Browsing by Author "Robinson, William M L"
Now showing 1 - 20 of 31
Results Per Page
Sort Options
- ItemOpen AccessAnalyses of the results from the island closure feasibility study for the Dassen/Robben and St Croix/Bird Island Pairs(2014) Robinson, William M L; Butterworth, Doug S; Furman, Liam BThe results from the Island Closure Feasibility Study are analysed using the GLMs as set out at the 2010 international stock assessment workshop. Estimates of residual variance for a random year effects GLM for the various penguin response variables are considered to be sufficiently precise to enable power analyses to be conducted to contribute to the evaluation of whether to transition to a full experimental closures programme, so that the feasibility study may be considered successfully concluded. For the Dassen and Robben Islands about 80% of the estimates of the fishing effect parameter λ are positive, with this same proportion maintained for those (about one sixth) of these estimates which are significant at the 5% level. Thus the preponderance of the evidence from these analyses is that the impact of fishing around these islands has been positive. The rather fewer instances available to analyse for the Eastern Cape colonies suggest a weakly positive effect at Bird Island, but a somewhat stronger negative effect at St Croix. The power analyses suggest that in cases for Dassen and Robben Islands where further data collection might render currently non-significant λ estimates significant at the 5% level within the next two decades, the likely period required for such further collection would typically be in the vicinity of five years. The advantage provided by continuing the closure programme itself seems however to be slight, as the natural variation over time in normal catches would be sufficient to provide the contrast to achieve such results with only typically two years’ extension to those five years. For the two Eastern Cape islands, it seems that results which are statistically significant at the 5% level seem unlikely to be achieved in the foreseeable future – a result which may be a consequence of the relatively low levels of sardine catches typically taken close to those islands.
- ItemOpen AccessAn analysis of the small scale surveys of anchovy abundance around Robben and Dassen islands from 2009 to 2013(2014) Robinson, William M L; Butterworth, Doug SThe results from the small scale hydroacoustic surveys of the abundance of anchovy around Robben and Dassen islands over the 2009-2013 period are analysed under the assumption of a Gaussian form for the trends in density at each island over the course of the winter months. Based primarily on AICc, the model selected from amongst a number of variants has the same trend in abundance with year for the two islands, compatible with the assumption used by Robinson (2013) in his GLM analysis of the impact of closures to pelagic fishing around these islands on penguin recovery, though the data have limited power to distinguish deviations from that assumption. The abundance estimates from the island surveys, though compatible also with the May recruitment survey trends, show appreciably larger variance. This raises the question of whether these small scale surveys merit continuation, unless it is possible to increase their frequency considerably during the winter months each year to improve the overall precision of the integrals over local abundance which they can provide.
- ItemRestrictedAssessing the adequacy of current fisheries management under changing climate: a southern synopsis(Oxford University Press, 2011) Plagányi, Éva E; Weeks, Scala J; Skewes, Tim D; Gibbs, Mark T; Blamey, Laura K; Soares, Muri; Robinson, William M L; Poloczanska, Elvira S; Norman-Lopez, AnnaClimate change is likely to have a significant impact on both target and non-target marine stocks worldwide, with the concomitant need for management strategies capable of sustaining fishing in future. We use several southern hemisphere fisheries to highlight the likely impacts of climate change at a range of levels, from individual to population responses, as well as ecosystem ramifications. Examples span polar (Antarctic krill fishery), temperate (west coast pelagic fishery, abalone and rock lobster), and tropical (Torres Strait rock lobster) commercially important fisheries. Responses of these fisheries to either past observed environmental changes or projected future changes are used to deduce some anticipated implications of climate change for fisheries management, including economic impacts and governance considerations. We evaluate the effectiveness of current single-species assessment models, management strategy evaluation approaches and multispecies assessment models as future management tools to cope with likely climaterelated changes. Non-spatial stock assessment models will have limited ability to separate fishery effects from the impacts of climate change. Anthropogenic climate change is occurring at a time-scale relevant to current fisheries management strategic planning and testing. Adaptive management frameworks (with their feedback loops) are ideal for detecting and adapting to changes in target stocks
- ItemMetadata onlyAssessing the adequacy of current fisheries management under changing climate:a southern synopsis(ICES Journal of Marine Science, 2012) Plagányi, Éva E; Weeks, Scala J; Skewes, Tim D; Gibbs, Mark T; Blamey, Laura K; Soares, Muri; Robinson, William M L; Norman-Lopez, AnnaClimate change is likely to have a significant impact on both target and non-target marine stocks worldwide, with the concomitant need for management strategies capable of sustaining fishing in future. We use several southern hemisphere fisheries to highlight the likely impacts of climate change at a range of levels, from individual to population responses, as well as ecosystem ramifications. Examples span polar (Antarctic krill fishery), temperate (west coast pelagic fishery, abalone and rock lobster), and tropical (Torres Strait rock lobster) commercially important fisheries. Responses of these fisheries to either past observed environmental changes or projected future changes are used to deduce some anticipated implications of climate change for fisheries management, including economic impacts and governance considerations. We evaluate the effectiveness of current single-species assessment models, management strategy evaluation approaches and multispecies assessment models as future management tools to cope with likely climaterelated changes. Non-spatial stock assessment models will have limited ability to separate fishery effects from the impacts of climate change. Anthropogenic climate change is occurring at a time-scale relevant to current fisheries management strategic planning and testing. Adaptive management frameworks (with their feedback loops) are ideal for detecting and adapting to changes in target stocks
- ItemOpen AccessChecking the penguin population model estimator for bias(2010) Robinson, William M L; Butterworth, Doug SA bootstrap approach is used to check the Robben Island penguin model estimator bias. The results show that there is little evidence for this. The distributions of estimates from the bootstrapping process are well-centred on the penalized maximum likelihood estimates.
