Understanding the structure and resilience of trophic dynamics in the northern Gulf of Mexico using network analysis

Network analysis is a framework that allows integration and evaluation of predator-prey interactions. In the present study, we synthesized diet composition information from 94 published studies (n = 12,335 unique predator-prey interactions) that reported food habits of teleost fishes in the northern Gulf of Mexico (nGOM). Using this information, we constructed 12 weighted trophic network models using three diet metrics and four levels of taxonomic resolution of predators and prey. We evaluated network resilience to simulated random and directed taxa loss by assessing changes in topological indices "complexity," "connectance," "efficiency," and "robustness" with respect to a priori minima. We found all networks were resilient to random removal of nodes. However, the response to directed removal varied depending on the index used to determine node importance. Directed removal simulations that targeted taxa with the greatest number of trophic interactions had the strongest impact on network topological indices. Using an additional simulation, we assessed how removal of taxa of commercial interest impacted the predation pressure on other taxa. We found a greater magnitude of predator diet shifts when Gulf menhaden (Brevoortia patronus Goode, 1878) were removed than when blue crab (Callinectes sapidus Rathbun, 1896) were removed, indicating predators of Gulf menhaden have a more limited portfolio of diet items than do predators of blue crab. Compared to previously described marine trophic networks, the network that describes the trophic dynamics in the nGOM is less connected and complex. This conclusion highlights the need for consistent reporting of stomach contents and improved understanding of the food habits of lesser-known taxa in the region.