--Jack A. Kaye (jkaye@mtpe.hq.nasa.gov), Office of Mission
to Planet Earth,
NASA Headquarter, Washington, DC.
--A. James Miler, National Centers for Environmental
Predication, National oceanic and
Atmospheric Administration, Camp Springs, MD.
The availability of ground- and balloon-based measurements of ozone distributions in the lower stratosphere (LS) and upper troposphere (UT), and their use in the validation of space-based measurements of ozone profiles in these altitude regions, were explored in a meeting held in Silver Spring, MD, in April, 1997. Some 30 scientists and managers from several U.S. government agencies, universities, and the private sector participated in the meeting, which was organized under the auspices of the Office of the Federal Coordinator for Meteorological Services and Supporting Research. The meeting was posed to help explore several issues, including:
Several scientific issues were emphasized at the meeting. In particular, the distribution of tropospheric ozone in the tropics was discussed at some length, spurred on by the satellite-based observations showing significant zonal asymmetries in tropical distributions of tropospheric ozone. The satellite determinations typically use data from NASA's Total Ozone Mapping Spectrometer (TOMS) instruments in one of two ways. In one, the TOMS data are used together with measurements of the vertical profile of stratospheric ozone so that tropospheric ozone is calculated as the residual difference between total ozone and stratospheric column ozone. In the other, some data-based assumption is made about the nature of stratospheric column ozone variability so that tropospheric ozone could be obtained by a similar subtraction method. There remain some important scientific questions on the exact characterization of the altitude and time dependence of these zonal asymmetries, as well as the relative contribution of dynamical and in situ photochemical sources of enhanced ozone. Direct comparison of the satellite-derived product with in situ data is complicated by the limited in situ measurements available from ozonesondes in the tropics, especially the lack of near-simultaneous observations of ozone profiles in the high-ozone (Atlantic) and low-ozone (Western Pacific) regions.
Considerable discussion also took place on the status of current ozonesonde networks in the U.S. and around the world. The main subjects of interest included data availability, which is seen to be a problem for certain stations, ozonesonde inherent accuracy, intercomparisons, and "infrastructure support" for continuing calibration and intercomparison activities. Particular attention was given to the Julich Ozonesonde Intercomparison Experiment (JOSIE), carried out in an environmental simulation chamber in Julich, Germany, so that performance of different in situ ozone measurement devices (notably balloon payloads) could be compared under realistic conditions. The results of this experiment, which was supported by the World Meteorological Organization and hosted by Forschungszentrum Julich, showed that there were appreciable differences between the measurements made by different instruments, and that it is possible that some instruments can get the "right" answer for the wrong reasons, especially where corrections for pump efficiency and background current are used, but may not be completely rigorous. The importance of continuing such realistic intercomparisons was stressed, and the need for reliable and long-term international funding to support these activities was made clear to the participants.
The increasing role which ground-based ozone lidars may play in the study of UT/LS ozone was also addressed. The need for measurements in both regions has led to development of different lidar systems for the troposphere and stratosphere, driven by the need to use different wavelengths for optimal measurement in each altitude region, given the different abundances of ozone. The stratospheric ozone lidars typically work from approximately 10-15 km up to 45 km, while the tropospheric ozone lidars work from approximately several kilometers above the surface up to 12-15 km. The lidar systems are particularly useful for the study of short-term meteorological variability of ozone distributions (such as tropopause folds) as they can operate routinely, in both day and night. They may provide a biased climatology, however, because of their inability to penetrate thick clouds. Airborne lidars were briefly discussed at the meeting; they have made significant contributions in both upward- and downward-looking modes.
The status of satellite measurements of ozone distributions in the UT/LS was also reviewed. The most direct measurement of the ozone profile in this region is that of the Stratospheric Aerosol and Gas Experiment (SAGE) instruments. It has been shown that the SAGE II instrument can see through the stratosphere and down into the middle troposphere as much as 50% of the time at mid-latitudes. The quantitative validation of the SAGE data does not extend below the stratosphere, however. The version of SAGE II data most recently developed (version 5.96) has improved correction for aerosol interference over the previous version, which has led to improved agreement with ozonesondes (notably those flown from Hohenpeissenberg) in the 20-25-km region (mostly 5% differences); in the 15-20-km region there are still problems (~10% differences), and below 15 km there are greater differences which remain to be evaluated. These are primarily associated with the lack of detailed knowledge of altitude registration of the satellite instruments at these levels.
The TOMS data that form the basis of the tropospheric ozone residual determination of column tropospheric ozone have become much better with the reprocessing of the data to version 7, which includes a calibration correction and improved treatment of cloud heights. The correction of the TOMS calibration (leading to lower TOMS ozone amounts by approximately 3%) has meant that the residual ozone has been reduced by a few Dobson Units everywhere, with larger changes in north sub-tropical Africa. The TOMS data still appear to have some inaccuracies for solar zenith angles greater than 75 degrees, however. At this point, there have been few attempts to derive tropospheric ozone column amounts from high-latitude TOMS data, so these errors are unlikely to have an impact for this particular issue.
