The quest for seal-safe fisheries—a comparison of alternative gear types in the Baltic Sea cod fishery

1 Introduction

In the 1970s, Baltic Sea seal populations were on the brink of extinction, but have since recovered to viable population densities (1, 2). The grey seal (Halichoerus grypus) population in the Baltic is increasing. In 2021, approximately 42,000 grey seals were counted; assuming a haul out-fraction of 70%, the total population estimate would be approximately 60,000 individuals (3). With increasing seal populations in the Baltic Sea, there is a growing interaction between seals and coastal fishers (4–8). The grey seal population has not only multiplied in numbers but also expanded its range from the northern Baltic further southwards (9). Although nowadays present throughout the Baltic Sea, the seal–fisheries conflicts are present in other regions across the globe as well, such as the eastern and western coasts of North America (10), Scotland (11), Australia (12), Chile (13), and Japan (14). One way seals interact is by raiding fishing gear, leading to both loss and destruction of catch, as well as by damaging gear (15, 16), causing economic losses for coastal fisheries (17, 18). Seal depredation in coastal fisheries causes both visible and hidden/partial losses in catch and gear. Visible losses include damaged fish, while partial losses include fish initially caught in fishing gear and later predated by seals along with fish that are scared away from fishing locations. A study from 2009 revealed partial losses of 36% to fish removed in cod (Gadus morhua) gillnet fishing (15), and later studies have indicated a loss rate of up to 45% in the southern Baltic area (19). Generally, fish loss is thought to have a large impact on the Baltic Sea coastal fisheries, and figures on seal-induced damage to gear and catches in the Baltic cod fisheries have increased by 100%–150% between 2006 and 2017 (20).

Not only do seals impact fisheries, but fisheries also directly affect seal populations, mainly through bycatch where seals become entangled in fishing gear and drown (21–23). Gillnet fisheries pose a particularly high bycatch risk (24). The magnitude of current bycatch and its influence on population growth are not known. However, bycatch rates of the past 25 years have not prevented the grey seal numbers to continuously increase (20, 25).

To reduce the seal–fisheries conflict, several mitigation strategies have been developed and implemented, with varying degrees of success (22, 26–31). One long-lasting and sustainable mitigation measure is the development and implementation of seal-safe fishing gear. Seal-safe gear makes it hard for seals to access the catch and will consequently minimize their reward and thereby decrease their motivation to raid fishing gear for food (27). Fishing gear such as pots, traps, and fyke nets can, in comparison to gillnets, more easily be made seal-safe by designing closed and rigid compartments where seals cannot access the catch and with modified seal exclusion devices that prevent seals and other marine mammals and birds from getting bycaught (22, 32). These passive gears are also seen as selective and sustainable and have been defined as “low-impact fuel-efficient fishing gear” (LIFE fishing gears) (33), not to be mixed up with the stakeholder organization “Low Impact Fishers of Europe”. Furthermore, they are also able to deliver live catches, allowing for the return of non-target species to the sea with a high chance of survival (33).

In the Baltic Sea, small-scale fisheries, including cod, have historically been economically, socially, and culturally important (6, 34, 35), with coastal fisheries, primarily conducted using passive gears such as gillnets (8, 36). As they are stationary and often left in the water for several hours or days and thus offer easily accessible food for seals and other fish-eating predators, seal interaction with gillnets is usually high. The increasing seal depredation and damages to fishing gear have, for decades, been perceived as two of the main threats to coastal fisheries, and the reason for local fisheries going out of business (15, 37). To allow for a future viable fishery in the Baltic Sea, there is an urge to develop improved fishing gear that may both withstand seal impact and keep bycatch rates of both target and non-target species, including birds and mammals, on a low level.

Cod has historically been an important species in the Baltic coastal fisheries (38), but since 2019, a general fishing ban has been in place for targeting the species (39) with a minor bycatch quota still allowed for the Total Allowable Catch (TAC) by the Council of the European Union (40); 41, 42). Still, when fisheries for cod are sustainable again, then there is a need for a continued development of fishing gear to overcome target conflicts before reopening fisheries. With cod being a predominantly benthic species (43, 44), gear used to target cod must be bottom set in order to work (30).

