The Mediterranean has been described as an ecosystem “under siege” (Coll et al. 2012). This applies particularly to the Mediterranean’s fisheries resources, which have been studied for a very long time, starting with Aristotle (Thompson 1947). This study has led to a rich literature, albeit one that is disparate, not least because it is written in a large number of languages ranging from Spanish, Italian, Arabic and French in the West and South to Greek and Turkish in the East.

Most importantly for the fisheries and their resource exploitation, however, is the fact that most Mediterranean countries do not fully monitor their fisheries in such a way that all withdrawals of biomass (i.e. landings plus discards) are accounted for over time, as is required for ecosystem-based management of fisheries (Pikitch et al. 2004). This is particularly true for small-scale fisheries, whose catches are suspected to be systematically underestimated throughout the world (Pauly 2006), the Mediterranean being no exception, and for recreational fisheries, which are rarely accounted for in official statistics. However, the scale of this underestimation is usually not known, and alternatives to the national/FAO datasets are rarely available.

Here, we document the concepts and methods used to ‘reconstruct’ the likely total withdrawals (catches) removed from the Mediterranean waters of Spain, France, Italy, and Turkey from 1950 to 2010. We present the estimated ‘real’ catches, and then discuss some of their implications for fisheries management.

The Mediterranean catch from the small-scale (i.e. artisanal, recreational and subsistence) and industrial fisheries of Spain (excluding the Balearic Islands, but including the Gulf of Cadiz because this Atlantic region shares similar ecological and socio-economic traits with its adjacent Mediterranean region), France (excluding Corsica), Italy (excluding Sardinia and Sicily), and Turkey (Fig. 1) were ‘reconstructed’ for 1950-2010 based on concepts in Pauly (1998) and the approach outlined in Zeller et al. (2007). Pauly asserts that:

- There is no fishery with ‘no data’, because fisheries, as social activities, will tend to impact on other sectors of an economy, such as employment and seafood supply, or through their fuel requirements. Thus, it is nearly always possible to estimate the ‘size’ of a fishery and its catches, even if approximately.

- When compiling statistics, it is easy to misinterpret the fields of questionnaires (or other forms) that are left empty because the catch of certain fisheries is not available. Thus, subsequent users of one’s statistics will interpret the empty fields as zeroes, and the notion of unknown but non-zero catches is lost.

Based on this, the approach of Zeller et al. (2007) for reconstructing catches consists of a six-step process, which may be summarized as:

1- Identification of existing reported catch time series:

a- Landing reported by FAO, by taxon and year.

b- National and regional data series by taxon and year.

2- Identification of sectors (e.g. recreational fishing), periods, taxa, gears, etc. not covered by (1), i.e. missing fisheries. This may involve extensive literature surveys and inputs by local experts.

3- Identifying alternative information sources for (2), via extensive survey of the peer-reviewed and grey literature and the input of local experts.

4- Development of data ‘anchor points’ in time for each missing item, and expansion of anchor point data to country-wide catch estimates.

5- Interpolation for time periods between data anchor points, often via per capita (or per-fisher) catch rates for non-commercial sectors.

6- Estimation of total catch time series, consisting of reported catches (1) and interpolated, country-wide expansion of previously missing data (5).

These basic rules were adapted to our four countries, each of which had different conditions and datasets available for the reconstructions of total catches. Country-specific technical details on how these six points were implemented are presented in Coll et al. (in press) for Spain, Bultel et al. (in press) for France, Piroddi et al. (in press) for Italy, and Ulman et al. (2013) for Turkey, which collectively form the basis of this contribution (see also Table 1 for summary statistics for the separate countries, and Table 2 for a summary of results for each country).

Besides the obvious results, i.e. the wide divergence between our reconstructions and officially reported data, we should discuss a few commonly-held views that we encountered during the reconstruction work leading to the four reports that form the basis of this contribution. The views concern the availability of alternative data, the definition of small-scale/artisanal fisheries and the relationship between subsistence and recreational fisheries. We also discuss, as an aside, the willingness of colleagues to participate in this reconstruction.

It was not really difficult to identify data or to formulate reasonable assumptions with which to fill in the gaps in official data. The reason here is that the official data are not incomplete because “there are no data”, but because the national agencies reporting officially in the four countries we report on here (and in the other 150+ maritime countries for which reconstructions are available) do not see it as their mandate to report on any more than industrial fisheries catches and the formal part of the artisanal sector, leaving the informal part of the sector, as well as the subsistence and recreational fisheries, undocumented. It is quite straightforward, using the scientific literature (peer-reviewed and grey), to infer on the catch and catch composition of these neglected small-scale, but often very substantial fisheries. There is thus no legitimate reason for the national agencies reporting to FAO to add to the confusion generated by illegal, unreported and unregulated (IUU) fishing.

The same applies to discarded by-catch, which practically none of the national fishery agencies in the world appear to include in catch statistics, although they are definitely part (and sometimes an important part) of fisheries catches. Only one Regional Fisheries Management Organization (RFMO), the Commission for the Management of Antarctic Marine Living Resources (CCAMLR) includes discards in its catch statistics, while the other RFMOs, as well as the FAO, publish studies on discarding but do not include them in their statistics (see e.g. Kelleher 2005; Zeller and Pauly 2005). We mention this here as the issue of by-catch retention, and hence the abolition of discarding, or at least its reduction, is a major issue for the reform of the European Common Fisheries Policy, which will have large implications for fisheries in European seas, including the Mediterranean fisheries (Sarda et al. in press).

