Understanding the implications of a changing environment on harvested bivalve populations using habitat suitability models

Habitat suitability models are useful to forecast how environmental change may affect the abundance or distribution of species of interest. In the case of harvested bivalves, those models may be used to estimate the vulnerability of this valued ecosystem good to stressors. Using literature-derived natural history information, rule-based habitat suitability models were constructed in a GIS for several bivalve species (Clinocardium nuttallii, Mya arenaria, and Tresus capax) that are recreationally and commercially harvested in NE Pacific estuaries. Spatially-explicit habitat maps were produced for Yaquina and Tillamook estuaries (Oregon) using environmental data (salinity, depth, sediment grain size, and burrowing shrimp density) from multiple studies (1960-2012). Habitat suitability values ranged from 1-4 (lowest to highest) depending on the number of environmental variables that fell within a bivalve’s tolerance limits. The models were tested by comparing the observed distribution of bivalves reported in benthic community studies (1996-2012) to the range of each suitability class. Results primarily showed that habitats of highest predicted suitability contained the greatest proportion of bivalve observations and highest population densities. Our model was further supported by logistic regression analyses that showed correspondence between predicted habitat suitability values and logistic model probabilities. We demonstrate how these models can be used as tools to forecast changes in the availability of suitable habitat for these species using projected changes in salinity and depth associated with environmental change scenarios. The advantage of this approach is that disparate, independent sets of existing data are sufficient to parameterize the models, and to produce and validate maps of habitat suitability. If the models are robust for multiple estuaries and bivalves, resource managers can transfer the approach to data-poor systems with only modest investment.