Numerous efforts to develop seal-safe and bycatch-reducing gear, targeting cod, have been made in recent years to improve cod catchability. Alternative fishing gears such as baited cod pots have been evaluated as a potential substitute for gillnets, demonstrating that catch per unit effort (CPUE) can reach comparable levels. However, efficiency is influenced by factors such as depth, soak time, and environmental conditions (45). Furthermore, modified seal-safe cod pot designs have been shown to reduce grey seal depredation by limiting seal access to the catch and reducing gear damage, which, in turn, decreases catch losses and can increase the number of marketable cod retained per unit effort (22, 46, 47). Optimized pot characteristics, such as entrance design and pot volume, have also been shown to influence cod entry rates and overall catch efficiency (22, 47). Furthermore, cod catch in pots has shown high seasonal variability (45) and cod caught in pots show a worse condition, i.e., more meager, in relation to gillnet-caught cod (29, 48). This calls for a broader comparison between alternative, seal-safe gear types.

The component of gear handling time has often been overlooked when evaluating gear performance. Since a fifth of all global catches (by weight) are landed by passive fishing gears, quantifying and estimating the passive fishing effort are particularly relevant for fisheries management (49) and are essential to understand fishers’ profitability and improving stock assessment (50). Handling time, soak time, and catch efficiency are highly relevant to accurately estimate the effect of the passive coastal fisheries and potential spatiotemporal changes in such maritime activities (49). This highlights the need for studies that compare multiple gear types while incorporating gear-specific handling times when evaluating effectiveness.

1.1 Aim

The aim of this study was to evaluate four passive fishing gear types—pots, pontoon traps, fyke nets, and gillnets—in terms of their catchability in seal dense fishing areas. Specifically, for each gear type, we measured (1) catch rates; (2) the proportion of seal-damaged catch; and (3) extrapolated catch rates based on handling time to estimate commercial viability for coastal cod fisheries in the Baltic Sea.

2 Method

2.1 Study area

Trials were conducted between 14 April and 2 June 2020, in the coastal waters of Ystad, Sweden, located in the southwestern Baltic Sea (Figure 1). In the study area, the shoreline is relatively straight, with a gently sloping, generally bare sand bottom, partly covered with gravel and rock at depths less than 10 m. The trials were conducted in collaboration with a commercial fisherman contracted by the Swedish University of Agricultural Sciences.

2.2 Gear and study design

All gear types were deployed at the same depth range, from 3 to 18 m (Figure 1). The fishing gears were distributed at distances and directions from one another such that pot bait plumes would not attract cod into adjacent gears. The pontoon trap, being a stationary gear, was lifted, emptied, and re-deployed at the same position throughout the study period while pots, fyke nets, and gillnets all were re-deployed at new gear-specific positions after collection (Figure 1). Soak times varied depending on local regulations on gear type, weather conditions, and optimum soak times, to maximize catch (Table 1).

Table 1. The different gears used in the study (Figure 2) along with their specifications.

For each gear deployment, we recorded the number of cod and total catch weight below and above the minimum conservation reference size (MCRS, 35 cm total length). Only cod above MCRS (35 cm total length) were kept, while cod below 35 cm TL were immediately released after counting. Length of damaged cod was estimated when precise measurements were not allowed. If present, we recorded the number of fish that appeared to be damaged by seals. Fish damaged by seals exhibit specific injuries characteristic of seal predation. These damages may include the gut being removed, only the head remaining in the net, large portions of the fish missing, or bite marks. Catch rates were collected as both number per unit effort (CPUE) as total catch number of cod day⁻¹ gear⁻¹ and weight per unit of effort (WPUE), which was calculated as the total catch weight of cod day⁻¹ gear⁻¹, for commercial catch (>35 cm TL), and the number per unit of effort, cod day⁻¹ gear⁻¹ (CPUE), for undersized catch (<35 cm TL). Along with gear-specific catch data, we recorded set date/time, retrieve date/time, GPS position, and depth using the ZebraTech deck logger (https://www.zebra-tech.co.nz).

