Remote sensing of the earth from space

  • Published on

  • View

  • Download


This article was downloaded by: [Universitaets und Landesbibliothek]On: 29 November 2013, At: 04:22Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: MortimerHouse, 37-41 Mortimer Street, London W1T 3JH, UKContemporary PhysicsPublication details, including instructions for authors and subscription information: sensing of the earth from spaceF. W. Taylor aa Atmospheric,Oceanic and Planetary Physics , Clarendon Laboratory, OxfordUniversity , EnglandPublished online: 20 Aug 2006.To cite this article: F. W. Taylor (1996) Remote sensing of the earth from space, Contemporary Physics, 37:5, 391-405,DOI: 10.1080/00107519608217544To link to this article: SCROLL DOWN FOR ARTICLETaylor & Francis makes every effort to ensure the accuracy of all the information (the Content)contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensorsmake no representations or warranties whatsoever as to the accuracy, completeness, or suitabilityfor any purpose of the Content. Any opinions and views expressed in this publication are the opinionsand views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy ofthe Content should not be relied upon and should be independently verified with primary sources ofinformation. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands,costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly orindirectly in connection with, in relation to or arising out of the use of the Content.This article may be used for research, teaching, and private study purposes. Any substantial orsystematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution inany form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at Physics. 1996, volume 37, number 5, pages 391-405 Remote sensing of the Earth from space F. W. Taylor Remote sensing is the popular name for the measurement ofphysical quantities at a distance, usually by quantitative spectroscopic methods. Its application to the study of the Earth's atmosphere und surface from satellites is a rapidly growingfield, with many applications, and new or improved techniques are constantly being developed. In this article we look at some of the basic principles and technical challenges involved, and at the beneJts which are accruing in a number of research areas in earth science, with particular emphasis on the use of remote sensing to understand global change. 1. Introduction The Earth, suspended in space, emits a varying stream of electromagnetic radiation in all directions at wavelengths ranging from the ultraviolet to the microwave. On the dark side of the planet, the photons which make up this energy flux have their origins in emission from the Earth's surface and atmosphere, and their intensity and wavelength distributions are complicated functions of the composition and physical state of the emitter. On the daylit side of the globe, there are additional photons of solar origin which have been reflected from the surface or from clouds, and scattered from aerosols or, at the shorter wavelengths, from the molecules of gas which make up the atmosphere. Overall about one third of the flux from the Sun, on average, is returned to space. The rest is absorbed, its energy reappearing later as long-wave infrared and microwave emission. Like the emitted flux, the backscattered and reflected photons also carry information about the planet as a physical system, and about the human environment below. The experimental techniques by which the rich informa- tion content of the outgoing photon flux is harvested and analysed has become known as remote sensing. The information is coded into the intensity as a function of wavelength, and so can be acquired by measuring pre-selected, and carefully calibrated, parts of the electromagnetic spectrum of the object of interest. This approach should not be confused with what we might call surveillance from space, using high-resolution photography and television, usually limited to the visible part of the spectrum. Such observations can be important scientifically as well as economically and strategically, and are often included under the remote sensing banner. However, remote sensing ~ Author '.F address: Atmospheric, Oceanic and Planetary Physics, Clarendon Laboratory, Oxford University, England techniques based on the use of imaging are usually of limited intrinsic interest to physicists and we will consider them only marginally in this article. On the other hand, where spectroscopy or radiometry is interpreted in terms of vertical projiles of quantities such as atmospheric temperature or ozone abundance, for example, the techni- que (sometimes called remote sounding in this case) involves advanced optical methods and sophisticated mathematical approaches to information retrieval, and we will be discussing these in some detail. The main purpose of this paper is to describe the measurement techniques and the interpretation methods of remote sensing, and to show how these can be applied to various important aspects of the behaviour of the Earth viewed as a physical system. Space techniques are the most interesting, because the combination of systemic, long-term global coverage and clever measurement and retrieval methods is what has made remote sensing such a valuable new tool in so many areas. Obviously, many of the methods can also be employed from ground- or aircraft-based platforms, and frequently they are. However, it is their use from satellites (and not only those orbiting the Earth) that has generated an information explosion which is revolutionizing our understanding of the behaviour of the planet and its surroundings. The most basic observations we can make are those of the energy budget of the whole Earth, from an investigation of the balance between the ingoing and outgoing fluxes of radiative energy which drive this formidable thermody- namic engine. Going beyond that, infrared spectroscopy and radiometry are well-known approaches to measuring the temperature and composition of atmospheric gases and solid surfaces, and are now routinely used from space platforms to observe the atmosphere, oceans and land surface. The stratospheric ozone layer, pollution in the 0010-7514/96 $12.00 0 1996 Taylor & Francis Ltd Downloaded by [Universitaets und Landesbibliothek] at 04:22 29 November 2013 392 F. W. Taylor Incoming Solar Radiation 0 5 10 15 20 25 WAVELENGTH/microns Ozone Water Vapour Carbon Dioxide Figure 1. Schematic mean spectra for the Sun and the Earth, approximated by black bodies at their effective temperatures of about 6000 K and 260 K respectively. The positions of the main absorption bands of the three most important minor constituents in the Earths atmosphere are indicated below. lower layers of the atmosphere, and the greenhouse gases which drive global warming, are all conveniently studied in this way. Mineralogy, vegetation (including agricultural products and some of the species in the sea), the icy cryosphere, and volcanism and its products are also candidates. Radar and lidar can be used to investigate land and ice topography, sea-state, and to infer wind fields. Operationally, remote sensing is used to observe the weather, snow cover, crop development and so on. In some cases, present remote sensing instruments and methods are too crude to replace older ways of obtaining data completely. For example, satellite measurements of atmo- spheric temperature and humidity near the surface are not good enough to supersede the traditional balloon-borne radiosonde for providing the input to weather-forecasting models, except of course where the latter is not available (which is, however, a good deal of the time over most of the globe). Current meteorological satellite sensors lack vertical resolution and can be confused by clouds. However, their use has already improved weather forecasting and further progress remains rapid, with advances being made which bode well for the future. 