CLAES V9, V8, V7 Data Attributes Summary

1. CLAES Data Product Versions

A summary of the data products, periods covered, number of day files, and availability of the three most recent and most used data versions V7, V8, and V9, is given in Figure 1 (below). Zonal mean maps for versions V7 and V9 can be browsed or downloaded from the CLAES web site.

Fig. 1 CLAES Data product versions

Data Version V7

Period: January 9, 1992 to May 5, 1993 [388 data days]

Location: CDHF, GSFC DAAC, CD-ROM
Zonal mean cross sections on CLAES web site

Products: Temp, O3 B8, O3 B9, ClONO2, CFC12, HNO3, N2O, CH4, N2O5, NO2, H2O, NO, CFC11, Aerosol (7: 780-1897 cm-1)

Validation: Validation papers for Temp, O3, ClONO2, HNO3, N2O, CH4, N2O5, CFC12, and Aerosol (JGR-d 1996 special issue)

Data Version V8

Period: October 25, 1991 to May 5, 1993 [459 data days].
[Added 58 data days between October 25, 1991, and January 8, 1992]

Location: CDHF, DAAC (EXCLUDING CH4, CF2CL2, H2O, NO)

Products: Same as V7, plus some experimental products

Validation: Comparison with V7, models, climatology

DATA Version V9 for Complete Data set

Period: October 25, 1991 to May 5, 1993 [461 data days]

Location: CDHF (as of December 21, 2000); DAAC (by May, 2001)
Zonal mean cross sections on CLAES web site

Products: Same as V7/V8: H2O is V7 and is available from January 9, 1992 to May 5, 1993 only

Validation: Comparison with V7, V8, models, climatology

Data version 7 (V7)

V7 represents a major upgrade to the first publicly released CLAES data set (V6). V7 was extensively validated and details of the validation results for a variety of V7 products were published in a special issue of JGR (D6, 1996). The V7 products were the first extensively used data from the CLAES experiment, and have since contributed to a wide array of investigations of stratospheric chemistry and dynamics.

V7 data cover the period January 9, 1992 to May 5, 1993. They are resident on the NASA Goddard Space Flight Center Distributed Active Archive Center (NASA-GDAAC), and are also available on CD-ROM. A zonal mean browser for V7 products is accessible on the CLAES web site.

Data version 8 (V8)

V8 added newly processed data for the period October 25, 1991 through January 5, 1992, and is resident on the GDAAC. V8 incorporated a number of improvements and refinements. These included:

As a result of these upgrades, artifacts in V7 data products CFCl3, N2O5, NO2, and aerosols which caused anomalous behaviour under specific atmospheric conditions were eliminated or significantly reduced.

Details of the validation results for V7 and V8 taken from the GDAAC documentation are attached as Appendix A to this discussion.

Data version V9

V9 is the most current data set and is the one recommended for scientific data useage, with the exception of H2O for which V7 is still recommended (see below). CLAES data were production processed by this latest algorithm for the complete operational period October 25, 1991, through May, 1993. The V9 data set is expected to be resident on the NASA Goddard Space Flight Center Distributed Active Archive Center (NASA-GDAAC) by May, 2001.

V9 processed data contain all of the generic improvements listed above for the V8 vs V7 processing. In addition, several V8 species which had continued to warrant upgrades, including NO, and NO2, have been improved in V9, and systematic biases in CFC12, and CH4 ,as well as a small global temperature bias, inadvertently introduced into the V8 data, have been mitigated in V9. The following section summarizes the results of V9 validation
 
 

2. V9 Validation Summary

The approach to validating the V9 data was mainly to compare with V7 which as noted above had itself had been extensively compared with correlative data, models, and climatology. By comparing selected V9 24-hour zonal mean altitude profiles with the V7 profiles used in previous correlative comparisons, allowed for V9 correlative data comparisons "by proxy". It is also instructive to compare V9 with V8 in some cases where the V8 data had regressed. The following is a summary of the results of selected comparisons between V9, V8, and V7 data, and correlative data: We note that of necessity, much of the correlative data in question is for northern mid latitudes (Appendix A).