2.3 Gear specifications

In this study, we compared four fishing gear types: pots, pontoon traps, fyke nets, and gillnets, with detailed explanation below (Figure 2 and Table 1). We tested two different pot models, used by Swedish fishers in the pot fishery over several years, demonstrating their ability to yield reasonable catches while being both practical to handle and of suitable size for use with smaller boats, along with showing high resistance against seals. A single hovering pontoon trap was used. Traditional pontoon traps, being a gear with high seal resistance, have been used primarily targeting salmon and whitefish but have also previously been used and evaluated for targeting cod (30). Modified fyke nets are the most newly developed alternative to traditional gears. Therefore, we included different models to evaluate their viability in cod fisheries regarding both catchability and seal resistance. Finally, two identical gillnets were used throughout the study. Individual gear types and models along with specifications are presented below. Because this study focused on evaluating the catchability of broad gear types, we did not evaluate differences in catch between the models of a specific gear type.

We deployed two separate pot strings with each pot string containing 10 pots, 5 of each model (Figures 2A1, A2), spaced at intervals of 50 m, resulting in strings with a total length of 450 m. The first pot model is rectangular with dimensions of 120 × 60 × 35 cm (length × width × height) and a volume of 252 dm³ (Figure 2A1, Table 1: A1). The second model is square shaped, with dimensions of 100 × 100 × 35 cm (length × width × height) and a volume of 350 dm³ (Figure 2A2, Table 1: A2). Pot frames of both models were made of 10-mm galvanized round steel and covered with a polyethylene green net with a mesh size of 22 mm (center-to-center knot). Rectangular pots were equipped with two entrances, one on each short side, while the square pots had four entrances, one on each side. Each entrance includes a funnel made of polyethylene with a mesh size of 22 cm placed vertically, allowing fish to enter the pot, but preventing escape. Pots were baited with approximately 250 g of herring (Clupea harengus) cut into 5-cm pieces, put in fine meshed bait bags and hung in the pots.

The leader net, wings, and adaptor of the pontoon trap were the same as previously used in the evaluation of bottom set traps in Ystad in 2017 and 2018 (30) (Table 1 and Figure 2B). The leader net used was 200 m with a mesh size of 100 mm (center-to-center knot). Trap height was 3 m, with a 35-mm (center-to-center knot) mesh size in the trap net and the adaptor. The pontoon trap used was equipped with a 4.5-m-long, 1.5-m-diameter, 35-mm (center-to-center knot) mesh size, hovering pontoon fish chamber (51) (Figures 1, 2B, Table 1: B).

Eleven fyke nets were used during the trials. The first main fyke model type has a 1.2 × 0.8 (width × height) stainless steel frame entrance narrowing into two additional entrance rings, with diameters of 40 and 30 cm, respectively, in order to further create hindrance for the fish attempting to turn back (Model Ry20, Table 1 and Figure 2C). The second main model had two types. One type was a single fyke net with 1.2-m square-shaped frames and only one seal-safe fish chamber (Model 120E, Figure 2E). Two others were single and a double fyke nets, with 1.4-m square sides, Models 140E and 140D, respectively (Figure 2D). The double fyke net had two identical seal-safe fish chambers on either side of the funnel-shaped entrance. These models were equipped with a 75-m-long leader. While model Ry20 was deployed in strings of four fykes for which two leader nets of 25 and 50 m, respectively, were used, the 120E and 140E/D fyke nets were deployed solitarily. All fyke nets were equipped with a 3-meter hose net made from black, knot-less 10-mm center-to-center nylon mesh for improved catch collection as hose nets have previously been implemented successfully in pontoon traps (30, 52).

