Assessing the Impact of Policy Changes in the Icelandic Cod Fishery

using a Hybrid Simulation Model

Sigríður Sigurðardóttir

1,3

, Björn Johansson

, Sveinn Margeirsson

and Jónas R. Viðarsson

Faculty of Ind. Eng, - Mech. Eng. and Computer Science, University of Iceland, Hjarðarhagi 2-6 107 Reykjavík, Iceland

Product and production development, CHALMERS University of Technology, 412 96 Gothenburg, Sweden

Matís, Icelandic Food and Biotechnology Research, Vínlandsleið 12, 113 Reykjavík, Iceland

Keywords: Hybrid Simulation Modelling, System Dynamics, Discrete Event Simulation, Fisheries Management, Life

Cycle Assessment.

Abstract: Most of the Icelandic cod is caught in bottom trawlers or longliners. These two fishing methods are

fundamentally different and have different economic, environmental and even social effects. In this paper

we present a hybrid-simulation framework to assess the impact of changing the ratio between cod quota

allocated to vessels with longlines and vessels with bottom trawls. It makes use of conventional bio-

economic models, discrete event modelling and provides a framework for simulating life cycle assessment

(LCA) for a cod fishery. The model was constructed in AnyLogic and consists of two models, a system

dynamics model describing the biological aspect of the fishery and a discrete event model for fishing

activities.

1 INTRODUCTION

1.1 Icelandic Cod Fisheries

Historically, the seafood sector has been the single

most important industry in the Icelandic economy

with cod fishery as its backbone. Even though other

industries have been growing larger during the

years, the seafood industry is still considered the

most important one. National accounts show that in

the year 2011, exported seafood accounted for more

than 40% of total exports, with cod explaining more

than 12% (Iceland, 2013a). Figure 1 shows value of

exported seafood as a percentage of total exports.

Moreover, it has been estimated that the contribution

of the fisheries sector and related industries, or the

so called fisheries cluster, to the GDP in the year

2010 is 26% (Sigfusson et al., 2013).

In the 1980’s, recruitments of cod began to

reduce drastically while at the same time fishing

effort remained higher than recommended by the

Marine Research Institute. Stock levels of cod

reached a critical level and to contain the situation a

harvest control rule was developed to determine total

allowable catch (TAC). In 1984, a comprehensive

system of individual transferable quotas (ITQ) was

introduced. In the beginning, quota was allocated

Figure 1: Ratio of seafood of total value of exports and

ratio of cod in total value of seafood in during 1990-2011.

based on vessel’s previous catch records. The ITQ

system resulted in an improved economic efficiency

of the fisheries as well as biological viability

(Arnason, 1993), (Arnason, 2006). The merits of the

quota system have however been heavily debated

since its establishment due to the consolidation of

quotas and the effect it has had on fisheries

communities short of quota (Eythorsson, 2000).

The Icelandic government has defined objectives

with its fisheries management system which are to

promote conservation and efficient utilisation of the

exploitable marine stocks of the Icelandic fishing

banks and thereby ensure stable employment and

1990 1995 1999 2000 2001 2002 2003 2004 2005 2006 20072008200920102011

Seafood Cod

294

SigurÃ

rardóttir S., Johansson B., Margeirsson S. and R. ViÃ

rarsson J..

Assessing the Impact of Policy Changes in the Icelandic Cod Fishery using a Hybrid Simulation Model.

DOI: 10.5220/0004488102940301

In Proceedings of the 3rd International Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH-2013),

pages 294-301

ISBN: 978-989-8565-69-3

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

settlement throughout the country (Ministry of

Fisheries and Agriculture, 2006).

1.2 Purpose of Study

Considering the aforementioned objectives, new

policies for managing the fisheries have to be

assessed in the three dimensions of sustainability;

economic, environmental and social. In this paper

we present a hybrid-simulation framework to assess

the impact of changing the ratio between cod quota

allocated to vessels with longlines and bottom

trawls. It makes use of conventional bio-economic

models, discrete event modelling (DES) and

provides a framework for simulating life cycle

assessment (LCA) for a cod fishery. The model was

constructed in AnyLogic and consists of two

models; a system dynamics model describing the

biological aspect of the fishery and a discrete event

model for fishing activities.

