1. Responsible management of fisheries requires an assessment of the
success of the management plan in achieving its objectives, together with
an assessment of the state of the fish stock and likely consequences of the
current and alternative management strategies. In many cases, the
management plan is intended to maintain the status quo. Trends in time
series of data, or values that fall outside the range of predicted outcomes,
may indicate that the status quo is not being maintained, or that significant
change has occurred within the system. Cost effective methods are
required to provide rapid feedback to fisheries managers that a major
perturbation has occurred, or that the system is changing, in order that
appropriate management action may be implemented.
2. Need to produce low-cost effective models for stock assessment and catch
prediction of Western Australian fisheries, especially those low-value
fisheries (eg., some finfish fisheries). With few biological assumptions and
simple mathematical forms, time series modeling may significantly reduce
modeling costs including research costs and computing costs. Time series
approaches may also significantly reduce model uncertainty, and therefore
may provide more reliable prediction results.
3. Need to reduce the risk of model failure through inadequate assumptions
regarding biological processes. Models currently used by Fisheries WA
involve often untenable biological assumptions, with the result that
predictions are conditional on the accuracy of the assumptions.
To reduce the risk of model failure through inadequate assumptions, it is
highly desirable to supplement the current models by applying techniques
such as time series methods that make few assumptions regarding the
biological processes and to compare predictions from the two modeling
approaches.
4. Need to incorporate economic data within the models. Little economic
modeling of Western Australian fisheries has been undertaken; and the changing relationships between supply and market demand (especially for the western rock lobster fishery) require study. As an extension of the VOM (Value Optimisation Model for western rock lobster) project (FRDC 97/104), the economic data collected in the earlier project require analysis to determine how these might best be incorporated into the existing models. This research project will address this need by using multivariate time series techniques to identify the relationships among the historical economic data, catch-effort data, exchange rate data, and other related data.
Given the above, time series modeling is seen as a highly valuable and strategic element of the research programme for Western Australian fisheries; the benefits of this project could be transferred Australia-wide to fisheries researchers allowing improved management advice.
Final report
The management of fisheries in Western Australia requires an understanding of the status of the fisheries stocks. For many species, the only available data are catch, effort and CPUE history. For these and other fisheries, time series methods may improve the stock assessment methods. Biological information is expensive to collect, and much of the information required for stock assessment methods such as age-structured models is simply unavailable, especially for low value fisheries. Time series analysis or control charting methods comprise a select few statistical techniques available for the purpose of stock assessment in these cases. Prediction may be improved using time series methods on catch and effort data with or without external data such as biological or environmental variables. Even when biological parameters can be estimated for a given model, time series methods may be superior as prediction tools.
The aim of this research was to apply time series methods on the western rock lobster fishery, several commercial finfish fisheries and the major tiger and king prawn fisheries, and determine how useful these techniques are for fisheries assessment and management.
