Identification of Potential Sea Turtle Bycatch Hotspots Using a Spatially Explicit Approach in the Yucatan Peninsula, Mexico
Abstract
A spatially explicit participatory approach was used to collect fishing effort and sea turtle bycatch data from local fishers at 15 ports in the Yucatan Peninsula, Mexico. These data were combined with satellite telemetry data to define potential bycatch hotspots. This is the first participatory and spatially explicit study on sea turtle bycatch rates in the region. Hawksbill turtles (Eretmochelys imbricata) were the most frequently caught bycatch species, followed by loggerheads (Caretta caretta) and green turtles (Chelonia mydas). Gillnets were the most dangerous for sea turtles, with the greatest incidence of dead turtles caught. Three particular bycatch hotspots were identified at the northeast, northwest, and southwest coasts of the peninsula. Identification of bycatch hotspots is recognized worldwide as a key element for protecting these endangered species, particularly in a region such as the Yucatan Peninsula that harbors critical habitats for ≥ 4 sea turtle species, 2 of them categorized as critically endangered (hawksbills and Kemp's ridleys [Lepidochelys kempii]). The spatially explicit participatory approach is versatile, easy to implement, and strategic for generating information under marine spatial planning for endangered species conservation.
Sea turtles are endangered throughout the world (International Union for Conservation of Nature [IUCN] 2017). The Yucatan Peninsula, located in the southeast of Mexico, is a key region for sea turtle conservation because it harbors the largest hawksbill turtle (Eretmochelys imbricata) nesting population in the western North Atlantic (Mortimer and Donnelly 2007). It also contains the only nesting beaches for loggerhead sea turtles (Caretta caretta) in Mexico (Lamont et al. 2012), and receives one of the largest nesting populations of green turtles (Chelonia mydas) in the Wider Caribbean (Spotila 2004; Dow et al. 2007).
There is an urgent need to monitor the in-water stages of sea turtles and the threats they confront particularly from fisheries, both of which are considered main causes of population reduction for sea turtles worldwide (Hamann et al. 2010; Wallace et al. 2011). The wide occurrence and intensity of fisheries on continental shelves, as well as their coincidence with areas where sea turtles occur, make this threat one of the greatest concerns (Wallace et al. 2013; Huang 2015). In this context, traditional practices such as keeping a gillnet soaked longer than its optimum efficiency time (i.e., inappropriate practices; Acosta 1994) and the use of prohibited fishing gear, are two of the main causes of high turtle bycatch rates (Kleiber and Boggs 2000).
Among the diverse types of fisheries around the world, artisanal fisheries are generally considered the most common, with intense activity in coastal zones that provides livelihoods for people directly involved in the industry and for local communities (Allison and Ellis 2001; Muallil et al. 2011). Based on the Food and Agriculture Organization of the United Nations (FAO 2015) definition, an artisanal fishery is characterized by using independent, small fishing boats with limited technology.
More than 6000 artisanal fishing boats operate in the north and west of the Yucatan Peninsula, and they are mainly concentrated in the coastal zones (Salas et al. 2006; Herrera-Pavón 2010; Munguía 2010). The fleets are grouped as multispecies targeted and multigear, with spatial and temporal dynamics dependent on the season of the target species (Salas et al. 2007).
Potential interactions between sea turtles and fishing activities are most likely to occur in zones where sea turtles are more abundant and where they stay the longest (Peckham et al. 2007; Wallace et al. 2011). This is the case of the Yucatan Peninsula, where important reproductive areas, feeding grounds, and migratory corridors have been identified for multiple sea turtle species (Cuevas et al. 2008; González-Garza 2009; Méndez et al. 2013; Shaver et al. 2013; Vázquez-Cuevas 2015). In this region, indirect efforts have been carried out to assess bycatch rates and illegal sea turtle consumption, and to describe artisanal fisheries (Labarthé-Horta 2005; Guzmán-Hernández and García-Alvarado 2006; Cuevas et al. 2009). These studies have indicated that inappropriate fishing practices with gillnet and artisanal longlines are the most dangerous fishing activities for sea turtles in the region.
