WAMSI Node 4.4.2-3 - Spatially explicit assessment models for the west Coast demersal scalefish fishery

Brief description

The capacity of management areas to sustain abundances of exploited species depends upon the dynamics of recruitment, and in particular the extent of dispersal between areas and the age classes contributing to dispersal. Like other marine fishes with pelagic larval stages (ca. the first 20-30 days), the dispersal of larval dhufish is likely to be profoundly influenced by oceanic currents. There is extensive evidence that the pole-ward flowing Leeuwin Current is responsible for the long-distance transport of marine fauna in the south-western Australian area (Hutchins & Pearce, 1994, Caputi et al., 1996, Beckley et al., 2009, Pearce & Hutchins, 2009). In addition, during the austral summer (December-February), when most spawning occurs in dhufish, wind driven currents drive flows along the west coast northwards and inshore of the Leeuwin current (the Capes Current; Pearce & Pattiaratchi, 1999, Feng et al., 2010), and these are believed to transport the larvae of marine fishes (Lenanton et al., 1996, Lenanton et al 2009b). The dhufish is characterised by inter-annual variation in recruitment (Lenanton et al., 2009), and it has been speculated that this may in part reflect variation in the strength of the prevailing currents.

The aim of the study, in including alternate hydrodynamic conditions, was to explore the potential influence of a range of realistic larval dispersal scenarios on dhufish population dynamics.

Lineage

Statement: A simulation model, known as ELFSim, was used.

ELFSim is a decision support software system designed to evaluate the implications of different hydrodynamic conditions in the WCDSF (West Coast Demersal Scalefish Fishery). We modelled the transport of larval fish with a Lagrangian particle tracking simulation nested within a global hydrodynamic model generated from the BlueLink Reanalysis (BRAN2; Schiller et al., 2008). BRAN is based on the Ocean Forecasting Australia Model (OFAM) of ocean circulation, which has a resolution of 1/10º in the Asian-Australian area. It assimilates observations from satellite SST, altimetric sea-level anomalies, and temperature and salinity profiles from the Argo float array (Schiller et al. 2008). Comparison of the outputs of BRAN with surface drifters indicate that it quantitatively reproduces the shelf circulation in the Australian area, with the observed and reanalysed sea level anomalies having a correlation of greater than 0.8 (Schiller et al. 2008). This analysis was facilitated by the ConnIe2 computer program (updated from Condie et al., 2005). ConnIe2 enables estimation of the probability that any two areas are connected by modelled ocean circulation over a specified period of dispersal. It is based on offline particle tracking within a three-dimensional hydrodynamic model and incorporates information from wind fields, temperature, salinity, sea level and tides (see Condie et al., 2005; Condie and Andrewartha, 2008; Schiller et al., 2008). Particles were released at 24 approximately equally-spaced locations between Kalbarri and Albany and within a transect composed of three 10 x 10 km cells running perpendicular to the shore. Although dhufish are known to spawn throughout their range (Hesp et al., 2002), little is known of the typical depth at which they spawn and there is no information on the depths at which larvae settle. Two types of model scenarios were examined. First, four different dispersal scenarios were examined that corresponded to the (strong) Capes Current dispersal scenario, a Weak Capes Current dispersal scenario, a Uniform dispersal scenario where all spatial locations have the same dispersal proportions regardless of the distance between the populations, and lastly a No dispersal scenario, where all recruitment to a population relies on the reproductive capacity of that population.

Second, six management scenarios were examined corresponding to possible long term effort that might occur in the WCDSF. These management scenarios include a Status Quo scenario which was based on current effort controls in the fishery which equated to 0.5 of 2006 effort levels across the fleets (i.e. current management arrangements). Three other multiples of this were examined, which include a decline in these effort levels by 0.5 (equivalent to 0.25 of 2006 effort levels), and increases of 50%, and 100% over the Status Quo as well to allow for receovery of the stocks and an increase in effort (and catches) in the future. For comparison purposes we also added a Status Quo scenario that had a 2% per annum increase in catchability, and a scenario that allowed the stock to recover from fishing at its highest rate (0 effort).

For each dispersal scenario, 10 independent initialisations of the model were generated, meaning that there were 10 different starting conditions of the model. Under each management scenario, each of these starting conditions was projected from 2011 to 2025 ten times, which gave ???100 replicated model conditions from which the results are shown.

The results are reported as the average of the final five years (2021-2025) of the projection.

Performance indicators presented include spawning biomass, catch, effort and CPUE in each of the four management areas of the fishery. Various comparisons therefore were made across management areas, management scenarios and dispersal scenarios.