Quick link to: EFIMAS home, WP4, CS9

Case study 9: Approach 2a: A spatially- and seasonnally- explicit fishery management model including fishing effort reallocation - Baltic cod

The model includes an economic description of the fishery since economic conditions impact in a feedback mechanism the fishing effort distribution in space and/or the effort allocation between fishing activities via change in fishermen behaviour. All used economic equations, classes and models are implemented in R and generally adapted to and/or extended from the FLR framework. This economic development of FLR has been done under the EFIMAS project.

Management regimes were defined to support the new F-adaptive approach advocated by the ICES WGBFAS (2007) and the EU Commission (2006) for the recovery of the Baltic Cod defining a HCR based on a step by step reduction in overall F until the ultimate Ftarget is reached, a level where the fishery is considered sustainable with high long-term yields. Here, the goal was to reduce the overall fishing mortality, either using a TAC system or using an effort control system, by 10 % compared to the year before, until F (4-7) is 0.3.

A spatio-temporal management regime for the Baltic Sea has also been enforced by the EU Commission since 1995. From January 2005, a MPA network in the Baltic Sea consisting in three large closures were enforced to restore the Cod stock which banned fishing in the main spawning areas of the Eastern Baltic cod stock. On this basis, two closure scenarios were tested: (i) the potential effect of the EU Commission closure proposal to protect spawning zones from all fishing activities, and (ii) a realistically-sized seasonal closure of the ICES subdivisions 25, 26 and 27 from the 1st of June to the 31st of September for all fishing activities.

Fishermen are likely to react to management implementation by changing their spatio-temporal pattern of effort impacting the underlying harvested populations in a different way. Such an effect occurs first at a time scale less than a year even if the regulation is yearly designed (e.g. TAC or TAE). Hence, testing these fishing reallocation effects require to develop a particular spatially- and seasonnally- explicit simulation frame.

Data and parameters

Stock data

ICES BITS survey was used to define a spatio-temporal age-disaggregated cod availability (Nielsen et al., (in revision)) (biological operating model OM1) or from results from a hydrodynamic model for the Baltic Sea to define population zones and migration patterns (biological operating model OM2). Population data are used from the ICES WGBFAS (2007) to fit stock-recruitment (SSB-R), maturity-at-age, and weight-at-age relationships. The eastern Baltic Cod decline is supposed to be the result of partly a change in the Baltic Sea abiotic conditions affecting the cod reproductive success and its growth rate. Accordingly, ‘good’ and ‘bad’ forcing environments were identified for favourable and unfavourable conditions for cod reproduction, respectively, and led to differentiate two set of values for population parameters depending on the years of inflow (‘good years’; Köster et al., 2005) from the North Sea to the Baltic Sea. The availability coefficients per age per quarter reflect the pattern of the abundance distribution over time between the different areas.

Fleet data

The exploitation part sub-model including the initial allocation of effort was conditioned from the monthly- and ICES square- disaggregated logbook data for the main participating countries in the Baltic cod fishery (Denmark, Sweden, Latvia, Poland, Germany), and the missing effort from other unreported countries was completed by calibrating on basis of the ICES WGBFAS (2007) eastern cod stock landings. The international fleet based landings and effort data dis-aggregated on month and ICES square was obtained under the cooperation between the EU-FP6 EFIMAS, PROTECT and UNCOVER projects. Sales values and cost structure for the Danish fleets per set of vessels sharing a common activity was also available from the Danish Institute of Food Economics (F.O.I.).


The simulation frame for fishery modelling was developed in R plugged to the FLR platform. The model consists of three sub-model components: (i) a multi-stock module that considers how the age-structured populations of fish stocks in different areas change; (ii) a multi-fleet module taking into account of the heterogeneity of the fishing practices; and, (iii) a management module that could examine both conventional management techniques and permanent or temporary closed areas and seasons. All these components operate on a spatial grid matching underlying data in monthly and spatially disaggregated observations.

population dynamic
The population model is age-structured population dynamic model with class change at the beginning of each year, SSB-Recruitment relationship, spatial distribution of abundance using season- and age-specific availability allocation key (Operating Model 1) or season- and age-specific migration pattern between population zones (Operating Model 2).