The gillnets used were the ones commonly used by fishers. Each of the two gillnets were 800 m long, containing eight individual net links. They had a height of 3 m and a mesh size of 55 mm (center-to-center knot).

2.4 Gear-type-dependent catch rate comparison

Catch comparison between gear types was evaluated by using each gear deployment as replicate. Furthermore, catch comparison in this study assumed that models within gear types were pooled and had equal catch rate, as the study focus was to evaluate differences between gear types and not between models. There is the assumption that there was no difference in catch rates between the rectangular and square pot types (Figure 2). Also, at this stage of catch comparison, no consideration was taken for gear-specific handling times.

2.5 Seal-damaged catch depending on gear type

The number of seal-damaged individual cod in each gear for each deployment was used to calculate the mean proportion of seal-damaged catch in each of the four gear types respectively.

2.6 Comparison of gear-specific catch rates depending on handling time

To predict daily catch rates, gear-specific catch (WPUE; kg gear⁻¹ day⁻¹) was extrapolated and used to calculate the potential catch per day in a full-scale commercial fishery for each gear type. The number of gears of each gear type that may be handled within a day (gear day⁻¹) was extrapolated from their gear/string-wise specific handling times collected using the ZebraTech logger data on set and retrieval. To standardize a fishing day, we used 7 h of fishing time, excluding the travel time to and from port. Handling time included retrieval of the gear, catch collection and handling, rearranging the gear, and, if needed, moving them to another fishing ground, along with resetting the gear that was measured for pots, fyke nets, and gillnets.

Based on gear-specific handling times collected during the study, the number of pots that could be handled within a fishing day ranged from 66 to 166, arranged in strings where each string contained 10 pots. For fykes, the handling ranged from 24 to 65 individuals if fykes were arranged in strings with 8 fykes in each string. Based on handling time data collected, handling time for gillnets ranged from 6 to 20 gillnets per day, with all gillnets handled individually. Pontoon traps were extrapolated under the assumption that a fisher would be able to handle and collect the catch from between four and six traps in a day (30). The minimum and maximum number of gears that could be handled were then multiplied by their individual catch rate and proportion of seal-damaged catch to estimate the potential daily minimum and maximum catch in a full-scale commercial fishery for each of the four gear types individually. With catch, proportion seal-damaged catch, and handling time being highly variable, an alternative estimation of potential daily catch was carried out using a bootstrapping procedure. Each vector of the variables catch rate (WPUE), daily handling time, and proportion seal-damaged catch for each gear type was bootstrapped using the sample function in R (53). Each bootstrapped variable was iterated N = 10,000 times, giving a new vector of predicted values for each variable and gear type. Bootstrapped vectors of gear-specific catch were then multiplied with vectors of handling time and proportion seal-damaged catch to predict the mean catch and standard deviation per day given a full-scale commercial fishery catch comparison.

2.7 Statistical analyses

All statistical analyses were performed using R (53). For the first evaluation of the effects on catch rate of different gear types, independently of handling time, on WPUE, a linear mixed model (lmm) was fitted using the “lmer” function in the lme4 package. In all models, the response variable was WPUE of collected catch (kg day⁻¹). The main predictor variable (fixed) was gear type (pot, pontoon trap, fyke net, and gillnet). The model also included soak time (days) and temperature (°C) as predictor (fixed) variables and string as a random factor. For the comparison of amount of seal-damaged catch, a model was fitted to a generalized linear mixed-effects model (GLMM), using the “lmer” function in the lme4 package. The proportion of seal-damaged catch was used as a predictor (fixed) variable along with soak time and temperature, while string was included as a random factor.

We compared the fit of the negative binomial type 1 (nbinom1), which assumes a linear increase in variance with the mean, against the negative binomial type 2 (nbinom2), which assumes a quadratic relationship, and Gaussian distribution, which assumes constant variance independent of the mean and normally distributed residuals. Model validation with residual check was performed using the DHARMa package in R. Model selection was guided by the Akaike Information Criterion (AIC) and inspection of residual patterns. Parameter estimation and inferences were performed on the full model, i.e., including all variables (both fixed and random). Statistical significance of explanatory variables was tested according to Satterthwaite’s method using the “lmerTest” function in R (54).

