Seasonal movements and habitat use of Cobia (Rachycentron canadum) in the Gulf of Mexico

doi:10.21203/rs.3.rs-7963325/v1

Download PDF

Research Article

Seasonal movements and habitat use of Cobia (Rachycentron canadum) in the Gulf of Mexico

https://doi.org/10.21203/rs.3.rs-7963325/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

Cobia (Rachycentron canadum) is a coastal migratory pelagic species that supports a popular recreational fishery along the U.S. Atlantic coast and throughout the Gulf of Mexico (Gulf). Limited knowledge of seasonal movements, stock boundaries, and the subsequent distribution of fishing mortality is a key source of management uncertainty for the Gulf migratory stock. Therefore, the goal of this study was to estimate the seasonal movement patterns and habitat use of Gulf cobia using conventional and satellite tagging.

Results

Between June 2020 and October 2024, 125 cobia ranging from 432–1302 mm FL were conventionally tagged off the coast of Texas, USA. Of these, 34 were tagged with pop-up archival transmitting (PAT) tags. Excluding six nonreporting tags and four PATs with deployment durations < 10 days, PAT tags released between mid-August and mid-April with deployment durations ranging from 20–180 days (mean = 73 d; SD = 43 d). Observed dispersal from the tagging region suggested cobia moved south-southwest (n = 14; mean = 173.4 km; range = 8.6–719.6 km) or east (offshore; n = 3; mean = 46.9 km; range = 35.8–56.2 km) between September and April. While movements of most individuals were restricted to the tagging region, three PAT tags reported pop-up locations off Mexico. Geolocated maximum likelihood tracks revealed complex movements, with total distances traveled ranging from 697.5–5256.7 km. Estimates of post-release mortality were 3.6% and 26.5%, depending on the treatment of tag attachment failure and non-reporting tags. Cobia displayed strong preferences for surface waters in summer and progressively reduced their time near the surface into winter. Diel- and reverse diel vertical migration behaviors were also observed. Seasonal fidelity to the tagging region was observed, and no cobia were detected north of Galveston, TX.

Conclusions

This study advances our knowledge of cobia seasonal movement patterns and habitat use in the western Gulf and suggests the South Texas Shelf serves as suitable overwintering habitat for Gulf cobia. Evidence of migratory contingents warrants expansion of cooperative tagging and genetic sampling to better understand stock dynamics and advance effective management strategies in the Gulf.

Satellite telemetry

cobia

pop-up archival transmitting tags

Gulf of Mexico

movement ecology

fish behavior

Migratory fish species encounter dynamic environmental conditions, experience varying levels of fishing effort, occupy multiple habitat types, and often cross multiple management jurisdictions during their movements, which can span thousands of kilometers [1, 2, 3]. Knowledge of these movement patterns and habitat use is essential for a better understanding of stock structure and for identifying the spatial and temporal scales at which fisheries should be managed to meet sustainability targets. Specifically, effective fisheries management relies on detailed knowledge of stock delineation and the distribution of fishing effort and mortality among different stocks, as each stock must be managed appropriately to optimize yield [4, 5]. Unfortunately, for many fisheries, a clear picture of stock structure, seasonal movement patterns, and fishing mortality is lacking, which presents complex challenges to accurate stock assessment and effective management.

In the southeastern U.S., the cobia Rachycentron canadum is a coastal migratory pelagic species that supports a popular recreational fishery (> 90% of total landings) and is also captured incidentally in commercial fisheries and as bycatch in shrimp fisheries along the U.S. Atlantic coast and throughout the Gulf of Mexico (hereafter “Gulf”) [6, 7]. Cobia are managed as separate Atlantic and Gulf migratory stocks by the Atlantic States Marine Fisheries Commission (ASMFC; authority delegated by the South Atlantic Fishery Management Council in Amendment 31; [8]) and the Gulf of Mexico Fishery Management Council (Gulf Council; [9]). Both population genetics and conventional tagging data suggest a transition zone separating the Atlantic and Gulf migratory stocks along the south Atlantic coast between Savannah, Georgia, and Cape Canaveral, Florida [10]. Additionally, stock substructure exists within the Atlantic stock, including the presence of genetically distinct inshore and offshore groups off the Carolinas and Virginia [10,11, but see 12]. Satellite telemetry data indicate cobia tagged in coastal Virginia overwinter near the continental shelf break off North Carolina [13].

