Spatial management strategies for deep-sea sharks

At least ten species of deep-sea (> 200m) dogshark (Squaliformes) have undergone major declines in the world's Oceans due to historical over-fishing. Two species were recently protected in Australia (Centrophorus harrissoni and C. zeehaani (Centrophoridae)) and a recovery plan was implemented. The key strategies of this plan are landing bans to prevent targeted fishing and a network of areas closed to all methods of fishing. This is the first closure network to be implemented in the world specifically for the recovery of a vulnerable deep-sea species. Success will take decades because these species are long lived (30+ years) and have only 1‚Äö-2 pups every three years and therefore remain vulnerable, even as bycatch. This thesis develops and applies a novel combination of field based survey methods and model based approaches to support management of these sharks or other vulnerable vertebrates that co-occur with more productive fisheries target species. These include, developing biologically meaningful criteria to measure performance against the broad conservation objectives of the plan ‚Äö- halt decline and support recovery‚ÄövÑvp, choosing appropriate locations and sizes for closures, understanding and limiting if possible the cost to industry due to lost fishery production, and improving monitoring and compliance data. Three principal datasets are used: 1. Genetic taxonomy data for Centrophorus species, used to resolve identification issues in observer records so they can be used to ensure targeted fishing for Centrophorus in Australia has stopped, ensure estimates of fishing mortality are reliable, and ultimately monitor recovery from fishing vessels. 2. Survey data including abundance (catch and effort) and population structure (sex ratios and size structure), used to select suitable locations for closures and provide a baseline abundance estimate to measure recovery. 3. Passive acoustic tracking data, used to measure the home range of C. zeehaani to inform closure size. The protected Centrophorus species are externally very similar to each other and some other non-protected Centrophorus species in Australia. My genetic approach found the 16S mitochondrial gene was able to distinguish six out of seven species from Australia and Indonesia. The remaining two species should not be confused because they occur in different geographic areas. The genetic identification method was tested on ten fin-clip samples provided by fishery observers and found all but one had been correctly identified by observers. This method will provide a means of non-lethal catch verification. Passive acoustic tracking was used to study the movements of C. zeehaani in the largest closure implemented for their protection. An array of 21 moored acoustic receivers monitored 71tagged individuals for 15 months. A general additive mixed model was applied to the data to test environmental variables (mainly light) as fixed effects on shark movement and individual variation as a random effect. Average along-slope range was 19.2 ¬¨¬± 12.2 km and the maximum distance recorded was 75 km over 15 months. Average depths ranged from 340 m at night to 640 m during the day with high individual variation. Detection depth was strongly correlated with seafloor depth. These results indicate a distinct daily movement pattern of synchronous diel migration with (night time ascent). Males tended to leave the closure and most did not return whereas the number of females detected did not vary significantly between months. The management implication of these results is that closures for C. zeehaani need to be 19‚Äö-75 km in size along the upper-slope, cover the 340‚Äö-640 m depth range and be located to protect resident females. A semi-quantitative management strategy evaluation (MSE) approach was developed and applied to C. harrissoni and C. zeehaani to identify and evaluate options for closures at the local and national scales, particularly outside the range of the telemetry data. Population structure (survey) data were used to identify areas where mating and pupping was likely to occur as leading criteria for locating closures. Commonwealth fisheries managers chose options that added new closures to the network and expanded some existing closures even though costs to industry of lost production were high. An individual-based simulation model of the movements of C. zeehaani was developed and applied to determine how long a depleted population would take to recover from its current status of 8% of un-fished numbers to a target of 20%. Individual movement patterns were based on tracking results and simulated across a spatial domain of three closures and fished areas with three different types of fishing gear and conditions. Key uncertainties were length of the female breeding cycle, natural mortality rate and spatial variation in population density. The base case (three year cycle, 2% natural mortality and survey based spatial variation in abundance) predicted recovery in 63 ¬¨¬± 3 years. Poor matching of closure locations to population density would delay recovery by an additional 31.9 years. Sensitivity testing predicted that the target would be reached 19.2 years earlier with a 2-year female breeding cycle or 16.5 years later with a four-year cycle. If natural mortality were half the base case estimate then the recovery target would be reached 13.5 years earlier or, significantly, if the natural mortality rate were double the base case estimate, recovery to the target would take 98.3 years longer than the base case. Improving handling practices for sharks or changing fishing methods on the continental shelf would not significantly affect the time for recovery but re-introducing trawling for orange roughy (Hoplostethus atlanticus) in deep waters would delay recovery by 45.9 years. Doubling the size of a closure where C. zeehaani is abundant would reduce recovery time by 9.9 years; halving closure size there would increase recovery time by 12.6 years. Such changes would have no significant effects where C. zeehaani is not abundant. The model can be used to evaluate the consequences of alternative management interventions and the risks associated with key uncertainties and can be applied to other shark species with telemetry data. Australia has implemented the first detailed recovery plan for a deep-sea species with spatial management as a key strategy. Decision makers were faced with conflicting conservation and resource use objectives and significant scientific uncertainties. This thesis has calculated the appropriate size of closures using linear models applied to telemetry data. Suitable locations for other closures were identified using demographic criteria developed from survey data. Population trends were simulated across the geographic range of a population over decades. Results of this thesis indicate that these species can be conserved but only with high costs of lost fishery production. Recovery will take decades, at least. The methods can be applied to plan conservation interventions for other long-lived deep-sea species.