Inter Calibration And Validation Of Atmospheric Ozone

Published Date: 02 Nov 2017

Introduction

Ozone (O3) is so rare in the atmosphere (there are only about three ozone molecules per 10 million air molecules) that if it was brought down to the Earth’s surface and compressed to standard temperature and pressure, it would form a layer about as thick as a small coin. In spite of this, the importance of ozone, particularly in protecting the biosphere from harmful effects of solar ultraviolet radiation, vastly exceeds what one might expect from this minor trace gas in the atmosphere. The measuring unit of ozone in atmosphere is the Dobson Unit (DU), which refers to a layer of ozone that would be 0.001 cm thick under conditions of standard temperature (0 °C) and pressure (one atmosphere or 1013.25 millibars). Traditional measurements of atmospheric ozone from the ground, using ultraviolet spectrophotometers, and from ozone sonde balloons were being made for many years before remote sensing satellites came to be developed and operated. However, these traditional methods only gave ozone concentrations at a few points on the surface of the Earth. The advantage of satellite remote sensing is that it provides frequent coverage of the whole Earth.

Ozone is produced naturally in the stratosphere and drifts downward by mixing processes, producing a maximum in its concentration near 25 km. It is also destroyed naturally by absorbing ultraviolet (UV) radiation (0.2-0.3 μm) and through chemical reactions with other atmospheric components. The natural balance between production and destruction leads to the maintenance of the stratospheric ozone layer which, however, is now subject to substantial anthropogenic impacts (e.g. Newman et al. 2009). Much smaller amounts of ozone, of the order of 10%, are residing in the troposphere (Figure 1).

One important reason for studying atmospheric ozone in great detail is to identify long-term trends in ozone concentration. Remote sensing played a vital role in the events leading up to the Montreal Protocol in 1987, which led to the banning of the manufacture and use of ozone-depleting substances throughout the world. Since then remote sensing has continued to play an important role in monitoring atmospheric ozone and studying the rate of recovery of the ozone layer in response to the implementation of the Montreal Protocol.

It is now slightly over 50 years since the era of satellite remote sensing began on 4 October 1957, when the former Soviet Union launched Sputnik 1, the world’s first artificial satellite (a 55-cm-diameter sphere that weighed 83 kg with four antennae coming off it) (Figure 2). In the intervening years since then, satellites have revolutionised the Earth sciences, including the study of atmospheric trace gases in general and ozone in particular. With the passage of time, satellite-flown instruments became more complicated, while larger satellites carrying more and more instruments were launched. One of the largest of these was ENVISAT (ENVIronmental SATellite) which was launched on 1 March 2002 by the European Space Agency (ESA), with a mass of 8211kg and dimensions in orbit, i.e. with its solar panels deployed, of 26 m × 10 m × 5 m. This was a polar-orbiting sun-synchronous satellite flown at an altitude of about 800 km, with an inclination of 98° and an orbital period of 101 min. It carried ten main instruments for monitoring the Earth’s atmosphere, biosphere, cryosphere, geosphere and the oceans (Figure 2) (http://envisat.esa.int/) and three of these were particularly relevant to atmospheric ozone. ENVISAT successfully reached its nominal 5-year mission lifetime, having orbited the Earth more than 26,000 times. Figure 2 shows Sputnik on the same scale as ENVISAT.

Recently, we have been seeing a trend to move away from large multifunctional Earth-observing satellites, like ENVISAT, to small satellites dedicated to one particular observational task. The advantages of this include simple and speedy design and manufacture, many more launch opportunities, low cost, non-competition between the requirements of different instruments, risk reduction, etc. Small satellites also provide opportunities for developing countries to become involved in technology transfer and the development of indigenous space-related capabilities. Thus, we foresee a role for both large and small Earth-observing satellites in the future (Cracknell and Varotsos, 2007).

The traditional ways to study the concentration of atmospheric ozone were by specially designed ground-based spectrophotometers and with ozonesondes. Nowadays, three main types of primary instruments are used for routine total ozone content (TOC) observations from the ground: the Dobson spectrophotometer, the Brewer spectrophotometer and the M-83/124/134 filter ozonometer. All of these instruments use the principle of differential absorption by ozone of solar UV radiation in the Hartley or Hartley-Huggins bands in the 300-460 nm UV spectral region. Total ozone is calculated by using the ratio of the intensities of two wavelengths of backscattered ultraviolet light where one is strongly absorbed by ozone while the other is absorbed very little. Ozonesondes are balloons like ordinary meteorological radiosonde balloons except that they carry instruments which sample the air as the balloon rises and measure the ozone concentration; the data are transmitted to ground by a radio transmission.