3 Results

3.1 Gear-type-dependent catch rate comparison

In total, 234 individual gear collections were used in the catch rate comparison (Pot strings; n = 12, Pontoon; n = 18, Fyke nets; n = 70, Gillnet; n = 26), (Table 2, Figure 3). Mean daily catch of commercial-sized cod in pots in numbers (CPUE ± SD) was 1.94 ± 0.89 cod pot string⁻¹ day⁻¹ and that in weight (WPUE ± SD) was 1.19 ± 0.49 kg pot string⁻¹ day⁻¹. In the pontoon, mean catch in numbers (CPUE ± SD) was 5.72 ± 5.07 cod pontoon⁻¹ day⁻¹ and that in weight (WPUE ± SD) was 3.60 ± 2.99 kg pontoon⁻¹ day⁻¹. For the fykes, mean catch in numbers (CPUE ± SD) was 1.12 ± 2.32 cod fyke⁻¹ day⁻¹ and that in weight (WPUE ± SD) was 0.8 ± 1.20 kg fyke⁻¹ day⁻¹. In the gillnets, mean catch in numbers (CPUE ± SD) was 13.12 ± 19.75 cod gillnet⁻¹ day⁻¹ with a mean weight (WPUE ± SD) of 11.59 ± 17.42 kg gillnet⁻¹ day⁻¹.When evaluating catch rate using the lmm, there was a difference in catch rate between gear types (F = 14.547, df = 3, p < 0.001), with estimates for the model intercept (β = 11.41 ± 5.36 SE, t = 2.129, p < 0.05). There was an effect of temperature on catch rate (β = −1.430 ± 0.650 SE, t = −2.200, p < 0.05) but no effect of soak time (β = 0.312 ± 0.537 SE, t = 0.581, p = 0.562) or individual string. There was also a difference in the interaction between gear type and soak time (F = 6.160, df = 3, p < 0.001).

Table 2. Total number of deployments per gear type.

3.2 Seal-damaged catch depending on gear type

In total, 140 individual gear collections were used in the evaluation of the amount of seal-damaged catch (Pot; n = 8, Pontoon; n = 18, Fyke; n = 45, Gillnet; n = 26). The mean seal damage proportion in pots was 0 pot⁻¹ day⁻¹, which was consistent with the pontoon trap that also showed a seal damage proportion of 0 pontoon trap⁻¹ day⁻¹ (Figure 4). For the fyke nets, the mean proportion of seal-damaged catch was 0.02 ± 0.07 fyke⁻¹ day⁻¹. Gillnets had the highest amount of seal-damaged catch, with 0.5 ± 0.41 gillnet⁻¹ day⁻¹. The statistical analysis showed an effect of gear type on the amount of seal-damaged catch (χ² = 22.235, df = 3, p < 0.001), a strong tendency in the effect of temperature (χ² = 3.371, df = 1, p = 0.06), while there were no effects of soak time (χ² = 0.581, df = 1, p = 0.45) or the random factor string.

3.3 Comparison of gear-specific catch rates depending on handling time

When extrapolated to a full day of fishing, predicted catches ranged from 12.92 kg day⁻¹ in pots (lowest) to 46.3 kg per day⁻¹ (highest) (Table 3). When bootstrapped, catch rate depended on gear type, with the largest catch predicted for a fyke net fishery with a catch of 37.2 ± 58.2 (kg gear⁻¹ day⁻¹), followed by gillnets 35.7 ± 75.4 (kg gear⁻¹ day⁻¹), pontoon trap 17.9 ± 14.6 (kg gear⁻¹ day⁻¹), and pots 9.7 ± 17.4 (kg gear⁻¹ day⁻¹) (Table 3). Results are visualized in Figure 5, and besides handling time and catch rate, the extrapolation also includes effects of seal depredation.

