Coastal and maritime activities have traditionally been important for the national economy and the historical, social, and cultural identity of Portugal (Leitão and Baptista 2017). The country has long relied on fishing as a major means of subsistence and many coastal communities depend almost exclusively on small-scale coastal and estuarine fisheries and related activities. However, over time substantial technological improvements and changes have been made in the fisheries. For example, in the around 1850, steam-powered vessels were introduced to the fishing fleets, resulting in a reduction of total fishers (Alves 1991). Additionally, fishers began to deploy for the first time an industrial gear, the otter trawl, which immediately created conflicts between the small-scale sector and this newly developing industrial sector (Baldaque da Silva 1891, Alves 1991). According to Hill and Coelho (2001), there was a decrease in the number of vessels in the Portuguese fishing fleet between 1989 and 1999, but this was compensated by an increase in vessel power. By 1996, 98% of the fishing fleet was motorized—a 2% increase from 1986. Today, in mainland Portugal, a variety of gears/métiers are used in the coastal fisheries, ranging from trawls to static gears such as gill nets and traps. Therefore, a wide variety of unwanted species are captured along with the target species (Gaspar et al. 2003, Gonçalves et al. 2007, Bordalo-Machado et al. 2009). Different types of gear often compete for the same resources (Borges et al. 2001). However, as different gear types and métiers target different organisms (Watson et al. 2006a, b), unreported catches such as discards also differ from métier to métier.

Discards refer to the part of the catch that is not retained on board during commercial fishing operations but is returned to the sea. Discarding of marine organisms is a widespread feature of commercial fishing operations. Discard patterns are affected initially by catch compositions, which are determined by environmental factors, the fishing gear and fishing tactics used, and ultimately by fishermen themselves when they decide which parts of the catch to retain. This decision is influenced by both market and regulatory conditions, and is constrained by space and time: storage space on board the vessel and sorting time (Catchpole et al. 2014). None of the historical accounts published between 1800 and 1950 on Portuguese fisheries address unreported catches, by-catch or discards. The first study for the purse seine was published by Borges et al. (2001) in 1997. The lack of information on by-catch and discards from this period might suggest that either most of the nearshore, artisanal catches were consumed or used and not discarded, or that discarding may have been low and utterly ignored. Brandão et al. (2000) describes how all fish were processed, salted, and dried by Portuguese women, indicating that fish discards may have been minimal between 1800 and 1950.

The way different gears operate suggests that long-term monitoring is required to improve our understanding of the factors affecting discarding and of the implications of the levels of discarding on the marine community structure (Hollingworth 2000, Kaiser and de Groot 2000). The knowledge of the overall quantity of species caught in coastal marine systems (including unreported catches) is key to understanding the indirect effects of removal of particular taxa from the system. In fact, over the past three decades, renewed interest in a more ecological approach to fisheries (an ecosystem-based management approach) has emerged.

The new European Union Common Fisheries Policy, which started to be implemented in 2014, sets out a gradual elimination of discards by reducing unwanted catches and ensuring that all catches are landed. Illegal, unreported and unregulated catches (IUU) are one of the most important topics in fisheries from both an economic and an environmental point of view (Alverson and Hughes 1996, Kelleher 2005, FAO 2010). The quantification and composition of the unreported catches and the understanding of the fate and impact of these unreported actions are key issues in fisheries (Zeller et al. 2007, 2011).

In Portugal it was estimated that an average of 123495 t/y (35.5% of the total catch) was unreported between 1938 and 2009 (Leitão et al. 2014). Overall, reconstructed total catches in Portugal amounted to just under 21.6 million t in 1950-2010, which is slightly more than twice the 10592310 t of landings officially reported by Portugal for the same time period. Discards contributed the most to the unreported catches, accounting for 7.6 million (i.e. 35%) t of total catches.

Many fisheries around the world have reached unsustainable levels and therefore deliver poor income to fishers. An effective fisheries management is urgently needed to improve the economic situation of fishing communities. Part of the solution is to reduce discards by finding market-based approaches that will increase the value for all by-catch fish (Leitão and Baptista 2017). The necessity of each country to manage all fisheries within their Exclusive Economic Zones (EEZ), a consequence of the United Nations Convention on the Law of the Sea (UNCLOS), led to attempts to find sustainable indicators for marine fisheries and ecosystems at the national level, including economic effects. However, information about unreported discard ban species that can comprise additional alternative value to the fishery sector is still scarce. Prohibited for the first time in some EU fisheries in 2009, economic-led high-grading is today illegal for all quota species, under amendments to fisheries technical measures enacted by the European Parliament and Council in March 2013 (Regulation (EU) No 227/2013). This means that fish that were discarded before should now have an economic value independently of their final use. Furthermore, sales of this fish will have to be accounted for and included in the country’s economy (Leitão and Baptista 2017). However, so far the amount of information regarding total volumes and species discarded is lacking. Underestimation of catches is especially important in countries where fishing fleets are highly diversified, the enforcement of fishery management is low, data availability is poor, and there is high demand for fish products in local markets (Coll et al. 2014). Estimation of unreported catches for Portuguese fishery was based on a fishery-by-fishery approach by Leitão et al. (2014). Herein we used information of unreported catches, for each commercial métier (from Leitão et al. 2014) and we reconstructed taxonomic profiles of unreported catches, namely discards by commercial fishing sector.

