97th ESA Annual Meeting (August 5 -- 10, 2012)

PS 58-169 - Using ground penetrating radar to characterize soil frost for ecological applications

Wednesday, August 8, 2012
Exhibit Hall, Oregon Convention Center
John R. Butnor, Southern Research Station, U.S. Forest Service, Burlington, VT, John L. Campbell, Northern Research Station, USDA Forest Service, Durham, NH and James B. Shanley, NH-VT Water Science Center, USGS, Montpelier, VT
Background/Question/Methods

Seasonal soil freezing is an important natural perturbation common in cold regions around the world. It is expected that future changes in climate will alter the temporal patterns and spatial extent of seasonally frozen ground influencing physical, chemical, and biological processes in soil. A thorough evaluation of ecological responses to seasonally frozen ground is hampered by our inability to adequately characterize the frequency, depth, duration and intensity of soil frost events. Ground penetrating radar (GPR) may be used to detect objects or interfaces where there are abrupt changes in soil electrical properties.  The electromagnetic waves generated by GPR transmitters move quickly in air and dry soil and attenuate in wet soils.  When soil freezes, its dielectric value drops creating a detectable interface. GPR has the potential to nondestructively quantify seasonal soil freezing in forests, though it requires further evaluation before it can be put to use in ecological applications.  Several key questions include: 1) What is minimum depth of detectable soil frost? ; 2) How accurate are frost depth estimates with GPR? ; 3) Can GPR antennas be configured to penetrate snowpack and still resolve soil frost level?  A series of test transects were established in Vermont (2) and New Hampshire (1).  Each site was fitted with a soil temperature logger (15 depths) and frost tubes (30) to make direct comparisons with GPR radargrams.  These sites were monitored with a SIR-3000 GPR system equipped with 400, 900 and 1500 MHz antenna in several configurations for the winter 2011-2012.

Results/Conclusions

Antenna frequency selection is a tradeoff between resolution and penetration.  Short wavelengths can resolve near surface reflections, but cannot penetrate deeply; the 1500 MHz antenna (high frequency) was able to detect frost penetration as shallow as 7-8 cm, but the uniformity of detection was marginal until depths of 10 cm were achieved.  The lower frequency 900 MHz antenna reliably detected frost at depths of greater than 15 cm.  The 900 MHz antenna is preferred, since the additional penetration from the longer wavelength should be more useful in penetrating snowpack above frozen soil. The snowfall in New England has been well below normal for winter 2011-2012. One period where deeply frozen soil (~25cm) was covered with snow, GPR was able to readily penetrate the snow and accurately detect the frost depth. We have noted that rain events, surface water films or surface thawing over frozen soil result in destructive interference.