Another issue (or rather non-issue) is the definition of artisanal fisheries, here understood to be a small-scale activity generating products that are sold, unlike subsistence fishing, conducted primarily for one’s own (or family) consumption, and recreational fishing, conducted primarily for enjoyment (see Griffiths et al. 2007). There is obviously a lot of overlap between these categories, but between countries, especially in the Mediterranean, there is far more coherence than one would think upon reading, for example, the literature on maritime anthropology (Pauly 2006). For the purposes of our reconstruction, we accepted each country’s definition of small-scale-artisanal fisheries, which generally amounted to using fixed or passive gear, and/or craft of 10-15 m, but excluding any active or dragged gear, which cannot be considered artisanal (Martín 2012).

There are also definitional problems concerning recreational (Cisneros-Montemayor and Sumaila 2010) and subsistence fisheries (Schumann and Macinko 2007), both of which are also neglected in official reporting systems, but which generate appreciable catches. In our case, their joint contributions ranged from 5% to 16% of total catches, with recreational fisheries, however defined, preponderant over subsistence fisheries (Table 2). In the course of completing the four historic catch reconstructions highlighted here, and others around the Mediterranean and elsewhere, in several cases we noted a distinct reticence by colleagues in government agencies to participate in work that might be seen as challenging their government, while this was less noticeable with university-based colleagues. We will not give details, but regret this situation, a rational response in the face of some governments’ actions (for the serious case of Canada, see Turner 2013). Other difficulties encountered during the catch reconstructions in the Mediterranean Sea include the facts that (1) much relevant documentation is not digitized and is only available in local languages, (2) access to digitized data in possession of regional and national administrations is not always publically available and requests to access these data are frequently ignored, and (3) the compilation of information from local stakeholders requires time to build trust and robust communication mechanisms. Therefore, the participation of local experts during the catch reconstruction is essential.

The main point of the contribution is, however, the disparity between the reconstructed catches and the official data. Overall, i.e., for our four countries and the entire 61-year period covered here, the reconstructed catches are 1.76 times the official data, but this global ratio hides more than it reveals. The ratio drops to 1.56 if discards are not considered; more interestingly, this ratio (with discards included) was 2.08 in the 1950s but dropped to only 1.78 in the 2000s (see above and Table 2 for details on a per-country basis), which is encouraging.

Currently, we do not present confidence interval or other measure of uncertainly with our reconstruction, although an approach combining the pedigree of (and hence confidence in) the data upon which they are based (Funtowicz and Ravetz 1990) with a Monte-Carlo approach (Ainsworth and Pitcher 2005; Tesfamichael and Pitcher 2007) may be used toward this purpose in the future. It is important to realize, however, that conventional confidence intervals, expressing how close one would be on average to a given result if one were to repeat its estimation a large number of times (i.e. dealing with its precision) are perfectly inadequate here, because we are dealing with the bias (lack of accuracy) caused by deliberate omission of a catch component.

Finally, we should briefly reflect on what, if any, are the management implication of a catch that is about twice as high as is conventionally assumed. One obvious implication is that the Mediterranean may be twice as productive, with respect to fisheries, than has been assumed so far, which may modify certain development plans, especially as they are now supposed to be ecosystem-based. However, real fishing effort has increased exponentially in the region and is substantially higher than nominal effort, thus painting a very different picture that indicates a process towards depletion of resources over time. Also, this result suggests that the institutional neglect that small-scale (artisanal, recreational and subsistence) fisheries suffer from may be actually based on the misunderstanding that small-scale fisheries must generate small catches. Our results demonstrate that from 2006 to 2010, from 13% (in Spain) to over 42% (in France) of total catches in the Mediterranean (discards excluded) are contributed by small-scale, mainly artisanal and recreational fisheries (see Table 2). The systematic underestimation of the small-scale fisheries of the Mediterranean is part of a global phenomenon (Pauly 2006) that will have to be overcome if the potential of these fisheries for sustainable exploitation of coastal systems is to be realized.

Increased tourism, seafood consumption and recreational fishing in the Mediterranean region will continue to exacerbate stresses on the coastal ecosystem and its resources. With many key commercial taxa showing a declining trend (Abdul Malak et al. 2011), it is necessary for fisheries managers to fully apprehend how much fish have been and are being removed and by whom, so that appropriate decisions regarding the future of these resources can be made.

The rationale of, and methods for, historic fisheries catch reconstruction are presented, along with some results for four Mediterranean countries: Spain, France, Italy and Turkey. Jointly, these reconstructions illustrate the fact that the government agencies of countries bordering the Mediterranean neglect small-scale (artisanal, recreational and subsistence) fisheries, and hence their contribution to local seafood supplies. Data availability for this reconstruction was generally good, and their obstacles are mainly perceived ones. Examples of such assumed, but easily overcome, obstacles were alleged scarcity of data, the assumed disparity of definitions of small-scale fisheries between different countries, and the absent quantification of uncertainty for reconstructed catch estimates.