There are few studies on reconstruction of unreported discards with higher taxonomic resolution, because of the enormous time required for obtaining sound fisheries information, processing the data and developing/applying accurate methodologies. After we estimated IUU in a previous study (see methods in Leitão et al. 2014), we reconstructed unreported discard profiles, because enhancing taxonomic information on discards is imperative for fisheries management within the new CFP and the landing obligation directive. Inevitably, reconstructions of catches are largely based on assumptions derived by analyses of recent data (e.g. discards studies, Table 1). For example, in the present work, the catch composition and discard ratios were mainly based on studies dating from after the mid-1990s but the reconstruction goes back to 1938. Furthermore, the recent studies may cover a small region, raising the question of whether the discards estimates apply to the whole Portuguese mainland. The Portuguese fishery is characterized by nearshore fisheries with the top rank preference in terms of species changing little over time (Almeida et al. 2015). In fact, small pelagics (Sardine pilchardus, trachurus spp. and Scomber spp.) and European hake, for instance, account for the greater proportion of the catches (landed and discards). These groups/species are data rich in the INE long-term database. Therefore, for the main métiers and traditionally consumed species, the rebuilding and taxonomic disaggregation of the discarded species (most of which match the landed species) should be considered more accurate, whereas in the case of species with smaller catches, rebuilding procedures may introduce larger estimation errors. However, the number of unreported species discarded is independent of the percentage in weight of the unreported discards estimated. That is, we can assume that qualitative analyses might be less affected than quantitative estimations in rebuilding methods. The average total number of marketable taxa landed per year is around 296 (Source: DGPA 1989-2009), with 225 taxa being discarded. Of the 225 species discarded, approximately half (109 taxa, 48%) are also landed or reported/discriminated at auction (according to the DGPA database). Therefore, this study showed that 89% of unreported bony fish and shark species are thought to have commercial value.

The compositions of unreported and landed catches do not vary much in terms of the main groups caught (Fig. 1, Table 2). In the Portuguese mainland fisheries the landings comprised mostly fish (84.5%), with cephalopods and crustaceans accounting for 2.9% and 1.3% of the total catches, respectively. For the period 1938 to 2009, the average landings of fish, cephalopods and crustaceans were 207419, 7162 and 3187 t (Fig. 1, Table 2). Fish accounted for 93.5% of the total unreported catches, with an average of 115493 t/y. The contributions of cephalopods and crustaceans to the unreported catches are minor compared with those of fish: 1.9% and 1.4% of total unreported catches, with averages of 2345 and 1754 t/y, respectively (Fig. 1, Table 2). The shark group accounts for 1.3% of total unreported catches in weight (1638 t/y). The recreational/subsistence fishery focuses mainly on fish, but there are no studies or data available on recreational fisheries/harvesting that target small invertebrates such as mussels, goose-barnacles or, more recently, sea urchins.

Between 1938 and 2009, sardine (S. pilchardus, 44.8%), horse mackerel (T. trachurus, 14.8%), hake (M. merluccius, 5.7%), chub mackerel (Scomber colias, 5.6%) and octopus (Octopus vulgaris, 2.3%) together accounted for an average of 73.2% of the landings (163826 t/year) and 36.8% of unreported catches (450411 t/year) (Table 3, Fig. 2). The unreported species included mainly S. colias (17.6%), Boops boops (9%), Trachurus picturatus (8.6%), M. merluccius (8.3%), S. pilchardus (7%), Liza aurata (7%), Micromesistius poutassou (6.9%) and T. trachurus (3.6%), all marketable species that together accounted for 68.2% of the total average annual unreported catches of approximately 84222 t (Table 3, Fig. 2; see also Supplementary Material Table S2). Species such as the two-banded sea bream (Diplodus vulgaris) and the Senegal sea bream (Diplodus bellottii) are often discarded in purse seine fisheries when they are small-sized but above the minimum legal size, as taking them to auction is considered not worthwhile (Gonçalves et al. 2008).

