S P A R C Stratospheric Processes And their Role in Climate A project of the World Climate Research Programme Home Initiatives Organisation Publications Meetings Acronyms and Abbreviations Useful Links

2.1. Introduction

No uniform data source exists to allow careful review of the current state of knowledge of long-term changes in the vertical ozone distribution. The data available for trend studies comes from different ozone measuring platforms using different sensing techniques with different spatial and temporal coverage. Therefore, in preparation for this assessment the current data quality status for the different measuring systems was evaluated by a group of experts at a workshop held at the Observatoire de Haute Provence in July 1996 (Harris, 1996). This study resulted in the definition of an inventory of the most reliable data sets with time records longer than 5-10 years (see Table 2-1).

It is obvious that only four measurement systems have records long enough to assess long term ozone trends: SAGE (I and II), SBUV and SBUV2, Umkehr/Dobson and ozone sondes. These measurements cover the lower stratosphere to the stratopause. The SAGE I and II satellite series extends from February 1979 to the present with a 3 year interruption from November 1981 to November 1984. The SBUV/SBUV2 satellites and ground-based Umkehr/Dobson instruments spanning the period from 1957 to the present provide a data set for examining middle and upper stratospheric trends. The ozonesonde network starting in the early 1960's extends to the end of 1996 and has generated a data set for potentially examining lower stratospheric/tropospheric ozone trends. There are also other ozone data records available over shorter time periods which can be used to validate the long-term measuring systems including the HALOE and MLS instruments operating on the UARS satellite, ground-based lidar, and ground-based microwave instruments. We report here on studies done to evaluate the validity of the long-term trends obtained by these observing systems.

A major objective of this chapter is to determine if there is sufficient confidence in the long-term measurement systems to use them for accurate determination of ozone trends in the stratosphere and troposphere. The purpose of this study is therefore to validate the quality of these data sets including quantification of errors and to determine if there are any limitations in altitude or latitude regarding use of the data for deriving global trends. Important data quality questions addressed in the study include:

• What are the lower and upper altitude limits of the different data sets to be used in validation or trend studies?

• What is the smallest long-term ozone change that can be verified and over what altitude range?

• Are there any evidences of instrumental drifts of the measuring systems with time?

• What special screening methods should be used for the different data sets?

• Should SAGE sunrise, sunset or both data sets be used for trend determinations?

• How should data offsets of SAGE I compared to SAGE II be handled?

Answers to these questions will provide the basis for recommendations needed to derive the ozone trends reported in Chapter 3.

The general philosophy of this study was first to do internal data consistency studies for the long-term ozone measuring platforms through intercomparisons and data analysis. Next the long-term ozone records were compared with data collected over shorter time periods (e.g. HALOE, MLS, Lidar and Microwave) to look for any possible seasonal or latitudinal biases, which would indicate systematic errors in the instruments or which point to inherent instrumental drifts.

Above 25 km SAGE I and II, SBUV, SBUV2, and the Umkehr/Dobson-network provide the major data sources to derive global ozone trends in the middle and upper stratosphere. The good spatial coverage of the SBUV and SBUV2 data are used to examine the effect of the poorer spatial coverage of SAGE on the derived trends. In addition, the HALOE and MLS data were compared to SAGE II over shorter time intervals to determine if there are any seasonal or latitudinal dependencies in SAGE II trends. Umkehr/Dobson, Lidar, and Microwave data were used to verify the accuracy of individual profiles measured by the SAGE I and II satellite instruments and to further assess the validity of trends.

An important issue in this study is the derivation of global ozone trends between the tropopause and 25 km which is the region with the largest ozone depletion. SBUV and SBUV2 and Umkehr/Dobson are not sensitive enough in this region which leaves SAGE I and II as the only global data set with temporal and spatial coverage sufficient to meet this objective. However, the coverage of SAGE I and II is still limited to between 60°S and 60°N. Further, the validity of SAGE measurements for Z £ 25 km needs to be carefully scrutinised due to possible errors in the SAGE data caused for example by high aerosol loading resulting from volcanic eruptions (e.g., Mt. Pinatubo) or ice crystal absorption due to high cirrus clouds (see also Cunnold et al., 1996a). Ozonesondes provide data to derive long-term trends in the lower stratosphere, but with very poor spatial coverage at a limited number of stations, mainly located on the continents in the northern hemisphere (NH). Ozonesondes alone therefore cannot provide global trends. However they provide valuable data sets for intercomparison with SAGE results to investigate any instrumental biases and drifts in indicated trends. In the troposphere ozonesondes provide the only data for use in deriving an ozone trend. The availability of other data to validate the sonde measurements is very limited and can only be done over very short time intervals. Only a few dedicated intercomparison campaigns have been performed in the past. However, over the last 5-10 years several studies have been performed to validate sonde data by comparison with other techniques like Lidar or UV-photometric measurements from aboard aircraft.

During the pre-satellite period before 1978/1979 ozone sondes and Umkehr/Dobson measurements provide the only data sets for study of ozone change. Therefore, it is important to characterise the quality of the sondes and Umkehr time series during this period (i.e. 1965-1979). There are indications that in the tropopause region up to 17-20 km ozonesondes are probably more reliable then satellites. We have attempted in this study to use the newest, latest, revised versions of the different ozone data sets which were made available before the end of June 1997 and which were stored on or linked to the dedicated SPARC temporary data base.

Coincidence criteria used for all comparisons shown in this report are ± 2° latitude, ± 12.5° longitude, and 2 days time unless otherwise noted. We recognise that these are rather loose criteria. It was found for example when validating the Nimbus 7 LIMS data, that even 3 hours could make an important difference in the temperature agreement; but because SAGE II uses the solar occultation experiment approach, coincidence opportunities are limited requiring adoption of the criteria used. When comparisons are done against SAGE II, SAGE II is always the reference used in calculating percentages with the formulation:

?(2.1)

where correlative refers to balloon sondes, lidar, and other techniques. Also unless otherwise noted, the error bars shown throughout this chapter are the 2 standard error mean bars (or 2 sem) calculated by dividing twice the standard deviation by the square root of the number of samples used in the comparison.