1.3 Fisheries Models

Most simulation research in fisheries management is

based on continuous multi-parameter models. Tools

that have been used previously for assisting in

fisheries management are for example the multi-

parameter models FLR (Fisheries Library for R) and

EcoSim. The FLR framework is a development

effort directed towards the evaluation of fisheries

management strategies (Kell et al., 2007). Ecopath

with EcoSim (EwE) is an ecosystem modelling

software suite that allows for spatial and temporal

modelling for exploring impact and placement of

protected areas and policy assessment. It is probably

the best known ecosystem model and has been

applied widely in fisheries around the world.

Atlantis (Fulton et al., 2011) is a modelling

framework developed to evaluate ecosystem based

management strategies. It consists of a number of

different linked modules: biophysical, industry and

socioeconomic, monitoring and assessment.

Many other modelling frameworks exist

including Gadget (Begley, 2004) and BEMMFISH

(Arnason, R. and Koholka, 2003

Most of these modelling frameworks allow for

great details in the biological aspect of fisheries

modelling but may lack overview in the three

aforementioned dimensions of sustainability. The

need for holistic modelling in fisheries has been

emphasized (Dudley, 2008). System Dynamics

(SD) is a good tool for creating holistic models and

understanding how things affect one another.

Dudley (2008) has demonstrated the benefits of

using SD for modelling fisheries and represented a

framework that can be adapted to most fisheries. A

number of system dynamics models in fisheries

exist. A SD model of individual transferable quota

system was constructed in order to differentiate ITQ

from total allowable catch effects and identify areas

where policy changes and management

improvements may be most effective (Garrity,

2011). Other SD models include a model for the

management of the Manila clam, a shellfish fishery

in the Bay of Arcachon in France (Bald et al., 2009),

a model of the management of the gooseneck

barnacle in the marine reserve of Gaztelugtxe in

Northern Spain (Bald et al., 2006) and a SD model

of the Barents Sea capelin (Yndestad, 2002).

Finally, a hybrid model combining SD and agent

based modelling has been constructed for

understanding competition and cooperation between

fishers (Bendor et al., 2009).

1.4 Combining DES and LCA

Life Cycle Assessment (LCA) standardized by ISO

14040:2006 and 14044:2006 (Finnveden et al.,

2009) is by far the most commonly used analysis

method for evaluation of environmental footprint.

LCA, however, holds drawbacks, which reduce its

preciseness and limit its value for producing reliable

results. The main associated problems with

traditional LCA analyses are (Reap J. et al., 2003)

 Use lumped parameters and site-independent

models.

 Static in nature and disregard the dynamic

behaviour of industrial and ecological systems.

 Focuses only on environmental considerations, not

economic or social aspects.

Hence, it can be beneficial to complement LCA

with other analysis tools, in order to effectively

combine environmental and economic analysis. An

example of such a combination is discrete event

simulation (DES) and LCA. Various different

examples of successful LCA-DES combinations

have been carried out and presented before (Thiede

et al., 2011), (Heilala, J. et al., 2009), (Solding, P. et

al, 2009); (Wohlgemuth et al., 2006).

2 BOTTOM TRAWLERS

VS. LONGLINERS

Nowadays most of the Icelandic cod is captured in

bottom trawls or with longlines. Use of gillnets used

to be more widespread than of longlines but that has

AssessingtheImpactofPolicyChangesintheIcelandicCodFisheryusingaHybridSimulationModel

295

Figure 2: Total landings (thousand tonnes) of cod by fishing gear 1993-2011.

changed as figure 2 confirms. In 2011 46% of the

total allowable catch for cod was captured with

bottom trawls and 32% with longlines (Iceland,

2013a), so around 78% of the total allowable catch

is under consideration in this study.

