Patterns of Abundance and Community Dynamics in Atlantic Coastal Sharks

Broad scale analyses of shark population and community dynamics are particularly challenging given the complex life history strategies employed and their vast migratory patterns. Consequently, studies are generally limited to analyzing small-scale, localized dynamics that can be examined from easily accessible, nearshore environments. In particular, fishery-independent shark surveys are frequently limited by spatial political boundaries, such that they only sample a discrete portion of a migratory coastal shark’s distribution. Given the age- and sex-structured movements of these species, a localized survey is likely unable to represent stock-wide changes in abundance, such that several small ranging surveys are treated as independent measures of abundance. Survey-based trends in abundance frequently display data conflict, likely due to high levels of uncertainty and variable timing in migrations. Similarly, sharks within communities interact, with the capacity of one species to alter the population size and growth rate of another species. However, these interactions have never been assessed at a wide geographic scale. In the current thesis, I used generalized linear models (GLMs) to estimate annual indices of abundance from eight species of Atlantic coastal sharks from six fishery-independent surveys along the U.S. east coast and within the Gulf of Mexico. These conflicting indices of abundance were input into a dynamic factor analysis (DFA) model with large-scale climatic indices and anthropogenic forces as covariates to produce simplified species-specific trends of abundance for each species throughout the sampled distribution. These common trends were then input into a multivariate, first-order autoregressive, state-space (MARSS-1) model to estimate interspecies interactions and density dependence. These broad-scale interactions were compared to localized interactions generated from conducting MARSS-1 analyses on GLM-based indices of abundance calculated from individual surveys. Resulting DFA common trends suggested that large coastal species followed similar patterns of abundance since
1975, where abundance was high at the beginning of the time series, declined into the early 1990s, was depressed for a length of time corresponding to age at maturity, and then showed initial signs of rebounding. The small coastal species showed more regional variability in abundance, likely due to separate Atlantic and Gulf of Mexico stocks for several of these species. Broad-scale community analysis results showed that seven out of ten coastal shark populations exhibited density dependence, and an additional seven interspecies interactions were identified that significantly influence the population growth rate of affected species. The localized, survey-specific MARSS-1 modeling results produced different results, suggesting that small scale results cannot be extrapolated across the entire stock. Nevertheless, results from these survey-specific models greatly assisted interpretation of the large scale results. Overall, by analyzing coastal shark population and community dynamics from a broader perspective, we can interpret broad trends in abundance and account for interactions that were previously unknown. These results may assist in assessment efforts by reducing conflicting information input into stock assessment models, and accounting for community relationships that may affect population growth rate of various species.