Spatial Analysis of Thaw Depth
ROSA T. AFFLECK AND SALLY A. SHOOP
INTRODUCTION
ditions, one would expect a very high degree of hetero-
geneity. The basic question is to define the spatial vari-
The objective of this work was to develop an
ability of depth of thaw and analyze the site's specific
approach to characterizing the spatial variability of
effects on this variability. The Alaska North Slope data
thawing soil. Results can then be applied to spatially
sets are from areas of continuous permafrost where the
distribute soil properties based on point data or one-
underlying frozen layer will remain year-round. The
dimensional models, or to fill sparse data sets with ter-
bulk of the measurements were taken at the end of the
rain properties. Results are also useful for probabilistic
summer when the thaw had penetrated to the maximum
analysis of the impact of the input distribution on the
depth. The other sites are located in areas without per-
confidence of predictive models, such as for predicting
mafrost where all frost will eventually be removed as
vehicle mobility or surface runoff. Thaw depth was used
the bottom of the frozen layer meets with a thawed layer
as the variable of interest because of its clear impact on
by late spring or early summer. The data sets were col-
terrain strength and hydrology. Other soil variables can
lected over a wide range of spacing and scales, all of
be analyzed similarly. A comparison among these data
which affect the resulting spatial analysis. This is an
sets also indicates the impact of other terrain parame-
initial effort to develop variation of thaw depth for a
ters, such as vegetation or radiation, on the thaw varia-
variety of conditions, but this work excludes the im-
bility for a particular terrain unit.
pact of surface energy balance and climatic conditions
A search and review of existing data sets having
on the thaw variability.
spatial thaw measurements yielded four major data sets
suitable for analysis. These data cover a variety of cli-
matic and terrain conditions, including permafrost sites
DATA SETS
in Alaska, temperate sites in Minnesota and Wiscon-
Alaska North Slope
sin, and a large-scale, controlled test section with no
Thaw depth (active-layer thickness) data on Alaska
solar input in the Frost Effects Research Facility (FERF)
North Slope were collected for Arctic climate-change
at the Cold Regions Research and Engineering Labora-
research headed by Dr. Frederick E. Nelson for Arctic
tory (CRREL) in Hanover, New Hampshire. Each of
System Science/LandAtmosphereIce Interactions
these data sets is described in detail below.
(ARCSS/LAII) sponsored by the National Science
Thaw is a complicated process, influenced by the
Foundation (Nelson et al. 1997). The National Snow
surface energy balance, terrain, and soil properties at
and Ice Data Center (NSIDC) managed and archived
each point. Thaw depth is the thickness of the thawed
these data sets.
layer from the surface to an underlying frozen layer,
Permafrost underlies the area. Nelson et al. (1997)
which typically occurs when temperature is above freez-
used Geographic Information Systems (GIS) to map
ing in spring. Terrain with a thawed layer can be very
and develop an integrated response of thaw depth us-
saturated due to snowmelt, rain, and water drawn to
ing air temperature, vegetation, soils and soil moisture,
the soil during the freezing process. Slopes (aspect and
and solar radiation for these sites. Figure 1 shows the
degree), types of vegetation, and soil properties affect
locations of the seven sites where thaw depth was meas-
the variation on thaw depth. With all these variable con-
to contents
Previous Page
