Quick link to: EFIMAS home, WP4

Case study 7: Swordfish fisheries in the Mediterranean

Introduction - Overview description of the case study

The main objective of this case study in the frames of the EFIMAS project is to evaluate different management scenarios using the FLR packages. As the fishery is managed through technical measures and control effort regulations, the scenarios that will be evaluated include effort reductions and seasonal closures.

Description of the fishery, stocks and management system

Swordfish (Xiphias gladius) is one of the most commercially important large pelagic species and it is heavily exploited in the Atlantic Ocean and the Mediterranean Sea. In the past twenty years due to the high market demand for swordfish, the fishing pressure on the different stocks of this species has increased dramatically. In the Mediterranean, catches showed a strong upward trend from about 4000 metric tons in 1976 to 20000 tons in 1988. Since then annual catches fluctuate from 12 to 15mt. The above increase in the reported catches could be, to a certain extent, attributed to the improvement of the fishery-statistic collection systems but it is also related to the intensification of fishing effort and exploitation of new fisheries. According to ICCAT records, the biggest producers of swordfish in the Mediterranean Sea in the recent years are Italy, Morocco, Greece and Spain. Genetic analysis has shown that Mediterranean swordfish form a unique stock separated from the Atlantic ones, and the management procedures adopted by ICCAT are in line with this finding.

Nowadays, the main fishing gears used are surface longlines and gillnets. Minor catches are also reported from harpoon, trap and recreational fisheries. Surface longlines are used all over the Mediterranean, while gillnets are mostly employed in Italy, Morocco and Turkey. However, following EU regulations and ICCAT recommendations for a general ban of driftnets in the Mediterranean, the size of the gillnet fleet has a clear decreasing trend.

Mediterranean swordfish fisheries are in principle single species fisheries having limited by-catches of other commercial species such as sharks, tuna and tuna-like species. They have been also reported incidental catches of cetaceans and sea-turtles in certain gillnet and longline fisheries respectively. In general, information on the composition and volume of by-catches is rather limited, while data on landings (biomass and size distribution) and effort are provided to ICCAT by the member countries.

While the most recent assessment carried out by the International Commission for the Conservation of Atlantic Tunas (ICCAT) indicated that recruitment showed little variation (CV~12%) with no trend over the past twenty years (Anonymous, 2007) in the same period, however, spawning stock biomass (SSB) has shown a decline between 24% and 38%, depending on the assessment model used. In addition, the main catch is of juveniles that have not yet spawned and assessment results clearly indicate growth overfishing and that at current levels of fishing mortality drastic stock declines could be seen within a generation (7-10 years).

Although till recently there had not been any Mediterranean-wide management measures for the swordfish stock, several countries have imposed technical measures in order to reduce the fishing pressure on the stock and the volume of juvenile catches. Those measures include: closed areas, closed seasons, minimum landing size regulations and license control systems. A minimum landing size regulation of 120 cm, enforced in the past within the EU countries, resulted in under-reporting of juvenile catches and proved not to be always practical due to the low size-selectivity of the fishing gears used. Investigations on the applicability of other methods for reducing juvenile catches, such as time and/or area closures have been suggested by ICCAT. Taking into account, that recruitment occurs in middle autumn – early winter, most discussions are focusing on the possibility of enforcing seasonal closures in that period. Recently, the ICCAT commission decided to implement a one month fishery closure during the recruitment period of 2008 and asked for the evaluation of the impact of the measure.

Description of the base case and scenario evaluations

Base case

The base case considers medium-term catch projections assuming the current exploitation pattern. The biological operating model was based on the output of the latest ICCAT age-based assessment (Anonymous, 2007). As no stock-recruitment (S/R) relationship exists for the stock two different approaches will be followed mainly differing on the assumptions about recruitment. In the first approach, a constant recruitment pattern with random variation is assumed, while in the second approach annual recruitment rates are obtained from an empirically estimated S/R model (Hilborn & Walters, 1992).

Case specific management issues to be addressed

Taking into account the management questions mentioned in chapter 1.2, the following issues have been considered:

1. a one-month fishery closure during the peak of the recruitment period. This measure has already been adopted for the Mediterranean for the year 2008.

2. a four-month fishery closure during the peak of the recruitment period.

3. a progressive capacity reduction reaching 20% of the fleet

Description of approaches for scenario evaluations


Summary of scenario evaluations

All different scenarios evaluated resulted to an increase in landings and spawning stock biomass compared to the base case scenario (Approaches 1 & 2) assuming either Beverton-Holt recruitment (BH models: Approaches 3, 5, 7) or constant recruitment (CR models: Approaches 4, 6, 8).

