High Frequency Passive Microwave Radiometry over a Snow-Covered Surface

Andrew Tait

Passive microwave data has long been used to differentiate between snow-covered and snow-free surfaces, as well as to estimate the depth and water equivalent of snowpacks. Most of the algorithms, past and present, employ passive microwave data in the frequency range of 18 to 37 GHz. The 37 GHz channel, in particular, has been shown to be quite sensitive to the presence of snow and to variations in depth and water equaivalent. However, little research has been performed on the applicability of relatively high frequency radiometric data for use in snowpack monitoring.

The Millimeter-wave Imaging Radiometer (MIR) records radiation emanating from the surface and atmosphere in nine frequency bands : 89 GHz, 150 GHz, 183.31 GHz, 183.33 GHz, 183.37 GHz, 220 GHz, 3251 GHz, 3253 GHz, and 3258 GHz. The frequencies were chosen to match those of the Advanced Microwave Sounding Unit-B (AMSU-B) to be included on the Earth Observing System (EOS) platform. The instrument was developed for atmospheric research with the three channels centred about 183 GHz and the 150 GHz channel designed to interpret the atmospheric water vapor profile. Other channels are less opaque and can provide some information from the surface. The instrument has been flown on several NASA ER-2 flights, two of which (the April 1995 Alaska mission and the February 1997 WINter Cloud Experiment (WINCE)) were specifically designed to address the problem of cloud detection over a snow-covered surface.

It is of interest to examine the MIR data to determine whether, in addition to cloud detection, these high frequency data can be used to identify varying surface snow conditions. The procedure used is as follows: 1) Identify cloudy and cloud-free scenes incorporating information from the Modis Airborne Simulator (MAS) (also flown on the ER-2 flights); 2) Determine whether a purely MIR-based cloud detection scheme is possible over a snow-covered surface; 3) Include a surface vegetation dataset and analyse the possible influence of changing vegetation type on the brightness temperatures; and 4) Compare completely snow-covered scenes with partly snow-covered and snow-free regions for cloudy and clear sky periods to determine whether varying snow conditions affect the MIR data.


USRA, NASA/GSFC, Greenbelt MD 20771
E-mail: atait@glacier.gsfc.nasa.gov

[ Return ]