2. Radiation measurements and energy budgets 2.1. Introduction Figure 1 shows the spectrum of the Sun, approximated as a 6000K black body, together with that of the Earth, approximated as a 250 K black body. The integrated energy under the two curves is equal, except for the very small contribution from internal sources of heat within the Earth. Obviously, the details of these two spectra are interesting and important: the solar spectrum deviates significantly from a black body at many wavelengths which drive important processes on the Earth, for example, in the ultraviolet bands responsible for ozone production. Although the Earth, on average, radiates about the same total energy to space as a 250 K black body, the terrestrial spectrum is rich in molecular vibration-rotation lines which contain most of the information about the atmosphere, and continuum fluctuations which represent weak and/or complex atmospheric bands plus surface and cloud emissivity variations. The most pronounced of these spectral features are those caused by absorption and emission in the bands of the principal atmospheric minor constituents, especially water vapour and carbon dioxide, as indicated at the bottom of the figure. 2.2. Solur irrudiunce The integrated energy from the Sun fluctuates slightly with time, with an amplitude of the order of 0.1%, but the intensity at some wavelengths varies much more than this, as a function of conditions on the Sun [l]. All of these variations can be of considerable interest since the response of the Earth is difficult to predict and, if it can be observed, is quite informative about the Earth-Sun relationship. For example, it has long been discussed whether weather conditions at the surface are influenced by short-term solar variability or not, and the debate continues. It is certainly the case that the abundances of certain important trace constituents, such as nitric oxide, vary dramatically in the upper atmosphere with solar activity (cf. section 5.2). The best series of measurements of the total solar input are from the active cavity radiometer irradiance monitor (ACRIM) series of instruments, which have flown on a variety of spacecraft since the Solar Maximum Mission in 1980, and are planned to continue to the year 2000 and beyond. The basic principle is to maintain the temperature of a black body, in the form of a cavity, at a constant value by electrical heating while a shutter over the entrance aperture facing the Sun is opened and closed. The thermodynamic state of the cavity remains constant and a measurement of the current required to compensate for the removal of solar heating gives a straightforward estimate of the latter. The more recent versions of this type of instrument [2] include many detailed refinements and achieve very high accuracies. 2.3. The solar spectrum The latest measurements of the solar spectrum from space have been made by instruments known as SUSIM (Solar Ultraviolet Spectral Irradiance Monitor [3]) and SOLSTICE (Solar Stellar Irradiance Comparison Experiment [4]) flying Downloaded by [Universitaets und Landesbibliothek] at 04:22 29 November 2013 Remote sensing of the Earth from space 393 on the Upper Atmosphere Research Satellite (UARS) in the last few years since 1991. The spectral region of main interest is the ultraviolet, since the variability is greatest here (of the order of 10% at some wavelengths) and these are the more energetic photons which drive photochemical reactions in the atmosphere. SUSIM and SOLSTICE are grating spectro- meters which, between them, cover the range from 119 to 420nm. Since the absolute intensity as a function of wavelength is required, calibration is a major issue; SUSIM uses a series of standard lamps, while SOLSTlCE has an innovative approach by which the input aperture is variable, allowing the observation of a number of standard UV stars. The results from instruments of this type [5] show, as might be expected, large variations, especially in the Lyman- a band, with the same frequency as the 27-day solar rotation and with sunspot activity, although other, smaller, periodic and episodic variations are also seen. The 27-day cycle produces a corresponding signal in the atmospheric tempera- ture at very high levels, and sunspot eruptions on the Sun produce particle fluxes which have been observed to strongly modulate the thermospheric nitric oxide abun- dance, over localized regions and for a short time, as already noted. It is not yet clear whether there is any measurable effect of solar fluctuations on the lower atmosphere. 2.4. Figure 2 shows schematically the principal ways in which the energy from the Sun is redistributed within the climate system, and finally re-emerges as long-wave infrared radiation. The individual components are variable of course, on a wide range of distance and time scales. For example, the difference between incoming and outgoing radiative energy varies markedly as a function of latitude because of the transport of heat, generally in a poleward direction, by the atmosphere and the oceans, and this transport has a seasonal dependence which is not the same every year. Climate models, like those now being extensively The energy balance of the Earth m Emitted to space Waqs and Eddies Reflected from 1 Absorbed 1 cloudsand uVvlM Adv&im of latent and sensible heat DY atmosphere I Emitted by atmosphere Convection I I I 1 by atmosphere 1 / Reflected from surface Absorbed by surface EARTHS SURFACE Emitted by surface Figure 2. Some important atmospheric processes (radiative in plain script, chemical in capitals, and dynamical in italics) involved in converting the sunfall into outgoing thermal radiation. employed to attempt medium-to-long term forecasts and to assess the likely progress of global warming, need to have the correct fluxes of reflected solar and emitted thermal radiation at the upper boundary as a necessary, if not sufficient, condition for realistic predictions. The extent and microphysical nature of cloud cover is one of the most important and unpredictable variables affecting these [6]. The current state-of-the-art in making the relevant observations is represented by the multi-satellite Earth Radiation Budget Experiment (ERBE). The measurements are simple in principle but actually very difficult in practice, mainly because the complete range of wavelengths has to be sampled at every outgoing zenith angle all over the Earth. The ERBE scanners [A have three broad spectral channels, covering the wavelength ranges from 0.2 to 5.0pm (i.e. effectively the entire solar spectrum), 5 to 50 pm (most of the Earths thermal infrared spectrum) and 0.2 to 50 pm (both). It is impossible to make the spectral response flat, i.e. independent of wavelength, within these broad ranges and the solar component, in particular, is difficult to calibrate, hence the redundancy in the selection of channels, use of overlapping coverage from multiple spacecraft and the painstaking use of reflective and emissive reference targets and empirical algorithms to relate the measurable quantities to the fluxes of interest. The angular dependence of the flux also has to be modelled, since the directional properties of the radiation fields are bound to be significantly undersampled by any realistic number of sensors. Overall, the uncertainty in each of the two components has been estimated [8] at about f 5 Wm-2, or about 2%. By compar- ison, the so-called cloud radiative forcing, a useful quantity in climate studies which is defined as the flux averaged over all conditions minus the clear-sky flux, is typically about 10 to 20% of the total 191. 3. Atmospheric temperature co,mposition and winds 3.1. Introduction The basic atmospheric variables leave their signature on the outgoing radiation field. From retrieved values for these the processes going on in the climate system, such as those mentioned in figure 2 can be investigated, at least implicitly. The atmospheric and surface fields of temperature, pressure, composition, and wind, control the photon 0ux directly through their effect on the intensity, shape and position of spectral lines. Most, although by no means all, of the detailed relationships involved are well understood in general from theoretical quantum mechanics and labora- tory studies. The line intensities in the lower atmosphere, for example, are readily understood in terms of Boltzmann statistics with the assumption of local thermodynamic equilibrium while the widths and shapes are mainly the result of collisional broadening. The line positions can be Downloaded by [Universitaets und Landesbibliothek] at 04:22 29 November 2013 394 F. W. Taylor derived from semi-classical models, matched to laboratory spectra to obtain the rotational and vibrational constants. Extracting the geophysical information we want from spectral radiance measurements is a process which has been likened to unscrambling an egg; formally it is known as inversion or the retrieval problem. Many different methods have been devised [lo] to recover all of the available information and no less (or no more!). Historically, the retrieval of temperature profiles was the first to be addressed, and remains one of the most important applica- tions of remote sensing. 