Objective
The objective of this project is to substantially improve and enrich existing bioeconomic models for the management of multi-species, multi-fleet, multi-gear fisheries that have been developed in Iceland and Italy in previous years as well as contributing to the development of an overall methodology for constructing and applying such models to the practical management of European fisheries.
The methodology combines bio-economic modelling, statistical estimation techniques, extensive computer programming, numerical calculation and maximization techniques.
An appropriate boiological model incorporating sub stock structures (e.g. cohorts), spatial movements (migrations), species interaction and environmental impacts specifying stock dynamics will be developed.
A corresponding economic model of the harvesting sector including the harvesting sector itself will also be developed. This model, incorporating several vessel types, will describe the selection of fishing grounds, fishery, fishing time, fishing gear, investment in vesseles and gear and the marketing of catches. Through investment (disinvestments) in vessels and fishing gear and the deterioration of these assets over time the economic model contains its own dynamics.
Some version of the models will also include the modelling of individual decision makers (fishing firms) maximizing their objectives and interacting with each other through markets for fish, vessels and possibly catch quotas. Superimposed on this complex may also be different fishery management regimes constraining the opportunity set of the fishing firms.
The biological and economic models are connected through the fishing mortality that the harvesting activity imposes on the biology and the corresponding harvest extracted from it. The project will naturally pay a special attention to this crucial link.
A significant research activity has been firstly devoted to the study and improvement of the catch-effort models. The features and of the results for the biological model developed for Italian fisheries have been analysed, and a statistical technique to assess the prediction reliability of logistic and biological models in fishery management applications developed 0. The technique, based on the use of the moving blocks approach, has been tested over a set of catch and effort Italian data. The results confirm that the biological model exhibits lower prediction reliability with respect to logistic models, due to a higher number of estimated parameters, but the differences between the prediction reliability of the two kind of models are lower than expected.
A relevant work on the use of Bayesian approach for fishery management has been done and extensively documented. Since policy decision making in fisheries management is often complicated by large uncertainties in scientific estimates related to the fish populations and the potential biological and economic responses to alternative policy options, Bayesian estimation methods have recently been developed and applied. They are able to provide a flexible framework for estimating parameters in fisheries models and giving policy advice to fisheries managers. Despite growing interest, Bayesian methods remain accessible to relatively few. This research has tried to provide some applications of these techniques using Italian fisheries data. It has considered several aspects of the Bayesian approach, such as: a review of some recent applications in fisheries and the calculation of Bayesian posterior probabilities by fitting a logistic model to relative abundance indices 0; the evolution of Bayesian methodology and its application in the provision of fisheries policy advice 0; a general discussion about usefulness and weakness of Bayesian decision theory in fishery policy formulation 0; the use of Bayesian approach for fitting a multi-species, multi-area, multi-gear bioeconomic model to catch biomass and effort data (together with a Bayesian statistical approach that uses prior information in order to fit a MSAG biomass dynamic model to catch-effort data that are disaggregated inappropriately) 0.
The case of parameter estimation for Italian fisheries has been specifically analysed 00. In the first paper, a Bayesian statistical approach that uses prior information is adopted to fit a biomass dynamic model to catch data for two species (ray and cuttlefish). An informative prior for the intrinsic rate of increase, r, is required in order to obtain reasonable parameter estimates. The estimates of r are very similar for each species across areas, and therefore estimation performance can be improved because there are fewer parameters to estimate. In the second paper, the main evidence is that, because of anomalies in the catch and effort data (due to the fact that the Italian Statistical Institute changed its method of classifying bottom trawl gear in 1983), only the application of models assuming that catchability has increased over time, together with informative priors, provide sensible parameter estimates. Different computational techniques have been also presented, with mathematical details. A computer program in Visual Basic for the application of Bayesian technique has been also produced.