Bottom trawls and longlines are very different

fishing gears and have different economic and

environmental impacts, and potentially social

impacts which are harder to quantify and measure

Economic Impact

Data from operating accounts of fishing companies

collected by Statistics Iceland reveal that the larger

vessels are more economically viable (Iceland,

2013b). During the years 2002-2007, the operation

of smaller vessels was unstable, partly due to

external factors such as high interest rates and strong

exchange rate of the Icelandic krona (Agnarsson,

2012).

Environmental Impact

When comparing bottom trawls and longliners in

terms of minimising environmental impact, the

longliner is a far better choice. In 2009, a life cycle

assessment was applied to compare the

environmental impact made when producing 1 kg of

frozen cod caught with a bottom trawl on the one

hand and a long line on the other. The conclusion

from that study was that a trawled cod has a higher

impact within all categories assessed such as climate

change, respiratory organics/inorganics, eco-

toxicity, acidification and fossil fuel (Guttormsd,

2009).

It has been reported that the distribution of corals

around Iceland began to decline when bottom

trawling was initiated (Institute, 2004). The biggest

drawback of longlines however are danger to marine

animals such as sea birds that get stuck in the hooks

of the longlines (Valdemarsen and Suuronen, 2003).

3 THE MODEL

A hybrid simulation model of the Icelandic cod

fishery was constructed to assess the difference

between the two fishing gears. The model consists of

a System Dynamics model that describes the growth

of the cod stock. Fishing activities were simulated

with a Discrete Event model. Figure 3 shows a

diagram of the model. The discrete event model

simulates fishing trips of four different vessel types.

Before a vessel starts a trip, it sends a query to the

SD model to see if there is still catch quota

available. If the total allowable catch is reached, no

further fishing trips are planned until the TAC is

updated for the following fishing year.

One of the key assumptions made in the model is

that every year, the vessels reach their catch quota.

This is a valid assumption as the system holds a lot

of fishing capacity and there is a demand for catch

quotas and landing records confirm that they are

always met (Marine Research Institute, 2004).

3.1 A System Dynamics Model

The SD model describes the dynamics of the

biological stock and provides the total allowable

catch.

Natural Biomass Growth Function

A simple biological model was applied to describe

the biomass of cod. It accounts for no age-structure

and the population dynamics are described

100

150

200

250

300

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

Other

Bottomtrawl

Danishseine

Handline

Gillnet

Longline

SIMULTECH2013-3rdInternationalConferenceonSimulationandModelingMethodologies,Technologiesand

Applications

296

Figure 3: A diagram describing the hybrid-simulation model and the interaction between the SD model and the DE model.

with a logistic function (Clark, 1985):

 1







(1)

Were  is the stock size of the fishable cod, K, is the

carrying capacity and r the intrinsic growth rate of

the stock.

Total Allowable Catch

The total allowable catch at a year y+1 is determined

with the following harvest control rule:









,





(2)

where a represents harvest rate, and B

4+,y

is the

fishable biomass at year y+1, which consists of cod

large enough to be caught (Ministri of Fisheries and

Agriculture, 2010).

3.2 Discrete Event Model

The discrete event model simulates fishing trips of

three different types of longliners and a bottom

trawler. Ideally the model would make use of

information from logbooks and use data on trip

basis, information such as duration of trip, distance

sailed, and amount of catch and oil consumption as

an input. In this study only public data on quota

allocation and landings were used and scaled over

the whole fishing fleet under consideration.

The model outputs are catch numbers, economic

performance and CO

equivalences.

Catch numbers for each vessel are estimated

with data over quota allocations published by the

Directorate of Fisheries (Directorate of Fisheries

2012).

Economic Impact

Economic performance is measured by multiplying

revenue with the ratio of net profit and revenue but

this information is available from Statistics Iceland

for different vessel types (see equation 3). Figure 4

shows the economic performance of the four

different vessels during 2006-2011. This shows

clearly how unstable the operating results have been

for the small vessels. Average numbers dating back

to 1997 were used in the model.

