Water content and matric potential of soil under different soil frost conditions

Eastern Hokkaido, where is one of the largest agricultural production regions in Japan, is characterized by low air temperature and relatively thin snow covers resulting in soil frost over the winter. However, the soil frost depth has been significantly decreasing since late 1980's due to an insulation from the cold air by a thick snow cover developing in early winter. In the current study, soil water movement under different soil frost conditions were monitored to obtain a knowledge of changes in hydraulic-regime of the agricultural production systems in the Eastern Hokkaido associated with the decreasing soil frost depth in the region. A paired soil plot experiment was conducted from Nov. 2005 to May 2006, where the frost depth was artificially enhanced by removing snow in the treatment plot and the natural condition was maintained in the control plot. The soil in the experimental field was classified as Andisol with much porosity and high drainability. In each plot, water content and matric potential were measured by TDR and thermally-insulated tensiometer, respectively. Changes in snow water equivalent volume (SWE) and soil-frost depth were manually recorded. The maximum soil-frost depth in the treatment and control plots resulted in 47 and 19 cm, respectively. In both plots, soil water content and matric potential in underlying unfrozen soil decreased with the progress of freezing front, and the direction of soil water flow between 90 and 100 cm changed from downward to upward after the onset of the soil freezing. It is of note that the matric potential at 90 cm in the treatment plot decreased down to -480 cm, while the matric potential at the same depth in the control plot was -200 cm at minimum. When the underlying unfrozen soil was most driest the soil water volume stored in a depth interval from 50 to 100 cm for the treatment and control plots was 189 and 212 mm, respectively. Further, the magnitude of upward hydraulic gradient between 90 and 100 cm in the treatment plot was 17 times greater than that in the control plot at maximum. The control plot with a thin frozen layer allowed infiltration of snow melt water, water content at the lower subsoil increased, and the hydraulic gradient changed to downward immediately after snow melting occurred, whereas a thick frozen layer in the treatment plot impeded the infiltration resulting in waterlogging being observed on the soil surface, which may induce surface runoff. Consequently, the results suggest that the resent decreasing soil frost depth inhibits lower subsoil from drying and facilitates infiltration of snow melt water.