Despite the information acquired by previous studies, no regional evaluation has been conducted and there has been no consolidated approach for assessing the overlap of sea turtle habitats and fishing activities. The Conservation Actions Program for Hawksbill turtles in Mexico (Secretaría de Medio Ambiente y Recursos Naturales 2009) and international cooperative plans in the Wider Caribbean (Secretaría de la CIT 2010; National Fish and Wildlife Foundation 2011) have emphasized this information gap. Both initiatives highlight the need to identify bycatch hotspots for sea turtle populations. In this sense, participatory mapping has proven to be a useful tool to collect spatial information derived from empirical knowledge on marine conservation, planning, fishing effort, and sea turtle bycatch distribution (Aswani and Lauer 2006; Moore et al. 2010; Bjorkland 2011; Brown and Raymond 2013; Strickland-Munro et al. 2015). This approach reinforces existing baseline information because it considers explicit data on fishing effort and the interactions with sea turtles. Participatory mapping has not yet been applied in the Yucatan Peninsula for the identification of sea turtle bycatch hotspots, nor has it been used to generate spatially explicit information regarding the fishing effort by artisanal fleets.
The objective of this study was to identify spatial hotspots of sea turtle bycatch based on fishing effort from artisanal fleets and their overlap with sea turtle habitats in waters adjacent to the Yucatan Peninsula, Mexico. We addressed this through participatory mapping with local fishers, and by collecting and integrating the local knowledge in a spatially explicit approach. Our hope is that these data will foster greater knowledge about where sea turtles are in most need of protection and what artisanal fleets and fishing gear present the greatest threat to their survival. Ideally these data will facilitate the development of robust conservation efforts directed at sea turtles of the Yucatan Peninsula.
METHODS
Study Area
The Yucatan Peninsula is located in southeastern Mexico, projecting into the Gulf of Mexico and the Caribbean Sea. This specific location between the Neotropical and the Nearctic biogeographic regions favors the establishment of highly productive ecosystems with species of ecological interest, such as whalesharks (Rhincodon typus), rays (Rhinoptera bonasus, Mobula birostris), seagrasses (Thalassia testudinum), and sea turtles (Wilkinson et al. 2009).
This region harbors all sea turtle species reported from Mexico, except for olive ridleys (Lepidochelys olivacea). Sea turtle species recorded in the area include green turtles, hawksbills, loggerheads, Kemp's ridleys (Lepidochelys kempii), and leatherbacks (Dermochelys coriacea). They nest in sandy beaches and feed, mate, and migrate along the seagrass, coral reef, and hard bottom habitats around the peninsula.
Questionnaire Design
The use of geographic information systems for acquisition of empirical local knowledge is a growing research field for social sciences in general, and particularly for marine conservation and planning (Aswani and Lauer 2006). In this context, implementation of questionnaires and workshops with local fishers for purposes of gathering essential data about their common livelihood activity and patterns and to translate these data onto maps and graphics is a robust strategy that is widely used in marine sciences, among other disciplines, because of the amount of information that can be drawn by its application (Strickland-Munro et al. 2015).
In this study, rapid-assessment interview tools (Moore et al. 2010) were adapted for local conditions, in terms of the features of local artisanal fisheries and the objectives pursued in this assessment, and an interview tool was elaborated for sea turtle bycatch and spatiotemporal fishing effort assessment.
We interviewed local fishers in 3 Mexican states—Campeche, Yucatan, and Quintana Roo—using a structured questionnaire (see the Appendix). The interviews were carried out at 15 ports representative of this region (Fig. 1). Fishers could identify and provide information about sea turtles because of their constant contact with multiple species at their fishing grounds, as well as their traditional consumption of sea turtles in the region.
Citation: Chelonian Conservation and Biology 17, 1; 10.2744/CCB-1263.1
The questionnaire was divided into 4 main sections: sea turtle species occurring in the study area, fishing gear, fishing effort quantification, and sea turtle bycatch estimation. The first section included questions about the species that fishers have seen in the study area and the frequency with which they observed each species. The second section included questions regarding the physical description of the fishing gears (length, height, mesh size) as well as where they use each gear type. The third section queried about the frequency of use for each gear type (season and number of months) as well as some general data regarding their crew. Finally, the last section asked for information on the catch rates and mortality of the species incidentally captured, although fishers were not able to spatially differentiate bycatch zones for each species.