AssessingtheImpactofPolicyChangesintheIcelandicCodFisheryusingaHybridSimulationModel

297

Figure 4: Profits as a ratio of total revenue by vessel type.

   ∙  ∙





(3)

Environmental Impact

The environmental impact of each of the fishing

gear was measured in CO

equivalences and based

on results from an LCA carried out in 2009. That

study showed that a one kilo of trawled cod had a

5.14 kg CO

equivalence while a long lined cod

added up to 1.58 kg CO

eq (Guttormsd., 2009). In

the same study, it was revealed that the hot spot in

the life cycle of cod is the fishing phase.

Social Impact

It is not an easy task to simulate social impact of

changing management policies. In this study the

only social factor taken into account are number of

jobs on each vessel. It might also be relevant to take

jobs on-shore into account since many of the

longliners do not have baiting machines on-board

and thus create jobs on land.

3.3 Model Validation

The model was validated using available historical

data as an input.

Biological growth

Stock assessment data from the Icelandic Marine

Research Institute was fitted to the logistic model

(equation 1) with a linear regression. With 57 data

points, the following fit was obtained:

Table 1: Results from fitting stock data to a logistic model

with linear regression.

Paramete

t-statistic

0.4700 6.6559

2654.44 2.5561

These results are not far from results obtained by

(Arnason et al., 1993). Moreover, running the model

with historical catch data as an input, results are

shown in figures 5 and 6.

Figure 5: Comparison of output from model simulations

and actual stock assessment data for fishable biomass.

There we compare our results from simulation runs

with data from 1955. The model gives good results

in comparison with data from the mid-eighties until

present times which is the period when the demersal

stocks of Iceland have been controlled under a quota

management scheme and the cod stock has been

quite stable. The model however does not account

very well for the fluctuations in the stock due to

overfishing in the years before the ITQs were

imposed. These fluctuations are very visible in the

graphs where there is a large gap between the blue

and the red line. This we find acceptable as in the

foreseeable future, the stock will without a doubt

continue to be controlled with catch quotas, and thus

maintain its equilibrium.

500

1000

1500

2000

2500

1958

1964

1970

1976

1982

1988

1994

2000

2006

Data

Simul

SIMULTECH2013-3rdInternationalConferenceonSimulationandModelingMethodologies,Technologiesand

Applications

298

Figure 6: Comparison of output from model simulations

and actual stock assessment data for landings.

Other results such as number of jobs, economic

performance and number of vessels were compared

to current numbers for validation purpose when

running the model with actual harvest rates from

historical data.

4 RESULTS

The main objective of the study was to use

simulation to determine the optimal ratio of quota

allocated to trawlers versus longliners with the

multi-objective aim of maximising profit and

number of jobs while minimising environmental

impact. The model was run multiple times over ten

years for different values of q which determines

division between quota allocated to bottom trawlers

and longliners. Figure 7 shows the results from these

runs. The results are displayed in such a way that for

each category, each value is displayed as a

proportion of the best possible outcome.

The best possible economic outcome is obtained

when the entire quota is allocated to bottom trawlers

whereas the best environmental outcome is at the

opposite end, where the entire quota is allocated to

longliners. The dashed line in figure 7 shows the

current allocation policy, which leans towards

maximizing profitability rather than minimizing

environmental impact. If the policy were to lean

more towards the intersection of the economic and

environmental components, we would get the best

possible outcome, assuming that the two

components have the same importance. The model

does not take into account jobs in baiting that are

created on-shore because of long liners.

By expanding the model boundaries, we are

likely to see even more positive effects of long liners

and a sharper contrast between longliners and

bottom trawlers in terms of social aspects. This also

leads to a more distribution of wealth which surely

would be accounted for as a positive social

Figure 7: Main results from running simulations with different quota allocations. To the far left we display the case where

the entire quota under consideration is allocated to bottom trawlers and the far right shows the opposite case with the entire

quota allocated to long liners.