TOC was measured from the ground at a few sites in the late 1920s and early 1930s by using Dobson spectrophotometers and records of TOC have been kept at various sites since then and they continue to be kept in parallel with satellite measurements. There are some contrasts to be noted. The ground-based spectrometers and the ozonesondes measure the concentration of ozone directly. The satellite-flown instruments generally are less direct and rely on algorithms which need to be validated to extract the ozone data.

Before considering satellite remote sensing of atmospheric ozone we should mention the use of aircraft. There have been a few studies of the atmosphere using instruments flown on aircraft to study ozone and related chemicals directly, most notably in the early study of the Antarctic ozone hole during the Airborne Antarctic Ozone Experiment (AAOE) in 1987. AAOE was mainly based on aircraft observations and explored the weaknesses of the existed theories about the chemical processes taking place in the ozone hole dynamics. In this context high flying aircraft were also used, a DC-8 flying at around 42,000 ft (=12,800 m) and an ER-2 flying at around 67,000 ft (= 20,400 m) in very hostile conditions right inside the ozone hole. In the frame of this experiment the data were also collected with ozonesondes launched at four Antarctic stations: Halley Bay, McMurdo, Palmer Station, and the South Pole. Considerable insight was gained into the chemical processes involved in the destruction of stratospheric ozone, see for instance chapter 5 of Cracknell and Varotsos (2012). However, in many cases of the satellite-flown instruments which we shall discuss in this article similar instruments have been flown on aircraft as well. There has usually been one of two reasons for this. The first is for calibrating and evaluating prototypes of the instruments before flying them in space. The second is for simultaneous flying of similar (nominally identical) instruments in aircraft and in space for the validation of the data from the space-borne instrument. However, there is another rather new aspect of airborne remote sensing which we should mention.

Just as remote-sensing satellites for environmental studies can be regarded as a spin-off development from military spy satellites, so we are now seeing a similar spin-off from unmanned military aircraft in a new mission called Global Hawk Pacific, (GloPac), and one or two other unmanned aircraft NASA (National Aeronautics and Space Administration) projects. NASA has staged environmental (manned) research flights from aircraft previously, but none has had the reach and duration of Global Hawk, a high-altitude, unmanned aircraft, see figure 3(a). The aircraft, which is distinguished by its bulbous nose and 35.4 m wingspan, can travel about 18,500 km in up to 31 hours, carrying almost 1 t of instrument payload. (http://gsfctechnology.gsfc.nasa.gov/GlobalHawk.htm). The Global Hawk aircraft used in the GloPac mission carried a suite of 12 instruments. The first operational flights of the Global Hawk for the GloPac were conducted in support of the Aura Validation Experiment (AVE) (Jiang et al. 2007). Aura is one of the NASA A-train of satellites (Figure 3(b)). This is a convoy of Earth observing spacecraft following one another in a polar orbit.

We now turn to the satellite remote sensing of atmospheric ozone and in particular of the total ozone content (TOC), the ozone vertical profile (OVP), atmospheric ozone circulation and ozone depletion.

2. Satellite remote sounding of the total ozone content (TOC) and ozone vertical profile (OVP)

There are three different principles involved in satellite remote sensing of ozone, two of which involve looking down at the Earth below the orbit of the spacecraft and one of which involves limb sounding that is observing solar radiation by looking towards the horizon from the satellite.

TOVS (TIROS Operational Vertical Sounder)

The first method, which has been used for many years by TOVS and its successors, involves looking directly down at the Earth’s surface and using Earth surface leaving radiation.

The TOVS is a set of three instruments, the High Resolution Infrared Sounder (HIRS), the Stratospheric Sounding Unit (SSU) and the Microwave Sounding Unit (MSU), providing 27 spectral channels which are all, except one, in the infrared and microwave region of the spectrum with one panchromatic channel (for more details see, for example, section 1.4 of Cracknell (1997). It scans a swath of about 2,000km width below the orbit of the spacecraft and its main purpose is to measure atmospheric profiles of temperature, pressure and humidity. There is one particular spectral channel, channel 9 of the HIRS, at a wavelength of 9.7 µm, which is particularly well suited to monitoring stratospheric ozone concentration; this is a (general) window channel, except for absorption by ozone, i.e. ozone is the only atmospheric constituent which absorbs radiation at this wavelength. The radiation received at this wavelength by the HIRS instrument at this wavelength was emitted from the Earth’s surface, but it is attenuated by the ozone in the atmosphere. The less ozone, the greater the amount of radiation reaching the satellite. A 3 DU drop in lower stratospheric ozone produces a measurable (circa 0.2°C) increase in the brightness temperature in this channel.

