Fast and behavior-selective exploitation of a marine fish targeted by anglers

Abstract

Harvesting of wild-living animals is often intensive and may selectively target heritable behavioral traits. We studied the exploitation dynamics and the vulnerability consequences of individual heterogeneity in movement-related behaviors in free-ranging pearly razorfish (Xyrichthys novacula). Using underwater-video recording, we firstly document a fast and high exploitation rate of about 60% of the adult population removed in just few days after the opening of the season. Subsequently, we tagged a sample of individuals with acoustic transmitters and studied whether behavioral traits were significant predictors of the vulnerability to angling. Tagged individuals revealed repeatable behaviors in several home range-related traits, suggesting the presence of spatial behavioral types. The individuals surviving the experimental fishery showed only localized and low-intensity movement patterns. Our study provides new insights for understanding the harvesting pressures and selective properties acting on behavioral traits of recreational fishing. Many fish stocks around the globe are today predominantly exploited by recreational fisheries. The fisheries-induced change in fish behavior described here may be therefore widespread, and has the potential to alter food-webs, profitability of the fisheries and to affect stock assessment by eroding catchability in the long-term.

Figure 1. Home range behavior defined by an Ornstein–Uhlenbeck process.

(A) Four discrete-time trajectories simulated to visualize the individual heterogeneity in the behavior of pearly razorfish, Xyrichthys novacula according to the home range model proposed by ref. . Four realistic combinations of the movement parameters (k and radius) in the range of empirically observed data (Simulation (Sim) 1: 0.001 min⁻¹ and 245 m, Sim 2:0.001 min⁻¹ and 387 m, Sim 3: 0.01 min⁻¹ and 245 m and Sim 4: 0.01 min⁻¹ and 387 m) were used to simulate 674 time-steps (of 15 min each) following the movement model. In all cases, the latitude and the longitude (in metres) of the center of the home range were 0. (B) Head map showing the effect of increasing the radius (in m) and the exploration rate (k in min⁻¹) on the average swimming speed of the individual according to the movement model considered here. The increase in both parameters corresponds to an increase in the average swimming speed (m/s) of the individual.

Figure 2. Location and dynamics of the recreational fishery of the pearly razorfish, Xyrichthys novacula.

(A) Map of the study area located in the NW Mediterranean. The central map shows the study area located in the marine protected area (MPA) of Palma Bay, Balearic Islands, Mallorca, Spain (MPA delimited by the isobath of the 30 m in relation to land). The central map also shows the location of the no-take MPA (ntMPA in dashed), the partial MPA (pMPA) and the location of the 21 omnidirectional acoustic receivers (black dots) used for the tracking the fish. The right panels show a detailed map and the location of the sampling points (as stars) where the underwater video cameras were deployed within the ntMPA (in blue) and the pMPA (in red). The habitat of the study area was composed by seagrass of Posidonia oceanica (PO), photophilic algae habitats (PA), fine-grain sand (FGS), medium-grain sand (LGS) and large-grain sand (LGS). Note how the suitable habitat for the pearly razorfish (FGS and MGS) is surrounded by PO restricting the movement within this area and limiting dispersal. The map was created by the first author of the manuscript using ArcGis 10.3 for desktop (http://desktop.arcgis.com/es/desktop/) and self-created base maps and shapes (B) Average and standard deviation of the number of recreational boats per unit of area (boats × km²) in the pMPA (exploited area of ~6.5 km²) visually censed (every 15 min) during several days in 2012 (red), 2013 (green) and 2014 (blue). Before the opening the fishery is closed and enforced according to the seasonal closure stipulated for this species in the area. By September, 1 (vertical dashed red line) the fishery is opened, inducing a harvesting peak in the fishery. The survival of the pearly razorfish was assessed during this harvesting peak considering the first 9 days (vertical dashed blue line). (C) Main effects and confidence intervals (2.5% and 97.5%) of the generalized mixed effects model (GLMM) fitted to test the effect of the interaction between the study areas (no-take or control in blue vs exploited in red) and period (before vs after) on the abundance of the pearly razorfish (approximated by nMax, the maximum abundance of fishes observed on a video frame, see Methods).

Figure 3. Miniaturized acoustic tracking of the pearly razorfish, Xyrichthys novacula in the wild using an array of omnidirectional receivers Chronograms of the number of acoustic detections per hour of the fish tagged and tracked in our study area in the years 2012 and 2013.

The plot shows the day of the tagging (vertical blue line), the opening of the fishery (red line) and the end of the monitoring period for assessing survival to the fishery (green line). Note the pattern of detections generated by the pearly razorfish characterized by the lack of detections during the night as the individuals bury in the soft bottom to avoid predation and to rest.

Figure 4. Individual heterogeneity of home range-related behaviors and the existence of behavioral types in the pearly razorfish, Xyrichthys novacula.

(A) Daily values (natural logarithm transformed) of the space use, center of activity and total distance travelled (TD) in the pearly razorfish revealed by aquatic telemetry. Each color represents an individual, and the Julian days were transformed and standardized to ordinary numbers for improved visualization. The sample size (days) was on average across all individuals 10.8 ± 3.5 days. The left panel shows the daily space use as 50% and 95% KUD in km², the latitude and longitude of the center of daily activity in geographic coordinates and the total distance travelled (in m) on a daily basis for each of the pearly razorfish tracked in 2012 and 2013. Note the small within- and the consistent among-indivuidual variability in all behavioural traits. (B) Repeatability for each trait and year and its Bayesian Credibility Intervals (BCI, 2.5% and 97.5%).

Figure 5. The fitness-consequences of harvesting on behavioral types in the pearly razorfish, Xyrichthys novacula.

Predicted survivorship from the Cox-regression model fitted to test the effect of radius and exploration rate k on the survival (as surrogated of fitness in the fished environment). The survivorship plots show the predicted survivorship of 21 simulated individuals from the minimum and maximum radius (in m) observed in our data with an increment of 20 m. In case of k, the plot shows the predictions for 25 individuals from the minimum and maximum of k (in min⁻¹) values observed in our data by incremental steps of 0.001 min⁻¹. The average effects in both cases were statistically significant as revealed by the Cox-regression.