fleet activities
The relationship between the spatially-disaggregated fishing effort and the resulting fishing mortality is the core of the operating model. Hence, the stock-specific area- and season- age- disaggregated fishing mortalities are computed from the partial F obtained for each fishing activity (i.e. a combination of a fleet and a gear). A linear link is assumed according to the following expression:

partialF(fleet, gear, species, area, season)= catchability(species, gear, -area-, -season-) x Effort(fleet, gear, area, season)

catchability(species, gear, -area-, -season-)=selectivity(species, gear) x efficiency(species, fleet, gear) x scale.factor (species, -area-, -season-)

A particular emphasize is put to break the total fishing effort into key allocations playing at different levels to act as entry points to test fishing effort re-allocation:

Effort(country,vessel.category,fleet,gear,season,area)= totNbVessels x propInCountry x propInVesselCategory x propInFleet x propInActivity x effort.per.vessel x propInGear x propInArea

The economic revenue are computed from landings using stock-specific fish prices and (spatial disaggregated) costs for fishing summed at the metier level or at the fleet level. Costs are generally divided into partly variable and fixed costs where variable costs varies depending on the activity of the vessel (e.g. fuel, ice, maintenance costs) while the fixed costs (e.g. crew share, sales costs) have to be paid irrespective the vessel activity. The costs per vessel given for each fleet segment assume average cost within each segment:

Revenue(country,vessel.category,fleet,gear,season,area)= sum over species of Landings(species, age) x Weight(species, age) x Price(species, age)

Profit(fleet,gear,season,area)= Revenue(fleet,gear,season,area) - vcost(fleet,gear) x Effort(fleet,gear) - vcost2(fleet,gear) x Revenue(fleet,gear,season,area) - fcost(fleet) x capacity(fleet)

The AHF-model (Hoff and Frost, 2007) suggests a function to model capacity change via investment/disinvestment dynamics which is reused here in a simplified way. The equations used to evaluate the dynamic fleet capacity change in the AHF model are described shortly in Appendix 4 of ECONOWS (2008). All used economic equations, classes and models are implemented in R and generally adapted to the FLR framework and can communicate directly with other FLR classes such as the FLEcon package.

Conditioning of Operating Model
Biological Parameters

The initial age-structured population number and the biological parameters come from the ICES WGBFAS (2007). The WG input data are also use to fit stock-recruitment (SSB-R), maturity-at-age, and weight-at-age relationships, split into two set of parameters depending on the presumed environmental condition(good years < 2000; bad years >= 2000). The species-specific catchability is calibrated using the scale factor obtained from the comparison between the simulated and the observed 2003 total landings in weight for the fleet of reference (i.e. the fleet having the efficiency factor at 1).

Fisheries & Fleets

To minimize the computation demand in the modelling procedure, some classes of vessels (or fleets) are defined assuming homogeneous features between vessels belonging to the same group. Then, fleets were defined as a combination of country, vessel size and the gear used (possibly a set of gears if the vessels are polyvalent), and specific spatio-temporal allocations of effort using a particular gear (fleet sub-segments i.e. metiers) were also explicitly modelled.

The effort allocation model has been initialised on Baltic Sea fishery data (Denmark, Latvia, Sweden, Poland, Germany) for the year 2003 i.e. the most recent common year between evolved countries we have data for. All in all, 25 fleets and 60 metiers along the 5 countries have been conditioned acting on more than 200 ICES squares belonging to the Baltic Sea and to the North Sea and simulated on a 15 years horizon time. When initialising the fishing effort for the year 2003, a raising factor was used from the ICES WGBFAS (2007) to adjust the initial landings by multiplying the catchability to take into account un-allocated landings (i.e. misreporting) from the logbook data.

To quantify the other factors than change in fish abundance which could impact the catch rates, a catch-effort standardisation procedure is applied running Generalized Linear Models (GLM.) to (i) model the relative catching power per (set of) vessels, (ii) model the fish selectivity per gear, (iii) model the sorting or discard behaviour per activity. Country, vessel size, and gear levels are required to take into account the gear-specific selectivity, the relative fishing power and the specific cost structure.