A map was included with each questionnaire, indicating the location of the nearby ports. The map depicted a grid of 6-km-diameter hexagons over the marine area in the map. It also included a terrestrial section with names of the most relevant localities and some landmarks to assist the fishers in locating their fishing grounds on the map. To obtain the spatial distribution of the fishing effort per gear type, the fishers marked every hexagon where they usually fish, specifying the gear employed (Appendix 1).
The questionnaire data were collected during 2012 and 2013 at landing sites of the ports and docks where fishers landed their catch, and interviews lasted ∼ 20 min/fisher interviewed. We attempted to interview ≥ 50 fishers/port, representing ≥ 5% of the registered boats in those locations (Comisión Nacional de Acuacultura y Pesca 2013). In some of the studied ports, however, the number of fishers using the gear type of interest and who were willing to respond to the questionnaire were fewer than the target number. Nevertheless, data collected in those ports were included because they helped complete the regional panorama. The interviewers had previous experience applying the same questionnaire, so they could adapt it in case of any unexpected situation.
Fishing Effort Assessment
From the fishing effort section in the questionnaire, we could evaluate the annual dynamics per fishing gear type in each port. The fishing effort was measured in square meters of net (for gillnets) and in thousands of hooks (for longlines) as reported by every interviewed fisher. Total fishing effort was the sum of the values reported by every fisherman for every specific hexagon.
All the fishing maps data were digitized and incorporated into a geographic information system. We used ArcGIS (Environmental Systems Research Institute [ESRI] 2010) to create a vector layer with an attribute table containing the collected data as attributes for each of the hexagons in the grid. We obtained average and standard deviations of fishing effort per gear type for each hexagon, and these metrics were included in the maps to provide context of the spatial variability of the fishers' responses.
To provide confidence estimation for the participatory maps, the interviewing team went back to the ports and randomly selected a set of 30 fishers for a second application of only the spatial section in the questionnaire (i.e., the map). The objective was to quantify the hexagons that were consistently selected as fishing grounds, taking into account 3 ranges of fishing effort for this assessment (low, medium, and high), based on Jenks classifications of fishing effort distributions (Chen et al. 2013). The new fishing effort polygons obtained from verification were overlaid on the maps from original questionnaires and compared, looking for those consistent hexagons to obtain a simple metric of spatial error. The mean percentage of correctly assigned hexagons gave us an intuitive metric of the spatial consistency of the data we obtained with this participatory process, a procedure modified and adapted from Green et al. (2000) for these semiquantitative spatial data.
We evaluated the differences in fishing effort between months by using a nonparametric variance analysis (Kruskal-Wallis H), including a post hoc contrast test.
Bycatch Assessment
The bycatch per unit effort (BCPUE) rate for each fishing gear was calculated as follows:
where CThexi is the number of captured turtles in hexagon i, and 1000s of hooks per fishing season and net in m2 per fishing season are the fishing effort at hexagon i by longlines (Eq. 1) and gillnets (Eq. 2), respectively.
We compared the mean reported bycatch between ports for each fishing gear type using an analysis of variance test (normality and homoscedasticity assumptions were accomplished, p > 0.05). For each hexagon of the corresponding fishing gear, we calculated the mean bycatch, the sum of the total number of turtles caught, and the total number of dead turtles per species.
Sea Turtle Occurrence in the Study Area
To evaluate the overlap of fishing areas with sea turtle aggregations, we used information collected from previous studies on habitat use by postnesting hawksbills (n = 16) and green sea turtles (n = 12) (Cuevas et al. 2008, 2012; González-Garza 2009; Méndez et al. 2013; Vázquez-Cuevas 2015). The tracking data were differentiated and classified as feeding and internesting phases, using the vector layer of recorded points and the female's nesting dates as references, their spatial behavior along the tracking period, and behavioral diving data collected by the electronic tags (Tucker 2010).