It appears that, strictly speaking, what is measured is the lower stratospheric ozone and a correction has to be applied to obtain the TOC. Images are now regularly produced from TOVS data giving hemispherical daily values of TOC. This means that there is now a long time series of such data available. TOVS data have been used to determine atmospheric ozone concentration from 1978 to the present (Neuendorffer, 1996; Kondratyev, 1998) and the standard deviation is approximately 7%. Nowadays, TOVS data are also used when the more reliable TOMS data are not available (http://www.theozonehole.com/2010oct.htm). Since TOVS ozone data are only sensitive to variations in the lower stratosphere, the long-term TOVS TOC trends are, at best, only indicative of lower stratospheric ozone trends. An advantage of TOVS 9.7μm radiation, which is infrared, over other systems that use solar UV radiation is that TOVS data are available at night time and in the polar regions in winter. The drawbacks are that when the Earth’s surface is too cold (e.g. in the high Antarctic Plateau), too hot (e.g. the Sahara desert) or too obscured (e.g. by heavy tropical cirrus clouds) the accuracy of this method declines.

Li et al. (2001) discussed the potential for using Geostationary Operational Environmental Satellite (GOES) sounder radiance measurements to monitor TOC with a statistical regression using GOES sounder spectral bands 1–15 radiances. The advantage is the high temporal frequency of the availability of the data. Hourly GOES ozone products have been generated since May 1998. GOES ozone estimates were compared with TOMS (Total Ozone Monitoring Spectrometer) TOC data and ozone measurements from ground-based Dobson spectrometer ozone observations. The results showed that the percentage root-mean-square (rms) difference between instantaneous TOMS and GOES ozone estimates ranged from 4% to 7%. Also, daily comparisons for 1998 between GOES ozone values and ground-based observations at Bismarck, North Dakota; Wallops Island, Virginia; and Nashville, Tennessee, showed that the rms difference is approximately 21 DU.

TOMS and (S)BUV

The second method also involves looking down from the spacecraft and, like the ground-based spectrophotometers, uses the wavelength dependence of the absorption of solar ultraviolet radiation in the atmosphere. There are two slightly different systems, namely the TOMS (Total Ozone Monitoring Specrometer) and the (S)BUV ((Solar) Backscatter UltraViolet) type. The TOMS has six UV wavelength bands from 312.5 to 380 nm (312.3, 317.4, 331.1, 339.7, 360 and 380 nm) and TOC is determined by utilising the Huggins band of the ozone absorption spectrum. Like the TOVS it is a scanning instrument. The first four wavelength regions are used in pairs, to provide estimates of ozone concentration by the differential absorption method, while the other two (free of ozone absorption) are used, to determine the effective background albedo. The TOMS instruments flown on the Nimbus-7 satellite (1978) and its successors have been used to measure global distribution of ozone. This instrument was followed by the TOMS flown on the Meteor-3 satellite from which data were gathered from August 1991 to December 1994. A detailed description of the adaptations to TOMS for Meteor-3 is provided by Herman et al. (1997) and the pre-launch and post-launch calibrations are described by Jaross et al. (1995). Of the five TOMS instruments which were built, four entered orbit successfully. There was a one and a half year overlap between Nimbus-7 and Meteor-3 and this permitted an inter-calibration between the two datasets. This allows the data from the two satellites to be used to form a continuous 16-year data set, which can be used to study ozone concentration trends from November 1978 to December 1994.

After an eighteen month period when the programme had no in-orbit capability, the Japanese ADEOS TOMS was launched on 17 August 1996 and provided data until the satellite which carried it lost power on 29 June 1997. For ADEOS/TOMS TOC, the absolute error was 3%, the random error is 2%, and the drift over the 9-month data record was less than 0.5%. The ADEOS/TOMS observations of TOC are approximately 1.5% higher than a 45-station network of ground-based measurements (McPeters and Labow 1996; Seftor et al.1997; Herman et al. 1997; Torres and Bhartia 1999).

The successor of ADEOS, Earth Probe TOMS (launched on 2 July 1996), failed on December 2006 and was subsequently replaced by the Ozone Monitoring Instrument (OMI). The historical and current daily and monthly ozone data are widely available at the NASA website http://ozoneaq.gsfc.nasa.gov/. Spatially, a good global coverage in TOMS (OMI) data is combined with a high resolution of l° by latitude and 1.25° by longitude.