3.2. Next to net energy flux, temperature is the most fundamental physical variable to which remote sensing offers ready access. The principle of remote temperature sounding is simply to measure the intensity of the thermal emission from the target, and to convert this to temperature by assuming Planck's function for the emission from a black body as the source function in the radiative transfer equation [l 11. When viewing the atmosphere vertically downwards, the emission can be thought of as originating from a series of layers, each of which contributes to the measured radiance (figure 3). The weighting of these contributions varies with the wavelength at which the measurement is made; layers nearer the top of the atmosphere contribute more in spectral regions where the atmosphere is more strongly absorbing, for example in a vibration-rotation band of an abundant infrared active species such as carbon dioxide. It is straightforward to show [ 101 that the weighting function is given by the derivative with respect to height of the atmospheric transmission function, which can by calculated on a computer, leaving the temperature of each layer as the principal unknown. Solving for the temperature profile, given a set of radiance measurements in a number of spectral intervals corresponding to different transmission functions, is an inverse problem of some complexity. The weighting func- tions are generally quite broad, and sets in which they strongly overlap are normally used to try to increase the (otherwise rather limited) information content of the measurements (figure 4). Statistical information on the probability of a given solution can be added, for example by using the 'first guess' field from a forecasting model, and vertical resolution can be traded off against the error in the retrieved profile at any height. Broadly speaking, nadir (downward-viewing) profiles obtained in this way obtain an accuracy of one or two kelvin on a vertical scale of around 10 to 15 km. The calculation of the weighting functions requires a knowledge of the composition of the atmosphere, at least the abundances of the principal absorbers in the spectral intervals chosen for the sounding instrument. This is one of Basic principles of temperature sounding ' 1 1 1 1 1 1 1 1 1 band Temperature Wavelength Mirror Condensing Spectral optics Filter 1 I 0 From Calibration Target EARTH Figure 3. (a) Illustrating the principle of remote temperature sounding (see text for discussion). Each vertical arrow corre- sponds to the flux of emitted infrared radiation in a different spectral interval within a C 0 2 absorption band. (b) Schematic diagram of one wavelength channel of a radiometer for remote temperature sounding. Absolute calibration of the radiances is essential; an on-board radiometric target which can be moved into the beam is shown. the reasons why wavelengths dominated by CO,, or in the microwave 02, are chosen in general, since these gases are nearly uniformly mixed. However, if a simultaneous set of observations is made at wavelengths dominated by an absorber of unknown concentration, for example water vapour, then a second retrieval can be performed in which the profile of the absorber amount is the unknown. This is the basis for composition studies by remote sensing, which are now commonplace and often quite refined, as the examples below will show. 3.3. Practical temperature sounding Modern atmospheric temperature sounding treatments Downloaded by [Universitaets und Landesbibliothek] at 04:22 29 November 2013 Remote .sensing of the Earth from space 395 -50 - E -40 6 - a, c -30 d 5 .- - .- 2 X 2 -20 g U ir 0 0. I 0.2 Weighting Function Figure 4. Weighting functions used for vertical temperature sounding by a six-channel radiometer on the satellite Nimbus 4. for the stratosphere and above mostly use limb scanning. This means that, instead of viewing downwards, they look sideways at the limb of the Earth and observe the atmosphere through a telescope. The main advantage of this is that the vertical resolution can be improved by about an order of magnitude, to around 1 or 2 km. Apart from the fact that the technique is not viable in the lower atmosphere, where clouds and turbidity interfere, the main disadvantage is that the horizontal resolution is reduced along the line of sight. (This is not so much of a problem as it first appears, however, since the smaller footprints of a nadir sounder are usually separated by the distance between successive orbital tracks, and in that case, the limb mode actually offers superior sampling of the gaps.) Also, the instruments tend to become huge: the ISAMS (Improved Stratospheric And Mesospheric Sounder, an infrared spectroradiometer [ 121) and MLS (Microwave Limb Sounder [13]) instruments on NASAs recent Upper Atmosphere Research Satellite, for example, have dimensions measured in metres and masses of hundreds of kilograms, in sharp contrast to the shoebox- sized instruments of the early years of remote sensing, in the 1960s. The MLS operates at frequencies in the GHz range, and requires an antenna 2.5 m across to avoid the diffraction limit on the beam width necessary to resolve the limb. ISAMS requires only a 25cm diameter mirror, but uses cooled detectors and optics to get high sensitivity. Cooling instruments in space is a fine art; which in this case was addressed using miniature mechanical refrigerators of an advanced design which avoided all rubbing or rolling parts to give high efficiency and a long lifetime. Here is an example of how getting good data on the Earth can involve mastering the application of other physics, in this case the thermodynamics of the Stirling cycle. Figure 5 shows an example of a temperature profile retrieved from ISAMS infrared radiances, measured in the y band of C 0 2 near 15pm wavelength. The resulting profile is compared in this figure with one measured at approximately the same time and place by a ground-based, upward-looking optical radar (lidar), which operates on the principle of measuring refractive index profiles, which can be related to density and therefore temperature in the analysis. The lidar profiles have higher vertical resolu- tion (around loom, compared to about 2 km for ISAMS), and more easily reveal the presence of waves and other interesting structures. On the other hand, they are obviously rooted to one spot, and cannot reveal the global extent of wave activity, or even the full characteristics in space and time of a single event, as fully as satellite sensors can. The combination of the two is particularly powerful and has been used, for example, to examine the global distribution of regions of strong gravity wave activity [14]. Satellite temperature maps have also proved valuable for studying the details of dynamical phenomena, like the stratospheric sudden warming-an event that tends to occur in the northern winter and which changes the circulation of the middle atmosphere dramatically in the course of a few days, with the temperature over the poles increasing by as much as a hundred degrees in a major event [15]. 