Investigation of the near-surface soil freeze-thaw cycle in the contiguous United States: Algorithm development and validation

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 remo...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Journal of Geophysical Research. D. Atmospheres 2003-11, Vol.108 (D22), p.n/a
Hauptverfasser: Zhang, T., Armstrong, R. L., Smith, J.
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung: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 (TB(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 TB(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 × 106 km2 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 × 106 km2 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.
ISSN:0148-0227
2156-2202
DOI:10.1029/2003JD003530