In addition to ozone, the TOMS instrument measures sulphur dioxide released in volcanic eruptions. These observations are of great importance in the detection of volcanic ash clouds that are hazardous to commercial aviation.

We now move on to the consideration of the BUV and SBUV instruments. The SBUV instruments are nadir-viewing instruments which are able to determine the TOC and OVP by measuring sunlight scattered from the atmosphere in the ultraviolet spectrum. They have more ultraviolet channels than the TOMS and many of these are at shorter wavelengths than in the TOMS, see Table 1. The SBUV instrument (onboard the NOAA-16) provided observations of the scattered UV radiation (252 to 340 nm). The data at the shortest wavelengths are employed for the estimation of the ozone concentration as a function of height, whilst those at the longer wavelengths are used to provide total ozone concentration (TOC) observations. The other spectral bands are at shorter wavelengths than on TOMS and this radiation is reflected back to the satellite by the upper layers of the atmosphere before it can ever reach the surface of the Earth. The data from these bands is used to estimate the OVPs (ozone vertical profiles) by using a maximum likelihood retrieval algorithm.

The ozone observations derived from satellite data have been performed by the BUV instrument since April 1970 (i.e. launch of Nimbus-4). The advanced SBUV instrument (launched in November 1978) has been replaced by the SBUV/2 (NOAA and TIROS) (Table 2). It should be noted that the stratospheric ozone began to be monitored operationally in 1985 (NOAA-9, NOAA-11, NOAA-14, NOAA-16, NOAA-17, NOAA-18 and NOAA-19).

The Shuttle SSBUV was designed and developed at the NASA Goddard Space Flight Center (GSFC) to calibrate the Nimbus and NOAA solar backscatter UV instruments. In late 1989, the Space Shuttle Atlantis carried the instrument for the first time, in an appropriate orbital flight path to assess performance by directly comparing data from identical instruments (SBUV) on board the NOAA spacecraft and the Nimbus-7. The SSBUV was flown on eight Shuttle missions between October 1989 and January 1996 and provided regular checks on the individual satellite instruments’ calibrations. Multiple inter-comparisons with ground-based instruments have improved data retrieval algorithms and therefore, satellite ozone measurements have become compatible with those of the network of ground-based measurements. The principal purpose was to compare observations from several ozone-measuring instruments on board NOAA-9, NOAA-11, Nimbus-7, and UARS; this was because of the degradation of the SBUV and SBUV/2 instruments in space. The quality of the in-orbit calibration depends on the flight-to-flight calibration repeatability in SSBUV. Given accurate measurements of the backscattered radiance, it is necessary to account for differences in solar-zenith angle and effective surface reflectivity (Heath et al. 1993).

Limb sounding

The third general system used for the observation of ozone – and of other trace gases in the atmosphere – is limb sounding. This involves looking towards the horizon from the spacecraft, rather than looking vertically downwards, and observing the light from the sun, or occasionally from a star. The reason for this is to obtain a longer path through the atmosphere and therefore a greater number of the trace gas molecules in the path of the radiation.

During the second half of the 1960s several attempts at satellite remote sounding of ozone and other minor gas components from Soviet manned spacecraft were first made, using a hand-held spectrograph and a complex of solar spectrometers functioning in a regime of occultation geometry (Kondratyev, 1972). An early set of limb sounding instruments was flown on UARS (Upper Atmosphere Research Satellite) which was launched on 15 September 1991. These instruments included the Cryogenic Limb Array Etalon Spectrometer (CLAES), the Improved Stratospheric and Mesospheric Sounder (ISAMS), the Microwave Limb Sounder (MLS) and the Halogen Occultation Experiment (HALOE). The decommissioned satellite re-entered the atmosphere on 24 September 2011, with considerable media attention, caused by NASA predictions that substantial parts of it could reach the land. In the event, it landed in a remote area of the Pacific.

The complementary instrumentation of ADEOS included the Improved Limb Atmospheric Spectrometer (ILAS), the Interferometric Monitor for Greenhouse Gases (IMG), and the Retro-reflector in Space (RIS) instrument. IMG is a nadir-observing Michelson-type Fourier Transform Spectrometer (FTS) designed to measure the vertical profiles of CO2 and H2O, TOC, and the concentrations of CH4, N2O, and CO in the troposphere. The RIS, which is used together with laser ground stations, supports vertical profile and/or column measurements of a small number of gases. Limb sounding instruments have now been flown, and still are being flown, on several later spacecraft.