Despite the species’ popularity as a sportfish, few stock assessments have been conducted, and only recently has cobia been assessed through the Southeast Data, Assessment, and Review (SEDAR) process. The most recent stock assessments for the Gulf and Atlantic migratory stocks were completed in 2020 (SEDAR 28 Update and SEDAR 58). While the assessment for the Atlantic Stock indicated the stock was not overfished or experiencing overfishing [7], the assessment for the Gulf Stock indicated that overfishing was occurring, and the stock was at risk of becoming overfished in the near future, with spawning stock biomass at reduced levels (2018 SPR = 0.21; [14]). Confidence in the update assessment for Gulf cobia remained low and management advice was greatly hindered by severe data gaps. In particular, a lack of defensible abundance indices and significant data deficiencies regarding fishing mortality, seasonal movements, and stock structure contributed toward high uncertainty in the stock assessment [6, 14]. Compounding these uncertainties, Gulf recreational anglers have expressed dire concerns about the condition of the stock during public testimony at Gulf Council meetings in recent years [15]. Collectively, anglers have reported decreased landings and infrequent sightings of cobia in times and places where they have seemed abundant in the past. Further corroborating this testimony, the Gulf Council recently surveyed Gulf anglers to more objectively assess angler sentiment toward the cobia fishery and found that declines in abundance and fewer large fish were commonly mentioned in the responses [16]. Thus, there is uncertainty in the status of the fishery and a clear need for updated data inputs for future stock assessments to provide the best possible management advice.

Poor resolution of seasonal movement patterns and stock boundaries, especially in the western Gulf, is a key contributing factor towards Gulf cobia assessment uncertainty. While sub-regional stock structure has been identified in the Atlantic stock (e.g., [10]), sub-regional stock structure has not been thoroughly evaluated in the Gulf [17]. No comprehensive genetic studies are available, and conventional tag-recapture data provide the only inferences on seasonal movement patterns. Analysis of long-term (1988–2014) cooperative tagging data indicates a significantly higher proportion of cobia recaptures in the Florida Keys from November to March, a significantly lower proportion of recaptures in the northcentral Gulf from December to April, and a significantly higher proportion of recaptures along the Florida Panhandle in April relative to other months, suggesting migration from the northcentral Gulf in the fall to overwinter in the Florida Keys and then migration northward in the spring to summer in the northcentral Gulf [18]. Seasonal movements are less clear in the western Gulf primarily due to limited tagging effort. For example, Dippold et al. [18] reported a relatively low number of cobia tagged and recaptured in Texas over a 27-year study period (227 tagged, 5 recaptured with ≥ 30 days at liberty). Cobia tagged in Texas were only recaptured in Texas and Louisiana (one fish), suggesting cobia in the western Gulf may exhibit restricted movements or represent a resident group of fish [18]. In comparison, a cooperative tagging program based in Texas during the late 1980s recorded four cobia tagged in Texas and recaptured between Tampico, Mexico, and Mérida, Mexico during the winter months (Steve Qualia, Corpus Christi, TX, personal communication), demonstrating that at least some fish may make long-distance seasonal migrations.

Given the limited data on cobia seasonal movements in the Gulf, several research recommendations stemming from the previous stock assessment and the recent SEDAR Stock ID Workshop [17] have been identified. These recommendations include: 1) evaluate potential substructure within the Gulf stock, 2) investigate overwintering habitat, including the potential for inshore-offshore migrations using conventional and satellite tagging, and 3) conduct tagging studies to determine the southern boundary of the Gulf stock, particularly with respect to Texas and Mexican waters [6, 17]. This information is key to future management as potential removals of U.S. Gulf cobia in Mexico are not accounted for in the current assessment framework. Considering these data needs, the primary goal of this study was to provide novel information on seasonal movements and habitat use of cobia tagged off Texas via analysis of satellite telemetry and conventional tagging data. We specifically evaluated dispersal from the tagging area, overwintering habitat, and vertical and thermal habitat utilization. Satellite tag sensor data was also used to infer the fate of each released individual and provide the first empirical estimate of the post-release mortality rate for recreationally-caught cobia in the Gulf.