Information gathered with satellite transmitters is expensive and, thus, limited data are available for sea turtles in many areas, including Mexican waters of the Gulf of Mexico. Only recently the total number of tracked turtles in telemetry projects around the world exceeded 30 individuals, including Mexico (Jeffers and Godley 2016).
We used the best available robust data of sea turtle locations to define the aggregation zones for green and hawksbill turtles (Cuevas et al. 2008, 2012; González-Garza 2009; Méndez et al. 2013; Vázquez-Cuevas 2015). We estimated the home ranges for each individual, separated in internesting and feeding phases, using a 90% kernel utilization distribution and a smoothing factor (h) calculated using the ad hoc method (Schofield et al. 2010). For these analyses, we used software Program R (R Development Core Team 2017) and the library adehabitat (Calenge 2006).
The home ranges of the tracked individuals were weighted based on the differences in conservation status per species (λsp, i.e., extinction risk category; IUCN 2017) and the behavioral phase (λmvt, based on the time these turtles remain in each phase, e.g., weeks at internesting areas, years at feeding grounds; Troëng et al. 2005; Gaos et al. 2012). Hawksbills are critically endangered with a general slightly positive historical trend on nesting indicators, and green turtles are endangered with an exponential increase of the number of nests in this region (Cuevas et al. 2013; Campbell 2014; IUCN 2017). Therefore, we proposed to weight hawksbills with a λsp = 0.7 and greens with λsp = 0.3, where hawksbills double the weight value of greens because they are critically endangered (Cuevas et al. 2013; Campbell 2014; IUCN 2017), whereas greens have a lower λsp owing to lesser extinction risk status and an increasing trend in annual nesting numbers.
The weight for each aggregation phase was divided, with a λmvt = 0.35 in the internesting phase and a λmvt = 0.65 in the feeding phase, with the latter phase doubling the first because individuals spend most of their lifetime at their feeding grounds (Broderick et al. 2001; Beggs et al. 2007). The weights of each level of the parameters summed to 1, as is recommended when assigning numeric values in multicriteria analysis (Malczewski 1999). The values were proposed for the quantitative evaluation, where the highest (feeding phase) was almost double the lowest (internesting), and they are a strategy to give a relative quantitative value to each behavioral phase (IUCN 2016).
Therefore, the home ranges were weighted as the product of both λs (Table 1) as it is described below.
Every hexagon was assigned a value of sea turtle use intensity (STUI), based on the equation:
where n is the total number of home ranges (HR; individuals per stage), λsp represents the weight defined for the correspondent species of the HRj, λmvt is the weight defined for the phase of that HRj; and HRj takes value 1 where the hexagon corresponds to an area used by the individual or 0 where the hexagon is an unused area. As a result, we obtained a map of sea turtle use intensity represented in hexagons.
Potential Bycatch Hotspots
Finally, based on the theory that evaluation of the interaction between any object of interest and a threat acting on it may be named “potential” impact, we identified potential bycatch hotspots for sea turtles, derived from overlaying the sea turtle use intensity map and the fishing effort maps (Carr et al. 2017; Thiault et al. 2017). This analysis does not intend to predict bycatch values; instead, we aimed to analyze in a unitless absolute scale (0 to 1) the magnitude of potential interaction given the level of spatial coincidence between sea turtle aggregation areas and the fishing grounds in the region, as it has been done for other organisms with similar data availability (Peckham et al. 2007; Travassos Tolotti et al. 2015), in contrast to an intended modeling analysis such as the effort by Eguchi et al. (2017).
The maps of potential bycatch hotspots were constructed by adding the weighted grades for sea turtle use intensity at each hexagon (Eq. 3) and then multiplying by the mean fishing effort at the same hexagon obtained from the interviews.
where PBhex is the potential bycatch value between sea turtles and fishing activities, STUIhex is the intensity of use by sea turtle in each hexagon, and FEhex is the fishing effort calculated at the same hexagon.
We generated maps of potential bycatch or interaction (accidental encounters with fishing gear that can result in injury and possible death [Lopes et al. 2016]) between sea turtles and two types of fishing gear (gillnet and longline), with the highest values representing the potential bycatch hotspots in the region. These maps were rescaled using a linear approach so they were between 0 and 1,
where Zhexi is the rescaled value (0 to 1) for potential interaction at hexagon i, Xi the value at hexagon i obtained by Eq. 4, and min and max represent the minimum and maximum value for Xi in the map.