An analysis of the results of the Italian biological model via classical statistical techniques has been also performed, in order to estimate the possible effects of area and species on model parameters and to derive consistent parameter estimates in those cases where time series catch and effort data are not enough informative. A linear regression approach has been first used. Although consistent outcomes can be found in a number of cases, many results are not satisfactory in terms of statistical significance. Some results were also obtained applying multi-variate techniques 0. The biological parameters refer to the physical characteristics of the fish dynamic population as performed by the Deriso-Schnute model and reflect the mathematical constraints of the model equations. Results obtained by hierarchical tree clustering algorithms and k-means clustering (applied to the 3290 biological parameters considered, i.e. seven parameters for each of the 470 equations) states that local stocks have their own level specificity. Results for the big pelagic species and, at some extent, for small pelagic species follow a better dynamic behavioural considering a national aggregation level. Best findings were obtained by considering aggregation of the results by the ISCAAT national groups-species stocks. Fishing mortality was found as the main rule of the dynamic behaviour. Catchability and pre-recruitment are the start points of the whole process.
Notable research activity has concerned the analysis and improvement of the economic part of the models, also in order to test different management regimes.
For the Italian model, a significant advance in the specification of reconversion costs has been achieved, in order to overcome some limits due to the use of inertia constraint. The original formulation of this constraint should have been able to take account of the costs linked to the production re-conversion, assumed as proportional to the differences between the starting and proposed fishing effort distributions, and to select optimal solutions corresponding to realistic re-distributions, thus considering the labour market and the sectorial constraints. Even though powerful, this way of modelling the inertia constraint contains two main limitations: (i) positive and negative changes in fishing effort produce the same value of the inertia term, although on the basis of an economic and social perspective they can not be considered as equivalent; (ii) this approach does not consider the cost transfers of fishing effort between different area/system combinations. A more precise specification of the cost of effort reconversion, connected to the reallocation of the fleet among areas and systems, has been therefore developed 00. A LP problem has been formulated to minimise the total re-conversion cost (connected to the transfers of fishing effort among areas and/or systems, as well as those effort units coming from or sent to the unemployment status). This method has been tested over a set of data obtained with the fishing effort optimisation model. Different cases have been analysed, considering several hypotheses on the unit costs of effort transfer between area/gears combinations and on the costs related the dismissal of effort (which have been linked to the importance of the fishing activity in the considered area and to the local unemployment level). The results have confirmed the capability of this technique in selecting the most suitable re-conversion solutions, considering structural and social constraints. The integration of the revised economic model within the whole bio-economic optimisation has also required substantial work on mathematical and computational procedures. In fact, within the previous approach there is not a direct link between the variables referring to different areas. Therefore, the optimisation problem can be solved separately for each area, and a substantial reduction in the dimension of the mathematical problem results. Within the new approach, the optimal re-conversion cost depends on the whole distribution of fishing effort over areas and systems. The resulting non-linear optimisation problem must be solved simultaneously over all areas, but this implies a significant increase in the number of iteration. Moreover, computationally efficient methods for the solution of LP problem must be adopted to keep the whole computational time within reasonable limits. With reference to numerical techniques and software implementation, for the IREPA model a new numerical algorithm for non-linear optimisation has been integrated within the catch-effort identification procedure, to improve the robustness of the identification process. A more powerful software platform has been adopted, in order to overcome some memory limitations and to allow the solution of larger optimisation problems. A reconversion of the user interface, from menu-driven to Windows style programs, has been also undertaken, in order to allow an easier use of the code and promote a wider diffusion.
Much time has been also spent on the improvement of the Icelandic model. The fundamental economic agents in the Nordic Fisheries Management Model (NFMM) are firms engaged in fishing and/or processing. The objective of the firms is to maximize total net profits subject to existing constraints. In the current version of the model, this task is specified as a LP problem. It is of considerable importance to investigate and formally test whether the linear specification adequately describes the true situation. An approach to test the linearity hypothesis and other important hypotheses about the economics of the fishing industry has been developed. Among other things, this methodology permits the testing of hypotheses concerning economics of scale and scope in production, the degree of factor substitution possibilities, geographical differences in cost and the possibility of technical change.
A great deal of work has been therefore allocated on improving the parameter estimates of the economic part of the model and relative functional structure and specifications. Particularly, some studies focused on the estimation of the production technology of Icelandic fish processing industry by means of the hybrid translog cost functions. The estimation was carried out for processing sectors separately (freezing industry, salt fish production industry, and fish meal industry), and it was found that the hypothesis of constant returns to scale in these industries cannot be rejected with reasonable statistical confidence 0. It also emerges that the technology exhibits substantial non-linearities and substitutabilities, so that the need for a more complicated representation of the production side of bioeconomic models comes out.