3.4. Infrared and microwave instruments also measure atmo- spheric composition. The principle is basically the same, as we have seen: the intensity of the emitted thermal flux depends on the temperature and the abundance of the emitter (i.e. on the probability per molecule of emission of a photon of given wavelength, and on the number of emitting molecules present). So, once the temperature has been determined from the emission by a molecule of known abundance such as C 0 2 , measurements of the same air parcel at a different wavelength in an emission band of a different species can yield information on the concentration of the second species. ISAMS used this approach to map the oxides of nitrogen, N20, NO, NO2, and N205, following the discovery that some members of this family act as catalysts to accelerate the destruction of ozone, posing the now- famous threat to the Earths shield against excessive amounts of solar ultraviolet radiation. Most members of the NOx family, as it is called, are very reactive and the amounts available to interact with ozone vary dramatically with space and time. Figure 6 shows some ISAMS measurements of nitrogen dioxide, as one example of the complicated patterns which these important trace constituents adopt, and which cannot be followed effectively Remote sounding of atmospheric composition Downloaded by [Universitaets und Landesbibliothek] at 04:22 29 November 2013 396 F. W. Taylor 11 ' I ' ' I ' I i ' ' ' ' I ' ' ' ' by traditional measurements from aircraft or balloons. The region at middle to high latitudes shows a very low abundance of the species, with a sharp boundary between this and the more normal values at lower latitudes. Studies have shown [16] that the NO2 has been removed by conversion to 'reservoir' species such as dinitrogen pentoxide and nitric acid in the long, dark polar night. An under- standing of the chemistry going on in these regions is crucial to predicting the fate of the ozone layer at all latitudes. Earlier, and cruder, instruments for measuring composition, especially ozone itself, work on the ultraviolet backscattering principle. Here, the differential absorption by ozone is measured at several UV wavelengths using solar radiation backscattered from the Earth as the source. The vertical resolution is poor, and the technique of course cannot be used at night, particularly at the winter pole where some of the most exotic chemistry takes place (including that which leads to the springtime ozone hole over the south pole). However, instruments of this kind such as BUV (Back- scatter Ultra Violet sounder) and TOMS (Total Ozone Mapping Spectrometer), offer sensitive monitoring of ozone with relatively simple hardware and have been in use since the earliest days of remote sensing from space (were it not 0.001 0.010 0.100 P n I 1.000 I o.oO0 Figure 5. Satellite (ISAMS) and ground-based optical radar (lidar) ver- tical temperature profiles. The shaded region is the lidar profile and the solid line ISAMS; the *lo uncertainty is indicated in both cases. 100.000 for a data processing error, TOMS would have detected the Antarctic ozone hole sooner than the ground-based Dobson spectrometers of the British Antarctic Survey). They are particularly valuable for monitoring world-wide trends in total ozone as a function of latitude. 3.5. Clouds and aerosols Clouds are the wild card of the Earth's climate system; they turn up in an inevitable but largely unpredictable way, and have a large impact. Quite apart from their role in radiative and latent heat transfer, precipitation and so forth, they interfere with remote sensing of the lower atmosphere and surface from space and place heavy demands on the methods used for the retrieval of information from these levels. For some measurements, the problem can be avoided by working at microwave and radar wavelengths, at which most clouds are transparent. Often, however, either this approach does not work for the data we need (spectroscopic observations of composition, for example, mostly have to be made in the infrared where their absorption bands lie), or else we want to obtain remotely sensed data on the cloud properties themselves. Downloaded by [Universitaets und Landesbibliothek] at 04:22 29 November 2013 Remote sensing of the Earth from space 397 0 20 40 60 80 Lat it ude/deg Figure 6. A cross-section of the abundance of nitrogen dioxide in the stratosphere from the equator to the North Pole, made by combining 45 succe59ive ISAMS profiles. The sharp drop towards high latitudes is caused by the conversion of NOz to NzOS and HN03 in the polar night. Clouds are useful, if limited, tracers of winds, as discussed further in the next section. Imaging provides valuable data on their coverage and movement, but little information about the height to which the observed winds relate and almost nothing on the cloud microstructure and optical properties. In principle, particle size and composition can be retrieved from multispectral observations, and by studying the reflection and emission properties of clouds at different angles. Tropospheric clouds are mostly too inhomogeneous for much progress to have been made with this so far, but good results have been obtained on stratospheric aerosols [17]. These are sulphuric acid droplets, mostly of anthro- pogenic and volcanic origin, which are dense enough to affect the radiative balance of the Earth significantly. They also provide surfaces on which heterogeneous chemical reactions can occur, further complicating the ozone deple- tion problem. In fact, it is now known that the Antarctic ozone hole is caused by a special kind of cloud called a polar stratospheric cloud (PSC) which forms only at very low temperatures (below about 195 K) and which contains frozen nitric acid (figure 7). Rather alarmingly, PSCs appear also over the northern polar regions, where their effect on the ozone layer is potentially more harmful to humans. Figure 8 shows such an event over Scotland and the North Sea, observed in thermal emission at 12.1 pm wavelength by the ISAMS instrument [18]. A huge tongue of ozone- threatening chloride oxide was observed at the same time at 1-5mm wavelength by the UARS Microwave Limb Sounder. In this case, the event was sufficiently localized and short-lived that the net effect on the ozone layer was relatively small, but the phenomenon is obviously of practical as well as scientific interest, especially if pollution continues to increase, and mean temperatures to fall, in the stratosphere, making such events more frequent and of longer duration. 3.6. Doppler wind measurements The movement of the atmosphere can be calculated approximately from measurements of temperature gradients, using the primitive equations of fluid dynamics, or by tracking clouds and other tracers such as water vapour. More direct wind measurements are also becoming possible by measuring the Doppler shift of spectral lines emitted from atmospheric gases, although at present the technique is only applicable to the middle and upper atmosphere where Downloaded by [Universitaets und Landesbibliothek] at 04:22 29 November 2013 398 F. W. Tuylor Figure 7. Aerosol optical extinction measurements over Antarctica in the Southern Hemisphere winter. The intense concentration over the pole is a large, long-lived polar stratospheric cloud, which plays a key role in producing the ozone hole in the Austral spring. The less dense cloud over the equator is the remnant of the plume from the eruption of Mt Pinatubo more than a year earlier. the lines are narrow, and easier to observe at the limb. Even this is something of a triumph, considering that accuracies of a few metres per second are being achieved from a spacecraft which itself is travelling, relative to the surface, at something like 7 km s-l. The secret is to be mounted on a very stable spacecraft, with accurate pointing. A simple calculation reveals that one degree of uncertainty in the pointing vector of the instrument introduces an uncertainty of around 50m s-l in the line of sight velocity, about the same as the mean magnitude of stratospheric winds. To obtain a useful accuracy of 1 ms-, pointing knowledge of around an arc minute is required. The Nimbus meteorological satellites of the 1960s and 70s wobbled in their orbits by several degrees, while modern platforms like UARS are orders of magnitude better and have demonstrated useful wind measurements with an interferometric instrument, WINDII, for the thermosphere, and a Fabry-Perot based spectrometer, HRDI, for the middle atmosphere. The real prize in the wind-measuring stakes is still to be won. Velocity fields in the troposphere, the lowest 10 or 15km of the atmosphere, would make a substantial difference to weather forecasting [19]. It is to this end we are most likely to see active remote sensing widely deployed for the first time, in the form of a 20 J C02 laser sending pulses of 10 pm radiation towards the surface and receiving the reflected power some 5ms later via a one-metre telescope. The returned signal will include a surface reflected component, plus peaks corresponding to layers of cloud and aerosol. The signals can be spectrally analysed to recover the Doppler shifts imposed by the motion of the target layers, and their relative heights above the surfacc calculated by timing the returns. Such instruments have been designed and extensively studied, and deployed in prototype form on the ground and on aircraft. 