20%

40%

60%

80%

100%

100%BT 87.5%BT 75%BT 62.5%BT 50‐50 62.5%LL 75%LL 87.5%LL 100%LL

Economic

Social(jobs)

Environmental

(CO2eq)

AssessingtheImpactofPolicyChangesintheIcelandicCodFisheryusingaHybridSimulationModel

299

5 DISCUSSION AND FUTURE

RESEARCH

In this paper we have presented the first steps in

combining a SD and a DE model resulting in a

holistic model of a system while looking at parts of

it in more detail. In this study, we used publicly

accessible data on landings and quota allocations,

which were scaled over the whole fleet under

consideration. The output of this work is a simple

model which can be improved by adding more

system details. Next step is to add more species but

the model is easily scalable in terms of number of

species. Another obvious step to make in terms of

improving the analysis is to expand the system

boundaries, for instance to include jobs throughout

the whole value chain.

We present a simulation framework which makes

it possible to combine LCA data with a hybrid DES-

SD model. Using logbook data, as an input to the

DE model, the fishing phase could be modelled in

more detail but the fishing phase is the part of the

life cycle of cod which has the most negative

environmental impact due to fossil fuel

consumption. With data from logbooks and more

detailed operational data, the model could be more

realistic and used for further scenario evaluation on

quota allocation. In terms of future research, it

would be possible to model agent based vessels for

finding company operations revenue and equilibrium

based on different quota allocations. Such a model

could be used to identify opportunities to minimise

environmental impact and reduce cost by simulating

alternative fishing routs for the vessels.

6 CONCLUSIONS

To conclude, the findings made from the combined

SD DE model shows and confirm the results in

terms of clarification of economic and

environmental impact of longliners versus bottom

trawlers. The model also shows a need for a larger

more complete modelling approach including

logbooks from the vessels for increasing accuracy on

catch and redirection of traffic to minimize cost and

environmental impact while maintaining job

opportunities.

ACKNOWLEDGEMENTS

The authors would like to acknowledge support from

Matís through the EcoFishMan project (FP7-KBBE-

2010-4) and ProViking through the EcoProIT

project and VINNOVA (Swedish Agency for

Innovation Systems). This work has been carried out

within the Sustainable Production Initiative and the

Production Area of Advance at Chalmers. The

support is gratefully acknowledged.

REFERENCES

Agnarsson, S., 2012. Útgerð smábáta : Krókabátar og

línuívilnun, Available at: http://hhi.hi.is/sites/hhi.hi.is/

files/C12_01_krokabatar.pdf (Accessed February 13,

2013)

Agriculture, M. of Fisheries and, 2010. Áætlun um stjórn

veiða úr íslenska þorskstofninum - nýtingarstefna

stjórnvalda. Available at: http://www.sjavaru

tvegsraduneyti.is/frettir/frettatengt/nr/10205 (Accessed

February 13, 2013).

Agriculture, M. of fisheries and, 2006. The Fisheries

Management Act. Available at: http://www.fisheries.

is/management/fisheries-management/the-fisheries-

management-act/.

Arnason, R., 2006. Property Rights in Fisheries: Iceland’s

Experience with ITQs. Reviews in Fish Biology and

Fisheries, 15(3), pp.243–264. Available at: http://

www.springerlink.com/index/10.1007/s11160-005-

5139-6 (Accessed February 4, 2013).

Arnason, R., 1993. The Icelandic Individual Transferable

Quota System : A Descriptive Account. , 8, pp.201–

218.

Bald, J. et al., 2009. A system dynamics model for the

management of the Manila clam, Ruditapes

philippinarum (Adams and Reeve, 1850) in the Bay of

Arcachon (France). Ecological Modelling, 220(21),

pp.2828–2837.