Finally, the potential bycatch maps were categorized in three classes defined by the natural breaks (Jenks) method (Chen et al. 2013), which is based on the natural inherent data grouping. The breaks are characterized by grouping similar values and maximizing differences between breaks.
RESULTS
Questionnaire
We interviewed 776 fishers in 15 ports along the Yucatan Peninsula (Table 2), with ≥ 30 fishers interviewed per port. There was a spatial variation in terms of the area used as fishing grounds by the fleets from different ports; fishers from Progreso, Yucatan, form some of the largest fleets in the region and they utilized the widest fishing ground area. On the other hand, the port with the smallest fishing grounds was Atasta, Tabasco, which is a very small port in the southwest of the peninsula. Neither the smallest fishing area, nor the widest, were associated with the least or greatest interaction rates with sea turtles, respectively.
Fishers recognized 5 sea turtle species occurring in the waters where they conducted fishing operations. The observation frequency for each species varied between ports along the study area, particularly for leatherback and Kemp's ridley turtles that were mainly reported from the northeastern and southwestern coasts of the Yucatan Peninsula, respectively. Loggerheads, hawksbills, and green turtles were frequently observed in all waters. Almost 50% of the interviewed fishers reported that they encounter more sea turtles currently than in the past decade.
Fishing Effort
The longline fishing effort was widely distributed all over the Yucatan shelf north of the peninsula (38% of hexagons in the study area; Fig. 2a). On the west side, in the Campeche Bay, a fringe was defined with the highest mean fishing effort estimated for a 1-yr period, ranging from approximately 10,000 to > 40,000 hooks.
Citation: Chelonian Conservation and Biology 17, 1; 10.2744/CCB-1263.1
In the case of gillnets (21% of hexagons in the study area), the largest area was recorded also on the west coast but further offshore than the longline fishing grounds (Fig. 2b). In addition, key areas on the same west coast and on the northeast corner were recorded with medium to high values for fishing effort. For these maps, a spatial consistency in fishing effort among hexagons of 87% was obtained through the validation by fishers from the same localities.
For both gear types, the areas with the highest values of fishing effort coincided with high standard deviation values. Most of the fishing activity in the region occurs in these areas, hence the variability in their values within 1 yr.
The annual fishing effort differed between months (H11 = 74.11, p < 0.05), with the greatest effort from February to July, followed by an abrupt drop in August. During September, October, and January the mean difference between months was 67.43%. Gillnets (with different local variants and mesh sizes, but all of them monofilament) are the most used fishing gear in the region (63.65%), followed by artisanal longlines (35%).
From August to January a clear decrease in fishing effort was reported for both longlines and gillnets. That period coincides with the start of the octopus (Octopus maya, O. vulgaris) fishery; therefore, the greatest fishing effort during that period is from an artisanal fishery that has no lethal impacts on sea turtles because octopus are caught with lines without hooks and that are constantly checked by fishers.
Bycatch
During a 1-yr period, Sabancuy and Champoton in the west coast, and Holbox Island and San Felipe to the north of the peninsula, were reported by fishers to have the highest values of bycatch rates with gillnets, although there were no significant differences (H11 = 802.95, p > 0.999), with mean values of 0.53 (± 0.65), 0.44 (± 0.57), 0.45 (± 0.60), and 0.50 (± 0.60) individuals/m2 of net/season, respectively. In the case of longlines, the port of El Cuyo at the north, and Isla Aguada and Sabancuy at the west, presented the highest bycatch values, although no significant differences were found (H11 = 907.28, p > 0.999), with averages of 0.70 (± 0.68), 0.72 (± 0.57), and 0.53 (± 0.56) individuals/1000 fishing hooks/season, respectively.
Specific areas of high reported rates of bycatch for longline fisheries were identified near the northwest corner of the peninsula (1.5–5.0), at the center of the north coast (1.5–5.0), and at a smaller one at the northeast corner of the peninsula (1.0–1.5; Fig. 3a).
