Resumen de Trophic cascade modelling in bathyal ecosystems
Different aspects of the bathyal ecosystem have been treated with the final aim to investigate trophic cascades driven by fishery (top-down control), a phenomenon that in bathyal systems has never been studied. This was exemplified by the in-depth study of the exploited soft-bottom continental slope off Catalonia (NW Mediterranean), where the fishery of the red shrimp Aristeus antennatus is carried out and the main structure of the trophic web is known.
In Part I the infaunal assemblages inside the canyon and on the adjacent slope have been related to environmental variables. A seasonal variability has been detected along the year related to the primary production, river efflux and the occurrence of the Levantine Intermediate Waters (LIW). All variables, used as proxy for these three causes, explain the seasonal variability in both assemblages. The canyon was found to be more influenced by seasons and the terrigenous provision than the open slope, that conversely was more related to the primary production and showed less evidence of seasonality. The canyon showed high variability in biomasses, taxonomic composition and trophic types, whereas the adjacent slope was dominated by sub-surface deposit feeders and in February by the carnovorous caudofoveats eating on foraminifera. This study was mandatory in order to parametrise the models in Part III.
Part II concerns a long term study of A. antennatus LPUE performed through both frequentist and Bayesian additive regression models. The variability explained by a total of six predictors captured the 43% of the total LPUE variability. The set of fishery-related variables (the daily trips performed by vessels, the gross registered tonnage and the factor vessel) was the most important source, with an explained deviance (ED) of 20.58%, followed by temporal (ED=13.12%) and finally economic variables (the ex-vessel shrimp price, ED=9.30%). We found that data derived from fishery, as well as independent data provided by scientific surveys, provide similar indices for the exploited species. The study also showed that it may be appropriate reduce the limit of the number of trips per month for a reasonable management. Finally has been shown that a mixed effect model permits to account for correlated data and if vessels hold random effects, there is no need to use them for standardization purposes.
Part III covers the ecosystem approach of this thesis. A total of 40 carbon flows among 7 internal and 6 external compartments, were reconstructed using linear inverse modelling (LIM). The total carbon flux to the community was 2.62 mmol C m-2 d-1, entering as vertical (5.2E-03 mmol C m-2 d-1) and advective organic matter (2.60 mmol C m-2 d-1). The influx was then partitioned between the total organic matter (TOM) in sediment, 87.05% and suspension feeding (12.95%) mainly by the macrobenthos (95.74%). The fate of carbon deposited in sediments was its burial, its degradation or the ingestion by metazoan deposit feeders. The ingestion of C in sediments by the fauna was the 36.34%, the rest was used by the prokaryotes and nematods (69.28%) or trapped in the sediment (32.19%). The total respiration was 1.89 mmol C m-2 d-1, of which the 83.75% was represented by sediments, including prokaryotes and meiofauna.
The dynamic simulation based on a system of ordinary differential equations predicts biomass trends during 5 years after perturbations induced by red shrimp fishery (top-down driver) and food supply (bottom-up driver). We only found very ephemeral indirect effects induced by fishery, persisting less than 10 days, that we considered not enough to demonstrate the occurrence of trophic cascade in the system. On the contrary we found effects driven by source limitation. We explain these results with the important role of detritus of allochthonous origin, that we hypothesised to be the component controlling the food web dynamics. A mechanism known as donor-control.
