Characterization of woody species distribution in riparian forests of Lower Michigan, USA using map-based models

Abstract

The goal of this study was to identify and characterize the range of variation in riparian forests across Lower Michigan, USA as a basis for assessing the utility of map-based information in the characterization of riparian environments. We obtained a regional sample of riparian forests and valley-bottom physiography from 94 locations throughout Lower Michigan and distinguished seven major riparian forest types using hierarchical clustering and NMDS ordination. NMDS ordination distances accounted for more than 70% of the distances in species space and achieved excellent discrimination among riparian types. We then evaluated a set of map-based variables indicative of regional climate, catchment hydrology, and valley character relative to ordination axes and interpretations of the autecology of principal tree species from each forest group. Map-based predictors accounted for 83% of the variation in sample scores along NMDS Axis 1 and explained 42% of the variation in Axis 2. Species and riparian types varied along two principal gradients, one associated with climate and geology along a north-south gradient, the second associated with flood duration and power. Map-based interpretations of regional climate and hydroperiod dynamics agreed closely with species-based interpretations of riparian character, although in certain cases, a similar biotic response arose from apparently distinct hydrogeomorphic contexts. Such dynamic patterns underscore the need for better and more explicit linkages between the controls of riparian hydrology and more proximal physical cues on biotic communities in order to understand the drivers of spatial variation in riparian ecosystem structure and composition.