 |
Stratospheric Processes And their Role in Climate
|
||||||||
| Home | Initiatives | Organisation | Publications | Meetings | Acronyms and Abbreviations | Useful Links |
|
Stratospheric Processes And their Role in Climate
|
||||||||
| Home | Initiatives | Organisation | Publications | Meetings | Acronyms and Abbreviations | Useful Links |
Proper understanding of the observed long term decrease in ozone is needed in order to be able to predict with confidence future changes in ozone and the implications of these changes on UV-B and climate. Accurate quantification of the observed changes is a pre-requisite for this understanding. The trends in ozone were last fully assessed in the WMO-UNEP Scientific Assessment of Ozone Depletion in 1994.
Changes in total ozone are well characterised. At mid-latitudes in the northern hemisphere the trends from 1979 to 1994 found in the combined SBUV/SBUV-2 record (Hollandsworth et al., 1995) are significantly negative in all seasons and are larger in winter/spring (up to 7%/decade) than in summer/autumn (about 3%/decade). In the tropics there seems to have been no statistically significant change in total ozone. Trends in the southern mid-latitudes are significantly negative in all seasons (3-6%/decade) and there is a smaller seasonal variation than in the north. From 1979 to 1991, the agreement between the trends derived from the SBUV, TOMS and the ground-based Dobson network records is better than 2%/decade.
The bulk of the loss in total ozone at mid-latitudes has taken place at altitudes between 15 and 25km. The available long term records for this altitude region are sparse with nearly all of those longer than 10 years in the northern mid-latitudes. In this region there is good agreement between the ozonesonde and SAGE trends above 20km. However there has been considerable debate about the magnitude of the trends below 20km with SAGE giving considerably larger trends than the ozonesondes (-20+/-8 vs 7+/-3 %/decade). SAGE also reports large trends (-25 to -30 %/decade) in the very low stratosphere in the tropics, although it should be noted that there is very little ozone at these altitudes so that the effect on the total column is small.
A joint initiative to resolve these issues has been instigated by SPARC (Stratospheric Processes And their Role in Climate change) and IOC (International Ozone Commission). A workshop was held at the Observatoire de Haute Provence from July 8-11 1996. 28 scientists from a number of countries attended to review the current state of knowledge about trends in the vertical distribution of ozone (see section 2). It was agreed that a report should be produced by the end of 1997, in time for use in the preparation of the next WMO-UNEP assessment planned for 1998 (see section 3). Successful completion of this report will involve a certain amount of work to be carried out in the next year.
A large part of the meeting was dedicated to discussing the current status of the data quality of the different measurement systems. Particular attention was paid to any changes in instrument performance over long time periods which might affect the ability to determine the true trends. In addition the participants discussed the ability of each of the systems to measure ozone correctly in the presence of volcanic aerosol.
After an overview of the current status of the different ozone profile measurement techniques, presentations of the results from individual research studies were presented. The attendees then split into three working groups to discuss the outline of the chapters to appear in the SPARC/IOC report.
Four measurement techniques have records long enough for long term trend studies: SAGE, ozonesondes, SBUV and Umkehr. Other techniques have been used for shorter lengths of time and are planned for use in the future, e.g. in the Network for the Detection of Stratospheric Change (NDSC). Lidar, HALOE, MLS and ground-based microwave instruments were specifically discussed at this meeting.
SAGE
The overview of the current status of the ozone measurements made by the Stratospheric Aerosol and Gas Experiment (SAGE) satellite instrument was presented by J. Zawodny (NASA/LRC, USA). The SAGE satellite data set comprises the longest term satellite measurements of the ozone profile including the lower stratosphere. The SAGE I instrument was operational from 1978 and 1981. SAGE II was launched in 1984 and is still working. SAGE is a solar occultation instrument making measurements at six wavelengths at sunrise and sunset.
