LANDSCAPE ANALYSIS OF PLANT DIVERSITY

Studies to identify gaps in the protection of habitat for species of c oncern have been inconclusive and hampered by single-scale or poor mul ti-scale sampling methods, large minimum mapping units (MMU's of 2 ha to 100 ha), limited and subjectively selected field observations, and poor mathematical and ecological models. We overcome these obstacles w ith improved multi-scale sampling techniques, smaller MMU's (< 0.02 ha ), an unbiased sampling design based on double sampling, improved math ematical models including species-area curves corrected for habitat he terogeneity and geographic information system-based ecological models. We apply this landscape analysis approach to address resource issues in Rocky Mountain National Park, Colorado. Specifically, we quantify t he effects of elk grazing on plant diversity, identify areas of high o r unique plant diversity needing increased protection, and evaluate th e patterns of non-native plant species on the landscape. Double sampli ng techniques use satellite imagery, aerial photography, and field dat a to stratify homogeneous and heterogeneous units and ''keystone ecosy stems'' (ecosystems that contain or support a high number of species o r have distinctive species compositions). We show how a multi-scale ve getation sampling design, species-area curves, analyses of within-and between-vegetation type species overlap, and geographic information sy stem (GIS) models can be used to quantify landscape-scale patterns of vascular plant diversity in the Park. The new multi-scale vegetation p lot techniques quickly differentiated plant species differences in pai red study sites, Three plots in the Ouzel Burn area (burned in 1978) c ontained 75 plant species, while only 17 plant species were found in p aired plots outside the burn. Riparian areas contained 109 plant speci es, compared to just 55 species in paired plots in adjacent forests. H owever, plant species richness patterns inside and outside elk exclosu res were more complex, One elk exclosure contained more species than i ts adjacent open range (52 species inside and 48 species outside). Two elk exclosures contained fewer species inside than outside (105 and 4 1 species inside and 112 and 74 species outside, respectively). Howeve r, there was only 26% to 48% overlap (using Jaccard's Coefficient) of plant species composition inside and outside the exclosures. One elk e xclosure had 13% cover of non-indigenous species inside the exclosure compared to 4% outside, but non-indigenous species cover varied by loc ation. We compared plant diversity patterns from vegetation maps made with 100 ha, 50 ha, 2 ha, and 0.02 ha MMU's in the 754 ha Beaver Meado ws study area using four 0.025 ha and twenty-one 0.1 ha multi-scale ve getation plots. Preliminary data suggested that the 2 ha MMU provided an accurate estimate of the number of plant species (-14%) for a study area, but the number of habitats (polygons) was reduced by 67%, and a spen, a unique and important habitat type, was missed entirely, We des cribe a hypothesis-driven approach to the design and implementation of geospatial databases for local resource monitoring and ecosystem mana gement.