Using multi-scale sampling and spatial cross-correlation to investigate patterns of plant species richness

Abstract

Land managers need better techniques to assess exotic plant invasions. We used the cross-correlation statistic, IYZ, to test for the presence of spatial cross-correlation between pair- wise combinations of soil characteristics, topographic variables, plant species richness, and cover of vascular plants in a 754 ha study site in Rocky Mountain National Park, Colorado, U.S.A. Us- ing 25 large plots (1000 m 2 ) in five vegetation types, 8 of 12 variables showed significant spatial cross-correlation with at least one other variable, while 6 of 12 variables showed significant spatial auto-correlation. Elevation and slope showed significant spatial cross-correlation with all variables except percent cover of native and exotic species. Percent cover of native species had significant spatial cross-correlations with soil variables, but not with exotic species. This was probably because of the patchy distributions of vegetation types in the study area. At a finer resolution, using data from ten 1 m 2 subplots within each of the 1000 m 2 plots, all variables showed significant spatial auto- and cross-correlation. Large-plot sampling was more affected by topographic factors than species distribution patterns, while with finer resolution sampling, the opposite was true. However, the statistically and biologically significant spatial correlation of native and exotic species could only be detected with finer resolution sampling. We found exotic plant species invading areas with high native plant richness and cover, and in fertile soils high in nitrogen, silt, and clay. Spatial auto- and cross-correlation statistics, along with the integration of remotely sensed data and geographic information systems, are powerful new tools for evaluating the patterns and distribution of native and exotic plant species in relation to landscape structure.