The U.S. National Vegetation Classification (USNVC) hierarchically classifies riparian communities of the southwest U.S. based on diagnostic species reflecting environments and biogeography that were generally derived from indirect ordinations and cluster analyses. This is particularly evident at the group and alliance levels of the classification. For example, Rocky Mountain-Great Basin Montane Riparian Forest Group versus Western Interior Riparian Forest & Woodland Group, among others. To test the efficacy of the USNVC riparian group and alliance concepts with respect to the putative environmental gradients and biogeographic patterns, we extracted plot observations in New Mexico from Natural Heritage program databases as well as herbarium specimen records from SEINet for 20 woody diagnostic species among six USNVC groups and 18 alliances. We then applied direct gradient analysis via general linear modeling and spatial species distribution modeling to evaluate concordance between environmental characteristics of individual species and the current classification concept. We were particularly interested in elevation and latitude/longitude limits of species and how alliances environmentally nested within their respective USNVC groups.
Results/Conclusions
Among the 20 diagnostic species, 13 were found to conform to central concepts for the groups and corresponding alliances in the USNVC classification. For example, Populus fremontii, Juglans major, and Fraxinus velutina were strongly diagnostic of Western Interior Riparian Forest while P. angustifolia, Alnus incana ssp. tenuifolia and Betula occidentalis reflect Rocky Mountain-Great Basin Montane Riparian Forest. There were transition zones in elevation among species such as between P. angustifolia and P. deltoides, but their central tendency was toward montane (>2,000 m) and lowland (<2,000 m), respectively. The hybrid of the two, P. acuminata, was found most commonly in the transition zone. Acer negundo was another transitional species by elevation that was considered diagnostic for both lowland and montane alliances. In a multi-species space there was a tendency for greater concordance with montane elements suggesting potential revision in the USNVC. Some species, while considered diagnostic for some groups, had wide elevation and geographic ranges that limited their diagnostic value (e.g., Salix exigua). Our results suggest that the group and alliance concepts for these riparian communities in New Mexico are reasonably coherent, but we would like to extend the analysis to a broader regional scale before recommending changes to the USNVC.