Global Satellite Observations of Negative Brightness Temperature Differences between 11 and 6.7 µm

Abstract

Global analyses of satellite spectral observations indicate the existence of negative brightness temperature differences between 11 and 6.7 µm (BT11 ?BT6.7) when cold scenes are viewed. Differences are typically greater than ?5 K for the Tropics and midlatitudes but can be smaller than ?15 K over high-altitude polar regions during winter. In July, more than 60% of the observations over the Antarctic Plateau had BT11 ? BT6.7 < ?5 K. In January, over Greenland, the frequency of occurrence is approximately 20%. Three factors are investigated that may contribute to these observed negative brightness temperature differences: 1) calibration errors, 2) nonuniform scenes within the field of view, and 3) physical properties of the observed phenomena. Calibration errors and nonuniform scenes may generate values of BT11 ? BT6.7 that are less than zero; however, these differences are on the order of ?2 K and, therefore, cannot fully explain the observations. A doubling and adding radiative transfer model is used to investigate the physical explanations of the negative differences. Simulations of satellite spectral observations for thick clouds produce negative differences that are comparable to those observed in the Tropics and midlatitudes. The magnitude of the differences is a function of cloud microphysics, cloud-top pressure, view angle, and the cloud optical thickness. The model simulations are also capable of producing large negative differences over high-altitude polar regions. Distinguishing clear and cloudy regions from satellite infrared radiances is a challenging problem in polar winter conditions. Brightness temperature differences between 11 and 6.7 µm provide a technique to separate cold, optically thick clouds from clear-sky conditions when strong radiation inversions exist at the surface. The presence of a cloud inhibits the development of this inversion and shields its detection using satellite radiance measurements. While physically reasonable, and in agreement with radiative transfer calculations, this technique has not been verified with ground nor with aircraft observations. Further evidence that the large negative values of BT11 ? BT6.7 are associated with surface inversions is presented by comparing the satellite observations with surface temperature measurements from an Antarctica automated weather station.