Investigation of the near-surface soil freeze-thaw cycle in the contiguous United States: Algorithm development and validation : GEWEX Continental-Scale International Project, Part 3 (GCIP3)

Tingjun Zhang,R. L. Armstrong,John A. Smith
2003-01-01
Journal of Geophysical Research
Abstract:[1] A combined frozen soil algorithm was developed and validated to detect the near-surface soil freeze/thaw cycle over snow-free and snow-covered land areas in the contiguous United States. The combined frozen soil algorithm consists of two parts. (1) Over snow-free land areas, a passive microwave remote sensing algorithm was used to detect the near-surface soil freeze/thaw cycle. (2) Over snow-covered land areas, a one-dimensional numerical heat transfer model with phase change was used to detect soil freeze/thaw status under snow cover. Using the Defense Meteorological Satellite Program's Special Sensor Microwave Imager (SSM/I) data, the passive microwave algorithm uses a negative spectral gradient between 19 and 37 GHz, vertically polarized brightness temperatures, and a cutoff brightness temperature at 37 GHz with vertical polarization (T B (37V)). SSM/I data and soil temperature data from 26 stations over the contiguous United States from 2 year period, 1 July 1997 through 30 June 1999, were used to calibrate the algorithm (year 1), to validate the algorithm (year 2), and to demonstrate freeze/thaw classification (both years). A cutoff brightness temperature of 258.2 K was obtained on the basis of a linear correlation (r2 = 0.84) between the soil temperature at 5 cm depth and the T B (37V). The combined frozen soil algorithm provides an accuracy for frozen soil detection of about 76% and an accuracy for the correct classification of both frozen and unfrozen soils of approximately 83% with a percent error of about 17%. The combined frozen soil algorithm was used to investigate the timing, duration and number of days, and daily area extent of near-surface frozen soils over the study area. The primary results indicate that the maximum area extent of frozen ground during the winter of 1997/1998 was about 4.4 x 10 6 km 2 or 63% of the total land area of the contiguous United States, while during the winter of 1998/1999. the maximum extent was about 5.2 x 10 6 km 2 or 74%. The duration of the soil freeze ranges from less than 1 month in the south to over 8 months in the Rocky Mountains. The actual number of days of soil freezing varies from a few weeks to more than several months. The number of near-surface soil freeze/thaw cycles varied from 1 to more than 11 during the winters of 1997/1998 and 1998/1999, while the average length frozen period varied from less than 20 days to more than 220 days.
What problem does this paper attempt to address?