2.1 Satellite instruments for the measurement of total ozone column

A convenient way to measure ozone abundance is to measure its total column, which corresponds to the total number of ozone molecules in a column from the Earth's surface to the top of the atmosphere. It is generally expressed in Dobson units¹. From the various satellite instruments that measure the total ozone column, probably the most famous is the Total Ozone Mapping Spectrometer (TOMS) that has been operated by the National Aeronautic and Space Administration (NASA) from 1978 until 2006. TOMS measures solar ultraviolet radiation scattered back into space by air molecules. Total ozone is retrieved from the different absorption of solar radiation by ozone at various wavelengths in the UV range. The first TOMS instrument began operation on board the Nimbus 7 meteorological satellite, which included also the Solar Backscatter Ultraviolet (SBUV) instrument, another ozone-measuring sensor. TOMS looked downward at several angles across the nadir track of the satellite to measure backscattered radiation at six wavelengths ranging from 312 to 380 nm. SBUV was designed to measure the ozone vertical distribution and total ozone. It looked in the nadir direction and measured the backscattered radiation at 12 wavelengths ranging from 250 to 400 nm. The concept of these instruments was based on the original Backscatter Ultraviolet (BUV) launched in 1970. The SBUV and TOMS instruments on board Nimbus 7 continued to make measurements until 1990 and 1994, respectively. The SBUV was followed by the series of SBUV/2 instruments launched on the NOAA polar orbiting series. The TOMS instrument was followed by Meteor TOMS on the Russian Meteor 3 satellite and the Earth Probe TOMS, which operated up to 2006 (see Fig. 1). The TOMS total ozone record has been taken over by the Ozone Monitoring Instrument (OMI), launched on the Aura satellite in 2004. OMI is the result of a Dutch /Finnish /American collaboration. It uses hyperspectral imaging in the visible and ultraviolet range for the retrieval of ozone and key air quality components such as NO₂, SO₂, BrO, OClO, and aerosol characteristics. OMI makes cross-track measurements like TOMS, but with an improved spatial resolution. It has also the capability to retrieve the ozone profile, like the SBUV instruments.

2.2 Validation and merging of total ozone satellite observations

The ozone vertical distribution can be measured from space by a variety of techniques that include nadir sounding in various parts of the solar radiation spectrum, solar occultation, limb scattered light and limb emission measurements. Solar and stellar occultation measurements are made by measuring radiation from the sun or a star as they set or rise. During these transitions, the instrument views the source through the atmosphere and the radiation is attenuated by scattering and absorption by minor constituents. Measurements made above the Earth's atmosphere are not attenuated and these data can be used to calibrate the instrument response. The fundamental measurement made by such instruments is a spectral slant path atmospheric transmission profile. These transmission profiles are then analysed in order to infer aerosol extinction and species density profiles of species like ozone, NO₂, and H₂O as a function of altitude. This technique provides high vertical resolution (∼1 km) and very small long-term drifts resulting from instrument calibration. However, spatial sampling is limited, and it takes approximately one month to sample the 60°N to 60°S latitude range in the case of solar occultation. This technique has been used by a variety of instruments such as SAGE I and SAGE II developed by NASA, HALOE on board the Upper Atmosphere Research Satellite (UARS) launched in 1991, and the series of Polar Ozone and Aerosol Measurement (POAM) instruments developed by the Naval Research Laboratory (NRL). Launched on board the Earth Radiation Budget Satellite (ERBS), the SAGE II instrument has provided the longest global ozone profile record, from 1984 to 2005. Comparisons with ground-based measurements showed that the quality of the latest version of SAGE II ozone profiles is excellent down to the tropopause, except after the eruption of Mt. Pinatubo in June 1991 when aerosol interference contaminated the measurements at the altitude of the volcanic aerosol cloud (Gleason et al., 1993). The HALOE instrument provided also a long ozone data record of about 14 years between 1991 and 2005. The POAM II was launched aboard the French SPOT-3 satellite in 1993 into a Sun synchronous polar orbit. The measurements were made at latitudes larger then 55°N and 63°S, which allowed them to be used for the study of the ozone loss in the Polar Regions. POAM II was followed by POAM III launched on board SPOT-4 in 1998. It operated until 2006. The Global Ozone Monitoring by Occultation of Stars (GOMOS) instrument was launched onboard the Envisat satellite in 2002 and is operated by the European Space Agency since then. It uses the stellar occultation technique and provides a global record of ozone profiles and other atmospheric compounds such as NO₂ and H₂O. The most recent satellite mission using the solar occultation technique is the Canadian Atmospheric Chemistry Experiment satellite launched in 2003. It includes two instruments, ACE-FTS a high spectral resolution infrared Fourier Transform Spectrometer (FTS), and a UV-visible-Near Infrared spectrophotometer known as MAESTRO.

