Effects of bottom trawl selectivity on populations, fisheries and biodiversity

Doctoral Dissertation uoadl:3228785 17 Read counter

Department of Biology
Library of the School of Science
Deposit date:
Mytilineou Chryssi
Dissertation committee:
Μεγαλοφώνου Περσεφόνη Καθηγήτρια, Τμήμα Βιολογίας, Ε.Κ.Π.Α., Επιβλέπουσα
Νικολαΐδου Άρτεμις Ομότιμη Καθηγήτρια, Ε.Κ.Π.Α.
Smith Christopher J. Διευθυντής Ερευνών, Ι.ΘΑ.ΒΙ.Π.Ε.Υ., ΕΛ.ΚΕ.ΘΕ
Κουτσικόπουλος Κωνσταντίνος Καθηγητής, Τμήμα Βιολογίας, Πανεπιστήμιο Πατρών
Μαραβέλιας Χρήστος Αναπληρωτής Καθηγητής, Τμήμα Γεωπονίας Ιχθυολογίας & Υδάτινου Περιβάλλοντος, Πανεπιστήμιο Θεσσαλίας
Ραΐτσος-Εξαρχόπουλος Διονύσης Επίκουρος Καθηγητής, Τμήμα Βιολογίας, Ε.Κ.Π.Α.
Τσερπές Γεώργιος Διευθυντής Ερευνών, Ι.ΘΑ.ΒΙ.Π.Ε.Υ., ΕΛ.ΚΕ.ΘΕ
Original Title:
Επιπτώσεις της επιλεκτικότητας του αλιευτικού εργαλείου τράτα βυθού σε ιχθυοπληθυσμούς, αλιεία και βιοποικιλότητα
Translated title:
Effects of bottom trawl selectivity on populations, fisheries and biodiversity
Selectivity is a fundamental feature of a fishing gear, which describes the probability of a fishing gear to select or retain specific species and/or specific sizes of a species. This is distinguished into the species- and size-selectivity, with the latter gathering the majority of the relevant research efforts to date. Selectivity provides crucial information for the proper management of fisheries and the identification of management measures since it is related to several important issues such as stock sustainability, discards, bycatch, biodiversity, fish behaviour and fisheries efficiency, viability and profitability. Therefore, selectivity studies should be examined along with these issues. The present PhD thesis aims to investigate the effects of the bottom trawl codend selectivity on fish populations, fisheries and biodiversity. At the same time, it is investigating the best scenario for the codend, in line with the current EU Mediterranean legislation, by testing different types of meshes in the codend. Finally, it is also seeking to identify the main behaviour patterns of fish that can support the improvement of the trawl codend selectivity. Although the experimental fishing for data collection was carried out in the fishing grounds of the Greek waters, this research aims to provide information useful for the Mediterranean bottom trawl fishery in general.
Data were collected during two experimental fishing surveys carried out in the trawl fishing grounds of the South Aegean Sea (Cyclades Plateau, Saronikos Gulf, Petalioi Gulf-S. Evoikos Gulf and N. Evoikos Gulf) in September-October 2014 and May-June 2015. Fishing depth extended from 50 to 310 m depth, where the Greek trawl fishery is usually operating. A professional trawler was hired to achieve conditions similar to those of the trawl fishery, equipped with a commercial trawl in which the codend was made of three different nets (40 mm diamond, 40 mm square or 50 mm diamond meshes, respectively). In total, 162 valid hauls were conducted using the well-known cover-codend method for the collection of data in the selectivity studies. Sampling was based on a three-fraction design including separately escapees (collected in the cover), landings and discards (the retained catch in the codend, sorted by the fishers onboard the vessel). This sampling design was an innovative scheme compared to the previously used two-fraction design (escapees/retained catch). Analyses, applying double bootstrapping procedures to account for between- and within-haul variation, in order to estimate uncertainty with a high degree of accuracy, were implemented using the SELNET software package.