Citation: Chelonian Conservation and Biology 17, 1; 10.2744/CCB-1263.1
Areas with exceptionally high bycatch-rate reports for gillnets were located very close to the shore along the northeast and northwest coasts of the peninsula, as well as in the southwest (Fig. 3b). For both longlines and gillnets, the standard deviation indicated that the higher the bycatch rates, the higher the rate of variability (Fig. 3).
There was a difference in the reported species caught at the north and west coast of the peninsula compared with the rest of the study area. This difference could be interpreted as an indication of the spatial distribution of the species. In general, hawksbill turtles were the most caught turtles in waters near the Yucatan Peninsula, particularly along the west coast, where they represented an average of 80.37% of all captured turtles. Off the north coast, hawksbills represented an average 49.50% of the sea turtles caught, followed by loggerheads (27.2%) and green turtles (21.50%). Off the west coast, green turtle individuals represented an average of 18.63% of captured turtles, followed by loggerheads with 5.37%. The least captured species in the study area was Kemp's ridley turtle with < 2% of captured individuals in the southwest corner of the region.
When we compared the total number of turtles caught dead between gillnets and longlines, the numbers differed between gears (U = 6.268E9, p < 0.05), indicating that gillnets captured most of the turtles. The number of dead turtles by gear type confirmed that gillnets, with the long soak times as used in several of these ports (> 10 continuous hrs), were the most dangerous fishing gear for sea turtles (U = 7.817, P < 0.05). The fishers were not able to report differentiated dead turtles per species because they reported total number of dead individuals.
Potential Bycatch Hotspots
Using the spatial concurrence of areas occupied by sea turtles (telemetry data) and fishing effort hotspots (from the interviews), we defined potential bycatch hotspots for sea turtles. The main hotspots for longlines were identified in the southwest and northeast coasts of the peninsula (Fig. 4a). These hotspots coincide with very important internesting grounds for hawksbills (southwest) and important feeding grounds for green turtles (northeast). For gillnets, the critical hotspots coincided with longline hotspots in the southwest and northeast coasts; however, a third hotspot was detected in the northwest corner, with a wide range of medium-intensity interaction around it (Fig. 4b).
Citation: Chelonian Conservation and Biology 17, 1; 10.2744/CCB-1263.1
DISCUSSION
Through the application of questionnaires to fishers and spatial distribution of the sea turtles (derived from telemetry studies), we could identify potential sea turtle bycatch hotspots along the northern and western coasts of the Yucatan Peninsula. Those zones with the highest values of interaction between sea turtle occurrence and fishing effort, for both fishing gear types, are located in the transition areas between waters of the Mexican states of Yucatan and Campeche, as well as in the northeastern and southwestern coasts of the Yucatan Peninsula.
Given the selected approach for representing the potential impact and interactions between sea turtles and fisheries, the areas with the greatest turtle-use occurrence (from satellite telemetry) and high fishing effort (from interviews) were highlighted as bycatch hotspots with the greatest potential impact (derived from the interaction of sea turtle occurrence and the threat). A similar approach has been implemented by other authors in the northern Gulf of Mexico for sea turtle hotspots (Hart et al. 2013) and sea turtles and fisheries (Hardy et al. 2014), the Atlantic Ocean (Huang 2015), the Caribbean Sea (Bjorkland 2011), the Pacific Ocean (Roe et al. 2014), and the Mediterranean Sea (Cambiè et al. 2012).
Fishing Effort
The artisanal fleets in this region are of the greatest importance to the fishing industry, and the extensive fishing effort they expend in the region is of concern for marine biodiversity conservation. Although the individual BCPUE per boat is small, the multiplicative impact these boats may have on sea turtle populations in the Campeche Bank is significant.
The fisheries along north coast of the peninsula primarily used longlines, in contrast to the west coast where gillnets are mainly used. We observed that 60% of the interviewed fishers on the west coast used gillnets more extensively, with < 20% of them using longlines. This observation agrees with those reported in previous assessments in the region (Guzmán-Hernández and García-Alvarado 2006).