There has been a long-recognised problem with the altitude determination of the SAGE I ozone measurements which has been recently examined by Wang and Cunnold (1996). They found empirical corrections through comparisons with SBUV and SAGE II which have a latitude dependence unlike previous adjustments. With these corrections the agreement in the upper stratosphere with trends derived from SBUV for the same time period is now good at all latitudes. The biggest improvement in the agreement is in the tropics where both records now show no significant trend at about 40km. There are currently no plans to revise the SAGE I measurements
The SAGE II data is currently being revised. Comparisons with other measurement systems have allowed the identification of possible improvements of the SAGE II ozone profiles. The most important of these, with regards to ozone trend determination, is the correction of an aerosol artefact in the ozone measurement (to appear in version 5.95). Other improvements include a correction for an electronic transient (version 5.94) and a drift in the 448nm channel (version 5.96), both of which will improve the NO2 measurements. In order to measure ozone in the upper troposphere, a correction for the water vapour interference (not needed in the much drier stratosphere) will be introduced to reduce the spectral interference on the ozone measurement in the 600nm channel. The vertical resolution of the transmission measurements (currently 1km) and the derived constituent mixing ratios will be increased to 500m. All these improvements will be implemented in version 6, due to appear late in 1996. A major reformatting of the data will occur, including a more elaborate error calculation and better indicators of quality assessment.
Ozonesondes
The ozonesonde data set comprises the most extensive and longest ground-based measurements of the ozone profile. Numerous comparisons between different types of sondes and other measurement systems have been carried out in the past few years. As well as the overview by H. Smit (KFA-Jülich, Germany) of the current status of the ozonesonde measurements, other presentations on the quality of the data and derived trends from individual stations were presented. The discussion focused on error sources (instrumental and procedural) which could induce apparent changes over time in the ozone measurements. Other issues raised were the ability of the sondes to measure ozone correctly above 26km (the pump correction factor) and the possible change in the tropospheric response of the Brewer-Mast sonde.
Recent chamber simulations of the performance of different instrument types, carried out as part of the GAW/GLONET programme, have helped quantify the characteristics of the different types of sondes and the differences between them.
It was generally agreed that the pump corrections currently recommended by the WMO (to allow for the changing efficiency with ambient pressure) may need to be revised. The correction varies for different types of sondes and also over time as the sonde technology evolves. The pump correction used could thus influence the trends calculated using the sonde data sets, and so is to be well understood and quantified.
The procedure of normalising the profiles with a total ozone column correction was also discussed. If the pump efficiency is not well corrected, this procedure could introduce altitude dependent trends. These issues are currently being examined and it was felt that within the next year, the quality of the trends measured by sondes will have improved.
SBUV
R. McPeters (NASA/GFDC, USA) presented an overview of the current status of the data sets of the two Solar Backscatter UltraViolet (SBUV) satellite nadir viewing instruments between 252 and 306nm. The SBUV instrument was operational from 1978 to 1990, and the first SBUV-2 instrument from 1989 to 1994. Two further SBUV-2 instruments are currently operating. The measurements from the original SBUV instrument will be revised into a version 7 using the experience gained in the revision of the TOMS v7 data.
The discussion centred on the accuracy of the profile retrieved at low altitudes. The vertical distribution of ozone is retrieved with a vertical resolution of 6-10km, and is constrained by the SBUV total ozone measurement. Since the vertical resolution is poor, the retrieval below the ozone maximum is sensitive to the error at higher altitudes and to tropospheric ozone. The 1994 WMO-UNEP assessment did not present SBUV trends at altitudes below about 25km (30mbar), but did give the trend in the column amount between the ground and 30mbar. A recent comparison of SBUV with CLAES and MLS has shown a longitudinal discrepancy which is thought to be due to differences in tropospheric ozone (Ziemke et al, 1996).
The combined SBUV/SBUV2 data set has been used to derive trends (Hollandsworth et al, 1995). The SBUV trends in total ozone are more negative than those calculated using the TOMS data by about 1%/decade. Comparisons with the trends derived from the SAGE stratospheric ozone column (above 16km) have shown a very good agreement (within 2%/decade) though the SBUV trend tends to be slightly smaller, especially in the tropics.