As a first approximation, the cost structure for other countries than Denmark was extended from the only available Danish data. The fish prices are fixed and has been expressed per gear and calculated from the Danish sales slip data.


TAC system
A one-year-lag Total Allowance Catches (TAC) system is modelled as follows: Each year, at the beginning of the year, (i) a stock assessment is performed from for the previous year running a XSA assessment assessing the yearly age-disaggregated fishing mortality and abundance at year minus 1; (ii) a Harvest Control Rule (HCR) is applied to decide on the target fishing mortality for the coming year, e.g. in ICES WGBFAS (2007) in the innovative approach framework, a specified HCR is to reduce F by 10 % compared to the year before until F (4-7) is 0.3; (iii) a two-years short-term forecast is performed for the current year (the year on which the assessment is performed) and the coming year using the assessed age-disaggregated abundances and applying twice the fishing mortalities exploitation pattern of the previous year reduced by 10 % each time, and geometric mean of the 3 previous recruitments to get forecasted recruitments; (iv) finally, the TAC is calculated for the coming year using the classical Baronov equation for which the targeted fishing mortality is known as decided from the HCR. Additionally, the final TAC value could be set to remain in a given interval (e.g. constrained bounds by 15%). The unofficial TAC is then split by country using the historical key allocation by country applied by the EU commission between the Baltic countries. Finally the official TAC is calculated removing the expected unofficial landings assuming a same level of misreporting across countries.

Effort control system
An effort control management is modelled to decide on the Total Allowed fishing Effort (TAE) from one year to the next. As for the TAC system, each year, at the beginning of the year, (i) a stock assessment is performed from for the previous year running a XSA assessment assessing the yearly age-disaggregated fishing mortality and abundance at year minus 1; (ii) a HCR is applied to decide on the target fishing mortality for the coming year. (iii) The total allowed effort for the coming year is then calculated using the linear link between the effort and the fishing mortality assuming constant catchability.

Some misreporting on catches could occur under the TAC system leading to apparent simulated landings lower than the true simulated landings. By contrast, the effort control is assumed to do not incentive for misreporting as fishermen are allowed to land and sell all legal-sized fish caught and then the apparent landings are considered to be the true landings. No effort misreporting is also assumed which suppose an efficient control on this aspect. The effort control system assumed that the landing of cod is actually only the driver for deciding on the TAE which seems a reasonable assumption for the Baltic Sea fisheries having reduced by-catches.

Spatio-temporal closure
Spatially- and temporally-explicit regulations are modelled specifying seasons, areas, years and fleets concerned by the regulations. If the regulation occurs at a given time step, the concerned fleets have obtained their initial effort distribution over areas modified in a way where effort is totally removed on closed areas and spatially re-allocated on the other possible fleet-specific fishing areas (i.e. spatial effort displacement). The model is especially designed to test presumable ways in effort reallocation in response to these closures.

effort reallocation model

A first assumed behaviour was tested under the TAC regime. Hence, if the (unofficial) TAC is exhausted for a given country then effort for this country is reallocated on the biological zones of species for which a country-specific quota remains to be caught. If any catches still occur on the closed stock for these given fleets (in case of overlapping harvested populations) then these catches are discarded. If no quota remains for all species then fleets stop fishing. If a quota is not reach before the end of the year for a given country then these remaining potential catches are lost as no effort increase (more than the 2003 effort level) is allowed in the simulation.

In the effort control management case, the regulation-induced effort re-allocation was tested at the short-term scale i.e. simulating the effort displacement in space toward higher LPUE areas. Equally, long-term effort-control-induced effort reallocation was tested simulating investment/disinvestment dynamics. An efficiency drift over years in response to the effort control is also tested.

Presumable ways in re-allocation or displacement of the fishing effort in response to spatio-temporal closure were investigated at the short-term scale by: (i) a closure-induced uniform spatial re-distribution of effort on the remaining fleet-specific area, and (ii) a closure-induced re-distribution of effort on remaining fleet-specific areas depending on the Landings Per Unit of Effort (LPUE) over the previous year in the same areas, i.e. each area among the remaining areas collects a part of the re-distributed effort in proportion to their relative LPUE.