4. Properties of the land, oceans and icecaps Compared to the atmosphere, radiometry and spectroscopy of the surface of the Earth from space is a relatively easy measurement task, because here we are dealing with a two, rather than three-dimensional object (or three rather than four, if you are one of those who likes to think of wavelength as a dimension). Also, the target is relatively warm and bright, providing strong signals. The interpreta- tion of the signals, however, is another matter. Surface temperature measurement is simplest over the ocean, because the infrared emissivity is quite well known, and is close to unity. However, the surface temperature contrasts being sought are quite subtle, measured in fractions of a degree kelvin, and can be masked by atmospheric effects, for example very thin clouds or patches of aerosol, which can be quite undetectable in the visible. The Along-Track Scanning Radiometer (ATSR) overcomes this problem by viewing the Downloaded by [Universitaets und Landesbibliothek] at 04:22 29 November 2013 a E Z d -? m 0 0 0 0 ci m D E z a 73 cb 0 03 0 0 cu CO i a 73 6 0 r. 0 0 cu m Remote sensing of the Earth from space a E cd 43 -;f cu Z d -? 0, 0 . . 07 0 0 cu m D E ci3 3 cu o! Z a -? co 0 03 0 0 cu 0, 11 E co 3 cu ? Z Q -? b 0 r-. 0 0 cu 0 399 Downloaded by [Universitaets und Landesbibliothek] at 04:22 29 November 2013 400 F. W. Taylor surface at two angles and analysing the difference between those views, which would be nearly zero if the atmosphere had no effect. Over land, one is more likely to assume a temperature from other information and examine the observed differ- ences in emissivity. By day, the reflectance properties at visible and near-infrared wavelengths show differences such as that between vegetation and bare soil, and careful choice of wavelength filters (for example, isolating the chlorophyll bands) enhances sensitivity to vegetation types and different stages of growth. The problem mainly is that multispectral data are sensitive to so many factors, including relatively uninteresting parameters such as fractional cover, shadow- ing and dampness, that detecting and classifying change is often easier than identifying the basic cause. Consequently, although remote sensing of the Earths land surface is a major industry, the techniques are more a matter of empirical models, and acquired skills such as image processing, than basic physics. Radar is a powerful tool to study the surface elevation and roughness. Apart from the obvious and important geographic applications to cartography and terrain char- acterization, this has applications in oceanography as a means of probing sub-surface currents and surface winds, and in climate studies for monitoring changes in the polar caps, of which more below. The basis of the technique for studying currents is altimetry, i.e. measurements of the time- of-flight of a radar pulse emitted from the satellite [20]. The pressure forces associated with major currents (and with tidal forces, and surface gravity waves) manifest themselves as surface height variations, relative to the geoid, of approximately 5 1 m. The latest state-of-the-art in satellite altimetry achieves measurements of the position of the surface relative to the satellite of perhaps 10 cm. For this to be useful in oceanography it is necessary to have knowledge of the geoid (the mean figure of the Earth as manifest by its gravitational field) itself to the same accuracy or better, again from satellite measurements. Polar ice sheet measure- ment by altimetry is less dependent on the geoid and more by the size of the measurement footprint. Small or slow changes in the shape and thickness of the ice can be masked by the effect of the variance which exists in the terrain within the footprint and which limits the range precision to around 30cm for smooth but undulating ice sheets, and a few metres over rough terrain. Improvements are expected from the use of laser altimeters, which have a much smaller footprint, especially if retro-reflectors can be placed at key points on the surface, for example on an icesheet or glacier whose thickness and movement is to be followed. The precision in that case can be centimetres or even millimetres. Devices of this type have already been flown to the Moon and Mars, and are in preparation for inclusion in NASAs Earth Observing System of satellites early in the next century. 5. Geophysical research using remotely sensed data So far we have considered what can be measured and, in outline, how it is done and the underlying principles. This section reviews some of the applications, particularly those which might be described as basic geophysical research. In being thus selective, it is worth emphasizing again that remote sensing is a multi-billion pound industry in which land usage, cartography, surveillance, and crop monitoring, to name but a few, are increasingly feeling the benefits. To give an example-more from Geography than Physics, because the methods employed are adaptive and descriptive rather than predictive-the worlds rice crop, which is a staple food for more than two billion people, can be followed through its growth and reproduction cycle by mapping from space using radar. Radar is not sensitive to cloud cover and this approach promises to replace the present highly erratic methods which are used to assess acreage and expected seasonal yield. The signals which correspond to different states of the crop are learned by comparison with ground truth, rather than from funda- mental principles, although quasi-physical modelling is a growing research field. 5.1. Oceanography In climate research, ocean/atmosphere coupling is a major issue, because the two fluid masses are coupled together and, in particular, the oceans exert a major influence on the behaviour of the atmosphere on time scales of seasons to decades. The coupled system exhibits a number of cycles which may be predictable, the most famous being the El Nino Southern Oscillation, known as ENSO. Modelling and analysing these requires data, and several of the most important parameters are readily available by remote sensing. For example, sea surface temperature controls the rate of heating of the atmosphere by the ocean, which of course has a much greater heat capacity (about 1200 times greater, in fact). Global measurements allow the sensible heat flux and evaporation across the interface to be calculated and circulation features, such as El Nino or the Gulf Stream, to be monitored as they change from day to day. Ocean currents are driven by the wind near the surface, which can be inferred from radar measurements of sea surface roughness, allowing the exchange of momentum to be calculated. 5.2. Meteorology In an ideal world, weather forecasting would now be done by gathering information the current state of the atmos- phere from satellite sensors of the type described above, and then solving the time-dependent equations of fluid motion, etc., on a large computer to obtain the predicted state some Downloaded by [Universitaets und Landesbibliothek] at 04:22 29 November 2013 Remote sensing of the Eurth from spclce 40 1 days hence. To a considerable extent, this is what is done, but the procedure is still much hampered by the limited quality of the data available (as well as), of course, by the complexity and inherently chaotic nature of the atmosphere itself). The temperature and composition (primarily humidity) profiles have too little vertical resolution, wind, precipitation and cloud microphysical data are almost non existent. We have hinted, above and below, at ways in which this might be improved in the foreseeable future, until predictability itself becomes the dominant problem (as it already is in long-term, or climate, forecasting). The main frontiers at present are in the areas of data assimilation directly into forecasting models, without going through the intermediate stage of retrieving profiles, and the development of more advanced microwave sounders which are less vulnerable to the cloud problem. In the longer term, active sounding using space-borne radars to detect precipitation, lidars to measure winds, and laser spectrometers for composition, can all be anticipated. 