The composition of landings varies considerably according to a number of factors, including the nature of the fishery, the type of fishing gear used, gear selectivity, tow duration, the target species and their price value, the depth of capture and the time of year (Oliver 1993, García-Rodriguez and Esteban 1999, Rochet et al. 2002). As expected, all the above factors also affect the discard species composition of each métier. In mainland Portuguese fisheries, the number of unreported/discarded species was higher in the multi-gear sector (N=184), followed by trawl (N=79) and seine (N=9) (Table 4). The large number of taxa in the multi-gear category is due to the differences between gears, fishing grounds and target species. Within the multi-gear métiers, the number of species discarded from trammel nets (N=120 taxa) was far greater than that of other static gears such as scabbardfish longlines (N=29), gill net and longline (N=22), artisanal dredges (N=18), trap (N=16) and demersal seine (N= 5) (see Supplementary Material Table S2). This is due to the greater diversity of trammel net catches compared with other static gear (Martins et al. 1992, Erzini et al. 2003) and can be accounted for by the species and size selectivity of trammel nets (Erzini et al. 2006, Stergiou et al. 2006). In the Algarve (south coast of Portugal), one exhaustive study showed that more than 900 species can be caught and discarded by the commercial fishery (trawls, purse seine and trammel nets): 69% are always discarded, 27% are frequently discarded and only 4% are occasionally discarded (Borges 2007). The number of taxa recorded above is far greater than those reported herein that were based on specific scientific literature. This finding might be related to the fact that scientific surveys are usually restricted to short time periods and are also limited in terms of the geographic area surveyed (scientific surveys onboard commercial boats allow exhaustive faunistic records to be obtained).

In multi-gear fisheries the unreported catches consisted mainly of L. aurata, S. colias, S. pilchardus, Trachinus draco and B. boops (Fig. 2, Table 3). Together, the latter species account for 76.4% of the multi-gear discards, with an average of 17935 t/y. The unreported multi-gear catches of S. colias, S. pilchardus and B. boops were mostly due to demersal seine and trammel net discards, while those of L. aurata were mostly due to demersal purse seine discards (see Supplementary Material Table S2).

As in the multi-gear category, B. boops, S. colias and S. pilchardus were the species most discarded by purse seiners (Table 3). Together with Belone belone, these species accounted for 79% of the unreported purse seine discards, with an average of approximately 19027 t/y. In purse seiners that use electronic equipment to detect the schools around which the seine net is set, the lack of success in determining the species and/or size composition of the fish in the school before setting the net is a major factor leading to high volume discards. In fact, the target species (sardine or horse mackerel) may also be captured and discarded when mixed with by-catch species, making the sorting of large catches uneconomical, and when the sizes caught are not suitable for the market or for canning (Borges et al. 2001).

In Portugal, the “trawling” category includes two different fleet components: deepwater trawlers that target crustaceans, and fish trawlers that operate mainly on the continental shelf (CEC 1993). Fishing trip duration is one of the most important factors influencing the proportion of the fish by-catch that is commercialized, and the quantity of by-catch landed is inversely related to trip duration (Clucas 1997, Costa et al. 2008). The main species unreported due to trawl discards differed from both the multi-gear (small-scale/artisanal) and purse seine fleets. For the trawls, T. picturatus (16%) and M. merluccius (13%) were the most discarded species, accounting for 29% of the unreported catches and approximately 19514 t/y. Together with S. colias (13%) and M. poutassou (13%) these species comprised more than half (55%) the unreported trawl catches. The occurrence of high concentrations of small, non-commercial species such as Capros aper and Macroramphosus scolopax accounts for the occasional high volume discards witnessed onboard trawlers (Borges et al. 2001). However, in this analysis different trawl studies were used and C. aper was the sixth most important species in terms of trawl discards.

Despite some overlap in the species that contribute most to unreported discards of different metiers, some significant differences were found. In fact, the discards of sharks were always higher than landings in all the time series, which is not surprising in view of discard rates for most species (see studies on trawl, Table 1). In fact, the catch ratio of IUU sharks/total sharks showed that 71.6% of the sharks are discarded without being reported (Table 2). The discards of sharks have increased in the last few decades although landings of sharks have not. This finding may also be related to discards of deepwater sharks, which were formerly used to produce liver oil, including during the Second World War. Compared with other sectors (see also Supplementary Material Table S2), trawlers (mainly crustacean trawls) discard considerable quantities of mainly deepwater sharks such as Scyliorhinus cannicula, Galeus melastomus, Etmopterus pusillus and Hexanchus griseus, which may have poor resilience to high levels of fishing mortality because of their life history characteristics (Stevens et al. 2000). In such deepwater communities with long-lived, slow growing, low-fecundity species, fishing activity with associated discard-related mortality may be expected to severely impact some populations of non-commercial species and in the long-term result in community changes (Kaiser and de Groot 2000).