A technique somewhat similar to the occultation technique is to observe the limb (edge) of the Earth's atmosphere. The Microwave Limb Sounder (MLS) experiments developed by the Jet Propulsion Laboratory (JPL) use this method to remotely sense vertical profiles of atmospheric gases, temperature, pressure, and cloud ice. The first MLS experiment was launched on the NASA UARS platform in 1991 and ceased operation in 2001. The second one (EOS MLS) is on the NASA Earth Observing System (EOS) Aura mission launched in 2004. Along with ozone it measures other important species involved in stratospheric chemistry such as H₂O, HNO₃, N₂O and ClO.

Nadir sounding instruments providing total ozone measurements such as SBUV and GOME can also provide information on the ozone vertical distribution. Both instruments measure in the near ultraviolet where ozone absorption is strong and change rapidly with wavelength. Rayleigh scattering² is also stronger towards shorter wavelengths and as atmospheric density increases. These effects provide a scanning function whereby shorter UV wavelengths are more strongly affected by the absorption of ozone at higher altitudes (e.g. Bhartia et al., 1996). The general problem of inverting nadir backscatter measurements to derive an ozone profile is ill-conditioned. One must apply some sort of constraint to achieve a physically reasonable solution. The most common way is to apply the optimal estimation method, which derives ozone by selecting a profile that has the highest probability of being the correct profile, based on a priori statistical information (Rodgers, 1990). Nadir sounding instruments provide a much better coverage of the ozone field than the occultation technique, but at the expense of a much lower resolution of the retrieved ozone profiles (several kilometres).

The full understanding of the various causes of ozone decrease at the global scale requires accurate long-term measurements of not just ozone but also a broad range of chemical species and long-lived tracers that influence the ozone budget in the stratosphere. To that aim several major atmospheric missions have been launched since the early 1980 s. A first insight of the atmospheric composition in the stratosphere was provided by the Limb Infrared Monitor of the Stratosphere (LIMS), which operated on board Nimbus 7 for several months in 1978/79 and by the Spacelab 3 Atmospheric Trace Molecule Spectroscopy (ATMOS) carried on the Space Shuttle in 1985. In 1991, a major satellite mission dedicated to stratospheric chemistry was launched: the Upper Atmospheric Research Satellite (UARS, Reber, 1993). UARS included several instruments for the measurements of stratospheric compounds: MLS, HALOE, the Improved Stratospheric and Mesospheric Sounder (ISAMS) and the Cryogenic Limb Array Etalon Spectrometer (CLAES), the two latter instruments both working in the infrared. UARS also measured winds and temperatures in the stratosphere as well as the energy input from the Sun. In addition to providing the first comprehensive picture of the atmospheric composition in the stratosphere, one of the major results of the UARS mission was the evaluation of chlorine activation in the Polar Regions. MLS measurements of chlorine monoxide (ClO) showed that chlorine in the lower atmosphere was almost completely converted to chemically reactive forms in both southern and northern polar vortices in winter, before the start of chemical ozone depletion (Waters et al., 1993). Fig. 2 shows enhanced ClO together with depleted ozone amounts in the southern polar vortex, as measured by MLS in 1991.

The ESA's Environmental Satellite (Envisat) was launched in 2002 with three instruments dedicated to atmospheric chemistry: GOMOS (described in section 3), the Michelson Interferometer for Passive Atmospheric Sounding (MIPAS) and the Scanning Imaging Absorption spectrometer for Atmospheric CHartographY (SCIAMACHY). MIPAS is a Fourier transform spectrometer operating in the near to mid infrared for the measurement of high-resolution gaseous emission spectra at the Earth's limb. SCIAMACHY is an imaging spectrometer, which aims at performing global measurements of trace gases in the troposphere and stratosphere. It has three different viewing geometries: nadir, limb, and sun/moon occultation, which yield total column values as well as distribution profiles in the stratosphere and (in some cases) the troposphere. These three instruments have provided since 2002 comprehensive data sets of stratospheric species, which were used for the study of ozone depletion in Polar Regions or the ozone budget in other atmospheric regions. The Envisat atmospheric composition data strongly contribute to assimilation systems such as those participating to the ESA PROMOTE project (http://www.gse-promote.org) in various areas (stratospheric ozone, air quality, UV monitoring and climate).