5.3. Stratospheric chemistry The stratosphere shot to fame in the early 1970s with the realization that certain oxides of nitrogen of largely anthropogenic origin were catalysing processes which threatened to reduce the thickness of the ozone layer. The discovery in the mid 1980s that the Antarctic ozone layer was being removed almost entirely in the spring lent credence to the theories and led to a massive series of studies in which remote sensing played an important part. Global measurements of the offending nitrogen oxides, and some of the even more malicious chloride compounds, are becoming available from missions like UARS and are helping to clarify what is happening. As usual, the problem becomes more complex with study, and is now known to involve heterogeneous chemistry on aerosols in an impor- tant way. This in turn highlights our current ignorance of the origin, abundance, properties and evolution of the aerosols themselves, which are of many different types and highly variable in size, surface areas and sulphuric acid content. Progress on understanding these was helped by the coincidental occurrence of the centurys largest volcanic eruption, Mt Pinatubo, a few months before the launch of a comprehensive set of instruments on the Upper Atmosphere Research Satellite. 5.4. Tropospheric chemistry The well-mixed layer near the surface of the Earth is experiencing a gradual change due to the accumulation of minor constituents such as carbon dioxide, carbon mon- oxide, chlorofluorocarbons, nitrous oxide, and methane. These are radiatively active gases which affect the energy balance at the surface, in part by radiating back absorbed radiative energy (the greenhouse effect), leading to a general warming. More reactive gases such as ozone are also changing, and have a profound effect on the chemical mix which affects not only the greenhouse but also leads to phenomena such as acid rain. A pioneering remote sensing experiment called Measure- ment of Atmospheric Pollution from Satellites, or MAPS, used an infrared gas-correlation technique to measure tropospheric carbon monoxide globally for the first time in 1976. The result was a revelation-high concentrations were found, not over the urban industrial centres of North America and Europe, as expected, but downwind of the forests of South America and Africa. Since then, the importance of biomass burning has been in the headlines. The global budgets and sources and sinks of CO and all the other important minor constituents in the troposphere remain to be observed and understood by experiments planned but not yet performed. 5.5. Climate change The Earths climate has changed dramatically in the past, and it is changing now. Action to ameliorate or adapt to these changes is hampered by ignorance; there are even those who believe that the Earth is alive and will clean itself up. A theme of this review has been that remote sensing has a big part to play in providing the global view and the detailed quantitative data which is allowing the problem to be approached scientifically. Firstly, it is necessary to understand the details of energy transport, in the atmosphere and ocean by radiation and transport. Models of the energy balance of the Earth which incorporate parametrized versions of the key processes need to match the observed outgoing flux of thermal radiation to space. The radiative and dynamical processes themselves can be elucidated by the type of measurement we have described. As a single example, let us look more closely at the question of cloud/radiation forcing. The most important properties of clouds, from the point of view of their effect on the global climate, are their height and their optical thickness. Both can be measured from space, although there is a difficult ambiguity for clouds of medium thickness, when sensed in the infrared (visible observations are not good at night, and microwaves tend to pass through clouds unattenuated). Basically, one cannot tell the difference, in a single measurement, between a thin, high (and therefore cold) cloud, and one which is thicker and warmer. In principal, multispectral or stereoscopic observations allow resolution of this problem; in practice, clouds are so inhomogeneous that they require more parameters to describe them than can ever be measured. It is hard to see how the problem will ever be solved completely. Satellite measurements are the best way of detecting and monitoring global temperature changes, as anyone who has Downloaded by [Universitaets und Landesbibliothek] at 04:22 29 November 2013 402 F. W. Taylor ' ' Lowertroposphere ' ' ' ' I .o , 0 Decadal trend = +O.O7'C Figure 9. Temperature trends from 1979 to 1991 in the lower stratosphere (about 15 km altitude) and the lower troposphere (about 5 km altitude) as measured by a microwave temperature sounder 1171. tried to piece together the global picture from a varied mass of local samples obtained in various ways, many of them unreliable, can testify. Even such simple concepts as the global mean surface temperature arc notoriously difficult to relate to actual measurements, even from space where cxtensive (if not complete) and consistent (if not necessarily accurate) coverage with a single instrument, or set of similar instruments, is possible. Microwave data is particularly suitable for monitoring temperature trends. because the flux is insensitive to the presence of clouds, which are the most variable feature. Eight versions of an instrument called the Microwave Sounding Unit or MSU have flown on the TIROS and NOAA series of operational satellites since late 1978, and from the data, temperature trends for the lower troposphere and the lower stratosphere have been produced [21] (figure 9). Interestingly. the trend during the 12 years rrom 1979 to 1991 is an insignificant 0.07 K decade^ I , much smaller than the prediction of around 0.3 K decade-' from the best climate models. It is beyond the scope of this article to discuss the reasons for this discrepancy. but it probably has to do with thc failure of the models (because o f a lack of understanding arising from a paucity of measurements) to include atmospheric aerosols properly. 5.6. As already noted, surface remote sensing is important, in that land use and crop type, and a whole range of other geopolitical variables, can be extensively monitored by surveys from space. In terms of methods employing basic physics, the most interesting possibilities are those which might use advanced spectral methods to address the mineralogy and physical state of the surface. with a view to studying desertification. for example, or even prospecting for oil or other valuable deposits. Little progress has been made t6 date. howcvcr. Among the The l m d surjiicc and wgclotiuti reasons are the contaminating and obscuring effects of the atmosphere, the relatively bland spectra of most minerals, the fact that material of different kinds is always mixed and layered in any scene, and the fact that the land, unlike the atmosphere, is relatively static and more amenable to in situ investigation. Remote sensing of land surfaces is currently more important where no alternative is available, for example on planetary bodies, especially relatively airless ones like the Moon, Mercury, Mars and the giant satellites of the outer planets (cf. section 6) . 