Soil freezing is a climatic disturbance that can occur sporadically during winter. Normally, the presence of a thick snowpack limits the occurrence of soil frost in northern forest ecosystems due to its insulating effect. However, it is anticipated that under a warmer climate, decreases in snow cover will increase the frequency and severity of soil freezing. Soil frost can limit the activity of fine roots, decreasing plant uptake, and result in elevated nutrient loss and changes in stream chemistry. In this study, we evaluated whether frequency of soil freezing has changed over the long term in a northern hardwood forest ecosystem at the Hubbard Brook Experimental Forest (HBEF), NH and whether soil freezing events have increased stream nitrate, calcium, magnesium and potassium loss following the event. To test the hypothesis that inter-annual patterns in stream chemistry were related to inter-annual differences in soil freezing severity, we determined four-year running averages for annual stream solute concentrations and calculated annual deviations from running averages. We examined temporal patterns in snow cover, soil frost depth, stream chemistry and climatic data (air temperature and precipitation) collected at the HBEF, from 1956 to 2014.
Although the annual average depth of snowpack at the HBEF has declined by 5 cm per decade over the past six decades (R2 = 0.26; p < 0.0001), we have not observed any significant change in frequency of severe soil frost (i.e., soil frost with depth > 10 cm). A relatively constant number of years with severe soil frost (i.e., 3-4 years per decade) were observed over the study period. We examined the cumulative effects of snow depth (by linear interpolating snow depth between weekly/biweekly observations) on soil frost depth. We found that for years with a cumulative snow depth > 1000 cm*day, severe soil frost was not evident. In contrast to our hypothesis, we did not observe any significant relationship between the soil freezing severity and stream chemistry variables. However, inter-annual deviations of K+ and NO3- concentration in streamwater from the running averages were related to inter-annual differences in precipitation quantity and air temperature. This pattern suggests that it would be necessary to incorporate the influence of climatic conditions on biotic mechanisms (e.g., root and microbial mortality, plant uptake) to assess nutrient loss following soil freezing events.