Umkehr
M. Newchurch (Univ. of Alabama, USA) presented an overview of the Umkehr technique which was first described in the 1930s. Measurements have been made routinely at some stations since the 1960s. Zenith sky observations are made at high solar zenith angles by Dobson spectrophotometers. The vertical distribution of ozone is inferred from the effect of the changing path of the sunlight through the atmosphere as the mean height of the Rayleigh scattering changes. The inversion procedure (similar to that used in the SBUV measurement) currently used (Mateer and DeLuisi, 1992) was discussed in great detail.
The dependence of the different Umkehr layers on the a priori profile and the total ozone amount used was reviewed since this subsequently places a limitation on which of the Umkehr layers should be used for trend calculations. The effect of aerosol on the retrieval errors was also discussed in detail. It has been demonstrated that for trend calculations, the errors induced are important mainly in the lowest and highest Umkehr levels (e.g., Newchurch and Cunnold, 1994).
Lidar
An overview of ozone measurements by DIAL Lidar was presented by S. Godin (SA/CNRS, France). This self calibrating technique is used to measure the stratospheric ozone profile under clear sky conditions with a resolution of 0.5km at 20km and 5km at 50km at a small number of stations. The ability of the lidar to measure ozone in the presence of aerosol was discussed. At the expense of vertical resolution and precision, the perturbation of the ozone profile due to aerosol can be reduced by measuring the Raman scattering of molecular nitrogen. Comparisons with satellite measurements and ozonesondes show good agreement. An intercomparison of retrieval algorithms is currently being carried out for the NDSC along with other quality control investigations as ozone lidar measurements are one of the core measurements.
The data record is still too short for trend calculations, but the lidar measurements are currently valuable for comparison with other measurement techniques.
HALOE
The current status of the HALOE ozone measurements was presented by Jim Russell. HALOE is an IR solar occultation instrument which has been in operation on UARS since 1991. Comparisons with other measurements (satellite, sondes and lidar) show good agreement in the upper stratosphere. As with SAGE, the ozone measurement is sensitive to the accuracy of the altitude determination and so is especially sensitive in the lower stratosphere. An improved altitude calculation is being introduced in the next version of the data.
MLS
L. Froidevaux (JPL, USA) presented the current status of the MLS ozone measurements. MLS is also on board UARS and measures the atmospheric emission of millimetre radiation. This technique has the advantage of high sampling frequency and coverage and is not affected by the presence of aerosols, ice clouds or temperature. Intercomparisons have shown the MLS ozone measurements at pressures of 46hPa and below to be systematically higher than other measurements, although no changes over time in the data quality have been identified. Ozone measurements at 100hPa are currently considered a research product but should be included in any intercomparisons performed for the SPARC-IOC report. A new version of the MLS ozone measurements will be available for assessment within a year.
The available ground-based microwave measurements and their suitability for intercomparisons was discussed, the main limitation being the sparse data record.
Introduction
At the outset it was recognised that the principle focus of the report would be the derivation of global ozone trends between the tropopause and 25km. One looks at the data sources that are currently available, only the SAGE data set could provide the temporal and spatial coverage to meet this objective. However, in order to be confident that the trends derived are correct, two questions need to be answered. The first is 'does SAGE measure ozone?'. The answer would seem to be obvious, however if we rephrase the question as 'are there systematic errors in the SAGE instrument or retrieval algorithm which could bias the derived ozone product?' then the answer is not so obvious. The second question is 'can one derive true global trends given the SAGE data coverage?'. The group considered both of these questions, and came up with the following broad strategy. This strategy can be broken up into two, one for the altitude region above 25km, and the other for below 25km. For the region above 25km we have other global long-term data sets which can also be used for trend analysis, these are from the SBUV/SBUV2 instruments, and to a lesser extent the Umkehr/Dobson network. There is also data from the LIDAR network, although this data set has sparse global coverage. It will, however, be used to verify the accuracy of individual profiles determined by the satellite instruments. In addition, data from the MLS and HALOE instruments on the UARS satellite, and from the LIMS instrument on Nimbus-7 will be compared with the SBUV, SAGE and Umkehr data over shorter intervals to examine any possible seasonal or latitude biases, which would indicate systematic errors in the instruments. The SBUV/SBUV2 data has good spatial coverage and will be used to test the effect of the poorer spatial coverage of SAGE on the derived trends. The overall strategy is to provide confidence that the derived trends above 25km are correct.