Biological variables


In all scenarios of both BH and CR models (Fig. 1) landings initially decreased in relation to the base case due to the reduction of fishing effort resulting from the applied management measures. In the medium term, however, as the stock responded to the effort reduction an overall increase was observed. In the 1-month closure scenario landings had increased by ~4% in both BH and CR models after the simulated 20 years of closure. The 4-month closure resulted to ~20% according to BH models, while a smaller increase of ~13% was estimated using CR models. Intermediate increases of ~7% were estimated in the 20% capacity reduction scenario using both CR and BH models.

Fig. 1: Evolution of median landings of the scenarios evaluated standardized to the base case scenario using Beverton-Holt modeled recruitment (left) or constant recruitment (right).

Spawning stock

The spawning stock biomass showed an increasing trend in relation to the base case from the beginning of the management period in all scenarios evaluated (Fig 2). In the 1-month closure scenario an increase slightly higher than 10% was predicted in both CR and BH models. An increase of 53% was estimated in the 4-month closure scenario by the CR model, while an even higher increase of 73% was estimated by the BH model. An increase between the values predicted by 1-month and 4-month closure models was estimated for the capacity reduction scenario reaching 18% and 26% of the base case in CR and BH models respectively.

Fig. 2: Evolution of median spawning stock biomass of the scenarios evaluated standardized to the base case scenario using Beverton-Holt modeled recruitment (left) or constant recruitment (right).

Economic variables

Analysis of economic variables considered only the Greek and Italian longline fleets. These fleets account for about 50% of the total Mediterranean swordfish production. Cost and revenue estimates were based on 2005 data (prices were considered constant over the examined period) from the aforementioned fleets.

Gross revenue

In both the 1-month and 4-month closure scenarios gross revenue initially decreased in relation to the base case following the yearly pattern of landings fluctuations in both CR and BH models (Fig. 3). The highest increases were achieved with 4-month closure scenarios reaching 20% and 11% for BH and CR models respectively. In the capacity reduction scenario a clear decreasing trend in gross revenue appeared in the first 5 years (when the capacity reduction took effect) followed by a slight progressive increase. By the end of the 20 years of simulation gross revenue was about 5% lower than the base case predictions for both BH and CR models.

Fig. 3: Evolution of median gross revenue of the scenarios evaluated standardized to the base case scenario using Beverton-Holt modeled recruitment (left) or constant recruitment (right).

Net revenue

Net revenue showed increasing trends in both the 1-month and 4-month closure scenarios but the relative increase was higher than that of gross revenue because of the elimination of the vessels variable costs during the closure period (Fig. 4). The highest increases in relation to the base case were again achieved with the 4-month closure scenarios reaching up to 66% and 47% for the BH and CR models respectively. The capacity reduction scenario resulted in net revenue levels closed to those estimated for the base case.

Fig. 4: Evolution of median net revenue of the scenarios evaluated standardized to the base case scenario using Beverton-Holt modeled recruitment (left) or constant recruitment (right).


Preliminary results of the management evaluation scenarios were presented in the 2006 SCRS/ICCAT Species Meeting. The FLR framework has been used for the development of a series of management scenarios during the 2007 ICCAT Assessment of the Mediterranean swordfish stock and at an ICCAT Intersessional meeting on Mediterranean swordfish organized in February 2008.

Tserpes, G. and Peristeraki, P., 2007. Effects of a seasonal closure of the Mediterranean swordfish fisheries on the stock production levels. ICCAT Collective Volume of Scientific Papers, 60: 2059-2062.

Tserpes, G., Tzanatos, E., Peristeraki, P., Placenti, V. and Kell, L., 2008. A bioeconomic evaluation of different management measures for the Mediterranean swordfish. ICCAT SCRS/2008/026.

  • Delivery Matrix by April 2006 (Note: See text from Meeting Minutes WP4 from EFIMAS Maastricht Meetings, September 2006).Case Study 7 Delivery Matrix


Hilborn, R. and C.J. Walters. 1992. Quantitative fisheries stock assessment. Chapman and Hall, London, 570 p.

Anonymous, 2007. ICCAT Mediterranean swordfish stock assessment session. SCRS/2007/016.


EFIMAS Contribution to the work

This work has been performed under the EFIMAS Project. The work uses data available in ICCAT SCRS/2007/016 document.


Coordinator: George Tserpes HCMR (Greece)


Panagiota Peristeraki HCMR (Greece)

Costas Papaconstantinou HCMR (Greece)

John Haralabous HCMR (Greece)

Christos Maravelias HCMR (Greece)

Evangelos Tzanatos HCMR (Greece)

Paolo Accadia CEMARE (UK)

Vincenzo Placenti IREPA (Italy)

Ramon Franquesa UB (Spain)

Jordi Guillen UB (Spain)

Ines Herrero UPO (Spain)

Ikerne del Valle UPV (Spain)

Kepa Astorkiza UPV (Spain)

Inma Astorkiza UPV (Spain)

Meeting Documents and Other Case Specific Work - working documents, models, analyses etc.

