HR: 08:15h
AN: B31E-02    [Abstracts]
TI: Principal Controls Governing Vegetation Dynamics Across Varying Climates in Ecoregions with Complex Topography
AU: * White, A B
EM: abwhite@uiuc.edu
AF: University of Illinois at Urbana-Champaign, 205 N. Mathews Ave., Urbana, IL 61801 United States
AU: Kumar, P
EM: kumar1@uiuc.edu
AF: University of Illinois at Urbana-Champaign, 205 N. Mathews Ave., Urbana, IL 61801 United States
AB: Remotely-sensed vegetation greenness, as measured through a photosynthetically-related index such as the enhanced vegetation index (EVI), can be extremely heterogeneous, both spatially and temporally. The driving factor behind this heterogeneity is the variability in the climate; however, a myriad of variables influence this heterogeneity, including: topography, which constrains the temperature, moisture availability, and amount of incoming radiation; soil properties, which influences the water availability, root density and length, plant morphometric characteristics, etc.; and land cover, which indicates the type of vegetation that is present. In this study, we investigate the dominant influences on the vegetation greenness across three different ecoregions (Level III CEC Ecoregions), as these regions represent natural divisions of the land with respect to vegetation dynamics and the influences on them. We employ a regression tree technique to extract patterns in the very large database created for this study and examine the patterns, or trees, for physical meaning. In all three ecoregions (the deciduous forest-dominated Blue Ridge, the evergreen forest-dominated Northern Rockies, and the shrubland-dominated Central Basin and Range), the topography is the most significant factor influencing the vegetation amount and vigor throughout the growing season. The meteorology, soil properties, and land cover follow in descending order. Thus, it is the local control on the meteorology that is the overriding influence, whereas the regional meteorologic patterns and local/regional soil properties are the next most influential. The land cover, which varies across a broad range of scales, is only influential when it differs from the dominant land cover type. The depth of the regression tree is an indicator of scale, as the greater the depth, the more the region is divided, thus the smaller the scale is. Patterns emerge in the relevance of each variable (varying somewhat from ecoregion to ecoregion and throughout the growing season): the land cover decreases exponentially with scale; the topography, meteorology, and soil properties are parabolically related to scale, although the vertex of the parabola is a minimum for topography whereas the meteorology and soil properties are inverted.
DE: 0439 Ecosystems, structure and dynamics (4815)
DE: 0486 Soils/pedology (1865)
DE: 0555 Neural networks, fuzzy logic, machine learning
DE: 1813 Eco-hydrology
DE: 1851 Plant ecology (0476)
SC: Biogeosciences [B]
MN: Fall Meeting 2005