More recent instruments

The TOVS, TOMS and (S)BUV series of systems and the early limb sounding systems have, between them, provided quite a lengthy archive of atmospheric ozone data, but there are now a number of newer instruments which are expected to continue supplying this data for quite some time in the future, some of which combine both looking vertically downwards and limb sounding, see Table 3. A large number of these instruments are described in chapter 2 of Cracknell and Varotsos (2012). A few of the more important ones include GOME (Global Ozone Monitoring Experiment) on the European Space Agency (ESA) satellite ERS-2 (1995 - 2011), the OMI (Ozone Monitoring Instrument) on the NASA Aura satellite (2004 - today) GOMOS, SCIAMACHY (SCanning Imaging Absorption Spectrometer for Atmospheric Chartography) and MIPAS (Michelson Interferometer for Passive Atmospheric Sounding) on the European satellite ENVISAT (2002 – 2012) and IASI (Infrared Atmospheric Sounding Interferometer) and GOME-2 on the first European polar orbiting satellite MetOp (2006 - today) and the TOU (Total Ozone Unit) which is one of the main payloads on FY-3 (FengYun-3), the Chinese second-generation polar-orbiting meteorological satellite series (2008 - today) (Meijer et al. 2004). We shall describe some of these briefly in turn.

GOME was launched on 20 April 1995 on board the ESA Earth Resources Satellite (ERS-2) in a polar sun-synchronous orbit. It is a nadir-viewing multi-channel spectrometer measuring solar irradiance and earthshine radiance (not solar irradiance) in the wavelength range 240-790 nm at moderate spectral resolution (0.2-0.4 nm). The instrument exploits the fact that near 260 nm the penetration depth of solar radiation into the atmosphere strongly increases with increasing wavelength, because at this specific wavelength the ozone absorption cross-sections have their maximum. When the radiation has wavelength greater than about 310 nm, it penetrates into the tropopause reaching the surface. Therefore, the measurement of the backscattered radiation in the UV-visible region provides information about the OVP. In general, GOME measures TOC, OVP and total column of several other trace gas constituents, including BrO, NO2 and ClO2 and also obtains information on clouds, aerosols, and surface spectral reflectance. It can measure the TOC at a higher horizontal resolution than TOMS, and thus complements the TOMS observations. A second GOME instrument, GOME-2, was flown on MetOp-1, the first European polar-orbiting meteorological satellite which was launched on 16 October 2006. Some early work on comparison of its TOC data with reliable ground-based measurement recorded by five Brewer spectrophotometers in the Iberian Peninsula was carried out over a period of a year from May 2007 to April 2008. This showed that GOME-2/MetOp ozone data has a very good quality; for details see Antón et al. (2009).

The Ozone Monitoring Instrument (OMI) is a nadir-viewing near-UV/Visible CCD spectrometer on board NASA’s Earth Observing System (EOS) Aura satellite. Aura flies in formation about 15 minutes behind Aqua in the A Train (see Figure 3(b)),.Among the Aura instruments are the HIRDLS (High Resolution Dynamics Limb Sounder), the MLS (Microwave Limb Sounder) and the TES (Tropospheric Emission Spectrometer). OMI observations have a daily global coverage and are made at a spectral region of 264–504 nm with a grid of 13 km x 24 km at nadir. The intercomparison between OMI, TOMS, GOME and SCIAMACHY shows that OMI exhibits improved spatial resolution for the routine monitoring of various trace gases from space. With regard to ozone, the OMI provides the near-real time mapping of ozone columns at 13 km × 24 km and profiles at 13 km × 48 km resolution and provides a continuation of TOMS and GOME TOC time-series.

The Aura satellite which carries the OMI has a direct broadcast capability to broadcast the measurements to ground stations at the same time as the measurements are being stored in the spacecraft's memory for later transmission to Earth. The Finnish Meteorological Institute's very fast delivery (VFD) processing system utilises this direct broadcast to produces maps of TOC and ultraviolet radiation over Europe within 15 minutes after the satellite overpass of the Sodankylä ground station in northern Finland. The VFD products include maps of TOC, ultraviolet index, and ultraviolet daily dose. The aim of this service is to provide up-to-date information on the ozone and ultraviolet situation for the general public and snapshots of the current situation for scientists. The accuracy of the VFD products compares well with standard off-line OMI ozone products as well as ground-based Brewer measurements (Hassinen et al. 2008, Ialongo et al. 2008).

