The hardbottom communities of Biscayne Bay, Florida, include populations of several commercial sponge species. These sponges have been subjected to harvesting, a major hurricane, trawling damage, and other stressors during the past 10 years. We developed a size-based matrix population model of the dominant commercial sponge species (Spongia graminea) as a tool to aid in assessing sponge population viability and potentially managing the sponge harvest. Repeated measurements of tagged sponges allowed estimation of growth, survival and fragmentation probabilities. Fecundity was estimated by solving for values of the reproductive size classes that matched the observed population growth when coupled with the observed transition probabilities. Three fecundity functions were applied to the sponge model: constant, a linear function of size class, and a function of sponge volume. All three models indicated a long-term decline in the sponge population (λ < 1), attributed to poor recruitment, and significant vulnerability to harvesting. The responses were similar when either 7 or 2 size classes were assumed to be reproductively functional, and when using different methods and different assumptions of size-class distributions to fit the fecundity values. Models fit with the assumption of an open population (recuitment from outside the population) and with fecundity values increased to achieve a λ of 1 were less sensitive to harvest, but did not match the observed population dynamics.