Table 3. Estimated potential daily catches (kg day−1) for the four gear types (pots, pontoon trap, fyke nets, and gillnets), assuming a full-scale commercial fishery.

Calculations are based on measured catch rates (WPUE), gear-specific handling times, and the proportion of seal-damaged catch. The “mean” column shows the average estimated daily catch for each gear type, derived from mean handling times and catch rates, including what is expected to be lost due to seal depredation, where “max” and “min” values represent the maximum and minimum potential daily catches.

4 Discussion

With seal-inflicted damage to catch and gear increasing in coastal fisheries, there is a need to find alternative gear to the commonly used gillnets. When developing alternative sustainable fishing gears, catch efficiency is a key consideration to implementation. Any new gear will be adopted on its own, if it can economically compete with the old gear, although that is however not always a preferable and sustainable way forward for declining fish stocks. Included in catch efficiency are environmental parameters (45, 55, 56), fishing strategy (33), and gear design efficiency (57, 58). However, gear handling time, a crucial factor when evaluating gear, is often overlooked (49). In this study, we compared four different gear types to use in coastal fisheries to evaluate their individual catch rates, ability to withstand catch loss from seals, handling time, and catch efficiency.

4.1 Catch rate

The catch rate of the four gear types used in the study varied depending on gear type and soak time, with traditional gillnets showing the highest catch rate (Figure 3). For the three alternative gear types, the hovering pontoon showed the highest individual catch rate, followed by fykes and pots (Figure 3). Despite no general effect of soak time on catch rate, an interaction between gear and soak time was observed. Predominantly, the hovering pontoon trap, as well as the fykes, showed increased catch with increased soak time. Both gear types utilize a leader net to lead the fish into the gear. With the hovering trap, the leader net used was 200 m while the fykes ranged from 25 to 75 m depending on the model and with a leader net height ranging from 3 to 1.2 m (Table 1). Earlier studies have investigated the effect on leader net length on catch rate, and no significant effect was found between 100- and 200-m lengths (30). However, there are no studies evaluating the effect of leader net lengths shorter than 100 on cod catch rate. Future studies should thereby focus on the optimization of leader net dimensions as this may also affect handling of the gear. Pontoon traps and fykes are often handled using smaller boats, typically between 5 and 8 m. As such, gear handling ease is an important factor in both fisheries. Apart from the lower gear costs, use of shorter leader nets facilitates the exchange of leader nets during prolonged fishing periods. Net exchange will also mitigate buoyancy issues due to algae overgrowth, ultimately resulting in a positive effect on catch rate.

Our results indicated that catches in pots were unaffected by soak time (Figure 3). In baited cod pots, catch rates have shown pronounced seasonal variability, with CPUE being significantly lower during the early season (April to June) and became markedly higher in late summer and autumn (August–November) compared to traditional gears (45). Consequently, studies conducted over a limited temporal window, such as a single season or, as in this study, weeks, may not capture the full range of seasonal variability in pot efficiency and can provide misleading estimates of average annual performance. Short-term trials conducted only during low-density periods could underestimate pot effectiveness.

A factor to the variability in pot catches may be related to bait attraction and can strongly influence pot efficiency. Previous studies have shown that the release rate of chemical attractants is initially high but declines rapidly, suggesting that a pot may be effective for only a few hours under standard conditions (59, 60). Also, the effective “active space” of the bait is determined by the bait’s chemical and the olfactory detection thresholds of the target species in combination with local environmental factors that influence bait plume spread, including tidal patterns and seabed topography (61–63). The water current speed affects both the initial concentration and the downstream dispersion of the bait plume; as faster currents can distribute attractants over a wider area, they also dilute the plume more rapidly, resulting in variable spatiotemporal plume distribution and detectability due to water current conditions (64). This suggests that factors beyond the initial release of bait attractants, such as fish behavior, site fidelity, or social attraction, may play a significant role in drawing into and retaining fish pots.

