Multiplex modeling of physical habitat for endangered freshwater mussels

Quantification of the potential habitat available for endangered freshwater mussels can be a challenging task, as habitat use criteria are very complex and often only low numbers of species observations are available. To address this problem in a riverine environment, we developed a concept of a multivariate, multi-scale, and multi-model (multiplex) habitat simulation through combining multivariate time-series analysis of complex hydraulics (CART and logistic regression), micro-scale (River2D), and meso-scale (MesoHABSIM) habitat models, to develop macro-scale management criteria. This concept has been applied and tested on the Upper Delaware River (USA) for the protection and enhancement of existing populations of Alasmidonta heterodon, an endangered freshwater mussel. The physical habitat conditions of approximately 125km of the Delaware River were described using digital aerial imagery and ground-based surveys. The temporal and spatial variabilities of complex hydraulics simulated by a River2D model at 1547 locations were statistically analyzed to select ranges of attributes that corresponded to mussel presence. We applied these criteria to the river's meso-scale hydromorphological unit mappings to identify suitable mesohabitats, which then served as a calibration data set for the coarser scale model. The final meso-scale model's predictions were hydraulically validated offering encouraging results. The meso-scale habitat suitability criteria defined moderately deep, slow-flowing, and non-turbulent hydromorphologic units as providing good conditions for A. heterodon. All three of the developed suitability models (descriptive statistics, CART and logistic regression model) indicated the species preference for hydraulically stable habitats.