The effect of frozen soil on snowmelt runoff at Sleepers River, Vermont

Soil frost depth has been monitored at the Sleepers River Research Watershe d in northeastern Vermont since 1984. Soil frost develops every winter, par ticularly in open fields, but its depth varies greatly from year to year in inverse relation to snow depth. During the 15 years of record at a benchma rk mid-elevation open site, the annual maximum frost depth varied from 70 t o 390 mm. We empirically tested the hypothesis that frozen soil prevents in filtration and recharge, thereby causing an increased runoff ratio (Streamf low/(rain + snowmelt)) during the snowmelt hydrograph rise and a decreased runoff ratio during snowmelt recession. The hypothesis was not supported at the 111 km(2) W-5 catchment; there was no significant correlation of the r unoff ratio with the seasonal maximum frost depth for either the pre-peak o r post-peak period. In an analysis of four events, however, the presence of frost promoted a large and somewhat quick:er response to rainfall relative to the no-frost condition, although snow cover caused a much greater time- to-peak regardless of frost status. For six years of flow and frost depth m easured at the 59 ha agricultural basin W-2., the hypothesis appeared to be supported. The enhancement of runoff due to soil frost is evident on small plots and in extreme events, such as rain on frozen snow-free soil. In the northeastern USA and eastern Canada, the effect is often masked in larger catchments by several confounding factors, including storage of meltwater i n the snowpack, variability in snowmelt timing due to elevational and aspec t differences, interspersed forested land where frost may be absent, and th e timing of soil thawing relative to the runoff peak.