5.7. During the course of the year, the snow and ice cover of the northern hemisphere varies by more than a factor of two. These changes are difficult to monitor from the ground, but can be observed from space, and microwave techniques allow not only the coverage but also the depth of the snow field to be obtained. The dielectric constants of snow, water, and ice are so different that the effect on microwave cmission is easily detected. Dry snow attenuates the emission from the surface underneath, mainly by scattering rather than absorption, and allows the depth of the layer to be estimated. Over the ocean, the coverage (although not the depth) of sea ice can be mapped by making use of the contrast between the emisskities of ice and water (at a wavelength of 1.55 cm, for example, the emissivity of sea water is 0.44. whereas that of ice is approximately twice this). 90% of the Earth's ice is in the Antarctic ice sheet, and if it melted the mean sea level would rise by 70m all over the planet. This, of course, I S not expected to happen in the foreseeable future, but the mere fact that it could, that the ice cap is not a static entity but is changing, and that present knowledge is not accurate enough to say whether it is growing or shrinking, all point to the value of mapping its morphology. Space-borne radars are now offering altimctry to accuracies of around one centimetre, and early estimates [22] of the thickening of the Greenland ice sheet have suggested an increase of as much as 1.6m in places in the period from 1978 to 1086. The cryoJpheve: snow and ice 6. Other Planets Let us not forget that physical systems highly analogous to that which concerns us so much on the Earth are present on Mars and Venus, and elsewhere. The giant hurricanes we strive to understand have giant cousins on Jupiter and Neptune. The processes are the same or similar, the physical laws identical; only the boundary conditions have been changed. Comparing another planet to the Earth is a kind of experiment, almost the only one we have in the physical world where altering the Earth is difficult and inadvisable (one such inadvertent experiment over Antarctica has already been discussed). All the major bodies of the solar Downloaded by [Universitaets und Landesbibliothek] at 04:22 29 November 2013 Remote sensing o f the Earth f rom space 403 system except Pluto have now been visited and we know in general terms what they are like. This is not the same as saying we understand how they behave, much less that we understand why they behave the way they do. Why is Venus so hot and so dry? How are its thick, sulphurous cloud layers created and maintained? Is the surface of Venus shaped by plate tectonics like the Earths? How was Mars able to have a thick atmosphere, and free water on its surface, in the past, when this would be impossible under the conditions we observe today? What mechanism maintains the Great Red Spot on Jupiter, and what is its relationship to the other Earth-sized dark spot on Neptune? To what extent do these other-worldly storms resemble terrestrial humcanes? These and many others are all questions which touch our understanding of the Earth more than a little. The 1; tonne, 6 m high Galileo spacecraft has recently (December 1995) been inserted into orbit around Jupiter. Its instruments are designed to determine the composition and geology of the surfaces of the large satellites, first pictured by Voyager ten years ago, to measure the abundances of elements and compounds in the plants atmosphere, and to monitor its meteorological activity over a two-year period. Another big. nuclear-powered, probe-carrying orbiter, Cassini, will embark on a seven-year voyage to Saturn in 1997. Saturn and its rings will receive intensive study as the spacecraft orbits, but the real prize this time is its intriguing satellite Titan. Titan is large, not much smaller than Mars, but boasting a thicker atmosphere even than that of the Earth with a surface pressure of around two atmospheres, and with the same primary constituent, nitrogen. Titan is ten times more distant from the Sun than Earth and so very cold. It has no free oxygen-nobody expects to find life-but a lot of methane and other hydrocarbons, some of which form oily clouds and probably rain, so that an Ocean of light oil on the surface is a real possibility. Most of what we know about Titan suggests an environment with many things in common with the primeval Earth, an interesting laboratory into which to drop the European-made probe (now called Huygens after the great Dutch astronomer of the seventeenth century). The Mars Surveyor series will permit new, high-resolution views of the surface of the red planet and a complete mapping of its atmosphere for the first time, characterizing the present climate and its circulation, maybe offering some real insight at last into the vast dust storms that cover the whole plant every few years. An advanced atmospheric sounder, using principles developed at Oxford and incor- porating our hardware, will scan the atmosphere vertically to measure dust and humidity profiles, and map temperatures structure from which winds are derived using assimilation models originally developed at the Met Office for weather forecasting. In addition to studying Mars, this approach will constitute a powerful test of the physical basis on which terrestrial weather and climate is currently predicted. 7. Future prospects for Earth observation Remote sensing is a young field and is continually being refined and improved. Larger, more stable spacecraft, for example, offer platforms for bigger, more sensitive instruments which can map increasingly finer detail on the surface and in the atmosphere. In additional to incremental improvements, we can expect to see some giant strides forward in the next decade or so, which promise progress on problems which were previously outside, or on the fringes, of space techniques. Perhaps the biggest breakthrough to look forward to is the further development of active remote sensing techniques. The use of radar is already well established, for example for studying the height and roughness of the ocean surface, as we have already seen. Mention has been made of the planned use of optical radar or lidar to improve surface altimetry and to obtain Doppler wind measure- ments. The use of space-based radars to detect and map bands of precipitating clouds is also likely to come of age soon. Looking further forward, occultation techniques using two or more satellites, in which one carries a source (probably again a high-powered laser) and the other a receiver, could in principle revolutionize research requiring temperature and composition measurements by providing very high vertical resolution and spectral sensitivity. One might envision that any species which is detectable in the laboratory by spectroscopy might be monitored in the atmosphere in this way. Experiments have already been implemented in which very large spectrometers on space- craft use the Sun as a source, but of course the measurements are then limited to times near local sunrise and sunset. Most recently, temperature profiling is being attempted using the satellites of the GPS navigation system as radio beacons for a receiving satellite which then retrieves vertical temperature profiles using the radio occultation techniques originally developed and refined for planetary missions. The next decade will see the infrared passive remote sensing technique developed almost to its limit. The experience obtained with the very large and sophisticated sensors on missions like UARS (figure 10) suggested that the next steps forward need to be in improving the spatial resolution, particularly in the horizontal dimensions, and in optimizing the spectral resolution and coverage. The former will permit small-scale features in the atmosphere to be better resolved, and the latter will permit the detection of more chemical species, including those present only in trace amounts. Two examples being prepared for launch on the Earth Observing System Chemistry Mission platform (EOS-Chem- 1) in 2002 or 2003 will illustrate these points. HIRDLS (High Resolution Dynamics Limb Sounder) is an infrared limb scanning radiometer designed to sound the upper troposphere, stratosphere, and mesosphere in order to measure tempera- ture, geopotential height gradients and concentrations of 03, Downloaded by [Universitaets und Landesbibliothek] at 04:22 29 November 2013 404 F. W. Taylor Figure 10. A modern remote sensing platform: the Upper Atmosphere Research Satellite, launched in 1991. The main body of the