Fish Tagging

Fish capture and tagging followed procedures approved by the Texas A&M University-Corpus Christi Institutional Animal Care and Use Committee (Animal Use Protocols #2020-04-001 & #2023-0007) and the Texas Parks and Wildlife Department (Scientific Research Permit #SPR-0303-279). Cobia were captured from nearshore waters off the south Texas coast using gear and methods representative of the U.S. Gulf recreational hook-and-line fishery, including the use of artificial lures or live or cut bait around artificial structures such as standing oil and gas platforms, floating aids to navigation, and commercial shrimp trawling vessels. Once hooked, fish were brought alongside the vessel, netted, and placed on a V-shaped measuring board, with their head and gills submerged in seawater within a cooler. Individuals were measured for fork length (FL; cm) and conventionally dart-tagged (Floy©, Seattle, WA, USA). Each tag included contact information, a unique identification number, and “reward” printed in English and Spanish for reporting recaptures. To engage recreational stakeholders and increase the potential for additional movement and post-release survival data, conventional dart tags were distributed to 20 charter captains and/or private anglers. Participants in this cooperative tagging program were supplied with a conventional tagging kit containing a tag applicator, dart tags (as previously described), tagging instructions, and data cards to record information (e.g., total length [FL]; cm) for each fish tagged. Additional data recorded for each cobia captured and tagged by the research team included fight time (time from being hooked until netted and brought onboard; min), handling time (from landing to release, including any in-water revival), and hooking location (categorized as “shallow” if the hook was visible in the mouth or jaw, or “deep” if embedded in the gills, esophagus, or gut).

Cobia measuring ≥686 mm FL (27 in) were also tagged with one of three models of pop-up archival transmitting (PAT) tags: MiniPAT (118 × ⌀ 38 mm, 61 g; Wildlife Computers, Redmond, WA, USA), Survivorship PAT (sPAT: 118 × ⌀ 38 mm, 61 g; Wildlife Computers, Redmond, WA, USA), and Mark Report PAT (mrPAT: 118 × ⌀ 28 mm, 44 g; Wildlife Computers, Redmond, WA, USA). Data from transmitting PAT tags were used to observe dispersal from the tagging location, identify overwintering habitat (November–February), and estimate the post-release mortality rate for cobia captured in the recreational fishery in the western Gulf. To obtain an unbiased estimate of post-release mortality, PAT tags were deployed on all live cobia captured and of sufficient size for tagging, regardless of apparent condition. Each PAT tag was rigged with a 136-kg (300 lb) monofilament tether (15.2 cm) coated with heat shrink tubing connected to a titanium dart anchor, which was inserted into the dorsal musculature of the fish posterior to the leading edge of the dorsal fin. PAT tags were coated with antifouling paint to reduce biofouling and maintain transmission efficiency to the Argos satellite system.

MiniPAT tags archive environmental data (depth, temperature, and light level) while attached to the fish. After a pre-programmed duration (122 or 180 d in this study) or upon a conditional release event (e.g., fish mortality), MiniPAT tags detach from their tether and float to the surface. Upon surfacing, the tag transmitted summaries of the archived environmental data at 12- or 24-h intervals and depth and temperature time series at 450- or 600-s intervals to the Argos satellite system, which also allowed the pop-up location of the tag to be estimated. MiniPAT tags use sensor measurements to make reasonable assumptions about the state of the deployment, and thus the tagged fish, and were programmed to release and begin transmitting if certain conditions were met. Specifically, MiniPAT tags monitor for three mortality conditions: (1) Floater (tag floating at the sea surface for 3 d, indicating fish mortality or premature tag detachment); (2) Sitter (tag remains stationary on the seafloor at a constant depth ± 1.5 m for 3 d, presumably attached to a dead fish); and (3) Sinker (tag sinks below 1700 m, presumably attached to a dead sinking fish).

sPAT tags are a more cost-effective alternative to MiniPATs, designed for short-term survivorship studies (up to 60 d). sPAT tags record and transmit a reduced set of environmental data, including daily minimum and maximum depth and temperature, daily delta (difference between minimum and maximum) light-level, and depth times-series at 600-s intervals for the 5 d prior to release. Like MiniPATs, sPAT tags can detect conditional release events (Floater, Sitter, or Sinker).

