Developing physical surrogates for benthic biodiversity using co-located samples and regression tree models: a conceptual synthesis for a sandy temperate embayment

Marine physical and geochemical data can be valuable surrogates for predicting
the distributions and assemblages of marine species. This study investigated the
bio-environment (surrogacy) relationships in Jervis Bay, a sandy marine embayment
in south-eastern Australia. A wide range of co-located physical data were analysed
together with biological data, including multibeam bathymetry and backscatter surfaces
and derivatives, parameters that describe seabed sediment and water column
physical/geochemical characteristics and seabed exposure. Three decision tree models
and a robust model selection process were applied. The models for three diversity
indices and seven out of eight infaunal species explained 32–79% of the variance.
A diverse range of physical surrogates for biodiversity were identified. The surrogates
are presented in a conceptual model that identifies the mechanisms that potentially
link to biodiversity patterns. While some surrogates may exert direct influence over
organisms to exposure and chlorophyll-a, for example, most pointed to complex
relationships between multiple biological and physical factors occurring in different
process domains/zones. The reliable bio-environment relationships identified from
co-located samples and conceptual models enabled a mechanistic understanding of benthic
biodiversity patterns in a sandy coastal embayment that may have implications for
marine environmental management.

Document type: