Search

Maturity-at-length is a key input to stock assessments when the management objectives are expressed in terms of the size of the spawning output relative to some reference level. Data for rock lobsters, Jasus edwardsii (Hutton, 1875), off Victoria, Australia, are used to estimate logistic relationships between carapace length and the probability of being mature. The analyses are based primarily on mixed-effects models in which the parameters governing maturity-at-length depend on year and region, fitted using a Bayesian hierarchical approach. Maturity-at-length differs among years and regions, and the length-at-50%-maturity increases from west to east and then remains relatively constant. However, the estimates for all years and regions are not equally precise, so there is value in using a mixed-effects approach to allow the years for which the dataset is large to ‘provide support’ for the years for which the data are sparse. The results provide the input needed to conduct assessments of rock lobster populations off Victoria and to form the basis for conducting population projections.

The evaluation of stock-assessment methods, monitoring programmes and harvest strategies for the South East Fishery using the Management Strategy Evaluation approach requires the development of a set of operating models. A generic framework for the development of operating models that includes the possibility of technical interactions among species is outlined. The operating model allows for discarding due to fish size, due to lack of quota and because of mismatches in the Total Allowable Catches (TACs) set independently for the species considered. The importance of technical interactions, mismatches in TACs, and the limits placed by constraints on landed catches and fishing effort is demonstrated by a set of illustrative simulations.

The performance of a set of harvest strategies is evaluated for four of the species in Australia's South East Fishery relative to the objectives for the Australian Fisheries Management Authority. Harvest strategies based on the Integrated Analysis method of stock assessment are found to outperform those based on trends in catch rate. Harvest strategies where the target level of fishing mortality aims at stabilizing the spawner biomass at some pre-specified fraction of its pre-exploitation level appear to outperform those that set Total Allowable Catch (TAC) based on estimates of F MSY. The harvest strategies considered are robust to many of the types of uncertainties considered. The factors that influence performance to the greatest extent include the amount by which landed catches are limited by market demands, the depletion of the resource when the harvest strategy is first applied, the variation in recruitment and productivity. Fairly tight limits can be placed on how much the TAC can be varied from one year to the next and any minimum TAC levels should be low. There appears to be little benefit from conducting assessments (and changing TACs) frequently.

The ability to estimate some of the quantities relevant to the management of South East Fishery species is examined using Monte Carlo simulation. The analyses are based on scenarios for four of the species of this fishery (spotted warehou, tiger flathead, jackass morwong and pink ling). Integrated Analysis was found to perform best out of six potential methods of fisheries stock assessment (Schaefer and Fox production models, age-structured production model, Integrated Analysis, ad hoc tuned VPA and ADAPT) that are, or have been, applied to data for South East Fishery species. However, its performance, particularly for spotted warehou, is nevertheless relatively poor. The sensitivity of estimation ability to many factors including model uncertainty and the extent of observation and process error is examined. Factors that influence estimation performance markedly include: violation of the assumption of no population spatial structure; time-trends in catchability; the depletion of the resource at present; and uncertainty about the value of the instantaneous rate of natural mortality. Little improvement in estimation performance can be expected without resolution of issues related to model structure. This highlights the importance of considering model uncertainty in future studies of the estimation performance of stock assessment methods.

Calls to use target, limit and threshold reference points to achieve precautionary management goals have increased substantially in recent years. For many fisheries there is little chance of conducting formal stock assessments, and indicators of stock status have to be based on changes in measurable quantities such as catch rates and percentiles of the length distribution of the catch. Use of such empirical indicators for management decisions also requires a choice of appropriate reference or trigger levels for the indicators. Monte Carlo simulation can be used to examine the performance of alternative indicators and associated reference points in terms of their ability to correctly identify the biological conditions that they were designed to measure. This study examines the performance of several potential empirical indicators for broadbill swordfish off eastern Australia, taking into account uncertainties in stock structure, current status, and future trends in fishing effort. Indicators based on catch rates are shown to be potentially very misleading. In contrast, indicators based on the mean length or mean weight of the catch perform better, because these quantities change in a more predictable manner with abundance. Nevertheless, reference points based on these quantities are frequently ‘triggered’ either too early or too late.

Information about the age-structure of the catch is one of the primary inputs to many of the methods of stock assessment currently applied to SEF species. Two methods for calculating the catch-, mean length-, and mean mass-at-age matrices from the data collected from the fishery are outlined. These methods are illustrated by applying them to data for blue grenadier, Macruronus novaezelandiae, and eastern school whiting, Sillago flindersi. The assessment and risk analysis results for blue grenadier are highly sensitive to the choice of method, whereas those for eastern school whiting are not. It is recommended that a method that allows for inter-annual variation in mean length- and hence mass-at-age should be the standard for SEF stock assessments but that sensitivity to alternative methods needs to be examined routinely.