The meaning of "endogenous optimisation" has been also clarified (by which the fishing firms maximise profits or more generally their objective functions within the confines of the model, i.e. by modelling the operating condition of the fishing firms allows the model to generate the relationship by allowing the firms to carry out their own maximisation within the model), and numerical experiments to determine the feasibility of actually computing endogenous optimisation fisheries models involving endogenous optimisation have been performed 0. It seems that the previous approach is perfectly feasible and extremely useful within the fisheries model (not only as a scientific tool but also for practical fisheries management), except that the extra maximisation loops add greatly to its computational requirements. The adaptation of the IFMM computer code (originally written in C++ for UNIX SUN-OS platform) to the PC Windows NT operating system is currently under development 0. Even if this conversion has found significant difficulties, the numerical analysis on the dynamic properties of the IFMM (together with the study of the related theoretical aspects) is well advanced. The results to data do not provide much evidence of the empirical relevance of complicated dynamics.
4. References
The following documents and papers have been produced during this project (the Discussion Papers are available from the Authors and from DG XIV-C-2):
[1] DP-1, R.Arnason P.Coccorese S.Olafsson V.Placenti G.Rizzo Comparison of the Icelandic (UI) and Italian (IREPA) Fisheries Management Models.
[2] DP-2, M.McAllister and G.Kirkwood Bayesian Stock Assessment: A Review and Example Application Using the Logistic Model.
[3] DP-3, M.McAllister G.Kirkwood and R.Arnason Fitting A Multi-Species, Multi-Area, Multi-Gear Bioeconomic Model to Catch Biomass and Effort Data: A Bayesian Approach
[4] DP-4, M.McAllister G.Kirkwood R.Arnason V.Placenti and G.Rizzo Fitting A Multi-Species, Multi-Area, Multi-Gear Bioeconomic Model to Catch Biomass and Effort Data: An Italian Example
[5] DP-5, M.McAllister and G.Kirkwood Can Bayesian Decision Theory be Useful for Fishery Policy Formulation?
[6] DP-6, M.McAllister Applications of Bayesian Decision Theory to Fisheries Policy Formulation: A Review
[7] DP-7, P.Coccorese V.Placenti G.Rizzo Inertia Constraint and Reconversion Costs in the Optimal Management of Fishing Effort: Some Results.
[8] DP-8, P.Coccorese V.Placenti G.Rizzo Analysis of a Biological Model for the Management of Italian Fisheries: Preliminary Results on Parameter Estimation and Model Prediction.
[9] DP-9, P.Coccorese V.Placenti G.Rizzo A Multi-Variate Analysis of the Endogenous Biological Parameters in the MOSES Model.
[10] DP-10, P.Coccorese V.Placenti G.Rizzo The Optimal Re-allocation of Fishing Effort under Unemployment Constraints.
[11] DP-11, M.McAllister V.Placenti G.Rizzo Use of Bayesian Methods for Parameters Estimation: a Review and Applications to Italian Multi-Species, Multi-Gear Fisheries.
[12] DP-12, G.Haraldsson The Icelandic Fish Processing Industry Estimation of Hybrid Translog Cost Functions.
[13] DP-13, D.Magnusson Development of the IFMM Computer Model: form a UNIX to a Windows NT Operating System Platform.
[14] DP-14, R.Arnason Endogenous Optimization Fisheries Models for Fisheries Management.
[15] DP-15, Yann Le Roch, Bio-Economic Fisheries Computer Models: an Overview of Existing Models.
[16] R.Arnason and Tryggvi B.Davidsson (Eds.), Essays on Statistical and Modelling Methodology for Fishery Management, The Fishery Research Institute, University of Iceland.
[17] V.Placenti G.Rizzo M.Spagnolo Bio-Economic Fishing Effort Optimization in Mediterranean Fisheries, EAFE Concerted Action, Bio-Economic Modelling Workshop, SEAFISH, Edinburgh, October 24-26, 1995.
[18] P.Coccorese V.Placenti G.Rizzo M.Spagnolo Bioeconomic Models for the Optimal Management of Mediterranean Fisheries: Analysis of Uncertainty in Model Previsions, FMU Symposium - Fishery Management under Uncertainty, Bergen, June 3-5, 1997.
[19] P.Coccorese V.Placenti G.Rizzo The Optimal Re-Allocation of Fishing Effort under Unemployment Constraints, EAFE Bioeconomic Modelling Workshop, University of Portsmouth, 17-18 December, 1997.