The NOAA series of polar-orbiting meteorological satellites which have operated continuously since 1978 carrying the TOVS and also the Advanced Very High Resolution Radiometer (AVHRR), and in some instances the SBUV/2, is finally about to be replaced by the NPOESS (National Polar-orbiting Operational Environmental Satellite System). The Ozone Mapper Profiler Suite (OMPS) is one of the five instruments to be included in the NPOESS. For many years two parallel and rather similar, but not identical, polar-orbiting meteorological satellite programmes were run by NOAA and the DoD (Department of Defense) in the USA. NASA, had also been involved, in various ways, in a variety of polar-orbiting environmental remote sensing programmes. NPOESS was planned as a tri-agency programme with the Department of Commerce (specifically NOAA) the Department of Defense (DoD, specifically the Air Force) and NASA. It was designed to merge the civil and defence weather satellite programmes, of the NOAA Polar Operational Environmental Satellites (POES) series and the US Department of Defense, Defense Meteorological Satellite Program, DMSP (Defense Meteorological Satellite Program), respectively, in order to reduce costs and to provide global weather and climate coverage with improved capabilities over the earlier systems. The main mission of NPOESS was the monitoring of the Earth's atmosphere, oceans, land hosting advanced sensors that were under operational-prototyping by NASA.

The NPOESS Preparatory Project (NPP) programme aimed to bridge the gap between the old and the new systems by flying new instruments on a satellite originally to be launched in 2005. However, the programme encountered technical, financial and political problems leading to a delay of the launch of NPP until 28 October 2011 and of the launch of the first NPOESS platform (C-1) until late 2014. NPP was successfully launched on October 28th, 2011. These would each be delays of five years from the original plan. This led to a review of the NPOESS and as a result of this review was that two polar-orbiting satellites serving military and civilian users would be pursued:

The JPSS (Joint Polar Satellite System) operating by NOAA/NASA with a performance period to 1 February 2015.

The Defense Department's portion is called DWSS (Defense Weather Satellite System).

The existing partnership with Europe through the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT), which operates the MetOp polar orbiting weather satellite programme, would continue.

The five-instrument suite for JPSS includes the Visible/Infrared Imager Radiometer Suite (VIIRS), the Cross-track Infrared Sounder (CrIS), the Clouds and the Earth Radiant Energy System (CERES), the Advanced Technology Microwave Sounder (ATMS) and the Ozone Mapping and Profiler Suite (OMPS). The OMPS is comprised of two sensors, a nadir sensor and limb sensor. Observations performed by the nadir sensor are employed to develop TOC measurements, while measurements from the limb sensor generate ozone profiles of the along-track limb scattered solar radiance. OMPS will make measurements used to generate estimates of TOC and OVP data. The JPSS algorithm is an extension of the TOMS Version 7 TOC algorithm (McPeters and Labow, 1996). It uses multiple triplets of measurements: one chosen for ozone sensitivity, a second chosen to give an estimate of cloud and surface reflectivity, and a third to estimate the variation of the reflectivity with wavelength. The triplets used in the final ozone estimates are selected so that the ozone-sensitive channels maintain sensitivity to the column as the solar zenith angle and column amount vary. OMPS will continue the daily observations performed by the current ozone instruments (SBUV/2 and TOMS), but with increased reliability. This data bank fulfills the U.S. obligation to watch the phenomenon of the ozone layer attenuation.

What we have described so far are the main instruments which have provided archived atmospheric ozone in the past, those which are currently providing this data and the instruments that are expected to be providing such data in the future. There are quite a lot of other instruments which have been only of ephemeral interest or which have only come on the scene recently and there are too many of them for us to describe them all. A comprehensive list is given in Table 3. Details of many of them will be found in several sections of chapter 2 of the book by Cracknell and Varotsos (2012); the general principles, however, have been covered by the instruments we have described here.

