Using microwave brightness temperature to detect short-term surface air temperature changes in Antarctica: An analytical approach

Abstract

Because of the importance of detecting climate change signals over Antarctica, we examined the dependence of the snow's effective temperature times its emissivity (brightness temperature) on surface air temperatures over Antarctica. A range of microwave frequencies and various time scales from daily to yearly is considered. Reasonable accuracy of the air temperature should be obtained without including the annual variation of the emissivity. We studied the sensitivity of the inferred air temperature to penetration depth, a determinant parameter for the effective temperature, and used a snow dielectric constant model derived from strong fluctuation theory. Below 10 GHz, the snow cover temperature and layering determine this depth, whereas above 20 GHz, the determinant factor is the snow crystal size; all parameters have an effect at intermediate frequencies. The penetration depth ranges for the Scanning Multichannel Microwave Radiometer (SMMR) channels are the following: 20–40 m at 6.6 GHz, 6–16 m at 10.7 GHz, about 1–5 m at 18 GHz, and 0.1–1.4 m at 37 GHz. Analysis of the effective temperature revealed that the snow effective temperature can explain short-term fluctuations at 37 GHz and annual trends for all SMMR channels. For short-term fluctuations (<10 days), calculations and measurements showed that the brightness temperature at 37 GHz starts to respond immediately to the air temperature changes, even though the time to reach the maximum is delayed by a day or less; furthermore, the amplitude of the resulting fluctuation was sufficiently large to be detectable in most cases. Indeed, short-term fluctuations of the air temperature can clearly be detected from significant variations of the air temperature over 3 days, but small amplitude fluctuations shorter than 1 day will hardly be detectable through the brightness temperature. This limits use of the brightness temperature at 37 GHz, and this is a point to consider for inversion models. Nevertheless, air temperature fluctuations over 2–3 days can be detected; thus, this technique could be used to observe the winter warming periods in Antarctica that often occur, even during cloudy periods.