- ItemOpen AccessComments on the benefit to penguins of fishing restrictions around Robben Island predicted by Weller et al. Robben Island penguin model simulations(2014) Robinson, William M L; Butterworth, Doug SThe following comments relate to results in a recent publication (Weller et al., 2014) that are pertinent to the discussions of the Island Closure Task Team, specifically the impact of fishing around Robben Island on the penguin population. Three scenarios were explored in that paper, each starting with 3500 adult penguins.
- ItemRestrictedEcosystem modelling provides clues to understanding ecological tipping points(Inter Research, 2014) Plagányi, Éva E; Ellis, Nick; Blamey, Laura K; Morello, Elisabetta B; Norman-Lopez, Anna; Robinson, William M L; Sporic, Miriana; Sweatman, HughEcological thresholds, associated with abrupt changes in the state and organisation of ecosystems, challenge both scientists and managers. Adaptive response to such changes, and planning for their occurrence, requires an understanding of the underlying drivers and system responses as well as appropriate monitoring. In addition to field studies, modelling can advance our ability to anticipate or deal with such major ecosystem shifts. Here, we used an existing multispecies model with smooth continuous functions that were modified to include thresholds representing 3 alternative scenarios of predator responses when prey numbers drop below a critical threshold: (I) no threshold-like response; (II) an abrupt decrease in breeding success by 90%, and (III) an abrupt halving of adult survival. Second, we analysed field observations from 3 independent marine case studies (abalone, starfish, penguins) for evidence of abrupt non-linear responses of predators to changes in abundance of principal prey. Third, we compared the model output with empirical results and tested (using both a statistical method and by fitting multispecies models) the 3 alternative response scenarios. With this approach, we found evidence for nonlinear changes in population parameters (such as survival rate) of predators as prey numbers declined below critical thresholds. As an example of the potential for this approach to inform management, we found that abundances of a range of marine predators become more variable as prey numbers decline, which may be a useful indicator that a system is approaching a tipping point.
- ItemOpen AccessEvaluating the accuracy of hake abundance index predictions using different smoothing techniques(2013) Butterworth, Doug S; Robinson, William M L; Ross-Gillespie, AndreaAlternative projection approaches based on linear and quadratic smoothing are applied retrospectively to abundance indices for hake to ascertain whether they can provide more accurate predictions of resource status one and two years ahead, in line with ideal needs for the empirical OMP used to set hake TACs. The results suggest that there is probably little if anything to be gained from attempting more complex formulations than three-year averages of abundance indices for input to a target-based empirical OMP for SA hake.
- ItemOpen AccessFitting both moult counts and tagging data to a population model for Robben Island penguins(2009) Robinson, William M L; Butterworth, Doug SA population model for penguins at Robben Island is fit to both moult count and tagresighting data. The fit is reasonable and provides an estimate of the over-dispersion of the resighting data. This is important for generating resightings in planned simulations for analysing the power of a future tagging programme to detect a change in mortality rate linked to the presence or absence of pelagic fish in the vicinity of Robben Island.
- ItemOpen AccessFull description of the Robben Island penguin-fish interaction model and base case Bayesian results(2011) Robinson, William M L; Butterworth, Doug SWork on the penguin–fish interaction model has progressed to the point where a reasonable base case for Robben Island has been achieved.
- ItemOpen AccessFurther results from future projections of a single or two sardine stock operating model, with initial results using a two-area directed sardine TAC(MARAM, University of Cape Town, 2013) De Moor, Carryn L; Butterworth, Doug S; Robinson, William M Lde Moor and Butterworth (2013a) showed some initial projections for future average directed catch and risk to the resources, assuming a no catch scenario, Interim OMP-13 v2, or a similar Management Procedure (MP) with re-tuned control parameters. Results were shown assuming either a single or a two sardine stock operating model.
- ItemOpen AccessGLMs relating penguin demographics to pelagic catches close to islands and to pelagic abundance(2010) Robinson, William M L; Butterworth, Doug SEarlier GLM analyses of the impact of pelagic fishing in the vicinities of Robben and Dassen Islands on the dynamics of penguins breeding on those islands are extended to cover a wider set of data series, and to consider relationships involving sardine and anchovy separately as well as together. This in turn allows the estimation of the change in penguin population growth rate to be expected from suspending pelagic fishing in the vicinities of these islands. Interpretation of results is confounded by poor precision which is a consequence of the shortness of the time series. Likely the most that could be said with some confidence is that the results of the analyses do not support the hypothesis that suspending fishing around Robben and Dassen Islands would enhance penguin reproductive success there.