satellite is nearly 10 m long; the satellite weighs over 6 tonnes and provides 1.6 kW of power to its eleven scientific instruments. Two of the instruments discussed in the text may be seen in this view-ISAMS is the cubical box below the middle of the spacecraft, while the MLS is the large dish mounted centrally near the right-hand end. The tall astromast at the top carries instruments for sensing the radiation and particle flux from the Sun. (Courtesy of NASA.) H20, CH4, N 2 0 , NO?, HNO3, N205, CION02, CFCl3, CF2Cl2, aerosol amounts, and locations of polar stratospheric clouds. These are the key elements that are needed to understand the dynamics of this region, including the roles of planetary and gravity waves in transporting and mixing momentum, potential vorticity, and chemicals. The goals are to make observations with horizontal and vertical resolution superior to that previously obtained, to observe the lower stratosphere with improved sensitivity and accuracy, and to use the data to improve understanding of atmospheric processes through data analysis, diagnostics and the use of the two- and three-dimensional models. In contrast to HIRDLS, the Tropospheric Emission Spectrometer (TES) emphasizes spectral resolution and coverage, combined with extensive cooling of detector and optics to retain high sensitivity. As its name implies, the primary goal is to view the troposphere, both vertically (to map pollutants near the surface) and at the limb, where it will also sound the stratosphere. TES is expected to revolutionize the understanding of tropospheric chemistry by bringing to the field the sort of global study of the large-scale behaviour of the key constituents (such as carbon monoxide and ozone), possible only using satellites, which has enhanced our knowledge of the middle atmosphere over the last 20 years. Before now, the measurement of minor constituents near the surface (especially with vertical resolution, i.e. profile measurements) was too difficult to attempt with much hope of success. Some of the techniques being applied (limb Downloaded by [Universitaets und Landesbibliothek] at 04:22 29 November 2013 Remote serzsing of the Earth ,froni space 405 scanning, mechanically cooled detectors, retrieval theory) were pioneered during programmes to study the strato- sphere. The primary measurement objectives of TES are the determination of chemical abundance profiles in the troposphere and lower stratosphere. The species to be measured include carbon monoxide, methane, ozone, water vapour, the NOx family, hydrogen sulphide, sulphur dioxide, hydrogen chloride, hydrogen fluoride, and various hydrocarbons. Since the data from UARS has become available, tropospheric has overtaken stratospheric chemistry as the area in which information on the global budgets of reactive species is most urgently required. In addition to its surface-viewing mode, TES has a limb-viewing capability which will cover the interesting tropopause region with high vertical resolution and simultaneous high spectral resolution and continuous spectral coverage. Simulations have shown that this will allow the measurement of minor constituents such as water vapour and ozone right through the tropopause, from 30 km altitude down to 5 or 6km assuming no high cloud, allowing problems of stratosphere-troposphere exchange to be addressed. The high spectral resolution of TES allows measurements in transparent 'microwindows' to penetrate right through the tropopause and down to as low as 5 or 6 km above the surface, at least when high clouds are not present. The TES instrument is a Michelson interferometer with a spectral resolution of 0*025cm-'. The problems of low signal-to-noise ratio associated with high resolution is overcome using multiple Stirling Cycle coolers of the type developed for ISAMS to achieve low (65K) focal plane temperatures. The interferometric approach, and the use of multi-element detector arrays, leads to very high data rates and to a particularly complex instrument, but retains the high species selectivity of ISAMS while giving the advantage of greater flexibility in selection of species for study. The same measurements can be made in different parts of the spectrum, which gives a large advantage in optimizing the vertical resolution and coping with the effects of thin or broken clouds, residual effects after image motion compensation, and surface emissivity effects. 8. Conclusion Hopefully, this brief overview has shown that space-based observations are contributing significantly to basic geophysi- cal research, aimed at a better understanding of our planet and its atmosphere as a physical system. Although we have not dwelt on this aspect, the role of remote sensing in operational areas such as meteorology and land use has been indicated. But currently it is perhaps in the area of countering the threat of the habitability of the Earth that more and better remote sensing expertise is needed. Global change is a serious problem and one which is very interesting scientifically. Remote sensing not only has a full role to play; it is probably the only way some of the necessary data can be obtained to monitor the problem and to understand the processes involved. Making new techniques available, and refining the existing ones, are high priorities for the future when the stakes will be even higher than they are now. References See, for example, Willson, R. C., 1985, Solar total irradiance and its spectral distribution, in The Encyclopaedia qf Physics, edited by R. M. Becanson (Van Nostrand Reinhold) 1135-43. Willson, R. C., 1979, J . Appl. Optics, 18, 179. Breukner, G. E., Edlow, K. L., Floyd IV, L. E., Lean, J. L., and Van Hoosier, M. E., 1993, Geophys. Res. Lett. Rottman, G. J., Woods, T. N., and Sparn, T. P., 1993. J . Geophys. RCJ. For example, London, J., Rottman, G. J . , Woods, T. N.. and Wu. F.. 1993, Geophys. Res. Lett., 20, 1315-1318. Ramanathan, V., 1987, J. Geophys. Res., 92, 4075-4095. Kopia, L. P., 1986, Rev. Geophys., 24, 400-406. Barkstrom. B. R., Harrison, E. F., Smith, G., Green, R., Kibler. J., Cess, R. and the ERBE Science Team, 1989, Bull. Amer. Met. Soc.. 70, 480-9. Harrison, E. F.. Minnis, P., Barkstrom, B. R., Ramanathan, V.. Cess, R. D., and Gibson, G. G., 1990, J. Geophys. Res., 95, 18687-18703. Houghton. J. T., Taylor, F. W., Rodgers, C. D., 1984. Remote Sounding of Amiospheres, (Cambridge: Cambridge University Press) 343pp. Goody, R. M. and Yung, Y., 1989, Atmospheric Radiation (Oxford: Oxford University Press). Taylor, F. W., Rodgers, C . D., Whitney, J. G., Werrett, S. T., Barnett. J. J., Peskett, G. D., Venters, P., Ballard, J., Palmer, C. W. P., Knight, R. J., Morris, P., Nightingale, T. and Dudhia, A,, 1993, J. Geophys. Res.. 98,799-10, 814. Waters, J. W., Froidevaux, L., Read, W. G., Manney, G. L., Elson, L. S., Flower, D. A., Jarnot, R. F. and Harwood, R. S. , 1993, Nature, 162, 597-602. Leblanc, T., Hauchecorne, A., Chanin, M.-L., Rodgers, C. D., Taylor, F. W. andLivesey, N., 1995, Geophys. Res. k t t . , 22, 1485-1488. Taylor, F. W., Rodgers, C. D., Remedios. J. J., Grainger, R. G., Lambert, A,, Lopez-Valverde, M., Goss-Custard, M. and Reburn, J., 1996, R. Soc. Chem., Faraday Discuss., 100, 353-369. Reburn, W. J., Remedios, J. J. , Ballard, J., Lawrence, B. N. and Taylor, F. W., 1993, Geophys. Res. Lett., 20, 1231-1234. Lambert, A,, Grainger, R. G., Rodgers, C. D., Taylor, F. W., Mergenthaler, J. L., Kumer, J. B. and Massie, S. T.. 1996, Global evolution of the properties of the Mount Pinatubo volcanic aerosols observed by the infrared limb-sounding instruments CLAES and ISAMS on UARS. J . Geophys. Res., submitted. Taylor, F. W., Lambert, A,, Grainger, R. G. and Rodgers, C. D., 1994, J. Atmos. Sci., 51, 3019-3026. Baker, W. E. et al., 1995, Bull. Amer. Met. Soc., 76, 869-886.[20] Chelton, D. B., Walsh, E. J. and McArthur, J. L., 1989. J . A t t ~ o s . Oceunic Techiiol., 6, 407-438. Spencer, R. W. and Christy, J . R., 1990, Science., 247, 1558-62. Zwally. H . J . , Brenner, A. C., Major, J. A,, Bindschadler, R. A. and Marsh, J. G., 1989, Science., 246, 1587-9. Further reading 1. B. J. Mason, Contemp. Phys. 34, 19 (1993) 2. B. J. Mason, Contemp, Phys. 36, 299 (1995). Downloaded by [Universitaets und Landesbibliothek] at 04:22 29 November 2013