Multispecies fisheries, such as the Mediterranean bottom trawl fishery, have always been a challenge in fisheries management. Trawl codend selectivity and discards are issues related to fisheries management, but separately studied and modelled to date. Under the aim to investigate the trawl selectivity and its effects on fish populations, discards and fisheries, an innovative approach was developed, for the first time, to simultaneously model two consecutive size-selection processes concerning the total population of a species entering the trawl codend. These selection processes are related to both the gear in the sea (codend size-selectivity) and the fisher onboard the fishing vessel (fisher size-selection). The model (overall selection model hereafter) was based on the new three-fraction sampling design (escapees/discards/landings), in contrast to the two-fraction design used in selectivity studies so far (escapees/retained catch). This offered the possibility to model simultaneously the escape probability through the trawl codend (resulting in escapes) and the subsequent discard and landing probabilities occurring due to the fisher’s selection process on the vessel deck (resulting in landings and discards). As a first step, three different trawl codends and three species were investigated in this initial case study for the evaluation of the model. The overall selection model, was based on the hypothesis that a fish entering the trawl codend can follow a multinomial distribution with three probabilities, the escape, discard and landing probability. The model described the escape and the landing probabilities as S-shaped curves leading to a bell-shaped curve for the discard probability, which is affected by both the gear and fisher selection. The model fitted well to the experimental data of the three species in the initial case study as well as the experimental data of 11 additional species afterwards, and provided at the same time parameters for selectivity, discards and fisher selection, useful in fisheries management. The proposed method can be considered a simple and cost-efficient approach to achieve such information and can be alternatively used if other kind of data (e.g. by observers onboard fishing vessels) are not available. The model was also found to be useful in predicting the size structure of discards and landings of a population entering the trawl codend or the estimation of the population entering the trawl codend when discards and landings are available, and testing different scenarios for potential technical measures in trawl fishing, which consist important information in fisheries management.
The overall size-selection model to estimate the trawl codend selectivity and fisher selection parameters was applied for five commercially important fish species (Merluccius merluccius, Mullus barbatus, Mullus surmuletus, Pagellus erythrinus, and Lophius budegassa). Commercial bycatch species and their exploitation pattern in the Mediterranean trawl fishery are little studied. Therefore, 9 additional bycatch fish species were also studied (Trachurus trachurus, Micromesistius poutassou, Boops boops, Serranus cabrilla, Lepidorhombus boscii, Helicolenus dactylopterus, Chelidonichthys lastoviza, Citharus linguatula, and Spicara flexuosa) by applying the overall size-selection model. For these purposes, three codends, made of 40 mm diamond (40D), 40 mm square (40S) and 50 mm diamond meshes (50D), were tested; the first one, is no longer allowed in the EU Mediterranean countries, but still in use in other Mediterranean areas, while the last two, are in use under several established EU Regulations. For each species, the model offered the possibility to simultaneously describe the escape, discard, and landing probability of the studied species for each codend. The results, useful for fisheries management, showed that the 40D codend was always detrimental for the stocks. The 40S codend was more appropriate for the sustainability of 5 of the studied species (M. barbatus, M. surmuletus, M. poutassou, B. boops, S. cabrilla), providing also a lawful catch along with greater compliance to the rules by the fisher, negligible discards and the lowest possible economic losses for the fisher. The 50D codend was more suitable than the 40S for the flatfish C. linguatula sustainability; however, because of the fisher selection for only the largest individuals for landing, even this codend was characterised by a high discard probability. Both the 40S and the 50D codends were appropriate for S. smaris. However, for most of the stocks, the 50D codend did not provide a better selectivity than the 40S, which is required for its use according to the Regulations of EC. Therefore, the prohibition of the use of the 50D codend should be examined in the near future. None of the codends was effective for M. merluccius, T. trachurus, C. lastoviza, H. dactylopterus, P. erythrinus, and L. boscii for maintaining catches above the minimum conservation reference size (MCRS) or the size at first maturity (LFM) of the species, although the escapement of the juvenile M. merluccius and P. erythrinus was favoured by the 40S and the 50D codend, respectively. All codends were harmful to L. budegassa as the majority of juveniles were retained by the codend, resulting in negligible escapees, high discard probability, and landings of a size much lower than the length at first maturity of the species. In almost all cases, the fisher behaviour for each species was consistent in the three tested codends, selecting for some species individuals smaller than the MCRS or LFM (e.g. M. merluccius, C. lastoviza), larger for others (e.g. T. trachurus, H. dactylopterus, L. boscii) or similar to MCRS or LFM if no juveniles were present in the catch (e.g. Mullus spp., M. poutassou, S. cabrilla). In general, fisher behaviour was more lawful when 40S was in use. Nevertheless, even in the case of the 40S codend, trawl codend selectivity improvement is still needed for many species. For some of the species, it was obvious that the only solution for their juvenile protection was the protection of the nursery grounds (e.g. L. budegassa) since the required increase in the mesh size in the codend would result in the loss of many other important commercial catches. It will be useful to conduct further studies for additional species in the future, as the trawl fishery in the Mediterranean has a multi-species nature.
Reducing juvenile catch and discards are important goals in fisheries policy. Therefore, special emphasis was given to hake (M. merluccius), which is an important commercial species with many undersized individuals, in both the discards and landings (indicated in the previous results on the gear and fisher selection processes). For this purpose, the overall size-selection model for hake was used to estimate the size structure (LFD) of discards and landings in areas and periods of high and low recruits’ occurrence, using the two legal codends (40S and 50D). In addition, various exploitation indicators, related to the gear and fisher selection were also examined. The results showed that both examined codends were inappropriate for the sustainable fishing of hake in areas and periods of high recruits’ occurrence; 50D performed worse than the 40S codend. Fishing with both codends in areas and periods of low recruits occurrence revealed the lowest discards rates and minimum sizes of landings close to MCRS. The exploitation indicators showed that the discard ratio (discards/retained catch) does not represent the real situation, due to the fact that the fisher selects undersized individuals (MCRS) in areas and periods of low juvenile occurrence. However, this may reach the value of 30%, when the presence of young individuals is high. This can explain why fishers select undersized hake as landings. The results provided useful information for fisheries management aiming to avoid juvenile catches and reduce discards. Spatio-temporal trawl fishing closures in areas and periods of high occurrence of hake recruits (nursery areas) were discussed as potential measures for the sustainability of hake stocks, since the improvement of the trawl selectivity seems difficult to be achieved for this species in the near future.
Fisheries is well known to have important effects on marine biodiversity. Therefore, in the present thesis, various parameters related to this topic, such as the abundance, species richness, diversity indices, species composition, trophic level and vulnerability index, were investigated to detect the effects of trawl codend selectivity on biodiversity. The aim was to identify differences between the fish assemblage entering the trawl codend and the fractions derived (escapees, retained catch, discards and landings), because of the overall selection pattern (gear and fisher selection) during trawl fishing. The three previously mentioned codends (40D, 50D, and 40S) were also examined and compared for this purpose too. In general, the trawl codend was found to produce a fraction of escapees that was always lower in abundance, richness, and vulnerability index, similar in diversity indices and trophic level, and different in species composition compared to the fish assemblage entering the codend. Fishers selected as landings a fraction that was always the lowest in species richness and diversity indices, and the highest in trophic level. In addition, fisherσ discarded also a fraction that was always the highest in diversity and vulnerability index, and the lowest in trophic level, compared to the other fractions, even when compared to the total fish assemblage entering the trawl. In this fraction of discards, the Elasmobranchs, species of high vulnerability and very low escaping probability, showed the highest percentage in numbers. Furthermore, it should be highlighted that a great number of species (~220) were identified in the fish assemblage entering each trawl codend from which approximately 83% were retained in the codend, indicating the multi-species nature of the Mediterranean trawl fishing. Moreover, about 60% of the total number of species escaped, but approximately 80% were discarded, and only 23% (~50 species) were landed. From the latter fraction, 6 species only (Parapenaeus longirostris, S. smaris, Illex coindetii M. merluccius, M. barbatus, M. surmuletus) constituted approximately 90% in abundance. The results indicated the low effectiveness of the gear, since from the great number of species entering the codend only a small proportion is finally marketable. On the other hand, for the escaping species and individuals, no information is available about their survival rate, which is important for the assessment of the effects of the selectivity of this fishing gear. Information gathered in this thesis on the condition and activity of species inside the trawl codend, permit the hypothesis of an important survival likelihood for some of them, with a low or null survival probability for others. Although the three codends did not differ significantly in the fraction of escapees in terms of diversity indices, trophic level, and vulnerability index, the 40S codend showed in general a significantly higher percentage in the number of escaping species and individuals (particularly for hake), and less differences in the species composition in relation to the total entering assemblage. Ιn addition, in terms of numbers, lower percentage of discards and higher of landings in the retained catch (0.6:1) was found in the 40S codend compared to the other two codends (0.9:1). Although the 40S codend showed less effects on the biodiversity of the fish assemblage in the sea, the results support proposals concerning an urgent modification of the trawl for the elimination of discards, and particularly modifications to enhance the escape of the highly vulnerable species included in this fraction (e.g. Elasmobranchs). In addition, it should be taken into account that diversity indices and trophic level of the fish assemblage entering the codend seem to change gradually over time (so not easily detected in the short term), because of i) the continuous non-removal (escape) of small-sized organisms, which are of lower trophic level and vulnerability, and ii) the continuous removal (catch) of the largest organisms, which are of higher trophic level and vulnerability. This may greatly affect overall biodiversity in the long term.
Finally, the investigation of the species behaviour in the trawl codend revealed different features for the 32 species studied. Some species (B. boops, Pagellus acarne, Pagellus bogaraveo, Dentex maroccanus, Mullus spp., Spicara spp., Trachurus spp., Engraulis encrasicolus) showed a very high percentage of dynamic condition and high activity, which indicated a high escape probability and survival rate after escaping through the gear in the sea. In contrast, some species (L. budegassa, Munida spp., P. longirostris, Sepia spp., Zeus faber, M. merluccius, Arnoglossus spp.) showed a high percentage of passive condition, and a high percentage (30-50%) of moribund individuals, evidence of a low probability of escape or survival after escape. For some species (M. merluccius, Serranus hepatus, Z. faber), the dynamic condition was greatly reduced (> 50%) during the process of trawling, while for others this was affected to a lesser degree (Centracanthus cirrus, S. cabrilla, P. erythrinus, Trachinus spp.). From the 32 species studied, 23 showed some escape activity (escape attempt, escape, entanglement, entanglement and disentanglement). Εscape activity seemed to take place during the whole process of the bottom trawl activity, however, this was generally reduced over time. Some species showed an increase in the escape activity in the final phase when the trawl was on the surface (B. boops, E. encrasicolus, Sardina pilchardus, D. maroccanus, P. erythrinus, P. bogaraveo, Spicara spp.). In most of the species, the area of escape activity was related to the area of their occurrence inside the trawl codend. Pelagic species were located in the upper area of the codend, while benthic species in the lower and lateral areas. Many demersal species were observed in the middle, lateral and lower areas of the codend. However, when the trawl was at the sea surface, many species, even the pelagic species, changed behaviour and were located mainly at the bottom of the codend. According to these results, if trawl modifications are planned, the behaviour of the species should be taken into consideration. Changes to the net in the upper part of the codend will favour the escape of pelagic species, while escape of benthic and demersal species would be enhanced by modifications to the bottom and side part of the codend. Furthermore, since the contact probability of the organisms with the net was found very low, it is necessary to identify modifications and interventions that can enhance this probability (e.g. guiding panels). Among the three codends tested, species in the 40S presented more escape activity than the 50D; and in the latter codend more than in the 40D. This is in agreement with the results on the gear selectivity based on the overall selection model.
In summary, based on all the above-mentioned results, it can be suggested that although the 40S codend seems more selective, with less impact on some stocks and the ecosystem, additional gear modifications are still needed to reduce the effects on both stock sustainability as well as biodiversity conservation. Further increase in mesh size in the trawl codend is not a promising modification without financial losses for fishers, due to the expected increase in escape of commercially important species (e.g. Mullus spp.). Investigation of other technologically innovative changes are needed to be defined, especially those that can more improve species- than size-selectivity, with changes that will support the escape of fish that would otherwise be discarded. This is expected to minimize both biodiversity loss as well as discards. Other modifications (e.g. escape device, guiding panelss) could also be tested to improve species- and size-selectivity, but their use needs to be acceptable (and practical) to the fishers and stakeholders.
The current doctorate thesis provided important information for the management of the Mediterranean bottom trawl fishery related to selectivity of the gear, selection pattern of the fishers and their effects on stocks and the ecosystem. Information on the behaviour of species was also extracted, which should be taken into account in future studies aiming to improve trawl selectivity. Given that the bottom trawl is an important and necessary fishing gear (providing 20% of total fishery production), this information can contribute to the future mitigation of the effects of this gear, leading to a more efficient but at the same time more environmentally friendly activity, which is also an important socio-economic activity for the fishing sector.
Main subject category:
selection models, discards, biodiversity, species behaviour, trawl impacts
Number of index pages:
Contains images:
Number of references:
Number of pages:
File access is restricted until 2025-08-05.

16 MB
File access is restricted until 2025-08-05.


15 MB
File access is restricted until 2025-08-05.