Natural variability of benthic species composition in the Delaware Bay

Biological monitoring of aquatic biota is used to assess the impact of changes in the environment. Critical to the development of a sound biological monitoring protocol is the judicious selection of organisms and organism characteristics to be monitored. Accurate interpretations of change necessitate description of the natural variability of the system. We introduce a state-space model for compositional monitoring data, and illustrate how one can incorporate spatial structure and covariates to assess natural variability. The methods are illustrated on benthic survey data from Delaware Bay, and applied to proportional composition at the genus level. The distribution of benthic macroinvertebrates in Delaware Bay depends significantly on salinity. There is residual spatial dependence in the data after accounting for the salinity effect.