In contrast, the relationship between soak time and catch rate, in gillnets, was the opposite, as catch rates decreased with increased soak time (Figure 3). Being a leading, non-baited fishing gear, similar to traps and fyke nets, increasing catches over time would be expected. This pattern is most likely attributable to seal depredation.

4.2 Seal effects on fisheries

The second objective was to evaluate gear-specific seal effects on catch. We found that gillnets were the most affected gear, with more than 50% of the catch exposed to seal depredation (Figure 4). While entangled, gillnets offer little to no protection for the catch, allowing seals to easily depredate directly on the entangled fish, while posts, traps, and fykes allow for protection of the catch by their partly rigid design. In this study, only catch showing visual damage was included as seal-induced catch losses; however, the depredation also includes partial losses in which entire fish are removed from gillnets. Our results show how catch loss are comparable to rates in a recent study on seal depredation in the same area (19). The results also indicate how the implementation of alternative gear is urgently needed in the southern Baltic coastal areas. In all the alternative gears, the number of damaged fish was significantly lower than in the gillnets. Fyke nets had cod showing signs of seal-induced damage, with 2% of the catch being affected, while in pots and hovering pontoon traps, damage levels on fish were non-existent in our study (Figure 4). The patterns of seal damage indicate how the evolution of alternative, seal-safe gear can heavily reduce the undesired feeding reward of seals from fishing gear. In the case of traps and pots, substantial amounts of effort have been put into the development of implementing seal safety. Pots have been made rigid (46) and, in some models, only collapsible when brought up on deck (47, 65), as seals otherwise will cause collapse in order to reach the catch. Another important aspect is the mesh size used in the pots; in this case, a netting with a mesh size of 22 mm (knot-to-knot) was chosen. As previous studies have assessed size selection of cod (66), recent experiences have shown that a smaller mesh size is needed to hinder seals from pulling cod from the pots (46). To allow for the escape of undersized individuals, the pots are instead equipped with escape windows (Figures 2A1:c, A2:c, Table 1). The netting used was stretched around the pot frame, as a looser netting is thought to induce more seal attacks and unintentionally creates pockets that allow for depredation (67). These key features combined allow for seal safeness, although the volume of the pots is small and, by that, would risk high depredation levels. For traps, a key to preventing seal-damaged catch is the large fish-chamber volume (4.5 × 1.5 m), which protects the catch despite a mesh size that could allow for depredation (Table 1). The fish may thus rest and protect itself in the center part of the chamber, where it is unreachable from seal attacks. In fyke nets, the construction is not built upon a rigid frame but on stainless steel rings, only connected by the netting. Fyke nets have previously shown to be affected by seal depredation (27). However, the fykes in this study have been equipped with entrance steel rings with polyester funnels (Table 1) to withstand depredation through entering the fyke. Additionally, the steel rings used in the frame are made from heavier material than traditional fykes to avoid collapse when attacked. The same 22-mm mesh netting, as in the pots, was used to hinder catch to be pulled from the fish chamber, from the outside. To reduce the risk of predation, the fyke net was accurately stretched at each deployment to avoid further collapse. The most probable explanation for the depredation is that despite being stretched, pockets or parts with loose netting may form in the gear, allowing for seals to reach the catch. With fykes being the latest gear to be developed for seal safeness, there is much room for further refinement. One improvement would be to use Dyneema netting (www.dyneema.com) to make it more robust, as earlier studies have shown that fishing gear made from Dyneema is more resistant to attacks by seals compared to traditional materials like nylon and polyethylene (68). Besides durability to seal attacks, replacing traditional mesh with Dyneema offers several other advantages. With Dyneema nets being lighter, drag in the water may be reduced, enabling easier handling, which enhances operational efficiency and reduces fuel consumption. Moreover, camera-based underwater observations of the fykes would facilitate further development, a method commonly used today when evaluating gear (57).

