Figure 3: Examples of infrared spectra.
From: Humidity trends imply increased sensitivity to clouds in a warming Arctic

Examples of ‘clear-sky’ infrared spectra calculated from profiles of temperature and humidity measured by radiosoundings at Barrow, Alaska (red) and Summit Station, Greenland (blue). The dashed lines are Planck functions corresponding to the near-surface air temperatures for the respective cases and represent hypothetical, optically thick clouds. The FIR and AW spectral regions are shown as dark-grey and light-grey shaded regions. The downwelling infrared CRE is the spectral integral of the difference between the curves (dashed minus solid for any case). (Note that because these spectra are for zenith views, the flux (in W m−2) is obtained by integrating the radiance over the hemisphere and over wavenumber (frequency) from 0 to 3,000 cm−1 (ref. 6)). The vertical arrows indicate conceptually the magnitude of the CRE in each window. (a) Spectra from 12 July 2012 (Summit) and 8 May 2008 (Barrow): similar near-surface air temperatures, but different PWV. The CRE is larger at Summit because of additional CRE in the FIR associated with the low PWV. (b) Spectra from 7 May 2011 (Summit) and 12 March 2007 (Barrow): similar PWV, but different near-surface air temperatures. The CRE is larger at Barrow because of additional CRE in the AW associated with the higher temperature. (c) Spectra from 21 April 2011 (Summit) and 18 May 2008 (Barrow), where the conditions approximate the Clausius–Clapeyron parameterization in Figs 4 and 6. These cases represent both the cold/dry and warm/moist limits of the range of conditions. The total CRE (CREFIR+CREAW) in c, represented conceptually by summing the arrows, is similar between the two cases due to the compensating effects described in the main text.