Swordfish price formation for Spanish longliners

* Modelling proposal for long term fleet behaviour. Data requirements and methodology. Presentation (Athens meeting): rum.pdf

* A price model for Mediterranean swordfish. Short report (Athens meeting): swordfishprice.pdf

A price formation exercise based on data from Spanish auctions was accomplished aiming to examine the fluctuation of swordfish prices caught by Spanish longliners in relation to a series of other factors (e.g. imports).

A double-log inverse demand function has been estimated by Ordinary Least Squares (OLS) for the Mediterranean-swordfish on monthly data from January-2001 to September-2006 (78 observations) from Spanish fish auctions. The problem of biased OLS estimations due to the identification problem, or simultaneous equations bias, that comes from the simultaneity between the variable own-price and the variable quantity is avoided (exogenous supply, that is, supply is only to a limited extent affected by prices. Fishermen supply what they catch and the market has to adjust thereafter) and it is not necessary to employ the method of moments for estimation (Two-Stage Least Squares or Instrumental Variable Estimation). The variables considered in the model are ex-vessel prices and quantities landed by the Spanish long line fleet operating in the Mediterranean sea, as well as other exogenous variables such as total expenditure and imported quantities of swordfish by Spain. These data do not gather any differentiation by commercial category associated to the length classes of the swordfish. Since attention is concentrated on a single equation, the only restrictions that can be made use of are those of homogeneity and negativity. Taking into account that homogeneity is the strongest restriction in the single demand equation context (in the sense of having the greatest chance of being rejected) and is therefore of the most interest, this restriction has been imposed prior to estimation. This implies accepting that demand is a function of relative prices and real total expenditure. The estimated function takes the form (p-values of each individual coefficient in brackets:



F(3, 74)= 2.2869 (0.00001)

Where: PMED: Ex-vessel prices of Mediterranean-swordfish caught by the Spanish long line fleet operating in the Mediterranean sea (€/kg). P: General price of swordfish (Weighted average price of total Swordfish in the market –from different origins and fishing systems-) (€/kg). QMED: Quantity of Mediterranean-swordfish landed by the Spanish long line fleet (kg). IMP: Quantity of swordfish imported by Spain (kg). PMED/P: Relative price of swordfish (price of Med-swordfish caught by the long line fleet and average price of total swordfish in the market) EXPEND/P: Real total expenditure in swordfish (taking into account total landings of different origins and fishing systems plus Net Imports of Spain).

R code: priceswordfih.r.zip

An application of behavioural models to the Spanish southern swordfish fishery

A model aiming to identify the elements that conduct fishers’ behaviour has been applied. This behavioural model concerns the fishers of swordfish located in Almeria (Andalusia).

While most applications of behavioural models in fisheries are related to an entry/exit/stay situation in a given fishery, our application is oriented to the selection of one of several alternative fisheries of a fleet. With respect to our case study, the vessels in the swordfish fleet have other alternatives for fishing. On the one hand, some of them can operate in the tuna fishery, though this is a seasonal fishery that operates mainly from April to July. On the other hand they can also operate in other coastal and more artisan fisheries. This alternative fishery is located closer to the coast and the main target species is silver scabbard fish (Lepidopus Caudatus), though many other species can be found in the catch.

Given that the swordfish fishery is assumed to be over-exploited it may be interesting to know which are the reasons that lead the skippers to take the daily decision of choosing one or another fishery. Behavioural models are probably the best way to analyse this. If we were able to modelise this and once we had the value of the significant variables for all boats, we would be able to predict which is the probability of each vessel of going to each of the fisheries, or, what is the same, which is the proportion of trips that each vessel will dedicate to each of the fisheries. Furthermore, we could also predict the effect that some changes in the law (either on volume or engine power of the vessels, on taxes…) would have on the proportion of time dedicated by the fleet to each of the fisheries. Therefore, this could be a good tool for authorities to lead a given fleet from an over-exploited fishery to other alternative fisheries.

As a conclusion we can confirm that behavioural models can be of help for predicting the behaviour of fishermen in a fishery and therefore to modelise effort in a bioeconomic model. On this occasion, given the multi-national characteristic of the swordfish fishery it is difficult to include these results in the bioeconomic model as the Spanish fishermen may behave completely differently than fishermen from other countries like Greece, Italy or France.

The manuscript presenting details of the findings is available for downloading.link to the document

George Tserpes 2007/01/30 14:54

efimas1/wp4/cs7/main.txt · Last modified: 2008/11/16 01:04 by admin
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