A formal approach to estimating selectivity patterns that allows for variability additional to Poisson sampling variability is outlined. This approach, which involves assuming that the catch of fish by a particular size of gill-net in a given size-class is distributed according to a negative binomial distribution, is applied to data for five of the species caught by experimental gill-nets in Australia’s South East Fishery. The data for two of these species (blue warehou and dogfish) are not compatible with the conventional assumption that the size of fish corresponding to maximum selectivity is linearly proportional to mesh size. Selectivity and population size-structure are found to depend on depth and habitat type for some of the species. It is necessary therefore to target future experiments to depth zones/habitat types in which the bulk of the catch is taken to estimate selectivity patterns for use in stock assessments.

The fishery can be divided into two subfisheries (‘spawning’ and ‘non-spawning’). Commercial catch rates for the ‘non-spawning’ subfishery declined from the late 1980s to 1997, whereas those for the ‘spawning’ subfishery exhibit no obvious temporal trend. An ‘Integrated Analysis’ assessment, of the feasibility of reconciling these differing trends, uses catch (landed and discarded), catch rate, length-at-age, and catch-at-age data and estimates of absolute abundance based on the egg-production method. It emphasizes uncertainty due to model assumptions and the data included in the assessment. Use of the discard data allows more precise estimation of the magnitude of recent recruitments. Spawning biomass is estimated to have declined from a peak in 1989–91 to 1999 although fishing mortality has consistently been

Current annual landings of southern rock lobster, Jasus edwardsii, from southeastern Australia are around 5000 tonnes valued at A$140 million. The Victorian component of this catch during the 1998-99 fishing season was 550 t, valued at A$18 million. During the past 20 years catch rates have declined from 0.8 kg per pot lift to 0.6 and 0.3 kg per pot lift in the western and eastern management zones respectively. The fishery has been managed with input controls during this period, but at the time of writing, the direction of future management is not clear. A size-structured model was developed to assess risk associated with both effort (input) and catch (output) controlled harvest strategies in each zone. The stock-assessment model was fitted to historical catch data (in weight and by number) from 1951, catch rates, and the length-frequency by sex. The uncertainty associated with the estimates of exploitable biomass and egg production was assessed according to Bayesian methods. The output of the assessment formed the basis for projections intended to determine the risk associated with different future levels of effort and catch. Reference points based on estimated biomass and egg production relative to the start of the data series in 1951 were considered.

The stock assessment process in the South East Fishery involves scientists, industry, fishery managers, economists and non-government organizations. A comparison with such processes in other countries, where stakeholder involvement ranges from government scientists only to involvement of scientists, industry and conservation non-government organizations, suggests that Australia is the only country in which fishery managers are active and integral stock-assessment participants. In Australia, as in several other countries, the form of advice is comparative (consequences of alternative decisions) rather than prescriptive. Although all approaches have advantages and disadvantages, the South East Fishery process has advantages that appear to clearly outweigh the disadvantages. The advantages include better communication among interest groups, improved ownership of and hence support for outcomes and better interactions among groups. The disadvantages include the cost of the process, lack of consistency among assessments, vulnerability of scientists and the frustrations of industry.

A spatially aggregated age- and sex-structured population dynamics model was fitted to standardized catch-rate data from the school shark resource off southern Australia. The model incorporates the peculiarities of shark populations and fisheries, including the pupping process and the selectivity characteristics of gill-nets. Estimates are determined by a Bayesian approach that incorporates prior distributions for virgin biomass, the parameter that determines productivity, and the variation in pup survival. Tests of sensitivity include changing the data series used, varying the value of adult natural mortality, and changing the prior distribution for the productivity parameter. The point estimates of the mature biomass at the start of 1995 range from 13% to 45% of the pre-exploitation equilibrium size, depending on the specifications of the assessment. The results are notably sensitive to the selection of a catch-rate series. Results suggest that the current fishing intensity will lead to further declines in abundance, that a reduction of ~20% in fishing mortality would achieve a 0.5 probability of not declining further, and that a reduction of 42% would achieve with a probability of 0.8 the management goal of not being below the 1996 mature biomass at the start of 2011. Extra keyword: CPUE.

From 1980 to 1995, annual catch rates of southern rock lobster (Jasus edwardsii) in Tasmanian waters declined from 1·6 to 0·9 kg per pot lift. From March 1998, the Tasmanian government intends to base management of this fishery on output controls using individual transferable catch quotas. This has necessitated the development of a stock assessment model that can be used to evaluate the relative merits of a variety of alternative catch limits in terms of the trade-off between protecting the rock lobster resource and achieving a high yield. Each of eight regions around Tasmania is assessed separately because of spatial heterogeneity between regions in lobster growth, size at maturity and catch rates. The stock assessment model is size-structured and is fitted to catch, effort and length–frequency data as well as to estimates of exploitation rate from experimental data. A Bayesian estimation framework is employed to estimate the quantities needed for risk analysis. Egg production differs markedly among the eight regions: from as low as 6% of the unexploited equilibrium level in the northern regions, where the growth rate is fast, to more than 80% in the south-west, where the growth rate is slowest.