The results showed that independently of the métier, common marketable species account for most of the unreported discarded catches. In fact, the species that most contribute to overall unreported catches are among the most frequently landed and are those that contribute most to overall landings. This point is important, because these species are considered choke species under the new Common Fisheries Policy landing obligation. So what are the implications of the landing obligation in relation to this finding and what is the value of this study? The identification of discarded species is a key factor for launching the debate regarding their use, particularly because most of them have quotas/total allowed catches (TACs). Until recently, the EU prohibited discards of fish with established quotas which could be legally landed (high-grading). However, it was legal to discard non-commercial fish and other organisms. As discussed by Leitão and Baptista (2017), it is difficult to know with certainty whether there will be any costs for fishermen if they land more fish than their quota for one or more species. In short, fish caught in excess of individual quotas can be marketed normally and “by-catch quotas” can be set as part of the fishing opportunities established by the EU council each year.

The difficulty of managing Portuguese fisheries can be largely attributed to their multi-gear nature, insufficient research (funding and lack of support for monitoring and analysis of non-target fisheries) and unreported catches, which affect stock assessment and management. Fisheries data collection, advice and management have traditionally been based on single-species approaches. However, ignoring interactions between métiers and species could lead to an undesirable situation in which fishing for one species may lead to discarding of another whose quota has already been exceeded. Moreover, the by-catch and discarding of non-target species may have negative consequences for non-commercial as well as commercial species due to influences on species interactions and consequent cascading effects throughout the trophic web (Harris and Poiner 1990, Hill and Wassenberg 1990, Yamamura 1997).

Borges (2007) state that the main reasons for discarding are economic restrictions (e.g. low or no commercial value of the species with no immediate market) and technical restrictions (fishing gear selectivity). Moreover, Bellido et al. (2011) reported that discarding may have a number of adverse ecological impacts on marine ecosystems, causing changes in the overall structure of trophic webs and habitats, which could in turn pose risks for the sustainability of current fisheries. Discarding is less frequently associated with legal/administrative restrictions such as quotas, minimum landing size and TACs. However, given the overfished state of many of the world’s most important stocks (Pauly et al. 2002, Leitão 2015), there has been great interest in documenting and finding solutions to the economic, political, and ecological implications of by-catch and discarding (Costa et al. 2008). Research on by-catch utilization is rapidly moving to the field of food and nutrition research, creating value-added fish products from by-catch or discarded fish: extracting gelatin from Alaska pollock (Theragra chalcogramma; Zhou and Regestein 2005) and shark cartilage (Isurus oxyrinchus, Cho et al. 2004), and even using these value-added products as alternatives to the use of mammalian gelatin (Karim and Bhat 2009).

There are still few summaries of estimates of unreported discards, especially with taxonomic detail. However, a complete review of IUU catches was made by Pauly and Zeller (2016). Overall, world results show that the taxonomic composition of unreported catches of the main target species vary considerably among areas, which is an expected result as fish assemblages, and target species, differ among regions. However, small pelagics and some demersal species are some of the most frequently reported species in the Mediterranean and Southern Europe (Coll et al. 2014, Pauly and Zeller 2016). Coll et al. (2014) showed that in Southern Europe unreported catches were due to (i) illegal catches of commercial species (undersized or with quotas, such as bluefin tuna), (ii) illegal fishing techniques (such as the Spanish driftnet fishery after the 1992 ban), and (iii) portions of misreported catches of protected species or species at risk (such as pelagic sharks). Illegal catches in the study area were mainly identified as juvenile commercial species such as juveniles of demersal species as hake or small pelagic fish such as sardines and anchovies. These results are similar to those found herein for unreported discards of species/groups. In Italy the main taxa discarded were clams (Bivalvia; 12.0%), sharks (Selachimorpha; 8.9%), jacks (Trachurus spp. 6.7%) and rays (Rajidae; 5.6%) (Piroddi et al. 2015). In fact, the worldwide unreported proportion is most often due to the discarded component (Pauly and Zeller 2016). It remains, however, to be determined whether worldwide total unreported discard species also match the commercial species most frequently landed and with the highest contribution to total catch, as in Portugal.

From an economic perspective, there are possibilities for making better use of some discarded species, thereby possibly reducing the pressure on target species (Leitão and Baptista 2017). Most of the discarded species in Portuguese fisheries have been shown to have economic potential. The critical fact is that sustainable management of fishing resources must take place in the ecosystem context, with a good understanding of all the possible effects of fishing activities (Borges 2007). Any effect on one stock, population or species may produce a change in another, resulting in readjustment in both populations (Hongskul 1979, Saila 1983, Kennelly 1995). Moreover, discard estimates are necessary, not only to evaluate the impact of fishing on non-commercial species but also on ecosystems as a whole (Alverson et al. 1994, Hall 1999), since they are not usually taken into account in stock assessments (Borges et al. 2005). Knowledge of unreported catches may change the way we assess the marine ecosystem, including the poorly understood trophic effects of fisheries in the marine environment, thereby improving our understanding of fishing trend variability and catch predictions.