Identifying indicators of the effects of fishing using alternative models, uncertainty, and aggregation error

The identification of indicators of the indirect effects of fishing is often an issue for fisheries management, particularly if just commercial catch data are available. Complex, intermediate, and simplified qualitative models were produced for a fishery case study off Western Australia to identify potential indicators of ecosystem change attributable to western rock lobster (Panulirus cygnus) extraction and bait input. Models of intermediate complexity were used to identify indicators because they produced the least aggregation error. Structural uncertainty was considered through a series of structurally different intermediate models. These alternate models consistently predicted that extraction of rock lobster may positively impact small fish of low economic value, such as old wife (Enoplosus armatus), footballer sweep (Neatypus obliquus), and king wrasse (Coris auricularis). These small fish were therefore identified as potential indicators of the effects of rock lobster extraction. Small crustaceans (amphipods and isopods) also displayed positive impacts attributable to bait input from the rock lobster fishery and were identified as potential indicators of bait effects. Monitoring of these indicators may aid the detection of ecosystem change caused by the rock lobster fishery.