Bald, J., Borja, A & Muxika, I., 2006. A system dynamics

model for the management of the gooseneck barnacle

(Pollicipes pollicipes) in the marine reserve of

Gaztelugatxe (Northern Spain). Ecological Modelling,

194(1-3), pp.306–315. Available at: http://linking

hub.elsevier.com/retrieve/pii/S0304380005005284.

Begley, J., 2004. An overview of Gadget , the Globally

applicable Area-Disaggregated General Ecosystem

Toolbox.

Bendor, T., Scheffran, J. & Hannon, B., 2009. Ecological

and economic sustainability in fishery management: A

multi-agent model for understanding competition and

cooperation. Ecological Economics, 68(4), pp.1061–

1073. Available at: http://linkinghub.elsevier.com/retri

eve/pii/S0921800908003406 (Accessed June 21,

2011).

Clark, C. W., 1985. Bioeconomic Modelling and Fisheries

Management, John Wiley & Sons Inc.

Dudley, R.G., 2008. A basis for understanding fishery

management dynamics. System Dynamics Review,

24(1), pp.1–29.

Eythorsson, E., 2000. A decade of ITQ-management in

SIMULTECH2013-3rdInternationalConferenceonSimulationandModelingMethodologies,Technologiesand

Applications

300

Icelandic " sheries : consolidation without consensus. ,

24, pp.483–492.

Fisheries, D. of, Allocated catch quotas 2012/2013.

Available at: http://www.fiskistofa.is/veidar/aflahei

mildir/uthlutadaflamark/ (Accessed January 31, 2013).

Fulton, E. a et al., 2011. Lessons in modelling and

management of marine ecosystems: the Atlantis

experience. Fish and Fisheries, 12(2), pp.171–188.

Available at: http://doi.wiley.com/10.1111/j.1467-

2979.2011.00412.x.

Garrity, E.J., 2011. System Dynamics Modeling of

Individual Transferable Quota Fisheries and

Suggestions for Rebuilding Stocks. Sustainability,

3(1), pp.184–215. Available at: http://www.mdpi.

com/2071-1050/3/1/184/ (Accessed May 20, 2011).

Guttormsd, B., 2009. Life Cycle Assessment on Icelandic

cod product based on two different fishing methods

Iceland, S., 2013a. Catch by fishing gear and species

1992-2011. Available at: http://statice.is/temp_en/

Dialog/varval.asp?ma=SJA09013&ti=Catch+by+fishi

ng+gear+and+species+1992-2010&path=../ Database/

sjavarutvegur/ af3Skip/ & lang=1&units=Tonnes

(Accessed January 30, 2013).

Iceland, S., 2013b. Export of marine products. Available

at: www.statice.is/Statistics/Fisheries-andagriculture

(Accessed February 13, 2013).

Iceland, S., 2013c. Financial accounts of fisheries 1997-

2011. Available at: www.statice.is (Accessed February

5, 2013).

Institute, I. M. R., 2004. Friðun viðkvæmra hafsvæða við

Ísland,

Kell, L. T. et al., 2007. FLR : an open-source framework

for the evaluation and development of management

strategies. , (i), pp.640–646.

Ragnar, A. et al., 2004. Optimal Feedback Controls :

Comparative Evaluation of the Cod fisheries in

Denmark , Iceland and Norway. North.

Sigfusson, T., Arnason, R. & Morrissey, K., 2013. The

economic importance of the Icelandic fisheries

cluster—Understanding the role of fisheries in a small

economy. Marine Policy, 39, pp.154–161..

Valdemarsen, J. W.; Suuronen, P., 2003. Modifying

Fishing Gear to Achieve Ecosystem Objectives. In

Responsible Fisheries in the Marine Ecosystem.

Rome: FAO.

Yndestad, H., 2002. System dynamics of the Barents Sea

capelin. ICES Journal of Marine Science, 59(6),

pp.1155–1166.

AssessingtheImpactofPolicyChangesintheIcelandicCodFisheryusingaHybridSimulationModel

301