Tropospheric ozone residuals using TOMS and data from the Solar Backscatter Ultraviolet (SBUV, SBUV/2) series of instruments have also been carried out. The SBUV data do not suffer from the spatial sampling limitations of the SAGE data as a source of stratospheric ozone information for use in residual calculations, but there are significant questions about the information content in the SBUV profiles, especially in the tropopause regions. The TOMS-SBUV residuals are of particular interest in attempts to compare space-based determinations of tropospheric ozone to surface-level ozone measurements made as part of air pollution networks, given the presence of daily measurements at all latitude regions in SBUV. Residuals have also been carried out using data from TOMS and the Microwave Limb Sounder aboard the Upper Atmosphere Research Satellite. These new methods, as well as other methods being developed for determining tropospheric ozone from TOMS data (e.g., those relying on the difference between ozone columns in clear and cloudy scenes) all require careful validation and intercomparison.
Some information on tropospheric ozone distributions is starting to become available from the Global Ozone Mapping Experiment (GOME) instrument flying aboard the European Space Agency's ERS-2 satellite. The publicly available GOME products include total column ozone and also column nitrogen dioxide. The GOME total ozone measurements appear to be good to 3% for solar zenith angles below 70 degrees, and to approximately 8% for higher solar zenith angles. Comparisons suggest the problems are associated with the GOME algorithm, as using GOME radiances at the TOMS wavelengths in the TOMS algorithm gave ozone values which agreed well with TOMS. Initial studies on GOME ozone profiles have been carried out, and there is some evidence that variability in tropospheric ozone can be seen in the lowest level of the profile, including, in one case, a region of enhanced ozone near the California-Nevada border that is thought to be a high ozone plume formed downwind of Los Angeles.
Some general issues were discussed at the meeting; in many cases definitive answers could not be reached, but a clearer sense of the questions was arrived at. These issues include the following:
There appears to be a problem with rapid and public availability of sonde data from some stations. The workshop participants called for all groups making regular sonde measurements to be sure that their data are made available quickly through the standard channels (for ozone profiles, the World Ozone Data Center in Toronto, Canada).
There was a strong sense that a minimum frequency of one sonde flight per week was required for a given sonde station to provide useful data for trend determination.
There was concern as to whether or not the available sonde data records are being examined sufficiently closely that previous errors, inconsistencies, etc., which may have been incorporated, would be located and corrected. It is hoped that the current ozone profile activity going on under the auspices of the Stratospheric Processes and their Role in Climate (SPARC) group of the World Climate Research Programme will encourage sonde stations and measurement agencies to do this work.
There was a clear sense that continuing calibration and intercomparison efforts for ozonesondes are important and should be supported. Given the role which sondes play in supporting the validation of space-based data, the suggestion that space agencies should provide some financial support for this effort was discussed. International efforts being organized to look at global measurement activities, such as those associated with the International Global Observing Strategy, should consider the status of such calibration and intercomparison efforts.
There was a clear recognition that our ozone measurements are lacking in certain well-defined geographical regions. In particular, there was a strong desire to have concurrent operations of sonde stations in both the Atlantic (Ascension Island) and Pacific (Samoa) oceans, which has not typically been the case. Such data could be very useful in understanding the zonal asymmetries seen in the satellite data. Additional data from Latin America could be very useful, especially in interpreting the TOMS-only tropospheric residual product derived by separating out TOMS column observations over the Eastern Pacific from those over the nearby Andes mountains.
Continuing efforts need to be made to ensure that the ozone profile needs of air quality agencies, like the Environmental Protection Agency, and those of global-change-oriented agencies, are coordinated, and that data obtained in support of one set of objectives will be routinely available to scientists involved in both.
The community needs to better understand the tradeoffs between ozonesondes and lidars, especially for tropospheric ozone.
A new philosophy of satellite validation for ozone measurement appears to be in order; this philosophy recognizes that validation is not a one-time effort done at the beginning of a flight period for a given satellite instrument, but one which continues over the course of a lifetime of a satellite instrument. Evidence from past satellite instruments suggests that this "lifetime" commitment to validation is very important (e.g., consider the increased rate of degradation of the Nimbus 7 TOMS instrument near the end of its lifetime, and also the increased inaccuracies in SAGE II ozone measurements which resulted following the Mt. Pinatubo eruption in 1991, when stratospheric aerosol layers became high for several years). Also, given the multiplicity of launches of space-based ozone measurements involving several countries and agencies, it is important the validation not be considered on a "satellite-by-satellite" basis, but as part of a total and international plan so that the full validation capability of all nations' ground- and balloon-based research communities can be applied to the total suite of space-based ozone measurements.
Further information about this meeting can be obtained from the authors of this article, who served as convenors of the meeting. The authors thank the participants of the meeting and, in particular, those scientists who helped develop the agenda (Jack Fishman, Jennifer Logan, Anne Thompson, Volker Mohnen, Guy Brasseur, and P. K. Bhartia).