The estimations of BCPUE are closely linked to fishing effort in the region. In all ports, it is evident that there is a decrease in the fishing effort of all reported fishing gear during the second half of the year. One of the main reasons for this is the start of the octopus fishing season, which runs from 1 August through 31 January. The octopus fishery is one of the most significant and best-paid fisheries in the region, and when the season begins, most of the fishers go after octopus using a gear known as “jimba,” which does not pose a threat to sea turtles (Salas et al. 2006). The first 3 mo of the octopus season are the most productive, and the risk of turtle bycatch is low because it coincides with the last half of the nesting season of sea turtles in the region (nesting season for both species extends from April through October, and the octopus fishing season starts in August), and the beginning of the period when turtles migrate to their feeding grounds. Also, it has been reported that the neritic waters north of the peninsula, where the octopus fishing occurs, are a crucial corridor for green and hawksbill turtles going from their nesting beaches to their feeding grounds, which usually occurs from July through November every year. In January, when the octopuses become scarce, the fishers return to their usual fishing activities, thus increasing the bycatch rate for sea turtles, which by then are established at their feeding grounds.
Many fishers reported that they encounter more sea turtles now than in the past decade. This could be interpreted as a positive result of conservation efforts on the beaches; however, despite the perception of an increase in sea turtle populations, in-water and on-beach protective and management efforts must be continued.
Bycatch
The mean bycatch estimates for the ports with the highest bycatch rates in this study are comparable to others around the world. Báez et al. (2014) reported bycatch values of 0.49 individuals/1000 hooks for longlines that are lower than the average values of 0.53–0.72 from the present study. Finally, Cambiè et al. (2012) reported a mean bycatch rate of 0.24 loggerheads caught per 1000 hooks in the coast of Ionian Calabria in Italy, which are values comparable to those reported in this study for longlines.
In the Yucatan Peninsula, Herrera-Pavón (2010) reported bycatch estimates for the eastern coast (Mexican Caribbean) of 0.0017 turtles/1000 m2 for gillnets and 0.08 turtles/1000 hooks; both values are lower than those reported in this study, considering the same units of effort. Herrera-Pavón (2010) also collected questionnaire data from fishers in an area close to our study area. However, key differences from this study are the spatially explicit information presented here, the inclusion of sea turtle tracking data, and the regional approach. On the other hand, Guzmán-Hernández and García-Alvarado (2006) reported values of 0.09 turtles/1000 m2 of gillnet in the western coast of the peninsula, and identified this fishing gear as the most dangerous for sea turtles in this region when soak times were > 6 hrs.
The spatial distribution of BCPUE for gillnets showed clearer patterns than for longlines along the study area, with well-defined hotspots close to the shore that coincided with areas where fishers reported extensive use of this fishing gear. In those hotspots close to the shore, chances of interaction with sea turtles are greater because these sites are important developing grounds for juveniles, as well as migratory corridors for postnesting females (Garduño-Andrade 2000; Cuevas et al. 2007).
Longlines seem to have a higher bycatch efficiency than gillnets (gillnets capture more turtles in similar set times); however, gillnets have a greater negative impact because of the larger number of dead turtles caught. This makes gillnets a more dangerous and lethal fishing gear compared with longlines. It is important to note that such greater mortality occurs when the net soak time is > 5–6 hrs; lesser soak times generally result in greater survival of bycaptured turtles as we observed in this assessment.
Potential Bycatch Hotspots
In the present study we were able to create maps for fishing effort and bycatch rates. A very decisive factor in achieving this was that, in this region, most of the fishers own global positioning system navigators and have at least basic knowledge of sailing using digital instruments; this is in contrast with poorer regions, where it is challenging to get a spatial representation of fishing grounds.
Several of the tracked individuals, per species, from different nesting beaches congregated at the same feeding grounds, suggesting the sample is representative of the location and spatial configuration of feeding grounds for these 2 species in the Bank of Campeche (for more details see Cuevas et al. 2008, 2012; González-Garza 2009; Méndez et al. 2013; Vázquez-Cuevas 2015). This fact had implications for the definition of hotspots because they marked the zones of greatest presence of these species and the fishing efforts in these zones may have greater impact than at other areas with less turtle occurrence (Wallace et al. 2013).