Introduction
Being designed to deal with broadly similar fishery situations, namely multi-species, multi-fleet, multi-gear, multi-region fisheries, the Italian and the Icelandic models are similar in many respects 0. However, there are also some important differences in design. Thus, the IREPA model is considerably more aggregative and less data demanding, so that it reveals less computationally demanding and easier to operate. Moreover, it includes statistical estimation as an integral part of its structure, which the UI model does not. The UI model is more disaggregate and detailed, at least potentially; therefore, it is also far more data demanding. Since the model is based on individual firm maximization, it generates behaviour endogenously. As a result, this implies a substantial increase in the computational requirements and in the output volume.
The IREPA model 00 is designed as an advisory tool for managers of particular fisheries under a given management regime. In this institutional context, the IREPA model can supply the fisheries managers with estimates of optimal fishing effort (per area and gear) computed by non-linear optimisation techniques, and calculate the outcome in terms of biomass, harvest and economic rents. Different catch-effort models can be used, both logistic (Schaefer and Exponential) and biological (Deriso-Schnute). Their parameters can be computed endogenously, from catch and effort time series, via non-linear least squares techniques. Biological reference points can be also considered and included in the optimisation procedure as constraints. Two alternative methods can be used to deal with the reconversion of productive structure, based on the "inertia" concept or on the optimisation of re-conversion cost via LP techniques. Different scenarios can be analysed by varying the relative weight of the economic, biological and social variables. The model can be also used for simulation, given the patterns of fishing effort.
The UI model is not an optimisation model, and cannot calculate optimal fishing patterns. However, given fishing effort, it is able to calculate the outcome in terms of biomass, harvest and economic variables. In this way, it can also provide assistance to the practical fisheries manager. The main strength of the model, however, lies in its ability to investigate the impact and economic efficiency of fisheries management regimes. Under new fisheries management directives the firms solve their new profit maximization problems and consequently alter their behaviour, possibly in quite unpredictable ways. Thus, the UI model is particularly well suited as an analytical and practical tool for designing efficient fisheries management systems. A fairly complete discussion on the Italian and Icelandic models is reported in 0, while a wide review on existing bio-economic fishery computer models is provided in 0.
Fields of science (EuroSciVoc)
CORDIS classifies projects with EuroSciVoc, a multilingual taxonomy of fields of science, through a semi-automatic process based on NLP techniques. See: The European Science Vocabulary.
CORDIS classifies projects with EuroSciVoc, a multilingual taxonomy of fields of science, through a semi-automatic process based on NLP techniques. See: The European Science Vocabulary.
- agricultural sciences agriculture, forestry, and fisheries fisheries
- natural sciences computer and information sciences software software applications system software operating systems
- natural sciences mathematics applied mathematics statistics and probability bayesian statistics
- social sciences sociology social issues unemployment
- natural sciences mathematics applied mathematics numerical analysis
You need to log in or register to use this function
Programme(s)
Multi-annual funding programmes that define the EU’s priorities for research and innovation.
Multi-annual funding programmes that define the EU’s priorities for research and innovation.
Topic(s)
Calls for proposals are divided into topics. A topic defines a specific subject or area for which applicants can submit proposals. The description of a topic comprises its specific scope and the expected impact of the funded project.
Calls for proposals are divided into topics. A topic defines a specific subject or area for which applicants can submit proposals. The description of a topic comprises its specific scope and the expected impact of the funded project.
Call for proposal
Procedure for inviting applicants to submit project proposals, with the aim of receiving EU funding.
Data not available
Procedure for inviting applicants to submit project proposals, with the aim of receiving EU funding.
Funding Scheme
Funding scheme (or “Type of Action”) inside a programme with common features. It specifies: the scope of what is funded; the reimbursement rate; specific evaluation criteria to qualify for funding; and the use of simplified forms of costs like lump sums.
Funding scheme (or “Type of Action”) inside a programme with common features. It specifies: the scope of what is funded; the reimbursement rate; specific evaluation criteria to qualify for funding; and the use of simplified forms of costs like lump sums.
Coordinator
84100 Salerno
Italy
The total costs incurred by this organisation to participate in the project, including direct and indirect costs. This amount is a subset of the overall project budget.