Aura, the last of the large Earth Observing System observatories, was launched by NASA in 2004. Aura was designed to make comprehensive stratospheric and tropospheric composition measurements from its four instruments, the High Resolution Dynamics Limb Sounder (HIRDLS), the Microwave Limb Sounder (MLS), the Ozone Monitoring Instrument (OMI), and the Tropospheric Emission Spectrometer (TES). As far as the ozone layer is concerned, one of the results of Aura was to show that HCl (measured by MLS) has started to decline in the upper stratosphere with a rate consistent with surface abundance decrease rates in chlorine source gases, taking into account a delay of about 6 years (Froidevaux et al., 2006). This result corroborates the success of the Montreal Protocol with respect to the cleansing of the stratosphere of the excess of chlorine loading due to human activities.

In addition to these three major satellite missions, one can also cite the ODIN satellite launched in 2001 as a result of Swedish, Finish, Canadian and French cooperation, and the ADEOS I and II satellites launched by Japan's National Space Development Agency. However, both ADEOS satellites were lost after less than one-year operation.

5.1 Total ozone trend estimation

5.2 Trends in the vertical distribution of ozone

In order to estimate long-term changes of ozone at a global scale as a function of altitude one has to use the longest satellite records of the ozone vertical distribution. The most recent ozone assessment used the SAGE I + II and SBUV ozone records and compared them with ground-based measurements from ozone sondes and Dobson spectrometer (see Fig. 4). The ozone vertical distribution can be retrieved from Dobson spectrometer measurements during sunrise and sunset, using the Umkehr method (e.g. Mateer and Deluisi, 1992). The vertical profile of ozone trends reflects the influence of halogen compounds on ozone as a function of altitude. This influence is largest in the upper and lower stratosphere. Around 40 km, the ozone loss is due to catalytic cycles involving chlorine radicals, while at 20 km it is caused by the hemispheric dilution of polar ozone chemical destruction and in situ destruction involving halogen and hydrogen compounds as well as stratospheric aerosols (WMO, 2007).

The evaluation of the extension of the Antarctic ozone hole is largely based on the compilation of the various satellite observations of the total ozone column. The ozone hole is commonly defined as the region where ozone amounts are lower than 220 DU. Since the 1980 s, the ozone hole has become a recurring seasonal feature and the year-to-year variation of its spatial extent is reported in various bulletins published on-line, e.g. by WMO⁴ or NASA⁵. Latest analyses have concentrated on the question of the stabilisation and decline of the ozone hole as a result of the Montreal Protocol. Fig. 5 presents the temporal evolution of the size of the ozone hole. It shows that over the last decade, the ozone hole has leveled off after the mid-1990s. Satellite observations have also visualized the effects of the first ever observed Antarctic major stratospheric warming in September 2002 (see Fig. 6). This warming caused a drastic reduction of the ozone hole's area and resulted in a much less severe ozone hole in that year. It was shown that this warming resulted from anomalously strong dynamical wave activity in the troposphere, which propagated into the Antarctic stratosphere because of favourable dynamical conditions. The triggers of these very unusual waves are unknown, and it is not clear whether the 2002 warming is a random event due to internal atmospheric variability or whether it can be related to long-term changes in climate. The Antarctic winter 2004 was also dynamically very active and had less ozone mass deficit than previous years. The higher levels of ozone in these two years were dynamically driven and not related to halogen chemical reductions (WMO, 2007).

In the Arctic, the winter stratosphere is much warmer than in the Antarctic, due to stronger dynamical activity in the northern hemisphere. Arctic ozone loss during spring is highly variable depending on dynamical conditions, but it is anyhow much less severe than in the southern hemisphere. Maximum ozone loss in terms of integrated columns has not exceeded 30% since the early 1980s. For current halogen levels, anthropogenic chemical loss and variability in ozone transport are about equally important for year-to-year Arctic ozone variability. It is thus difficult to disentangle chemical and dynamical processes affecting ozone. A rather successful method is to use both Chemistry Transport Models (CTM) and satellite observations to infer chemical ozone loss (Deniel et al., 1998; Singleton et al., 2007). CTMs are atmospheric models that use meteorological analyses for the simulation of transport processes. This method, using CTMs passive ozone (not affected by chemical processes) and ozone fields from satellite observations, is illustrated in Fig. 7. It shows ozone loss in a very cold winter (2004/2005), computed from the SLIMCAT CTM (Chipperfield, 1999) and various satellite observations. All satellite instruments provided similar results, with a maximum inferred loss of 2–2.3 ppmv near 450 K (16 km). In addition to estimating ozone depletion, this technique is very useful for testing the understanding of chemical processes taking place in the Polar Regions through the comparison with the ozone loss simulated by the model.

