Describing species-environment relations for macrobenthic and microphytobenthic community structure using constrained ordination and predicting environmental variables from species composition

The human potential to change the environment and the associated biological communities is well documented. Yet, it is difficult to identify species responses to a suite of environmental variables (acting additively or antagonistically), and additionally, identify anthropogenic impacts. Here, a new constrained ordination technique called Canonical Principal Component Analysis of Hypergeometric Probabilities (C-PCA-H) is described which uses both species and environmental variables to explain patterns of species variation based on the environment. In this method, the species data are transformed that the ordination diagram represents the main patterns of variation in the Chord Normalized Expected Species Shared (CNESS) distances among samples. CNESS is a family of dissimilarity metrics that can be made more or less sensitive to the rare or the dominant species of a community. C-PCA-H is applied to two benthic surveys. The first consists of estuarine benthic invertebrate data and environmental variables measured by the Environmental Monitoring and Assessment Program (EMAP-VP) in the Virginian Province. Salinity, temperature, and depth are the main variables governing species composition. C-PCA-H is used to distinguish natural from anthropogenic impacts and to identify pollution-tolerant and sensitive species. The second survey consists of estuarine benthic invertebrate data and environmental variables measured by the Massachusetts Water Resources Authority (MWRA) to assess the effects of the secondary treatment sewage outfall in Massachusetts Bay. C-PCA-H reveals that depth and grain size are the main variables governing species composition. Long-term changes in community structure are also shown. Three species are identified as pollution indicators. Additionally, past conditions in Boston Harbor are assessed by inferring past environmental variables from fossil diatoms using the Weighted Averaging method. Modern benthic diatoms were sampled from Massachusetts Bay, and analyzed using C-PCA-H. Depth, salinity, and nutrients are the main variables governing species composition and hence appropriate for paleoreconstructions. Fossil diatoms were extracted from a dated sediment-core. Past depth, salinity, and nutrient concentrations are reconstructed from the 1850s to the 1990s revealing key periods in the history of Boston Harbor. Such paleoreconstructions beyond the decadal time-frame offer a methodology to identify natural variability which is vital for the effective management of coastal systems.