A hierarchical analysis of vegetation on a Mojave Desert landscape, USA

Abstract

To examine four elements of the hierarchical structure of desert communities, we analyzed plant species composition and 13 environmental variables at 126 sites within a 755 000-ha Mojave Desert landscape, southwestern USA. By a coarse, six-group level (out of 17 groups) in cluster analysis, four generalized community types emerged: widespread, low-elevation communities with Larrea tridentata or Ambrosia dumosa; communities on unique soils (e.g., gypsum) indicated by Atriplex spp.; higher elevation/rugged terrain communities including Coleogyne ramosissima; and disturbance-associated communities such as Bebbia juncea-Hymenoclea salsola. Based on indicator species analysis (ISA), there was no clear level of the community classification that optimized discriminating among communities, because each of four measures of ISA peaked at different hierarchical levels. Three general types of indicator species were identified based on whether their value for discriminating among communities peaked at coarse (e.g., L. tridentata), intermediate (Atriplex hymenelytra), or fine (Krameria grayi) levels of the community hierarchy. Environmental variables differed in their relationships to the hierarchy, with some (e.g., pH) not differing among communities at any level and others, such as rooting depth, differing among communities at multiple levels. Hierarchical analytical techniques can help identify structural patterns within arid land plant communities, species distributions, and vegetation–environment relationships.