3. Inter-calibration and validation of atmospheric ozone data sets

We now move on to consider the results obtained from all the various instruments that we have described and the ground-based instruments which are used to validate and augment the satellite-derived data. As we have mentioned before, the satellite-flown instruments do not directly give the values of the required ozone concentrations. It is necessary to use an appropriate retrieval algorithm for each type of instrument and these various algorithms have to be validated, and if necessary refined, by comparison with data from ground-based instruments and ozone sondes. Also, if there are successive versions of instruments that are nominally the same, e.g. the several TOMS instruments, it is necessary to intercompare these datasets so that one consistent dataset can be constructed with no artifacts that arise from the transition from one instrument in the series to another one. Great importance is attached to the question of temporal changes in atmospheric ozone concentration. Therefore in looking for long-term trends in atmospheric ozone concentration one needs to have accurate long-term datasets. The problem of instrumental bias has been discussed throughout the whole history of ozone measurements, but it acquired a special importance after satellite observations started. Even at the early stages of satellite TOC observations very serious attention was paid to this problem. It is important, therefore, to reconcile the various different datasets obtained from ground-based measurements, ozonesondes and satellite data. The most reliable information about the spatial and temporal variation of the concentration of atmospheric ozone will be obtained by making use of all sources of data from ozonesondes, from ground-based instruments and from satellite-flown instruments. Not only is it necessary to intercalibrate different instruments that are operating at any given time, but to construct a long-term archive it is necessary to intercalibrate different versions of the same instrument, such as Dobson spectrometers, that have been operated over different periods. A very great deal of effort also has to be devoted to the intercalibration of ozone datasets derived from different ground-based instruments. The whole question of intercomparisons between different atmospheric ozone datasets is discussed at great length in chapter 3 of the book by Cracknell and Varotsos (2012). The results of ozone measurements come in four groups, (i) point measurements of TOC, (ii) point measurements of OVPs, (iii) the dynamics of atmospheric ozone and (iv) ozone depletion. We shall consider these separately. However, before we do this, we need to consider the question of inter-calibration of different data sources and the validation of ozone extraction algorithms.

Ozonesondes provide direct measurements of ozone concentrations as a function of height in the atmosphere (OVP) and therefore also provide estimates of TOC. But these measurements are only made at a few geographical locations and often also somewhat infrequently. Dobson spectrophotometers, and other ground-based instruments, have been used for many years and provide quite long-term datasets, but the question of the intercomparison of data from the various early laboratories is very important and it has been discussed by Staehelin et al. (2009). These laboratories between them provided data for more than 4 decades before the beginning of anthropogenic ozone depletion; they included Arosa (Switzerland), Oxford (United Kingdom), Lerwick (United Kingdom), Tromsø (Norway), Svalbard (Norway), Hohenpeissenberg (Germany) and Vigna di Valle (Italy). However, Dobson spectrophotometers and other ground-based instruments, like ozone sondes, only provide data for rather few geographical locations. On the other hand, as we have already noted, satellite-flown instruments provide frequent data over enormous areas, approaching global coverage. However, they do not provide direct measurements of ozone concentrations; algorithms have had to be developed to determine TOC and OVP from the spectral data recorded by these instruments. All these instruments need some algorithm to calculate the OVP or TOC from the actual measurements made by the instruments. We may loosely describe this as calibration of the remote sensing instrument. Then, once an algorithm has been established for a given instrument the results obtained using the algorithm need to be validated. Ozonesonde data and data from ground-based spectrophotometers therefore play a key role in algorithm development and validation. Ground-based measurements constitute a key component of the Global Ozone Network, both on their own account and by providing the ground truth for satellite-based instruments. However, the uneven geographical distribution of the existing ground-based network gives rise to a spatial sampling error, when attempts are made to determine the global distribution of ozone. The benefit of ground-based instruments is that it is easy to maintain them in good condition, while the benefit of satellite-based instruments is that they provide better temporal–spatial coverage and resolution.

One crucial parameter for the reliability of the ozone measurements is the selection of the ozone absorption coefficients for the in-field ozone observations. Usually, knowing the ozone absorption cross sections in the ultraviolet and infrared spectral range, with an accuracy of better than 1%, is of the utmost importance for atmospheric remote-sensing applications. For this reason, various ozone intensity intercomparisons and measurements have been published these last years (e.g Gratien et al. 2010). However, the corresponding results proved not to be consistent and thus have raised a controversial discussion in the community of atmospheric remote-sensing. In this context, Gratien et al. (2010) have reported a new laboratory intercomparison of the ozone absorption coefficients in the mid-infrared (10 μm) and ultraviolet (300−350 nm) spectral regions. This intercomparison confirmed that the IR and UV cross sections recommended in the literature are in disagreement of about 4%.

The validation of some of the satellite-flown instruments that we discussed in section 2 is discussed in some detail in chapter 3 of Cracknell and Varotsos (2012). There we consider the intercomparison of Dobson, ozonesonde and satellite data and – as an example – we consider in some detail the case of the Dobson spectrophotometer of the University of Athens. Scatter diagrams for TOMS data and Athens Dobson spectrophotometer data and for SBUV data and Athens Dobson spectrophotometer data from March to December 1991 are shown in Figures 4 and 5. Then, as an example of the validation of ozone data obtained from a satellite-flown instrument, we consider in some detail the validation of data from MIPAS which was flown on ENVISAT (Ridolfi et al. 2007). This involves data from a wide variety of other instruments. Finally, we present other intercomparisons between data sets obtained recently from various ozone monitoring systems. We also consider a few examples of what is involved in the intercalibration of different ozone data sources and the validation of satellite-derived ozone datasets. In particular we discuss the intercalibration of Dobson spectrophotometer data with the validation of TOMS, SBUV, OMI and SCIAMACHY data.