4.3 Handling time

The third component of catchability evaluated was gear handling time, which is often neglected in gear development (30, 45, 69). The four gear types used in this study present rather different characteristics when it comes to handling. Both fykes and pots are relatively small individual units, often handled linked together in strings. While pots are each connected to a specific string, often with a 50-meter spacing in between each pot, fyke nets are tied together either by cod end to cod end or by cod end to leader net end, shaping a string by the gears themselves. Each pot string contains various numbers of individual units, optimized for the fishers’ individual conditions. We have used a within-string count of up to eight individual fykes and 10 pots, respectively. The pontoon trap was deployed solitary. For gillnets, for which fishing vessels are often optimized, each set within this study was one 800-m net.

4.4 Comparison of gear-specific catch rates depending on handling time

While gillnets initially showed the highest catch rate per unit and day, it is essential to evaluate how handling time affects the total potential catch over a full day for the different gear types. With handling time often overlooked in passive fishing gear development, it is a neglected but crucial factor for fishers’ profitability and for accurately estimating fishing effort, which is emphasized in this study. Despite the relatively high daily catch in gillnets (catch per net), the high proportion of seal-damaged catch in gillnets (Figure 4) heavily decreases the actual commercially viable catch. When accounting for handling time and seal depredation, the results indicate that fyke nets tangent gillnets in yielding the highest average daily (Figure 5, Table 3). The results on extrapolated catch rates, combined with the near-absent levels of seal damage in pots, pontoon traps, and fyke nets, support the conclusion that these gear types represent viable alternatives to traditional gillnets in areas with high seal interaction levels. Although showing the highest extrapolated catch, the variability, particularly in fyke nets and gillnets, shows large standard deviations in the bootstrapped analysis, indicating an uncertainty in the potential daily catches, and emphasizes that the results may vary depending on conditions. This study was conducted during the spring in the vicinity of Ystad in the southern Baltic Sea. During this season, it is commonly known from fishers’ experiences how cod move from deeper into shallower areas for feeding. The greater variability in the fyke nets may have been caused by the regular relocation, much more frequently than pots and the pontoon trap (Figure 1). The higher variability in the gillnets may be caused by a larger deployment range, but more likely due to the larger extent of seal-induced damage to the gear (Figure 4). This study was carried out during a limited time during spring, which concurs with earlier studies that have shown high variability in gear-dependent catch also depending on fish stock, season, individual fisher efficiency, and, most relevant, location. Also, this study focuses on cod as target species, for which there currently is only a bycatch quota. To be able to allow for commercial viability, similar studies should include other target species, but with that comes a probable evaluation of entrance design in the gear. Further studies should include more fishers situated over a larger geographical area and over a larger time frame, as both catch composition and seal damage are supposed to vary in time and between locations.

5 Conclusion

In this study, we conclude how alternative passive fishing gears, specifically fyke nets, pots, and pontoon traps, present viable and more seal-safe alternatives to traditional gillnets in the Baltic Sea cod fishery. While gillnets showed the highest initial catch rates, the extensive seal damage significantly diminishes their commercial effectiveness. When factoring in handling time, bootstrapping showed that fyke nets demonstrated potential for the highest average daily catch. We address that future studies should focus on optimizing the design and material of alternative gears, particularly fyke nets, by investigating variations in leader net length and height, and exploring the use of more robust materials to enhance seal resistance. Moreover, there is a need to more precisely refine gear designs to minimize bycatch of undersized cod and non-target species, contributing to more sustainable fishing practices. Furthermore, we address the importance of evaluating these alternative gears across a broader range of environmental conditions, seasons, and geographical locations to assess their consistent performance and commercial potential. Finally, we stress the need to assess the economic feasibility and fisher acceptance of transitioning to these alternative gear types since commercial fishing, not just in the Baltic Sea but also in other areas where seal-inflicted damage in fishing gear, is causing economical loss in coastal fisheries.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Michele Luca Geraci, Istituto per le Risorse Biologiche e le Biotecnologie Marine (CNR/IRBIM), Italy

Reviewed by: Michael Dähne, University of Rostock, Germany

Andreas Winter, Falkland Islands Government, Falkland Islands