Figure - Management procedure evaluation loop combining FLR sub-models as “lego blocks” implying (i) an Operating Model generating the stock and fleet dynamics, (ii) a Management Model applying regulations from stock assessement and/or from Harvest Control Rules, and (iii) a Effort Reallocation Model modelling the response of fishermen to regulations and/or change in stock availability depending on decision choice variables. A particular emphasis is put to test different assumptions about presumable ways in effort reallocation. Notice that the operating model is seasonally disaggregated while the management model is on a yearly basis


This model acts as a spatially- and seasonally-explicit framework for fishery modelling to support variously sophisticated sub-models concerning effort displacement and fishermen behaviour response to regulations (and/or change in stock availability over time). Hence, additional rules for regulation-induced effort reallocation, spatially-explicit or not, could be easily developed and simply added to take into account additional or alternative hypothesis. As this frame uses the open-source FLR-platform suggesting a common language shared with other developers and users, a collaborative work could be carried out here between fishery scientists to check the existing model code as well as to add some additional features and improvements to the current version.

Concerning the Baltic cod fishery, information are needed about environmental conditions driving the system (indicators on whether inflow will happen, etc.) because the whole system is driven by environmental determined recruitment and fish growth characteristic under both regulation systems, however, only the direct effort regulation system could reach precautionary limits in the horizon time under bad environmental conditions (given the assumptions). Hence, with respect to evaluation of effort regulation compared to TAC regulation under the adaptive approach, effort regulation seems to be more biological robust than a TAC system because independent of the assessment uncertainties, but, in the evaluation of the effort control system this result is exclusively dependent of the assumptions of how effort is re-allocated between fleet, areas, and seasons and on the assumptions in relation to fleet specific (constant) catchability for cod for the individual fleets.

The regulations with existing closures in relation to the cod recovery plan have a slight effect on a year scale and redistribute the landings on a monthly scale. If the closures are extended to a larger area comprising SD 25, 26, 27 in the spawning period then there were a more significant effect with increased SSB and landings/revenue both on monthly and yearly basis irrespective of the preliminary effort reallocation assumptions tested here. However the assumptions about effort reallocation in response to closure should be refined as soon as economic data would be available. Equally, the importance of the biological operating model design mitigating the evaluation of closure effect should be further investigated.

All in all, the closure designs tested here were not alone enough to reach the spawning stock biomass management targets and should rather be implemented in combination with other regulations such as effort reduction, even if interaction between effort control and closure could lead to minimize the efficiency of closure. In addition, the positive balance of landings over years due to the closure effect in comparison to the baseline scenario should not be accompanied by an increase in total effort or in fleet capacity (from new entrants attracted by higher CPUE) to be able to protect the spawning biomass surplus driving the stock until the reference targets.

Bio-economic data has been used and simulated forward in time to test the short-term as well as the long-term effort re-allocation (side-) effect on stock evolution in response to regulations depending on spatially-explicit and heterogeneous fleet-specific economic features. In a socio-economic perspective, the scenarios should be further evaluated regarding their bio-economic effects in implementing such management regimes i.e. evaluate their capacity to drive the multi-fleet fishing activity toward individual and/or global economic targets on which managers may draw their decisions for management.

Effort Control

Figure 1 A- Effort Control effect on recruitment, SSB and landings under adverse environment scenario; B- with reallocation assumption (i.e. remove effort first on the fleet area with lower CPUE) relative to the baseline scenario in A.

Figure 2 A- Spatio-temporal closure (2006 EU commission proposal design for closure of the cod spawning areas) effect on recruitment, SSB and landings under adverse environment scenario with uniform spatial effort reallocation; B- with reallocation assumption (i.e. redistribute on the fleet area with higher CPUE) relative to the baseline scenario in A.

