Life histories, functional traits, and habit connectivity can explain and predict population and ecosystem dynamics. Ecological interpretations and ecosystem management decisions, however, are often limited by modeling approaches based on common species, those for which information about ecosystem roles are established. Rarer or less studied species often lack such information, and are typically ignored. Using fish species within estuarine-coastal mosaics of the northwest Gulf of Mexico, I propose a new method for predicting community and ecosystem dynamics without requiring a priori information for all species. I hypothesized that species with similar ecosystem dynamics should have similar and correlated population trends,, then used time-series correlation among species abundance trends (1985-2013) for 136 fish species, ranging from common to rare, using data from the Southeast Monitoring and Assessment Program in the Gulf of Mexico (SEAMAP-GOM). Here, I present and interpret classifications from post hoc ordination on the resulting covariances to infer roles and traits regardless of a priori knowledge.
Results/Conclusions
Traits known predominantly for common species provided information that generated predictions about rare species. Post-hoc ordination on covariation among fish species suggests that information about ecosystem roles lacking for many species may be inferable without need for a priori classification or information. Non-metric multidimensional scaling (NMDS) discriminated among life history traits such as adult size and reproductive effort, as well as schooling behavior and estuarine use. Trait classification and gradients included phylogenetic correlations but distinguished differently adapted close-relatives. Classifications for species with known traits were highly consistent. I propose that: 1) within-group classifications for rare species are robust and imply shared traits; 2) ecological predictions and ecosystem-based management decisions are possible using these inferences to fill gaps in existing knowledge; and 3) predictions can be directly tested as ongoing research on these species provides new information.