Assessment of the southern rock lobster (Jasus edwardsii) resource in Tasmania is based on a size-structured population dynamics model. One of the most important inputs to this model is the set of matrices that represent the season-specific probabilities of a lobster growing from one size-class to another. These matrices are estimated from tag–recapture data within a maximum-likelihood estimation framework. Measures of precision are determined from the asymptotic variance–covariance matrix. Various alternative models are contrasted for one site in the south-east of Tasmania, and a best model is selected by the likelihood ratio test. The growth model used is based on a generalization of the von Bertalanffy growth equation. Growth rates differ markedly among regions around Tasmania, being slowest in the south and fastest in the north. Growth of legal-size males is noticeably faster than that of legal-size females. It is shown that ignoring the effects of selectivity can lead to biased estimates of growth rate. An extension to the method is presented and applied that estimates size-specific selectivity in an attempt to eliminate this bias.

The degree to which a stock is depleted is one of the most important quantities in fisheries management because it is used to quantify the success of management and to inform management responses. However, stock depletion is extremely difficult to estimate, particularly with limited data. Using the RAM Legacy database, we developed a boosted regression tree (BRT) model to correlate depletion with a range of predictors calculated from catch data, making the model usable for many fisheries worldwide. The most important predictors were found to be catch trends obtained from linear regressions of scaled catch on time, including regression coefficients for the whole catch time series, the subseries before and after the maximum catch, and in recent years. Eight predictors explain about 80% of variation in depletion. There is a correlation of .5 between measured levels of depletion and the predictions of the BRT model. Predictions are less biased when the stock is fished down below half of the carrying capacity. The BRT model outperforms comparable existing catch‐based depletion estimators and could be used to provide priors for depletion for data‐poor stock assessment methods, or used more directly to provide estimates of the probability that depletion is below a given threshold value.

Assumptions about the future productivity of a stock are necessary to calculate sustainable catches in fisheries management. Fisheries scientists often assume the number of young fish entering a population (recruitment) is related to the biomass of spawning adults and that recruitment dynamics do not change over time. Thus, managers often use a target biomass based on spawning biomass as the basis for calculating sustainable catches. However, we show recruitment and spawning biomass are not positively related over the observed range of stock sizes for 61% of 224 stocks in the RAM Legacy Stock Assessment Database. Furthermore, 85% of stocks for which spawning biomass may not drive recruitment dynamics over the observed ranges exhibit shifts in average recruitment, which is often used in proxies for target biomasses. Our results suggest that the environment more strongly influences recruitment than spawning biomass over the observed stock sizes for many stocks. Management often endeavours to maintain stock sizes within the observed ranges, so methods for setting management targets that include changes within an ecosystem may better define the status of some stocks, particularly as climate changes.

The crown-of-thorns starfish (COTS) Acanthaster planci is one of the largest causes of coral cover loss on Australia's Great Barrier Reef. Numerous trophic-mediated and anthro- pogenic-related hypotheses, singly or in combination, have been proposed to explain COTS out- breaks, but the evidence remains inconclusive. There is, however, an urgent need for effective control and mitigation strategies. We develop a model of intermediate complexity for ecosystem assessments to describe the trophic interactions between juvenile and adult COTS and 2 groups of coral (fast- and slow-growing corals). By fitting to available data, the impact of prey-switching behaviour by COTS and its effects on the system are quantified. Simulations are used to evaluate the effects of (1) predation by large fish on adult COTS, (2) predation by benthic invertebrates on juvenile COTS, and (3) manual removal of adult COTS. The results highlight the effectiveness of invertebrate predation at reducing juvenile COTS numbers and suggest that manual removal is unlikely to be an effective control method except on a small scale. Our model focuses on the importance of trophic interactions in COTS control and provides an initial platform for future work which foresees the estimation of the effects of key predators, no-take areas, anthropogenic nutrient enhancement, and reef connectivity in mediating the spatio-temporal heterogeneity of COTS outbreaks.

Three approaches are commonly used to fit surplus production models to observed data: effort-averaging methods; process-error estimators; and observation-error estimators. We compare these approaches using real and simulated data sets, and conclude that they yield substantially different interpretations of productivity. Effort-averaging methods assume the stock is in equilibrium relative to the recent effort; this assumption is rarely satisfied and usually leads to overestimation of potential yield and optimum effort. Effort-averaging methods will almost always produce what appears to be "reasonable" estimates of maximum sustainable yield and optimum effort, and the r2 statistic used to evaluate the goodness of fit can provide an unrealistic illusion of confidence about the parameter estimates obtained. Process-error estimators produce much less reliable estimates than observation-error estimators. The observation-error estimator provides the lowest estimates of maximum sustainable yield and optimum effort and is the least biased and the most precise (shown in Monte-Carlo trials). We suggest that observation-error estimators be used when fitting surplus production models, that effort-averaging methods be abandoned, and that process-error estimators should only be applied if simulation studies and practical experience suggest that they will be superior to observation-error estimators.