For the region below 25km a different strategy has to be used, as the SBUV/SBUV2 and Umkehr instruments are not sensitive in this region. For this altitude interval we must now rely on the ozone sondes and LIDAR measurements for the long-term comparisons. We will again compare the ozone profiles derived from SAGE II with those from HALOE and MLS, as a function of season to ascertain any seasonal differences in the retrieved ozone data. Individual profiles from the sondes, LIDAR, etc., will provide further validity of the derived profiles.
Two additional results could come from the analysis discussed above. If the comparisons above 25km are positive, then one could consider extending the Umkehr data to earlier years, which would increase the time period over which trends could be obtained. If the comparison below 25 km is successful, then one could consider extending the sonde data to earlier years, and to the derivation of tropospheric trends.
The SAGE I ozone measurements should play an important part in this assessment as their use potentially extends the SAGE measurement back an extra 5 years, even though the SAGE I instrument only operated for 2-3 years. The recently revised LIMS ozone measurements should used to help characterise the SAGE I measurements.
The quality of the SAGE data (and of the other satellite instruments) depends critically on the quality of the NMC temperature analyses. The NMC analyses are being revised which will improve their quality and consistency over time. It will be important to have access to these revised data sets to be available for use in this assessment.
Contents
The report will be divided into four chapters. Chapter 1 will discuss the instruments to be used in the analysis, and their algorithms' characteristics with the aim of developing error estimates for the ozone measurements. Chapter 2 will compare the different data sets, and provide estimates of the quality of the data products. Chapter 3 will discuss the trends in ozone that can be derived from the data. Chapter 4 will provide a synopsis of the lessons learned from the overall study.
A team will be organised to write each chapter, headed up by two or more co-chairman. The co-chairmen for Chapter 1 are C. Rodgers (Oxford University, UK) and D. Hoffman (NOAA/CMDL, USA); for Chapter 2, J. Russell (Hampton University, USA) and H. Smit (KFA-Jülich, Germany); and for Chapter 3, W. Randel (NCAR, USA), R Stolarski (NASA/GSFC, USA), and D. DeMuer (BMI, Belgium). The chairmen for Chapter 4 have not been designated, the responsibilities have currently been assumed by N. Harris (EORCU, UK) and R. Hudson (Univ. of Maryland, USA).
Timetable/assignments
At the end of the session, a tentative timetable was agreed to, and writing assignments were made within each chapter working group. The timetable calls for a draft of each chapter to be written prior to the full meeting next July, which might be held in Cambridge, UK. It was the opinion of the participants, that whereas some of the communication between the participants can be done by e-mail etc., a face-to-face meeting of each group will be necessary sometime in early 1997. Financial support will be required by some group members to attend these meetings.
There are two large scientific meetings before next July, the Quadrennial Ozone Symposium in Italy this September, and the First SPARC General Assembly in Melbourne, Australia this December. Only a small number of the group participants are attending these meetings, nevertheless subgroup meetings are being arranged.
It was decided to set up a password restricted home page for the study, which would not only contain the most up-to-date copy of the draft report, but which would also contain data sets which were peculiar to the study. It was also agreed that these data sets would be considered privileged and only for the use of the study, however, any data used in the final report would be made freely available on the release of the report. Access to the satellite data sets would be made through this home page. The draft report will be considered a living document. Representatives from NASA/GSFC and NASA/Langley, kindly agreed to provide the home page, and to organise the data base.
In the presentations given at the meeting, a number of improvements to the SAGE I and SAGE II algorithms were identified, which should lead to a significant change in the quality of the retrieved ozone densities. Given the central nature of the SAGE data in the report, it is considered vital that these data be made available to the study in the near term in order that these improvements can be validated and documented prior to the final version of the report.
N. R. P. Harris, R. D. Hudson and C. Phillips