Retrieving Profile Temperatures in a Frozen Topsoil Near the TFS, Alaska, Based on SMOS Brightness Temperatures at the 1.4-GHz Frequency

In this paper, the method previously proposed in earlier work for measuring the temperature profile in a frozen topsoil using multiangular brightness temperature observations in the L-band has been experimentally tested. At a frequency of 1.4 GHz, full-polarization multiangular brightness temperature data were obtained from the Soil Moisture and Ocean Salinity (SMOS) satellite land product of Level 1C, with the SMOS footprint being centered at the Toolik Field Station (TFS), Alaska. The SMOS data covered the period from January 1, 2010 to December 31, 2011. Retrieval of the temperature profiles in a frozen topsoil was based on the semiempirical emission model L-MEB and the temperature-dependent dielectric model for an organic-rich tundra soil. The soil samples measured to develop the dielectric model were collected at the TFS site. For winter seasons, the retrieved temperature profiles in the 16.0-cm topsoil were validated relative to the temperature profiles measured in situ. As a result, the values of root-mean-square error and determination coefficient of the temperatures retrieved at the depths of 0.6, 8.7, and 16.0 cm, relative to the respective temperatures measured in situ, were found to be 2.8 °C, 4.9 °C, and 6.4 °C and 0.62, 0.42, and 0.26, respectively. The sources of error and possible improvements of the proposed retrieving algorithm were discussed. The major result of this study is the demonstration of the potential possibility for remote sensing of the temperature profile in a frozen arctic topsoil using the SMOS multiangular brightness data.