Testing the robustness of a linked-ecosystem trophic model of the Great Barrier Reef, World Heritage Area

Abstract

This presentation summarises the results of an international collaboration between the University of British Columbia, Aquatic Ecosystems Research Laboratory, and the Department of Primary Industries and Fisheries, Queensland. The project represented a unique opportunity to test the robustness (c.f. validation) of a large scale ecological model of one of Australia's iconic natural wonders; the linked-ecosystems of the Great Barrier Reef. (GBR) The original ECOPATH EwE©, GBR ecosystem models were based on survey data of 1000 taxa from a 10,000 sq nautical mile cross-shelf transect of the far northern Great Barrier Reef. Although large, this area is only a small proportion of the GBR in its entirety. In 2007 however, we completed comprehensive surveys of the entire reef as part of the CSIRO, AIMS, QM, DPI&F, ReefCRC; “Mapping seabed biodiversity of the Great Barrier Reef”. These surveys provided a larger dataset to test the models against; i.e., to validate them. Not only is validation of large-scale ecological models rare anywhere in the world, but the validated GBR models would potentially provide one of the few tested ecosystem based management (EBM) tools available. Two methods of testing the models were trialled: 1. Various scenarios of fishing effort profiles, time-scales, and tuning the model for species vulnerabilities were trialled to attempt to understand the dynamics of the model. Following this painstaking process, visual comparisons were run for the temporal simulations of biomass profiles against the ensuing 10 year time-series of logbook CPUE for species groups taken in the Trawl, Reef-line, and inshore Gill Net fisheries (as the estimate of “observed” changes in relative biomass of those species groups). 2. Spatial expansion of the GBR Linked-ecosystem model to a more realistic GIS basemap of the whole GBR. Habitat type from the smaller GBR Linked-ecosystem cross-shelf simulations were assigned to depth contours that were available in the larger GBR GIS Map. Visual comparisons were then possible between the models predicted spatial distributions of species groups over the whole GBR, against those observed from survey data obtained by the “Mapping seabed biodiversity of the Great Barrier Reef” project. Within a tightly controlled set of constraints the current model does provide useful predictions of the cumulative impacts of the Trawl, Reef-line, and inshore Gill Net fisheries on the “virtual” GBR cross-shelf. The expansion of the model to the full GBR will require a re-specification of the basic model; however, useful predictions can be made of the impacts of fishing if these are treated as hypotheses that need independent testing. The latter is probably a general caveat that should be applied to the output from any model.