Data assimilation is another important application for satellite observations of ozone and other stratospheric chemical species. Data assimilation is the combination of information from observations and models of a particular physical system in order to get the best possible estimate of the state of that system. Data assimilation methods were largely developed for operational weather forecasting, but in recent years, they have been applied to an increasing range of Earth Science disciplines and, more specifically, in Atmospheric Chemistry. There are many reasons for the assimilation of ozone observations:

• • data assimilation provides an improved mapping of ozone fields from multiple data sources;
• • it enables short-term ozone forecasts for the determination of surface UV radiation;
• • it is one of the best approaches for the estimation of tropospheric ozone at global scale, from the combination of satellite observation of stratospheric ozone with meteorological information provided by the model;
• • ozone data assimilation provides improved wind analyses as ozone measurements in the lower stratosphere contain information about the wind field, owing to the long photochemical lifetime of ozone in that region.

The assimilation of satellite ozone data has thus been the focus of several research groups in the last decades, with studies showing constant improvement in the state of the art. They have demonstrated the importance of having both total column observations as well as high-resolution profile information (e.g. Rood et al., 1989; Swinbank and Lahoz, 2003). As an example, the European Centre for Medium-Range Weather Forecasts (ECMWF) 40-year Reanalysis (ERA-40) is based on the use of a single, variational data assimilation scheme, which has processed observations from September 1957 to August 2002 in a consistent manner. For atmospheric ozone and water vapor, ERA-40 represents one of the largest continuous data sets available (Oikonomou and O’Neill, 2006). A new “ERA-Interim” dataset is in progress, with improved scheme for data assimilation and extended temporal coverage (Simmons et al., 2007).

With the expected decline of ozone depleting substances (ODS) linked to the Montreal Protocol, the next step is to predict the timing of the full recovery of the ozone layer. Owing to their generally long lifetime, the removal of ODS from the atmosphere will take several decades even with continued compliance with the Montreal Protocol. Ozone is expected to respond to these changes in ODS, but the timing and sensitivity of the response will depend on other factors, such as the increase in greenhouse gases, in particular CO₂, that induce a cooling in the stratosphere, increase of the stratospheric water vapor content or increased emission of source gases such as N₂O. Chemistry Climate Models (CCM) are important tools for predicting the future evolution of the stratosphere and the ozone layer, taking into account these changing conditions. These models include a full representation of dynamical, radiative and chemical processes in the atmosphere and their interactions. The feedback between chemical and dynamical processes is a major difference between CCMs and CTMs and such interdependencies are required to simulate the future evolution of ozone in a changing climate (Austin et al., 2003; Dameris et al., 2005).

While these models have substantially improved their climate simulations, there are still flaws in many models, including cold biases in the Southern Hemisphere and a lack of realistic tropical zonal wind oscillations. In that respect, satellite observations are important resources for model evaluation. A key question is how the current understanding of the stratospheric climate system fits together with temperature and constituent observations. This question is addressed by comparing past observations of ozone and other trace gases with various model simulations. In order to achieve a better understanding of model biases, quantities that are important for the simulation of ozone fields, like transport diagnostics, e.g. the mean age of air and the propagation of the annual cycle in water vapor, are validated against observations. The mean age of air can be inferred from observations of conserved tracers, as for example CO₂, SF₆ or HCl and HF. The age of air varies from 1 year in the equatorial lower stratosphere to 6 years in the high stratosphere. The annual variations in water vapor mixing ratios propagate vertically into the tropical stratosphere. This phenomenon, called “water vapor tape recorder”, is used to test the capacity of the models to simulate vertical transport in the Tropics. The comparison with water vapor observations from HALOE in a recent assessment of various CCMs simulations has shown a large spread in the results, indicating large differences in the representation of transport processes in the models (Eyring et al., 2006).

Spatial observations have been a major contributor to the assessment of the ozone layer and the understanding of ozone balance as a function of latitude and height since the early 1980s. They have provided a comprehensive picture of the variability of ozone and atmospheric compounds that play a key role in stratospheric chemistry. The various satellite instruments developed in the last decades for the observation of the ozone layer have paved the way for the development of new sensors dedicated to other scientific issues such as air quality and climate change.

According to the latest estimates (e.g. WMO, 2007), recovery of the ozone layer is expected to occur around the middle of the 21st century as a result of the decrease of ozone depleting substances. Due to influence of future increases in greenhouse gas concentrations on the ozone layer and the reverse effect of future ozone changes on climate, particularly in the Southern Hemisphere, it is of utmost importance to continue the observation of the ozone layer with dedicated satellite instruments. Scientists recently raised concerns about the potential lack of global satellite ozone observation at the medium term (see the Athens statement at the 20th anniversary of the Montreal Protocol⁶). Indeed, the list of planned satellite observing systems shows that the capability for measuring ozone profiles will end at about 2015. These profiles characterize ozone in the lower and upper stratosphere, where the chemical loss is largest. The lack of such measurements might hinder the ability of the scientific community to assess the success of the Montreal Protocol with respect to the recovery of the ozone layer.