were confirmed and extended to include the region 4 -12 p.Unexpectedly high absorption was observed in this latter region.(ii) These results inspired a reexamination of all available previous work, from which it was concluded that the extinction arises from two separate effects with different frequency and temperature dependence.They are i) a foreignbroadened component, roughly proportional to the foreign gas pressure and probably positively temperature dependent (0.5 to 1% per °C); and ii) a self-broadened component roughly proportional to partial vapour pressure and with a strong negative temperature dependence (2% per °C at 30°C).On the basis of several assumed line shapes, the accumulated wings of lines in the 50 la rotation band were computed at several points in the range 12 -40 p; the Lorentz shape gave a very good fit (1. 10%) to the foreign-broadened component, but none gave a satisfactory fit to the selfbroadened component. *The 15 11 CO2 band was found to give no detectable absorption (< 1% for 43 atm -cm of CO 2' the equivalent of 1.4 km of air at sea level) outside the range 12.7 to 17.8 11.The horizontal path experiment has already been reported in a joint paper with F. Saiedy and P. A. Sheppard, entitled 'On the Atmospheric Infrared Continuum', J.O.S.A. 51, 466 (1963); a copy is attached to this thesis.During this research a further paper, on the related topic of terrestrial heat balance, entitled 'Heat-balance measurements from an earth satellite -an analysis of possibilities' was published in Quart. J. Roy. Met. Soc., 87, 231 (1961); a copy is attached. 11.13 See Addendum for amended version. ,Chapter 1?1. IntroductionThe existence of a continuous absorption throughout the atmospheric infrared window between 8 and 13 microns has been knoinfor a long time, Elsasser (1938) probably being the first to draw attention to it. It has been observed many times since by all workers in the study of atmospheric radiation, both in transmission observations using the sun and artificial sources, and in emission studies from ground and space 2.iii) Since the advent of artificial earth satellites, several infrared emission experiments have been proposed, or in some cases already been carried out, to determine the earth's surface temperature distribution, cloud top height, ozone and carbon dioxide distribution, and temperature structure in the upper atmosphere.All of these, except possibly the last, require correction to some extent for the continuum.Much discussion has centred on the question of the origin of the continuum absorption; there is no doubt that it is largely due to wcter vapour, but the role of aerosol is less certain. That it is necessary to appeal to something other than water vapour appears to be due to the fact that measurements of continuum absorption by different workers do not seem to be consistent when compared on the basis of the existing picture of the water vapour continuum alone. Whilst it is not disputed that aerosol absorption in the 8 -13j. region can be significant under extremel...

No abstract

No abstract

No abstract

Artificial earth srttellites offer a unique opportunity t o exploit the possibility of deducing temperature profiles on a global scale from measurements of radiance in several narrow spectral intervals in A strongly absorbing band of an atmospheric gas whose mixture is uniform. In the earth's atmosphere the 4.3-micron and 15-micron bands of carbon dioxide and the 5-mm. band of oxygen may be used; only the 15-micron band is considered in detail, although the procedures are applicable to the other bands. The Woblem considered is the numerical solution of the integral form of the radiative transfer equation from measurements in a finite set of spectral intervals. It is shown that, by a suitable approximation of the Planck radiance, the radiative transfer equation can be reduced to an integral equation of the first kind. After a discussion of the kernel, which is associated with the transmittance of the gas, the equation is changed to a finite set of equations which is amenable to numerical solution. The solution is limited to about six pieces of information, which may be expressed as points along the vertical profile, or as coefficients of an expansion; the limitation in information is manifest in the transmittance curves for the several spectral intervals, the errors of measurement, and the approximations employed. However, even in this limited case the formal solution of the set of equations is unstable. I n this process the amount of smoothing remains small, so that the inherent properties of the temperature profile are not affected. Several possible 351 352