88th Annual Meeting (20-24 January 2008)

Monday, 21 January 2008: 11:00 AM
Historic climate impacts on seasonal soil frost in the Midwestern United States
223 (Ernest N. Morial Convention Center)
Tushar Sinha, North Carolina State University, Raleigh, NC; and K. A. Cherkauer
Soil frost is the state when soil temperature is equal to or below 0°C and soil moisture has begun to freeze. Seasonally frozen soil, that which thaws completely in the summer and snow accumulation play an important role in the hydrology of the Midwestern United States. Snow melt and rain over a frozen ground surface can lead to an increased runoff response due to restricted infiltration rates. Soil freezing and thawing also weakens soil bonds, increasing the risk of soil erosion. Climate projections suggest that the Midwest will experience warmer air temperatures and a shift in precipitation from the summer into the winter and spring; this is likely to change regional hydrology especially in the winter and spring. Understanding how the region has responded to such changes in the last century can provide insight into how it may change in the future.

In this study, significant trends in soil frost variables such as freeze-thaw cycles and soil frost days are identified using available historical observations of soil temperature and their relationships is compared with other climatic variables. Observational records of soil temperature at multiple soil depths are available since 1966 at a limited number of sites in the Midwestern United States. Trend analysis of seven sites located in Indiana, Illinois and Minnesota indicates significant increases in mean maximum and mean minimum winter soil temperatures since 1966. Significant reductions in the annual number of days with soil frost are also indicative of a warmer climate. Extension of the analysis in both time and space is achieved by using the Variable Infiltration Capacity (VIC) large-scale hydrologic model. Soil parameters are calibrated and evaluated by comparing subsets of the available observed soil temperature and moisture records with simulated values. The calibrated model is then run over a six state region in the north-central Midwest using observed meteorological forcing data from 1915 through 2006. Model predicted variables such as soil temperature, soil ice content, freeze-thaw depths and snow accumulation are analyzed spatially and temporally over the study domain.

Supplementary URL: