Hostname: page-component-89b8bd64d-5bvrz Total loading time: 0 Render date: 2026-05-07T11:20:06.120Z Has data issue: false hasContentIssue false
  • English
  • Français

The influences of temporal-spatial parameters on CPUE of the Atlantic bluefin tuna purse seine fishery in eastern Mediterranean

Published online by Cambridge University Press:  15 February 2024

F. Saadet Karakulak Open the ORCID record for F. Saadet Karakulak [Opens in a new window]  and
Tevfik Ceyhan Open the ORCID record for Tevfik Ceyhan [Opens in a new window]
F. Saadet Karakulak
Affiliation:
Faculty of Aquatic Sciences, Istanbul University, Istanbul, Türkiye
Tevfik Ceyhan*
Affiliation:
Faculty of Fisheries, Ege University, Izmir, Türkiye
*
Corresponding author: Tevfik Ceyhan; Email: tevfik.ceyhan@ege.edu.tr
Rights & Permissions [Opens in a new window]

Abstract

In this study, we applied generalized additive model to investigate the influence of spatial temporal variables and vessel length on catch per unit-effort (CPUE) of Atlantic bluefin tuna (ABFT) purse seine fishery using catch and effort data from commercial logbooks and field surveys from 1992 to 2006. The vessel lengths of sampled purse seines ranged from 20 to 64 m. The number of ABFT caught within each operation varied between 1 and 2000. A total of 386 CPUE values for ABFT were calculated 0.05 and 60 t ⋅ (haul day)–1 with mean CPUE of 5.51 ± 0.54 t ⋅ (haul day)–1. Although the sea surface temperature had little influence on the CPUE, the effect of time and spatial variables, vessel length and salinity was found as significant. In conclusion, the spatial dynamics of ABFT should be considered if the impact of fisheries on the ecosystem is to be reduced.

Keywords

catch per unit-effortgeneralized additive modelMediterraneanspatialtemporalThunnus thynnus

Information

Type
Research Article
Information
Journal of the Marine Biological Association of the United Kingdom , Volume 104 , 2024 , e6
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
Copyright © The Author(s), 2024. Published by Cambridge University Press on behalf of Marine Biological Association of the United Kingdom

Introduction

Catch per unit-effort (CPUE) is able to obtain a relative index of the abundance of a fish stock especially for the closed populations in the presence of successive removals (Nishida and Chen, Reference Nishida and Chen2004; Skalski et al., Reference Skalski, Ryding, Millspaugh, John, Kristen and Joshua2005). Furthermore, CPUE could be described in the data section for trends in absolute abundance, and trends in mean age or mean body weight (Martell, Reference Martell, Jørgensen and Fath2008).

Despite being a substantial general acceptance about using the relative abundance indices based on CPUE as information in fish stock assessment, the raw data could be admitted as problematic in plain view (Maunder et al., Reference Maunder, Sibert, Fonteneau, Hampton, Kleiber and Harley2006). CPUEs are affected by a variety of factors such as years, season, area of fishing and environmental factors (depth, sea surface temperature, etc.). Maunder and Punt (Reference Maunder and Punt2004) argued that one of the most applied fisheries analyses is standardization of CPUE to remove the effect of these factors. The standardized CPUE could provide information about the effect of fishing on stocks. As Ducharme-Barth et al. (Reference Ducharme-Barth, Grüss, Vincent, Kiyofuji, Aoki, Pilling, Hampton and Thorson2022) stated, the fisheries-dependent CPUE will remain a common and informative input to fisheries stock assessments because of the cost and lack of availability of fisheries-independent surveys.

The Atlantic bluefin tuna (ABFT) holds a prominent position as one of the most highly sought-after fish globally, particularly due to its integral role in luxury sushi and sashimi markets. Its economic significance has further escalated since the advent of tuna farms in the Mediterranean during the mid-1990s. Additionally, as an export product, it holds a significant position within the Turkish fishing sector. The spatial dynamics of ABFT in the Mediterranean Sea are influenced by various factors, including seasonal migrations, feeding behaviour and reproductive activities (Druon et al., Reference Druon, Fromentin, Aulanier and Heikkonen2011). Bluefin tuna are highly migratory species that undertake extensive movements within the Mediterranean region. They exhibit a distinct pattern of migration, moving from their spawning grounds in the Gulf of Mexico and the Mediterranean to feeding grounds in the eastern Atlantic Ocean and back. These migrations are primarily driven by changes in water temperature, availability of prey and reproductive requirements (Block et al., Reference Block, Dewar, Blackwell, Williams, Prince, Farwell, Boustany, Teo, Seitz, Walli and Fudge2001, Reference Block, Teo, Walli, Boustany, Stokesbury, Farwell, Weng, Dewar and Williams2005; Varela et al., Reference Varela, Carrera and Medina2020).

The International Commission for the Conservation of Atlantic Tunas (ICCAT) currently assumes the existence of two discrete stock units which are in the Western and in the Eastern Atlantic with adjacent seas. Management measures, such as the establishment of marine protected areas and fishing quotas, have been implemented to conserve and sustainably manage ABFT populations in the Mediterranean. These measures aim to protect important spawning and feeding grounds, reduce fishing pressure and ensure the long-term viability of this iconic species. However, ICCAT (2020) reported that there was an obscurity period between the mid-1990s and 2007 on the catches of ABFT from the East Atlantic and Mediterranean due to unreported fishing. Thus, all kinds of information about catches for this period have great importance.

There are many statistical methods that can be used to standardize CPUE. Hinton and Maunder (Reference Hinton and Maunder2004) summarized the methods as general or generalized linear models (GLM), general additive models (GAM), neural networks, regression trees and others. In the eastern Atlantic and Mediterranean, the studies about the standardization of CPUE for bluefin purse seine fleets generally used just one variable (CPUE~year) and/or were based on GLM approach (Karakulak, Reference Karakulak2004; Zarrad and Missaou, Reference Zarrad and Missaou2018). Additionally, models with multiple variables have been employed to standardize the CPUE of bluefin tuna in baitboat (Rodríguez-Marín et al., Reference Rodríguez-Marín, Arrizabalaga, Ortiz, Rodríguez-Cabello, Moreno and Kell2003), trap (Addis et al., Reference Addis, Secci, Locci, Cau and Sabatini2012) and longline fisheries (Kimoto and Itoh, Reference Kimoto and Itoh2017). However, there is a lack of ABFT CPUE standardization. Therefore, our focus in this paper is on addressing this issue. Overall, these manuscripts contribute to the understanding of ABFT fisheries by examining catch rates, standardized CPUE and historical catch trends in different regions. They offer valuable insights that can inform the management and conservation efforts for the ABFT.

Including spatial variables in the GAM for the standardized CPUE of ABFT may provide important insights into the spatial dynamics of the species and improve the accuracy of the model predictions. By including spatial variables as covariates in the GAM model, it allows for the incorporation of this autocorrelation structure, leading to more robust and reliable estimates of CPUE and identification of important spatial hotspots, areas of high productivity or critical habitats for ABFT. This information can guide spatially targeted management measures and conservation strategies to ensure sustainable exploitation and protection of the species. There are some studies on the CPUE standardization methods which have given more consideration to spatial and temporal correlations (Nishida and Chen, Reference Nishida and Chen2004; Arrizabalaga et al., Reference Arrizabalaga, Dufour, Kell, Merino, Ibaibarriaga, Chust, Irigoien, Santiago, Murua, Fraile, Chifflet, Goikoetxea, Sagarminaga, Aumont, Bopp, Herrera, Marc Fromentin and Bonhomeau2015; Perryman and Babcock, Reference Perryman and Babcock2017; Grüss et al., Reference Grüss, Walter, Babcock, Forrestal, Thorson, Lauretta and Schirripa2019).

The objective of the present study is to assess the impact of spatial, temporal variables and vessel length on the CPUE of ABFT in the purse seine fishery. While previous studies have utilized GAM to standardize CPUE and examine variable influences (Damalas and Megalofonou, Reference Damalas and Megalofonou2012; Zhou et al., Reference Zhou, Campbell and Hoyle2019; Tosunoglu et al., Reference Tosunoglu, Ceyhan, Gulec, Duzbastilar, Kaykac, Aydin and Metin2021), this study represents the first application of GAM to the CPUE indices of purse seine fleets targeting ABFT in the eastern Mediterranean. By incorporating spatial variables into the GAM model, we aimed to gain valuable insights into the spatial dynamics of ABFT, enhance predictive accuracy and facilitate informed decision-making for the management and conservation of this species.

Material methods

Catch and effort data of the Turkish purse seine fishery of ABFT were collected from commercial logbook, made by the skippers and by field surveys between 1992 and 2006. The database includes current date (day, month and year), fishing vessel measurements (length of overall, gross tonnage and horsepower), fishing set position (latitude and longitude), number of fishes caught by vessel and total catch. The sea surface temperature (SST) and the salinity data of this study were retrieved from E.U. Copernicus Marine Service Information (2023a, 2023b).

The CPUE was calculated from three parameters as below for each fishing vessel:

$${\rm F} = {\rm H \cdot D}$$
$${\rm CPUE} = {\rm B\ }\cdot {\rm F}^{{-}1}$$

where F is the fishing effort, H is the number of hauling, D is the fishing day and B is the biomass of ABFT.

The effect of variables on CPUE was examined by means of GAM techniques (Hastie and Tibshirani, Reference Hastie and Tibshirani1986). The additive models and their generalizations make available to use many non-parametric models which are essential in regression analysis, when the linearity assumption does not engage well (Friedman and Stuetzle, Reference Friedman and Stuetzle1981; Hastie and Tibshirani, Reference Hastie and Tibshirani1986; Amodio et al., Reference Amodio, Aria and D'Ambrosio2014). Additionally, the advantages (i.e. interpretability, flexibility and regularization) of GAM made general addictive models more appropriate for this dataset. In this study, GAM with Tweedie family and log link function was used (Tweedie, Reference Tweedie1984; Dunn and Smyth, Reference Dunn and Smyth2005; Wood et al., Reference Wood, Pya and Säfken2016). Tweedie distributions are one of the family of distributions that include gamma, normal, Poisson and their combinations, in spite of being based partly on the Poisson family. In this modelling, the variance power (p), which parametrizes the Tweedie distribution, was set to 1.01 being that p must be greater than 1 and less than or equal to 2. P values that fall in this range are analogous to Poisson and gamma distribution (Tweedie, Reference Tweedie1984; Dunn and Smyth, Reference Dunn and Smyth2005). Moreover, restricted maximum likelihood was applied as a maximum likelihood-based smoothness selection procedure. Isotropic smooths on the sphere (Wahba, Reference Wahba1981) and univariate penalized cubic regression spline smooths (Wood, Reference Wood2017) were also chosen.

Thus, the form of the GAM used was

$${\rm CPUE}\sim {\rm a} + {\rm s}( {{\rm Y}, \;{\rm bs}{ = }^{\prime \prime}{\rm a{d}^{\prime \prime}}} ) + {\rm s}( {{\rm M}, \;{\rm bs}{ = }^{\prime \prime}{\rm a{d}^{\prime \prime}}} ) + {\rm s}( {{\rm lon}, \;{\rm lat}, \;{\rm bs}{ = }^{\prime \prime}{\rm so{s}^{\prime \prime}}} ) + {\rm s}( {{\rm LOA}, \;{\rm bs}{ = }^{\prime \prime}{\rm a{d}^{\prime \prime}}} ) + {\rm s}( {{\rm SST}, \;{\rm bs}{ = }^{\prime \prime}{\rm a{d}^{\prime \prime}}} ) + {\rm s}( {{\rm sal}, \;{\rm bs}{ = }^{\prime \prime}{\rm a{d}^{\prime \prime}}} ) + {\rm e}$$

where, a is the intercept, Y is years, M is months, lon is longitude, lat is latitude, LOA is length of overall, SST is sea surface temperature, sal is salinity and s indicates the smoother function of the corresponding independent variable. In the model, bs indicates the (penalized) smoothing basis to use, cs represents the cubic regression spline smooths, sos represents the univariate penalized cubic regression spline smooths. Finally, e is a random error term.

The test of whether the basis dimension for a smooth is adequate (Wood, Reference Wood2017) were done by k-index (the estimate of the residual variance based on differencing residuals) and p-value, computed by simulation. Low p-value may indicate that the basis dimension has been set too low, especially if the reported estimated degrees of freedom (edf) is close to upper limit on the degrees of freedom associated with an s smooth (k'). The procedure automatically selects the degree of smoothing based on the Generalized Cross Validation (GCV) score, which is a proxy for the model predictive performance. However, given the limited number of observations, the model was constrained to be at maximum a quartic relation. Consequently, the maximum degrees of freedom for each smoothing term, indicated by the number of knots (k), were set to 49 for the spatial variable and 39 for the remaining variables (i.e. k = 50 for the spatial variable and k = 40 for the rest of the variables in the GAM formulation). A log link function was assumed, and deficiencies of the fitted model were diagnosed by QQ plot of the deviance residuals and means of randomized quantile residual plots (Foster and Bravington, Reference Foster and Bravington2013; Pedersen et al., Reference Pedersen, Miller, Simpson and Ross2019). Statistical inference was based on the 95% confidence level. The model fitting was accomplished using the ‘mgcv’ library (Wood, Reference Wood2003, Reference Wood2004, Reference Wood2011, Reference Wood2016, Reference Wood2017; Wood et al., Reference Wood, Pya and Säfken2016), and the ‘tidyverse’ (Wickham et al., Reference Wickham, Averick, Bryan, Chang, McGowan, François, Grolemund, Hayes, Henry, Hester, Kuhn, Pedersen, Miller, Bache, Müller, Ooms, Robinson, Seidel, Spinu, Takahashi, Vaughan, Wilke, Woo and Yutani2019), ‘gratia’ (Simpson, Reference Simpson2022), ‘sf’ (Pebesma, Reference Pebesma2018), ‘ggspatial’ (Dunnington, Reference Dunnington2021), ‘scales’ (Wickham and Seidel, Reference Wickham and Seidel2022) and ‘rnaturalearth’ packages (South et al., Reference South, Michael and Massicotte2024) were also required under the R language environment (R Core Team, 2022).

Results

The sampled Turkish ABFT purse seines ranged from 20 to 64 m (average: 44.9 ± 0.4 m) in length (LOA) and 320 to 3675 hp (average: 1724 ± 32.8 hp) in machine power (Table 1).

Table 1.

Summary of the sampled fishing vessels

The number of ABFT caught within each operation varied between 1 and 2000. Meanly 99.79 ± 9.57 BFT were caught. During the catching operations the recorded weight of individual ABFT ranged from 8 to 350 kg. A total of 386 CPUE values for ABFT were calculated, which ranged between 0.05 and 60 t ⋅ (haul day)–1 with mean CPUE of 5.51 ± 0.54 t ⋅ (haul day)–1. Moreover, the smallest median was 0.3 t ⋅ (haul day)–1 for July and the maximum median was recorded as 4.8 t ⋅ (haul day)–1 for June (Figure 1). The range of variability around the median value in June was between 2 and 20 t ⋅ (haul day)–1. After 2001, fishing activity shifted to the southern part of Turkey and was mainly performed in May (Figure 2).

Figure 1.

The CPUE values of Turkish ABFT fishery in the eastern Mediterranean by months.

Figure 2.

Spatial distribution of the ABFT fishery by year and months.

The diagnostic information regarding the fitting procedure and results indicates that the basis dimensions used for the smooth terms were appropriate (Table 2). Additionally, Figure 3A, which presents the deviance residuals plotted against the approximate theoretical quantiles of the deviance residual distribution, demonstrates that the model's distributional assumptions were mostly satisfied. While there were three values that appeared to be lower and did not fit the model perfectly, we did not consider them as outliers and chose not to remove them. This decision was based on their association with the fishery season during 1992 and 1994, which represents an early period. Furthermore, Figure 3B shows that the response data were independent, and thus the residuals appeared reasonably well-distributed. The proportion of explained deviance was calculated to be 80.2%. Hence, we can confidently argue that the model fit very well, and there were no significant confidence intervals observed.

Table 2.

The result of basis dimensions of model

k’ = upper limit on the degrees of freedom associated with an s smooth, edf = estimated degrees of freedom, k-index = ratio of neighbour differencing scale estimate to fitted model scale estimate, p = p-value.

Figure 3.

(A) QQ-plot of residuals (black). The grey line indicates the 1–1 line. (B) Means of randomized quantile residuals.

The analysis of the deviance table revealed that all variables, except SST, were found to be significant (Table 3). Furthermore, Figure 4A displayed an undulant pattern in the line. Negative effects were predominantly observed on the CPUE in June and July (Figure 4B). Despite the presence of wide confidence intervals in certain parts of the plots, Figure 4C and 4D depicted relatively stable lines. While a negative effect was evident around 20 °C, the remainder of the line appeared to be stable in Figure 4E. These figures corroborated the findings presented in Table 3. The effect of spatial data on CPUE was greatest along the Turkish coast and eastern Mediterranean (Figure 5). Consequently, the results indicated that the year, month, LOA, salinity and area were significant explanatory factors, whereas SST had little influence on the CPUE.

Table 3.

Analysis of deviance table for the GAM model fitted to the CPUE data of the BFT purse seine fleet

*<0.05; df, degrees of freedom; F, F-value; p, p-value; s(x), smoother function of the corresponding independent variable; lon, longitude; lat, latitude; LOA, length overall; SST, sea surface temperature.

Figure 4.

GAM estimated effect of years (A), months (B), LOA (C), salinity (D), SST (E) on CPUE for ABFT PS fishery (grey area corresponds to the 95% confidence intervals of the estimates).

Figure 5.

GAM estimated effect of spatial data on CPUE for ABFT PS fishery.

Discussion

Purse seine became the main catching method for ABFT in Turkish fishery since the 1950s. Initially purse seiners mainly targeted the ABFT under the sardine purse seines for feeding or small schools off the coast. Due to the decrease in anchovy fishing in the 1990s, purse seine fishermen started to target tuna and exported it to far Eastern countries (Mert et al., Reference Mert, Oray, Patrona, Karakulak, Kayabaşı, Gündoğdu and Miyake2000). With the start of tuna farms in Turkey in 2003, there have been significant changes in tuna fishing time and areas, and fishermen have targeted more breeding populations (Karakulak et al., Reference Karakulak, Öztürk, Yıldız, Daniel, Gavin and Alejandro2016). In this study the temporal and spatial variations in the tuna fishery between 1992 and 2006 were analysed.

The study presented comprehensive CPUE indices ranging from 0.05 to 60 t ⋅ (haul day)–1, with a mean CPUE of 5.51 ± 0.54 t ⋅ (haul day)–1. In comparison, Zarrad and Missaou (Reference Zarrad and Missaou2018) reported standardized CPUE values of the Tunisian purse seine fleet ranging from 1.4 to 6.6 t ⋅ day–1. The discrepancies in CPUE values can be attributed to the month restrictions imposed in Tunisia. The fishing season was two months until 2009, and since then, it has been reduced to one month. Additionally, Karakulak (Reference Karakulak2004) calculated CPUE values for the Turkish fleet ranging from 3.22 to 7.32 t, with an average of 5.58 t per year. It is notable that CPUE values in Turkey were lower during the 1990s when tuna fishing predominantly occurred in autumn and winter. However, a shift towards fishing in May, particularly during the breeding period, as indicated by Karakulak et al. (Reference Karakulak, Oray, Corriero, Deflorio, Santamaria, Desantis and De Metrio2004) and Corriero et al. (Reference Corriero, Karakulak, Santamaria, Deflorio, Spedicato, Addis, Desantis, Cirillo, Fenech-Farrugia, Vassallo-Agius, Serna, Oray, Cau, Megalofonou and Metrio2005), resulted in a significant increase in CPUE data. Furthermore, Yalçın (Reference Yalçın2022) argued that factors such as shooting duration, closing the lacing duration, current speed at 10 m, setting speed, sinking speed and length-height ratio play crucial roles in the success of ABFT catching by purse seining, in addition to the physical variables of the sea.

The month trend of standardized CPUE showed the slight decrease between July and August in this study. The Eastern ABFT stock spawns from June to July in the Mediterranean Sea then migrates to their feeding areas (Piccinetti and Piccinetti-Manfrin, Reference Piccinetti and Piccinetti-Manfrin1970; Richards, Reference Richards1976; Dicenta and Piccinetti, Reference Dicenta and Piccinetti1978). Some of them move to areas between NW coasts of Africa and Norway and even trans-Atlantic areas (Mather et al., Reference Mather, Mason and Jones1995; Block et al., Reference Block, Dewar, Blackwell, Williams, Prince, Farwell, Boustany, Teo, Seitz, Walli and Fudge2001; Rodríguez-Marín et al., Reference Rodríguez-Marín, Arrizabalaga, Ortiz, Rodríguez-Cabello, Moreno and Kell2003). Temporal distribution of bluefin tuna catches indicated that the higher probability of catching of ABFT is related with the species’ spawning period in the eastern Mediterranean. Furthermore, the fishing grounds in Mediterranean changed from juveniles to spawners and fishing has been performed in spawning areas after 2000s (Gordoa et al., Reference Gordoa, Rouyer and Ortiz2019). As well known, the purse seiners became the main provider of live fish to farms in Mediterranean. Therefore, the fishing strategy for ABFT was only set up to catch larger fish in spawning areas for the Turkish purse seine fleet. This is also the same for other fleets of Mediterranean countries. The tagging studies carried out in the Eastern Mediterranean reported that ABFT do not prefer to migrate through the Gibraltar straight. In contrast, ABFT stay in the Eastern Mediterranean and also move to the Aegean Sea for feeding (de Metrio et al., Reference de Metrio, Oray, Arnold, Lutcavage, Deflorio, Cort, Karakulak, Anbar and Ultanur2004, Reference de Metrio, Arnold, de La Serna, Cort, Block, Megalofonou, Lutcavage, Oray and Deflorio2005). The ABFT catches between October and April during the 1990s also indicate that the Aegean Sea is an important feeding area after ABFT have spawned in the Eastern Mediterranean (Karakulak and Oray, Reference Karakulak and Oray1995; Oray and Karakulak, Reference Oray and Karakulak1997; Druon et al., Reference Druon, Fromentin, Hanke, Arrizabalaga, Damalas, Tičina, Quílez-Badia, Ramirez, Arregui, Tserpes, Reglero, Deflorio, Oray, Saadet Karakulak, Megalofonou, Ceyhan, Grubišić, MacKenzie, Lamkin, Afonso and Addis2016).

In this study, we only put one fishing vessel measurement (LOA) into the model to avoid the unstable model-fitting process and overfitting. Inclusion of all quantities of fishing vessel as explanatory variables would not improve the predictive ability of the model (Maunder and Punt, Reference Maunder and Punt2004). However, the standardization procedure showed that length of fishing vessel has a relatively minor explanatory effect on the CPUE in the purse seine fishery. Moreover, Rodríguez-Marín et al. (Reference Rodríguez-Marín, Arrizabalaga, Ortiz, Rodríguez-Cabello, Moreno and Kell2003) stated that the results related with vessel size should be interpreted with caution, because there are several possible underlying effects such as technological improvements in fishing gear, electronics and fishing power. We believe that tuna purse seiners had become more accustomed to the new technology, and more efficient at identifying large schools of ABFT, year by year.

With respect to spatial interaction effects, the area around Antalya Bay had higher value than the area around Gokceada Island. These two areas are quite different in terms of fishing strategy and fishing season. Furthermore, there was an area of positive value near shore along Turkey. These results indicate area effects on CPUE vary according to area. We also found the high effect on coastal areas. Damalas and Megalofonou (Reference Damalas and Megalofonou2012) explained the same situation with the reflecting foraging of ABFT, as prey usually congregate near land or on seamounts and banks. Moreover, the thermocline is a key factor in depth distribution of tunas (Abascal et al., Reference Abascal, Peatman, Leroy, Nicol, Schaefer, Fuller and Hampton2018). Actually, the distribution of ABFT has been affecting from many biotic and abiotic environmental variables such as SST, salinity, chlorophyll-a, etc. We found that the salinity was a significant explanatory factor, whereas SST had little influence on the CPUE in this study. Arrizabalaga et al. (Reference Arrizabalaga, Dufour, Kell, Merino, Ibaibarriaga, Chust, Irigoien, Santiago, Murua, Fraile, Chifflet, Goikoetxea, Sagarminaga, Aumont, Bopp, Herrera, Marc Fromentin and Bonhomeau2015) reported that ABFT does not exhibit a well-established temperature preference but rather demonstrates a wide temperature tolerance range (approximately between 1 and 20 °C). On the other hand, salinity can have a notable impact on the large-scale spatial distribution of ABFT (Reygondeau et al., Reference Reygondeau, Maury, Beaugrand, Fromentin, Fonteneau and Cury2012; Fromentin et al., Reference Fromentin, Reygondeau, Bonhommeau and Beaugrand2014). Additionally, Teo and Block (Reference Teo and Block2010) stated that ABFT CPUE increased substantially during the breeding months in areas with negative sea surface height anomalies and cooler SSTs. We believe that the reason SST did not significantly affect CPUE in our study could be attributed to the development of tuna fattening in Turkey and the fact that fishing is limited to a specific area during the breeding period.

Akyüz and Artüz (Reference Akyüz and Artüz1957) documented the length distribution of ABFT caught in the Bosphorus and Marmara Sea from 1955 to 1956, which ranged from 120 to 330 cm, with an average length of 228.9 ± 2.8 cm. On the other hand, Karakulak and Oray (Reference Karakulak and Oray1995) reported the size distribution of tuna caught in the Aegean Sea during the 1990s, ranging from 50 to 240 cm, and in the Levantine Sea, it ranged from 120 to 230 cm.

Since the initiation of tuna farming activities, the predominant method of capturing tuna in Turkish waters has been through cage farming. In accordance with the ICCAT recommendation (Rec. 19-04), the use of stereo-camera technology has become obligatory for estimating the number, weight and size/age distribution of the captured fish. Ortiz et al. (Reference Ortiz, Karakulak, Mayor and Paga2021) conducted a study examining data obtained from both farms and camera images, revealing a significant shift in the size distribution of fish caught by the Turkish tuna purse seine flotilla. While larger fish (>200 cm SFL) were once abundant in previous years, since 2017, medium-sized tuna (120–140 cm SFL) have become more prevalent, resulting in a notable increase in catch. The proportion of medium-sized fish now accounts for 60% by weight (80% numerically), while the proportion of large fish stands at 5%. Evaluation of fishing activities uncovered a doubling in the number of fishing boats and a fivefold increase in the number of fishing operations. Several factors contribute to this shift, including rising water temperatures in the Levantine Sea, where fishing is intense, the proliferation of Lessepsian species, ongoing gas/oil exploration studies in the Eastern Mediterranean and the overall increase in fishing activities. Further investigation is required to understand the underlying reasons for the decline in the population of larger fish.

In this paper, we utilized a generalized additive model to standardize the CPUE of ABFT in the Eastern Mediterranean purse seine fishery, incorporating spatial data as a variable. Given the economic importance and stock status of ABFT, there is a pressing need for a better understanding of its habitat and spatial dynamics. We strongly believe that considering the spatial dynamics of ABFT is crucial if we aim to mitigate the impact of fisheries on the ecosystem and implement effective remedies.

Data availability

The data that support the findings of this study are available from the corresponding author, upon reasonable request.

Author contributions

F.S. Karakulak: collecting data, conceptualization and editing. T. Ceyhan: writing the draft text, modelling, conceptualization, review and editing processes.

Financial support

There was no external financial support provided for the research conducted in the preparation of this manuscript.

Competing interests

None.

Ethical standards

Not applicable.

References

Abascal, FJ, Peatman, T, Leroy, B, Nicol, S, Schaefer, K, Fuller, DW and Hampton, J (2018) Spatiotemporal variability in bigeye vertical distribution in the Pacific Ocean. Fisheries Research 204, 371–379.CrossRefGoogle Scholar
Addis, P, Secci, M, Locci, I, Cau, A and Sabatini, A (2012) Analysis of Atlantic bluefin tuna catches from the last Tonnara in the Mediterranean Sea: 1993–2010. Fisheries Research 127-128, 133–141.CrossRefGoogle Scholar
Akyüz, EF and Artüz, I (1957) Some Observations on the Biology of Tuna (Thunnus thynnus) Caught in Turkish Waters. Rome: General Fisheries Council for the Mediterranean.Google Scholar
Amodio, S, Aria, M and D'Ambrosio, A (2014) On concurvity in nonlinear and nonparametric regression models. Statistica 74, 85–98.Google Scholar
Arrizabalaga, H, Dufour, F, Kell, L, Merino, G, Ibaibarriaga, L, Chust, G, Irigoien, X, Santiago, J, Murua, H, Fraile, I, Chifflet, M, Goikoetxea, N, Sagarminaga, Y, Aumont, O, Bopp, L, Herrera, M, Marc Fromentin, J and Bonhomeau, S (2015) Global habitat preferences of commercially valuable tuna. Deep Sea Research Part II: Topical Studies in Oceanography 113, 102–112.CrossRefGoogle Scholar
Block, BA, Dewar, H, Blackwell, SB, Williams, TD, Prince, ED, Farwell, CJ, Boustany, A, Teo, SLH, Seitz, A, Walli, A and Fudge, D (2001) Migratory movements, depth preferences, and thermal biology of Atlantic bluefin tuna. Science 293, 1310–1314.CrossRefGoogle ScholarPubMed
Block, BA, Teo, SLH, Walli, A, Boustany, A, Stokesbury, MJW, Farwell, CJ, Weng, KC, Dewar, H and Williams, TD (2005) Electronic tagging and population structure of Atlantic bluefin tuna. Nature 434, 1121–1127.CrossRefGoogle ScholarPubMed
Corriero, A, Karakulak, S, Santamaria, N, Deflorio, M, Spedicato, D, Addis, P, Desantis, S, Cirillo, F, Fenech-Farrugia, A, Vassallo-Agius, R, Serna, JM, Oray, Y, Cau, A, Megalofonou, P and Metrio, G (2005) Size and age at sexual maturity of female bluefin tuna (Thunnus thynnus L. 1758) from the Mediterranean Sea. Journal of Applied Ichthyology 21, 483–486.CrossRefGoogle Scholar
Damalas, D and Megalofonou, P (2012) Discovering where bluefin tuna, Thunnus thynnus, might go: using environmental and fishery data to map potential tuna habitat in the eastern Mediterranean Sea. Scientia Marina 76, 691–704.Google Scholar
de Metrio, G, Arnold, GP, de La Serna, JM, Cort, JL, Block, BA, Megalofonou, P, Lutcavage, M, Oray, I and Deflorio, M (2005) Movements of bluefin tuna (Thunnus thynnus L.) tagged in the Mediterranean Sea with pop-up satellite tags. ICCAT Collective Volume of Scientific Papers 58, 1337–1340.Google Scholar
de Metrio, G, Oray, I, Arnold, GP, Lutcavage, M, Deflorio, M, Cort, JL, Karakulak, S, Anbar, N and Ultanur, M (2004) Joint Turkish-Italian research in the eastern Mediterrenean: bluefin tuna tagging with pop-up satellite tag. ICCAT Collective Volume of Scientific Papers 56, 1163–1167.Google Scholar
Dicenta, A and Piccinetti, C (1978) Desove del atún (Thunnus thynnus) en el Mediterráneo occidental y evaluación directa del stock de reproductores, basado en la abundancia de sus larvas. Collective Volumes of Scientific Papers ICCAT 7, 389–395.Google Scholar
Druon, J, Fromentin, J, Aulanier, F and Heikkonen, J (2011) Potential feeding and spawning habitats of Atlantic bluefin tuna in the Mediterranean Sea. Marine Ecology Progress Series 439, 223–240.CrossRefGoogle Scholar
Druon, J-N, Fromentin, J-M, Hanke, AR, Arrizabalaga, H, Damalas, D, Tičina, V, Quílez-Badia, G, Ramirez, K, Arregui, I, Tserpes, G, Reglero, P, Deflorio, M, Oray, I, Saadet Karakulak, F, Megalofonou, P, Ceyhan, T, Grubišić, L, MacKenzie, BR, Lamkin, J, Afonso, P and Addis, P (2016) Habitat suitability of the Atlantic bluefin tuna by size class: an ecological niche approach. Progress in Oceanography 142, 30–46.CrossRefGoogle Scholar
Ducharme-Barth, ND, Grüss, A, Vincent, MT, Kiyofuji, H, Aoki, Y, Pilling, G, Hampton, J and Thorson, JT (2022) Impacts of fisheries-dependent spatial sampling patterns on catch-per-unit-effort standardization: a simulation study and fishery application. Fisheries Research 246, 106169.CrossRefGoogle Scholar
Dunn, PK and Smyth, GK (2005) Series evaluation of Tweedie exponential dispersion model densities. Statistics and Computing 15, 267–280.CrossRefGoogle Scholar
Dunnington, D (2021) ggspatial: Spatial Data Framework for ggplot2. https://paleolimbot.github.io/ggspatial/. Accessed online 23 September 2021.Google Scholar
E.U. Copernicus Marine Service Information (CMEMS) (2023a, June 25) Marine Data Store (MDS) Global Ocean Biogeochemistry Hindcast.Google Scholar
E.U. Copernicus Marine Service Information (CMEMS) (2023b, June 25) Marine Data Store (MDS) Global Ocean Physics Reanalysis.Google Scholar
Foster, SD and Bravington, M V. (2013) A Poisson–Gamma model for analysis of ecological non-negative continuous data. Environmental and Ecological Statistics 20, 533–552.CrossRefGoogle Scholar
Friedman, JH and Stuetzle, W (1981) Projection pursuit regression. Journal of the American Statistical Association 76, 817–823.CrossRefGoogle Scholar
Fromentin, J-M, Reygondeau, G, Bonhommeau, S and Beaugrand, G (2014) Oceanographic changes and exploitation drive the spatio-temporal dynamics of Atlantic bluefin tuna (Thunnus thynnus). Fisheries Oceanography 23, 147–156.CrossRefGoogle Scholar
Gordoa, A, Rouyer, T and Ortiz, M (2019) Review and update of the French and Spanish purse seine size at catch for the Mediterranean bluefin tuna fisheries 1970–2010. ICCAT Collective Volume of Scientific Papers 75, 1622–1633.Google Scholar
Grüss, A, Walter, JF, Babcock, EA, Forrestal, FC, Thorson, JT, Lauretta, MV and Schirripa, MJ (2019) Evaluation of the impacts of different treatments of spatio-temporal variation in catch-per-unit-effort standardization models. Fisheries Research 213, 75–93.CrossRefGoogle Scholar
Hastie, TJ and Tibshirani, RJ (1986) Generalized additive models. Statistical Science 1, 297–310.Google Scholar
Hinton, MG and Maunder, MN (2004) Methods for standardizing CPUE and how to select among them. ICCAT Collective Volume of Scientific Papers 56, 169–177.Google Scholar
ICCAT (2020) 2020 SCRS Advice to the Commission. Available at https://www.iccat.int/Documents/SCRS/ExecSum/BFT_ENG.pdf (Accessed 28 April 2022).Google Scholar
Karakulak, FS (2004) Catch and effort of the bluefin tuna purse-seine fishery in Turkish waters. Fisheries Research 68, 361–366.CrossRefGoogle Scholar
Karakulak, FS and Oray, IK (1995) A preliminary report on the investigations of bluefin tuna (Thunnus thynnus, L. 1758) caught in the Turkish waters. ICCAT Collective Volume of Scientific Papers 44, 140–143.Google Scholar
Karakulak, S, Oray, I, Corriero, A, Deflorio, M, Santamaria, N, Desantis, S and De Metrio, G (2004) Evidence of a spawning area for the bluefin tuna (Thunnus thynnus L.) in the eastern Mediterranean. Journal of Applied Ichthyology 20, 318–320.CrossRefGoogle Scholar
Karakulak, FS, Öztürk, B and Yıldız, T (2016) From ocean to farm: capture-based aquaculture of bluefin tuna in the eastern Mediterranean Sea. In Daniel, DB, Gavin, JP and Alejandro, B (eds), Advances in Tuna Aquaculture from Hatchery to Market. Cambridge, MA: Academic Press, pp. 59–76.CrossRefGoogle Scholar
Kimoto, A and Itoh, T (2017) Simple update of the standardized bluefin CPUE of Japanese longline fishery in the Atlantic up to 2016 fishing year. ICCAT Collective Volume of Scientific Papers 73, 1957–1976.Google Scholar
Martell, SJD (2008) Fisheries management. In Jørgensen, SE and Fath, BD (eds), Encyclopedia of Ecology. Oxford: Academic Press, pp. 1572–1582.CrossRefGoogle Scholar
Mather, FJ, Mason, JM and Jones, AC (1995) Historical document: life history and fisheries of Atlantic bluefin tuna. NOAA technical memorandum NMFS-SEFSC 370, 165p.Google Scholar
Maunder, MN and Punt, AE (2004) Standardizing catch and effort data: a review of recent approaches. Fisheries Research 70, 141–159.CrossRefGoogle Scholar
Maunder, MN, Sibert, JR, Fonteneau, A, Hampton, J, Kleiber, P and Harley, SJ (2006) Interpreting catch per unit effort data to assess the status of individual stocks and communities. ICES Journal of Marine Science 63, 1373–1385.CrossRefGoogle Scholar
Mert, I, Oray, IK, Patrona, K, Karakulak, FS, Kayabaşı, Y, Gündoğdu, M and Miyake, P (2000) Historical review of Turkish bluefin tuna fisheries and their development. ICCAT Collective Volume of Scientific Papers LI, 813–826.Google Scholar
Nishida, T and Chen, D-G (2004) Incorporating spatial autocorrelation into the general linear model with an application to the yellowfin tuna (Thunnus albacares) longline CPUE data. Fisheries Research 70, 265–274.CrossRefGoogle Scholar
Oray, I and Karakulak, S (1997) Some remarks on the bluefin tuna (Thunnus thynnus L.1758) fishery in Turkish waters in 1993, 1994, 1995. ICCAT Collective Volume of Scientific Papers XLVI, 357–362.Google Scholar
Ortiz, M, Karakulak, FS, Mayor, C and Paga, A (2021) Review of the size distribution of caged eastern bluefin tuna (Thunnus thynnus) in Turkish farms 2014–2020. ICCAT Collective Volume of Scientific Papers 78, 159–169.Google Scholar
Pebesma, E (2018) Simple features for R: standardized support for spatial vector data. The R Journal 10, 439.CrossRefGoogle Scholar
Pedersen, EJ, Miller, DL, Simpson, GL and Ross, N (2019) Hierarchical generalized additive models in ecology: an introduction with mgcv. PeerJ 7, e6876.CrossRefGoogle ScholarPubMed
Perryman, HA and Babcock, EA (2017) Generalized additive models for predicting the spatial distribution of billfishes and tunas across the Gulf of Mexico. ICCAT Collective Volume of Scientific Papers 73, 1778–1795.Google Scholar
Piccinetti, C and Piccinetti-Manfrin, G (1970) Osservazioni sulla biologia dei primi staid giovanili del tono (Thunnus thynnus L.). Bolletino Di Pesca, Piscicoltura e Idrobiologia 25, 223–247.Google Scholar
R Core Team (2022) R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing.Google Scholar
Reygondeau, G, Maury, O, Beaugrand, G, Fromentin, JM, Fonteneau, A and Cury, P (2012) Biogeography of tuna and billfish communities. Journal of Biogeography 39, 114–129.CrossRefGoogle Scholar
Richards, WJ (1976) Spawning of bluefin tuna (Thunnus thynnus) in GRE Atlantic Ocean and adjacent seas. Collective Volumes of Scientific Papers ICCAT 2, 267–278.Google Scholar
Rodríguez-Marín, E, Arrizabalaga, H, Ortiz, M, Rodríguez-Cabello, C, Moreno, G and Kell, LT (2003) Standardization of bluefin tuna, Thunnus thynnus, catch per unit effort in the baitboat fishery of the Bay of Biscay (Eastern Atlantic). ICES Journal of Marine Science 60, 1216–1231.CrossRefGoogle Scholar
Simpson, GL (2022) gratia: Graceful ‘ggplot’-based graphics and other functions for GAMs fitted using ‘mgcv’. https://gavinsimpson.github.io/gratia. Accessed online 23 September 2021.Google Scholar
Skalski, JR, Ryding, KE and Millspaugh, JJ (2005) Analysis of population indices. In John, RS, Kristen, ER and Joshua, JM (eds), Wildlife Demography. Cambridge, MA: Academic Press, pp. 359–433.CrossRefGoogle Scholar
South, A, Michael, s and Massicotte, P (2024) rnaturalearthhires: High Resolution World Vector Map Data from Natural Earth used in rnaturalearth. R package version 1.0.0. https://github.com/ropensci/rnaturalearthhires. Accessed online 23 September 2021.Google Scholar
Teo, SLH and Block, BA (2010) Comparative influence of ocean conditions on yellowfin and Atlantic bluefin tuna catch from longlines in the Gulf of Mexico. PLoS ONE 5, e10756.CrossRefGoogle ScholarPubMed
Tosunoglu, Z, Ceyhan, T, Gulec, O, Duzbastilar, FO, Kaykac, MH, Aydin, C and Metin, G (2021) Effects of lunar phases and other variables on CPUE of European pilchard, Sardina pilchardus, caught by purse seine in the eastern Mediterranean. Turkish Journal of Fisheries and Aquatic Sciences 21, 283–290.CrossRefGoogle Scholar
Tweedie, MCK (1984) An index which distinguishes between some important exponential families. In Applications and New Directions—Proceedings of the Indian Statistical Institute Golden Jubilee International Conference.Google Scholar
Varela, JL, Carrera, I and Medina, A (2020) Seasonal feeding patterns of Atlantic bluefin tuna (Thunnus thynnus) in the Strait of Gibraltar. Marine Environmental Research 153, 104811.CrossRefGoogle ScholarPubMed
Wahba, G (1981) Spline interpolation and smoothing on the sphere. SIAM Journal on Scientific and Statistical Computing 2, 5–16.CrossRefGoogle Scholar
Wickham, H, Averick, M, Bryan, J, Chang, W, McGowan, L, François, R, Grolemund, G, Hayes, A, Henry, L, Hester, J, Kuhn, M, Pedersen, T, Miller, E, Bache, S, Müller, K, Ooms, J, Robinson, D, Seidel, D, Spinu, V, Takahashi, K, Vaughan, D, Wilke, C, Woo, K and Yutani, H (2019) Welcome to the Tidyverse. Journal of Open Source Software 4, 1686.CrossRefGoogle Scholar
Wickham, H and Seidel, D (2022) scales: Scale Functions for Visualization. R package version 1.3.0, https://github.com/r-lib/scales. Accessed online 21 September 2021.Google Scholar
Wood, SN (2003) Thin plate regression splines. Journal of the Royal Statistical Society: Series B (Statistical Methodology) 65, 95–114.CrossRefGoogle Scholar
Wood, SN (2004) Stable and efficient multiple smoothing parameter estimation for generalized additive models. Journal of the American Statistical Association 99, 673–686.CrossRefGoogle Scholar
Wood, SN (2011) Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. Journal of the Royal Statistical Society: Series B (Statistical Methodology) 73, 3–36.CrossRefGoogle Scholar
Wood, SN (2016) Just another Gibbs additive modeler: interfacing JAGS and mgcv. Journal of Statistical Software 75, 1–15.CrossRefGoogle Scholar
Wood, SN (2017) Generalized Additive Models. Generalized Additive Models: An Introduction with R, 2nd Edn. Boca Raton, FL: Chapman and Hall/CRC.CrossRefGoogle Scholar
Wood, SN, Pya, N and Säfken, B (2016) Smoothing parameter and model selection for general smooth models. Journal of the American Statistical Association 111, 1548–1563.CrossRefGoogle Scholar
Yalçın, E (2022) The effects of variables on the success of Thunnus thynnus fishing in the Eastern Mediterranean Sea. Fisheries Science 88, 677–691.CrossRefGoogle Scholar
Zarrad, R and Missaou, H (2018) Update on CPUE bluefin tuna caught by Tunisian purse seines between 2009 and 2016. ICCAT Collective Volume of Scientific Papers 74, 3457–3461.Google Scholar
Zhou, S, Campbell, RA and Hoyle, SD (2019) Catch per unit effort standardization using spatio-temporal models for Australia's Eastern Tuna and Billfish Fishery. ICES Journal of Marine Science 76, 1489–1504.CrossRefGoogle Scholar
Figure 0

Table 1. Summary of the sampled fishing vessels

Figure 1

Figure 1. The CPUE values of Turkish ABFT fishery in the eastern Mediterranean by months.

Figure 2

Figure 2. Spatial distribution of the ABFT fishery by year and months.

Figure 3

Table 2. The result of basis dimensions of model

Figure 4

Figure 3. (A) QQ-plot of residuals (black). The grey line indicates the 1–1 line. (B) Means of randomized quantile residuals.

Figure 5

Table 3. Analysis of deviance table for the GAM model fitted to the CPUE data of the BFT purse seine fleet

Figure 6

Figure 4. GAM estimated effect of years (A), months (B), LOA (C), salinity (D), SST (E) on CPUE for ABFT PS fishery (grey area corresponds to the 95% confidence intervals of the estimates).

Figure 7

Figure 5. GAM estimated effect of spatial data on CPUE for ABFT PS fishery.