HR: 1340h
AN: H53E-0526 [Abstracts]
TI: Quantifying Spatial Heterogeneity and Co-variance of Surface Hydrologic Parameters at the Plot Scale in
a Semi-arid Landscape
AU: * Truschel, A
EM: anthony.truschel@colorado.edu
AF: Department of Geological Sciences, University of Colorado, Campus Box 399, University of Colorado at
Boulder, Boulder, CO 80309-0399
United States
AU: Bhark, E W
EM: ebhark@intera.com
AF: Stoller-Navarro Joint Venture/INTERA Inc., 7710 W. Cheyenne Avenue, Las Vegas, NV 89129
United States
AU: Small, E E
EM: eric.small@colorado.edu
AF: Department of Geological Sciences, University of Colorado, Campus Box 399, University of Colorado at
Boulder, Boulder, CO 80309-0399
United States
AB:
In arid to semi-arid ecosystems, the partitioning of precipitation into infiltration, runoff, and interception loss is
strongly influenced by the spatial arrangement of vegetation and related (near-)surface hydraulic properties. This
partitioning is critical to a variety of hydrological and ecological processes and the interactions between them. On
semi-arid hillslopes, we study the spatial patterns of infiltration and runoff, including vegetation mosaic geometry,
microtopography, and soil hydraulic properties, from the centimeter to tens-of-meters scale.
Approximately 8,000 measurements were collected in creostebush shrubland and black grama grassland hillslopes in the
Sevilleta National Wildlife Refuge, central New Mexico. Measurements were collected over plot-scale (100 sq. m.) domains,
the surface area of which was defined prior as a representative elementary area (REA) with respect to mosaic geometry, in
sampling schemes designed for semivariogram estimation. All measurements were colocated.
Auto- and cross- relationships in the form of experimental semivariograms are first characterized between properties out to
the maximum measurement scale. These relationships are used to define spatial scales of directional heterogeneity within and
between variables. Next, geostatistical models are fit to the experimental data, at appropriately smaller scales, for model
parameter estimation. Model parameters are used 1) to identify surface properties that control features of the
semivariogram, 2) for the comparison of spatial features between the data sets, and 3) for improved estimation of the
measured properties for use in related flow simulation. Results show marked attributes of the semivariogram that can be
directly related to spatial patterns of surface properties. Similarly, the strong spatial correlation between properties
appears as distinct features on the cross-semivariogram. Corresponding (cross-)semivariogram model parameters indicate that
significant improvement is likely for the geostatistical estimation (e.g., kriging) of a primary variable using one or more
colocated secondary datum.
DE: 1813 Eco-hydrology
DE: 4815 Ecosystems, structure, dynamics, and modeling (0439)
SC: Hydrology [H]
MN: Fall Meeting 2005