The most significant bycatch hotspots, identified by overlying the telemetry data and fishing effort from interviews, were observed along the northwest, northeast, and southwest coasts of the peninsula owing to the intensive use of gillnets and longlines over important sea turtle habitats in these areas. Other authors have reported those same areas as crucial feeding grounds for different sea turtle species. Blumenthal et al. (2006) reported both the northwestern and northeastern coasts as important feeding grounds for green turtles migrating from the Cayman Islands. Shaver et al. (2013, 2016) also reported that both the northeastern and northwestern coasts of the peninsula are important feeding areas for postnesting Kemp's ridley turtles migrating from Tamaulipas, Mexico, and from Texas, United States. In other studies, Hart et al. (2012) and Foley et al. (2014) reported the same latter coasts of the Yucatan Peninsula as important feeding grounds for postnesting loggerheads migrating from western Florida, United States.
Furthermore, the southwestern area of the Yucatan Peninsula was highlighted as an important internesting area for adult female hawksbill and green turtles (Cuevas et al. 2008; Méndez et al. 2013). It is also an important development area for immature hawksbill turtles, as reported by Bjorndal et al. (2017).
Vázquez-Cuevas (2015) emphasized the need of protecting these areas, which are considered critical habitats for multiple life stages of at least 4 species of sea turtles. Nesting populations arrive in these areas from the north, south, and west Gulf of Mexico, as well as from the western Caribbean, making them sea turtle biodiversity hotspots.
All these reports confirm the great ecological value and conservation importance of both the northeastern and northwestern coasts of the Yucatan Peninsula. In accordance with these reports, the results of the present study support the conclusion that these zones are sea turtle hotspots, which is in potential conflict with the high fishing effort and resulting elevated bycatch rate.
Final Remarks
Regarding the great value of data collected from a participatory approach, as reported by Moore et al. (2010), the acquisition of spatial data for compiling participatory maps of bycatch hotspots is a key methodological tool for better understanding the interactions between sea turtles and fishing activities. The approach implemented in the present study proved to be versatile because it was easy to adapt to several locations from which the questionnaire data were collected, and the amount of data was sufficient for robust analyses, including the identification of hotspots. Thus, we strongly recommended this approach in any evaluation of threats on endangered species.
This study is one of the first in the region to widely investigate fishery bycatch in Mexico, and the first to integrate mapped information for the Yucatan Peninsula. This spatially explicit approach that identifies sea turtle hotspots resulting from the interaction with artisanal fisheries is key in decision-making and management for the conservation of these endangered species.
This information is a basis for the identification, lobbying, and implementation of official and voluntary strategies to diminish sea turtle bycatch in this region. The spatially explicit tools that this study provides create a practical and easy approach that could be used as an instrument for decision-making and collaborative work with local communities.
The process of working together with local fishers to identify potential areas to be comanaged requires a regional panorama of the actual interaction status between sea turtles and a sample of the artisanal fishing industry. This will facilitate and support the process of identifying key priority areas where the different strategies should be implemented, presenting as management and conservation alternatives the delimitation of no-take areas, temporal restrictions, and implementation of low-bycatch fishing practices for the protection of sea turtle populations in this region.
Study area indicating the 15 ports (stars) where a spatially explicit participatory approach was used to collect fishing effort and sea turtle bycatch data from local fishers in the Yucatan Peninsula, Mexico.
Fishing effort of (a) longlines and (b) gillnets as determined by interviewing local fishers at 15 ports in the north and west coast of the Yucatan Peninsula, Mexico.
Bycatch per unit effort participatory mapping, as determined by interviewing local fishers for (a) longline (caught turtles · 1000s of hooks−1 · fishing season−1) and (b) gillnet (caught turtles · m−2 · fishing season−1) in 15 ports in the north and west coasts of the Yucatan Peninsula, Mexico.
Maps of intensity of potential impact on sea turtles by (a) longlines and (b) gillnets, as determined by interviewing local fishers at 15 ports in the Yucatan Peninsula, Mexico. (Color version is available online.)
Quiestionaire applied for Campeche Area
Quiestionaire applied for Campeche Area
Contributor Notes
Handling Editor: Jeffrey A. Seminoff