mrPAT tags, the smallest of the three PAT models, are designed for large-scale dispersal and movement studies. These tags detach after a pre-programmed duration (57–131 d in this study) and float to the surface, providing a pop-up location estimate via the Argos satellite system. If a tag remains on the sea surface for two consecutive hours before the programmed release date, it is considered to have been released prematurely and will initiate transmissions. Based on the deployment date, mrPAT tags were programmed to release and begin transmitting data in the winter to track cobia movements and dispersal to overwintering habitats. Additionally, mrPAT tags transmit daily minimum and maximum temperature and daily delta tilt (change in tag orientation) from the last 100 d of deployment, which can be used to investigate thermal habitat use and infer fish survival (Jensen and Graves 2020).

Post-release Mortality

Post-release (or discard) mortality (PRM) has been estimated to occur shortly after release because of acute injury [19, 20]. However, delayed mortality due to the loss of ability to feed or infection can occur several days to weeks following release, especially if the fishing gear was not completely removed from the animal [21, 22, 23]. Although increasing the observation period beyond several days allows the incorporation of delayed mortalities, there is a risk of confounding results with natural mortality unassociated with the initial capture event or additional capture-related events after the capture event of interest (e.g., fish recaptured two weeks later, but it sustains significant injuries that lead to mortality). To minimize potential bias and maintain comparability with previous studies (e.g., [13]), only cobia that appeared to perish during the first 10 d after tagging were considered post-release mortalities. The PRM rate was estimated using two well-established methods (e.g., [24, 25]) for PAT-tagged fish. The first method included only mortalities definitively inferred from the tag data (e.g., Sinker or Sitter) using the expression:

PRM Rate = \(\:\frac{Mortality}{Survived+Mortality}\) (1)

The second and more conservative method yielded the highest possible mortality estimate, in which PAT tags that did not report (DNR) or experienced tag attachment failure (TAF) in the first 10 d after tagging (e.g., Floater) were also considered mortalities using the expression:

PRM Rate = \(\:\frac{Mortality+TAF\:\le\:10\:Days+DNR}{Total\:Number\:of\:Tags\:Deployed}\) (2)

The second method is based on the consideration that a PAT tag may fail to report from a live or dead cobia for multiple reasons, including mechanical failure of the release mechanism, biofouling, or destruction of the tag during an attempted or genuine predation event [26]. We included this method because eliminating non-reporting and premature releasing tags decreases the precision of the estimated PRM rate and may artificially lower estimates of PRM [24, 27]. Confidence intervals for these two estimates of PRM were calculated using software developed by Goodyear [28], which used PAT tag data to bootstrap the proportion of inferred mortality, survival, and non-reporting tags over 10,000 simulations, assuming no tagging-induced or natural mortality during the 10-d survival observation period.

