Abstract

Background/Question/Methods Overfishing highly threatens marine ecosystems with over 80% of marine taxa depleted. Scientific research has demonstrated that overfishing decreases the abundance and body size of individual species and accelerates their maturation. However, we still lack understanding and predictive power on how these impacts propagate through the trophic network, affect other species and alter the structure and dynamics of the whole community. We help develop such understanding by evaluating how extinctions of species harvested by artisanal fisheries affect the structure of food-webs using the intertidal and subtidal ecosystems of Central Chile as study systems. We compiled data on artisanal commercial loads to map the harvested species and evaluate the inter-dependency between harvested and non-harvested species. We simulated species extinctions from most connected to least connected species and from more harvested to least harvested species and compared the topological characteristics between the original networks with the networks that resulted after the simulated extinctions. Results/conclusion Fishermen harvested 17.8% and 23% of species in the intertidal and subtidal food webs among which 31.6% and 11.7% are among the 30 most connected species of each network. The 67.3% and 80% of non-harvested species in the intertidal and subtidal networks shared at least one link with harvested species. Both food webs followed an exponential degree distribution, suggesting that the networks are vulnerable to the removal of most connected species. When we simulate this removal scenario, the intertidal food-web is more robust to species extinctions than the subtidal food-web with secondary extinctions starting after 31 and 8 species removals and with 50% of species loss after 22% and 34% of species were removed, respectively. The intertidal food-web, however, looses >45% of its links when its harvested species were removed. Despite the strong interdependence between harvested and non-harvested species in both food-webs, neither connectance nor modularity significantly changed when all the harvested species were removed. This null change is explained by the high redundancy of links in both networks. Our work shows how the selective extraction of species by fisheries might spread through the food-web and highlights the limitations of static approaches (lack of population dynamics) to evaluate fisheries impacts.