2.1. Temperature: Reduced global cold temperature bias of ~2K in V8

Comparisons are shown between V8 and UKMO (Figure a), V9 and UKMO (Figure b), and V9 and V7 (Figure c). V9 warms by about 2K. The bias arose from a combination of effects associated with the approach to radiance fitting, pressure registration, and strong temperature gradients. Reducing the temperature bias also acts to decrease the vmrs of CH4, N2O, and to a lesser extent CFC12, bringing all of these gases in better agreement with correlative data. 2.2 NO2: Mitigated 30-40% low bias Comparisons between V7, V9, and ISAMS are given for Daytime NO2 (Figure d) and Night time NO2 (Figure e). V9 has increased substantially and is now within about 15% of LIMS/ISAMS at peak and exhibits better agreement with other data including HALOE. This was accomplished largely by the use of refined instrument spectral functions constrained by the reduction of apparent diurnal dependence in aerosol in this spectral channel. No obvious discontinuity is apparent between 1991 data and the rest of the data 2.3 CFC12: Corrected for 20-40% high bias introduced into V8 Comparisons are shown between V8 and V7 (Figure f), V9 and V7 (Figure g), and between V9, V7 and midlatitude correlative data (Figure h). V9 is now within about 10% of correlative. This was mitigated largely by the use of new temperature dependent spectral absorption parameters, and benefits from the increased temperature. 2.4 CH4: Corrected for 30-35% high bias introduced into V8 Comparisons are shown between V8 and V7 (Figure i), V9 and V7 (Figure j), and between V9, V7, and correlative data (Figure k). V9 CH4 is now within about 10% of correlative, although still 15-20% high near 35 km. Bias arose largely from incorrect handling of the N2O5/Aerosol boundary. 2.5 N2O: Relatively minor changes in V9 V9 reduces wrt V7 below 30 Km (Figure l). This brings V9 within ~ ± 5% of correlative data below 30 km vs ~ +10% for V7 (Figure m). 2.6 HNO3: Relatively minor changes in V9 V9 reduces near peak vs V7 for vmr ³ 10 ppbv, (Figure o). This brings V9 within ~ ± 5% of correlative data near peak vs ~ +10% for V7 (Figure p). 2.7 ClONO2: Relatively minor changes in V9 V9 increases wrt V7 above 27 km bringing it into somewhat better agreement with correlative data ( + 5-10%) vs ~ + 10-20% for V7 (Figure s). V9 also increases by ~ 10% wrt V7 below 21 km, but still within ~ 5% of correlative. There are fewer anomalous spikes apparent in the global field data.

1991 data appear continuous across the 01-04-92 boundary.

 2.8 N2O5: Offset subtraction software included N2O5 is available on 6 UARS standard pressure levels between 10 mb and 1.47 mb. The V9 data retain the major improvements seen in V8, i.e. extended altitude range, removal of artifacts in the 3.16 mb level data, and supplied software routines to subtract latitude dependent offsets. 2.9 O3B9: Relatively minor changes in V9 vs V8 Retains improvements seen in V8, mainly decreased tropical values below 20 mb (Figure t) bringing better agreement with correlative data, and elimination of the factor of 2.4 numerical error in the algorithm-generated profile errors. V9 now ~10% near peak whereas V7 ~15%;V9~ 10% below 25 km whereas V7~ 18%. (Figure u).

An apparent anomaly in V9 global data is that on the first day of the YAW cycle the peak O3 mixing ratio is about 10% higher than adjacent days in the YAW period. This feature is present but not so apparent in V7 because some 1st days in the yaw cycle were not successfully processed. In the newly processed 1991 data, days before 10-31-91 show anomalies near the equator possibly due to the very high Pinatubo tropical aerosol loading during this period.

 2.10 NO: Improved calibration Note that only daytime retrievals are available. V9 NO is reliable only from ~ 5 to 0.5 mb (See Appendix A). Below 5 mb the cataloged values are scaled climatology. An improved approach to calibration of the typically very low radiances associated with NO has reduced systematic artifacts in retrieved profiles. V9 increases vs V7 throughout the useable altitude region of 5 to 0.5 mb, and especially near the peak at ~2 mb (Figure v). This brings the V9 data into better agreement with ISAMS data for example (Figure w). 2.11 CFC11: Relatively minor changes in V9 Retains improvements seen in V8 but reduces slightly vs V8 below ~25 km (Figure x). 2.12 AER: 790 and 780 cm-1 extinction coefficients now in good agreement Extinction values for AER790 have decreased bringing them into much better agreement with AER780 2.13 H2O: Use V7 only; data begins Jan 9, 1992 Figure y is a qualitative pictorial guide to the utility of V7 H2O. For a more detailed discussion of H2O data quality see Appendix A. Overall, use of night time quantities is recommended. Best confidence is in the summer-fall hemisphere for latitudes above 30 deg and altitudes lower than 10 mb, and in the winter hemisphere for mid latitudes (3 to 55 deg) below 10 mb. The data are also reasonably good for all latitudes and seasons at altitudes above 1 mb.

The CLAES H2O data are not recommended for scientific use anywhere between 10 and 1 mb, at tropical latitudes (30N-30S) below 1 mb, and at polar winter latitudes below 1 mb, especially in the southern hemisphere.

Appendix A

(Summary of V7 Data Quality and Summary Comparisons Between V8 and V7 for Several Products)