- ItemOpen AccessIllustrative outputs of the age-structured model of African penguin populations for linking to the pelagic OMP testing process(2008) Robinson, William M L; Plagányi, Éva E; Butterworth, Doug SGiven the move towards adopting an ecosystem approach to fisheries in the pelagic sector, the new joint OMP needs to be tested in the light of not only the risk parameters as considered previously, along with catch statistics, but also parameters denoting risk to the African penguin population(s) Spheniscus demersus. Penguins have been chosen as a key predator species to consider because of their conservation status, and because of their potential sensitivity to changes in pelagic fish abundance and distribution as a consequence of their land-based breeding sites. A model of penguin dynamics has been developed for use as a penguin Operating Model to be coupled to the pelagic fish OMP. This paper summarises the base-case penguin model and proposed method for use in evaluating the impact on penguins of predicted future pelagic fish trajectories under alternative harvest strategies (OMPs).
- ItemOpen AccessIsland closure feasibility study power analysis(2014) Robinson, William M L; Butterworth, Doug SThis paper indicates how it is planned to extend the general linear model (GLM) analyses by Robinson (2013) to estimate the power of the Island Closure Experiment. Statistical power reflects the probability that an experiment will detect an effect if it exists.
- ItemOpen AccessIsland closure feasibility study power analysis results for Dassen and Robben islands(2014) Robinson, William M L; Butterworth, Doug SThe power analysis for penguin responses to fishing around Dassen and Robben islands has been completed. Response variables considered are chick condition, active nest proportion, fledging success, chick growth, foraging path length, and foraging trip duration. Although complete results are given here only for the first two of these variables, overall there are clear indications regarding the residual variances associated with these variables. Those variances are now estimable for Dassen and Robben islands with sufficient reliability to allow a decision on whether an experimental closure programme could yield definitive conclusions regarding the impact of fishing close to island colonies on penguin demographics within a realistic time span. Thus the purpose of the feasibility study has been achieved and the study can be concluded, with the island closure experiment commenced for these two islands. Definitive results from the experiment can be expected for both islands with about 2–4 years’ additional data.
- ItemOpen AccessPenguin island closure feasibility study analysis results update: random effects models applied to both Western and Eastern Cape islands(2014) Robinson, William M L; Butterworth, Doug SThis Addendum updates results from previous papers using random effects instead of fixed effects models for the year factors. Results are now given for Bird and St Croix islands in addition to those for Dassen and Robben islands. The results from random effects models are qualitatively unchanged from those given earlier from the fixed effects models, except that the periods required to obtain statistically significant results are extended somewhat as a result of the removal of the negative bias in the residual variance estimates for the earlier fixed effects models.
- ItemOpen AccessPenguin model update: an improved penguin mortality relationship with sardine biomass and results of sensitivities to the base case(2011) Robinson, William M L; Butterworth, Doug SWork on the penguin–fish interaction model has progressed to the point where a candidate base case for Robben Island was presented in document FISHERIES/2011/SWG–PEL/40. However, one concern there was that the likelihood could always be increased by increasing the value of the power parameter n in the biomass–mortality relationship. This indicated that the functional form of that relationship was perhaps not the most appropriate. Here, a new piecewise-linear form has been implemented and proposed as a new base case as it seems generally satisfactory.
- ItemOpen AccessPenguin population models for Robben Island(2010) Robinson, William M L; Butterworth, Doug SAs an initial illustrative exercise, a population model for Robben Island penguins, which includes dependency of reproductive success and survival rate on pelagic fish abundance, is fitted to moult count information for the colony. The results indicate a strong dependence of adult penguin survival rate on sardine abundance west of Cape Agulhas. However, the logistic transformation used to ensure respect of biological constraints on these demographic parameters leads to some problems in extending the approach to Bayesian estimation. Further work will explore use of the beta distribution and the incorporation of further data for the penguin population in the fitting process.
- ItemOpen AccessPenguin survival estimates from tag data using a multinomial likelihood(2010) Robinson, William M L; Butterworth, Doug SA simple population model for Robben Island penguins is considered which incorporates fitting to both moult counts and tag data. The latter are now fit using a multinomial likelihood which is the method used in program MARK. Probability intervals on survival rates are more reliable than those obtained previously using an over-dispersed Poisson likelihood. Furthermore, the incorporation of a prior for relative undercounts of juveniles in the moult counts generally moves penalized likelihood estimates for adult survival rate away from an upper constraint boundary.
- ItemOpen AccessProjections of the Robben Island African penguin population(2012) Robinson, William M L; Butterworth, Doug SThe Robben Island penguin population model was described in the working group document FISHERIES/2012/JUL/SWG–PEL/35. This paper shows projected penguin numbers under various scenarios concerning the future abundance and distribution of sardine.