Figure 3 Under the reallocation scenario on the higher catchability areas, mapping by ICES square the spatio-temporal closure effect (CP 2006 design) on fleet total effort distribution in gain/loss terms relative to the no management scenario (blue: gain; green: loss) over closured months (JUN to SEP) in the year (As no time displacement is modelled then other months are not shown)


Bastardie, F., Nielsen, J.R., and Kraus, G. 2008. Management Strategy Evaluation framework for the Eastern Baltic cod fishery to test robustness of management against environmental conditions and fleet response scenarios. (Submitted ICES J. Mar. Sci.) submitted paper

EFIMAS ECONOWS Report, Final Version 2008.
Final report ECONOWS 26/06/2008

Nielsen, J.R., Bastardie, F., Nielsen, J.N., and Pedersen, E.M.F. (2008, In revision). Whole fishery selectivity, fishing patterns, and fleet catchability dynamics in international Baltic Sea cod fisheries – from observed spatio-temporal patterns in resource availability and fleet specific selection, relative fishing power, and fisherman sorting behaviour. (In revision) ICES J. Mar. Sci.

Links to Other Work

(See under acknowledgements)


EFIMAS ECONOWS Report, Final Version 2008.
Final report ECONOWS 26/06/2008

EU Commission 2006. Proposal for a Council Regulation establishing a multi-annual plan for the cod stocks in the Baltic Sea and the fisheries exploiting those stocks. EU Proposal 11984/06 PECHE 238 – COM(2006) 411 final. FLR. Fisheries Library in R. http://flr-project.org

Hoff, A. and Frost, F. 2007. Modelling Economic Response to Combined Harvest and Effort Control in Fishery. http://www.univ-brest.fr/gdr-amure/eafe/eafe_conf/2007/hoff_frost_eafe2007.pdf

ICES WGBFAS Report 2007. ICES Advisory Committee on Fishery Management. ICES CM 2007/ACFM:15 Kell, L. T., Mosqueira, I., Grosjean, P., Fromentin, J-M., Garcia, D., Hillary, R., Jardim, E., Mardle, S., Pastoors, M. A., Poos, J. J., Scott, F., and Scott, R. D. 2007. FLR: an open-source framework for the evaluation and development of management strategies. ICES Journal of Marine Science, 64: 640-646.

Köster, F.W., Möllmann, C., Hinrichsen, H.-H., Tomkiewicz, J., Wieland, K., Kraus, G., Voss, R., MacKenzie, B.R., Schnack, D., Makarchouk, A., Plikshs, M., and Beyer J.E. 2005. Baltic cod recruitment – the impact of climate and species interaction. ICES Journal of Marine Science, 62: 1408-1425. Motos, L. and Wilson, D. (editors). 2006. The Knowledge Base for Fisheries Management. Developments in Aquaculture and Fisheries Sciences Series, 36. Elsevier. ISBN-13: 978-0-444-52850-6.

Nielsen, J.R.*, Sparre, P.J.*, Hovgaard, H.*, Frost, H.*, and Tserpes, G.* 2006. Effort and Capacity Based Fisheries Management. Chapter 7: p. 163-216. In: Motos, L. and Wilson, D. (editors). 2006. The Knowledge Base for Fisheries Management. Developments in Aquaculture and Fisheries Sciences Series, 36. Elsevier. *Authorship equal.

R Development Core Team 2007. R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria. http://www.r-project.org/


This work has been made through cooperation between the EU-FP6-502516-EFIMAS Project, the EU-FP6-PROTECT Project and the Danish National DFFE-Sustainability Package-Project (Effort Regulation in the Baltic Sea), Danish Ministry of Food, Agriculture and Fishery.

We would like to thank for valuable discussion with Per J. Sparre on this in relation to the collaborative parameterization of the FLR, ISIS-Fish and TEMAS Evaluation Frame on the Baltic cod fisheries case study. Furthermore, we would like to thank for the contribution of fleet based catch and effort data for all Baltic institutes participating in EFIMAS.

EFIMAS Contribution to the work

The major financing of this work originates from the EU-FP6-EFIMAS Project.


Francois Bastardie (DTU-Aqua), J. Rasmus Nielsen (DTU-Aqua), Gerd Kraus (DTU-Aqua)

efimas1/wp4/cs9/appr2/main.txt · Last modified: 2008/12/18 14:08 by fb
Except where otherwise noted, content on this wiki is licensed under the following license:CC Attribution-Noncommercial-Share Alike 3.0 Unported
Recent changes RSS feed Donate Powered by PHP Valid XHTML 1.0 Valid CSS Driven by DokuWiki