radiances from enhanced SO2 following the Mt. Pinatubo eruption to CH3CN, leading to unrealistically large CH3CN amounts. For this reason, only data after January 1992 should be used in scientific studies. Note: The data for Dec. 13-20, 1992 (UARS days 348-355) show a significant enhancement (~factor of 100) in CH3CN in a confined region over the south- eastern US and nearby ocean. This is believed to be a genuine enhancement and is currently being investigated. CAVEATS FOR CH3CN THE PARAMETER FILES SHOULD BE EXAMINED AND ONLY PROFILES WITH MMAF_STAT='G', 't', or 'T' AND QUALITY_ClO=4 SHOULD BE USED FOR SCIENTIFIC STUDIES. Profiles with MMAF_STAT set to 'T' (which occurs quite infrequently) may have slight biases at pressures larger than or equal to 46 hPa. Moreover, data with negative estimated uncertainties (quality field in Level 3 data files) should not be used (although such values are rare in Version 5 data, for the vertical range recommended above). PARAMETER #9: GPH (Geopotential Height) Geopotential height is a new data product from MLS. It is retrieved at 100 hPa and extrapolated to other surfaces assuming hydrostatic balance among tangent pressures, temperatures, and tangent heights. ESTIMATED PRECISION AND ACCURACY OF GPH -------------------------------------------------- | | | Typical | | | | | single profile | Estimated | | UARS | Pressure | precision (1) | accuracy (2) | | index | (hPa) | (km) | (km) | |--------------------------------------------------| | 30 | 0.010 | 0.22 | 1.6 | | 29 | 0.015 | 0.20 | 1.6 | | 28 | 0.022 | 0.18 | 1.6 | | 27 | 0.032 | 0.16 | 1.6 | | 26 | 0.046 | 0.15 | 1.6 | | 25 | 0.068 | 0.14 | 1.6 | | 24 | 0.10 | 0.13 | 1.5 | | 23 | 0.15 | 0.13 | 1.5 | | 22 | 0.22 | 0.12 | 1.5 | | 21 | 0.32 | 0.13 | 1.5 | | 20 | 0.46 | 0.13 | 1.5 | | 19 | 0.68 | 0.12 | 1.5 | | 18 | 1.0 | 0.11 | 1.5 | | 17 | 1.5 | 0.10 | 1.5 | | 16 | 2.2 | 0.10 | 1.5 | | 15 | 3.2 | 0.09 | 1.5 | | 14 | 4.6 | 0.08 | 1.5 | | 13 | 6.8 | 0.08 | 1.5 | | 12 | 10 | 0.07 | 1.5 | | 11 | 15 | 0.07 | 1.5 | | 10 | 22 | 0.07 | 1.5 | | 9 | 32 | 0.08 | 1.5 | | 8 | 46 | 0.07 | 1.5 | | 7 | 68 | 0.07 | 1.5 | | 6 | 100 | 0.07 | 1.5 | -------------------------------------------------- Notes: (1) Precisions (1-sigma) are estimates obtained by calculating the minimum monthly rms variability of individual measurements at 5S-5N latitudes during Oct. 1991 to Sep. 1992. Users should disregard the uncertainty values quoted in the Level 3 files. (2) The GPH accuracy is an estimate with 95% confidence that takes into account systematic error at 100 hPa and the contribution from integrated temperature uncertainty. The systematic error (~1.5 km) is obtained from unexpected discontinuities and trends of the GPH time series in the 5S-5N latitude band. The time series exhibits a ~1 km discontinuity around UARS yaw maneuver days and an unexpected trend of 0.5 km over six years. At the surfaces above 100 hPa, additional error is considered by integrating the effect of temperature accuracies on the GPH measurements. KNOWN ARTIFACTS AND SYSTEMATIC EFFECTS IN GPH The accuracy in a single GPH profile is dominated by a pointing bias that is slowly varying with orbital angle and can be assumed constant during the time when the profile is measured. Users can significantly reduce the effects of the bias by only studying the relative changes of GPH (e.g., perturbations about the zonal mean). These perturbations may yield scientifically useful results within the precision specified. Users should consult the MLS team for using such approaches on the GPH measurement. Some unexpected discontinuities and trends in the GPH product are mentioned in note (2) above. CAVEATS FOR GPH THE PARAMETER FILES SHOULD BE EXAMINED AND ONLY PROFILES WITH MMAF_STAT='G', 't', or 'T' AND QUALITY_TEMP = 4 SHOULD BE USED FOR SCIENTIFIC STUDIES. The MLS 63-GHz radiometer was turned off on June 15, 1997 (UARS Day 2104) to conserve spacecraft power. The GPH data after that day should not be used.