Corlett and Monks (2001) made a comparison of total column ozone data retrieved from the Global Ozone Monitoring Experiment (GOME), the TIROS Operational Vertical Sounder (TOVS), and the Total Ozone Mapping Spectrometer (TOMS) for 1996, 1997, 1998 and 1999. A statistical analysis and a spatial difference analysis were performed on a range of temporally and spatially averaged datasets. An analysis of globally averaged column ozone values showed a consistent offset with TOVS and TOMS values being consistently higher than GOME by ~10-13 DU averaged over the 4-yr period. A 4%-5% drift was noted between the years 1996/97 and 1998/99 in the magnitude of the offset. The drift was identified as an increased offset of + 25-30 DU between TOVS/TOMS and GOME occurring over latitudes above 70°N during 1998/99, and is a result of TOVS/TOMS ozone columns being higher during 1998/99 than in 1996/97. Seasonal and latitudinal trends were noted in the global differences. In particular, TOMS and TOVS ozone values are consistently higher than GOME in the southern hemisphere from 30°-90°S. TOVS and GOME ozone columns show good agreement between 20°S and 20°N, with TOMS values approximately 10-15 DU higher than both TOVS and GOME in the same region. All three sensors show reasonable agreement between 20° and 60°N. However, there is no agreement above 60°N, where TOVS columns are higher than TOMS columns that in turn are higher than GOME columns. Results from a spatial difference analysis indicated further differences between GOME and TOVS ozone values that were not obvious from the global or latitudinal analysis owing to cancellation effects, including an area over Indonesia where GOME columns are higher than TOVS columns.

4. Observations of TOC

TOC is characterised by significant temporal and spatial variability. At any given location TOC varies rapidly from day to day, see Figure 6. As far as temporal variability over the longer term is concerned, it consists of large periodic and aperiodic components (Cracknell and Varotsos 1994, 1995; Varotsos, 2005; Varotsos et al. 2005). Periodic variations have timescales ranging from day-to-day changes, through seasonal and annual variations to an 11-year solar cycle. Aperiodic variations include the irregular quasi-biennial oscillation (QBO) with a period of roughly 26-30 months, the irregular El Niño Southern Oscillation (ENSO) with a period of 4-7 years and other interannual signals (Varotsos 1989).

A very important feature of global TOC distribution is the strong latitudinal gradient with lower values over the equator and tropics and higher values over mid and high latitudes. This gradient is characterised by a well-pronounced annual cycle, reaching a maximum in spring and a minimum in autumn. The amplitude of this annual cycle is a function of latitude, with a maximum at about 60° north and south latitude. In the tropics seasonal variations are small, with ozone maxima in summer. Such a latitudinal distribution results from the relatively long lifetime (months to years) of ozone in the lower stratosphere and the Brewer-Dobson circulation that transports stratospheric ozone from the tropics toward the poles and downwards at high latitudes. Annual ozone variation was most clearly pronounced in the subpolar stratosphere of both hemispheres at the 1.5, 8.0 and 40 hPa levels, where the contribution of the annual components to total ozone variability exceeds 80%. The minimum amplitude of annual variations takes place in the mid and upper stratosphere near 10°N. The maximum amplitude of averaged semi-annual variations was observed near the 2-3 hPa levels in subpolar latitudes, decreasing towards mid-latitudes. Semi-annual ozone variations in the tropical middle stratosphere are comparable with, or even stronger than, the annual variations.

Superimposed in the annual variations which have, presumably, been present since pre-industrial times are the recent ozone depletion due to human activities. There are two aspects to ozone depletion arising from the escape of manufactured ozone destroying substances into the atmosphere. The first is a general depletion all over the the globe; the extent of this varies according to latitude, being quite small in the tropics rising to several percent per decade in high latitudes, see figures 7 and 8. Secondly, there is the dramatic "ozone hole" which appears in the Antarctic spring and persists for a few months and more recently it is appearing in the Arctic spring too (e.g. Varotsos 2002, 2003). This is a massive ozone loss, falling from around 400 DU to below 230 DU and taking several months to recover. We postpone the discussion of ozone depletion and of the longitudinal variation of TOC until section 6.