Geolocation Estimation

Horizontal movements of cobia tagged with sPAT and mrPAT tags classified as survivors (> 10 days at liberty) were restricted to known tagging and quality pop-up locations (the first Argos location class 1 to 3 after transmitting, with an associated error estimate of ≤1500 m), and dispersal was calculated as the net displacement (minimum haversine distance; km), geodesic bearing, and azimuthal direction between these two locations. For classified survivors tagged with MiniPATs and the single recovered sPAT (Fish 27), which permitted access to the entire archival record, location estimates were reconstructed using the tag manufacturer’s proprietary geolocation processing estimator (GPE3). GPE3 incorporates tag-derived observations of twilight, sea surface temperature, and dive depth with corresponding reference data to generate two 12-h, 0.25° × 0.25° gridded probability surfaces per day using a discretized hidden Markov model (sensu [29]). Geolocation estimates were constrained by GPS-quality deployment locations and Argos-derived pop-up locations when available. Geolocation estimates for two individuals (Fish 7 and 25) without acceptable pop-up locations (see Results) were constrained by the estimated tag release date determined using transmitted depth data. Published records of adult cobia swim speed were unavailable; therefore, the diffusion parameter (i.e., the standard deviation of a normal distribution of diffusion rate for each fish that governs the allowable distance moved per day) was applied iteratively to the longest track (Fish 11) by estimated total distance traveled (5256.7 km) with an acceptable Argos-derived pop-up location (location class 3 [error < 250 m] within 16 h of tag detachment) and examined at increments of 0.25 m s^–1 from 0.5 to 3.0 m s^–1. GPE3 outputs were compared using the provided quality score as an indicator of goodness of fit, with 2.25 m s^–1 producing the best quality score and realistic track, aligning with the range of optimal (2 body lengths [BL] s^–1) to critical swim speeds (3 BL s^–1) for captive juvenile cobia [30]. This diffusion parameter was then applied to the remaining cobia tracks for consistency. Maximum likelihood locations and tracks were calculated using a forward-backward algorithm and represent points centered on an interpolated 0.025° × 0.025° grid smoothed with a cubic spline having the highest likelihood score for a given 12-h probability surface. Maximum likelihood location uncertainty was calculated using the 95% cumulative probability contour of each 12-h probability surface, resampled to a 0.025° × 0.025° grid using bilinear interpolation in the ‘raster’ package [31] in R [32]. Pop-up and maximum likelihood locations for cobia classified as survivors reporting between November and February were used to identify overwintering habitat.

Monthly residency distributions (RDs) were estimated by averaging the 12-h probability surfaces generated by GPE3 on a 0.25° × 0.25° grid. Grid cell values represented the percentage of time each fish spent in that location [29]. Overlapping monthly core use areas (50% RD) were calculated at a 25-km (0.25°) resolution across all tracks and months. To mitigate potential biases resulting from the inclusion of short tracking periods, monthly RDs were excluded for cobia tracked for < 10 d within a given month.

Vertical and Thermal Habitat Utilization

Daily minimum and maximum depth and temperature data transmitted from MiniPAT and sPAT tags were aggregated by month to characterize vertical and thermal habitat utilization patterns. In addition to transmitted data, four MiniPATs and one sPAT were physically recovered, providing the entire archived time series that was used in place of transmitted records. Daily minimum and maximum temperatures transmitted by mrPAT tags were also aggregated by month and included in analyses of thermal habitat. To minimize potential movement bias resulting from the capture and tagging event, all analyses excluded data from the first 24 h post-release and the PAT tag release date for individuals classified as survivors (> 10 days at liberty).

MiniPAT tags transmitted time-at-depth (TAD) and time-at-temperature (TAT) data summarized in 12- or 24-h histograms. When available, archival records were used instead of transmitted records and analyzed using the same TAD and TAT bin structure. TAD and TAT data were pooled into 24-h intervals, and only days with complete 24-h records were averaged by month and by individual fish. It was not possible to examine diel patterns from transmitted TAD and TAT data summarized into 12- or 24-h histograms, and transmitted time-series data (sampled at 450- or 600-s intervals) was sparse, comprising 0–45.6% of days with complete 24 h of data throughout the deployment periods. Therefore, archived time-series data from recovered tags sampled at higher resolutions (1 to 3 s sampling intervals) over the entire deployment period were used to investigate diel patterns in depth and temperature. For comparative analyses among recovered tags, archival time-series data were down-sampled to the longest sampling rate (3 s). To divide each 24-h period into day or night, sunrise and sunset times were derived from tag-recorded light levels using GPE3. TAD and TAT were calculated in 15-min increments over 24-h periods, aggregated by month, and combined across individuals to produce monthly averages. The thermal structure of the water column was visualized along the entire track of each archival record using linearly interpolated temperature-depth time series.

Data Recovery

From June 2020 to October 2024, 125 cobia were conventionally dart-tagged off the coast of Texas, USA, ranging in size from 432 mm (17.0 in) to 1302 mm FL (51.3 in), with a mean (± SD) of 820 ± 175 mm FL. Of these, 72% (90 of 125) were below the legal minimum size limit of 914 mm (36 in) FL. Eleven tagged cobia were recaptured and reported, yielding a recapture rate of 8.8% (11 of 125). Among the recaptured fish, five exceeded the minimum size limit and were harvested, while one (Fish 21), also above the limit, was voluntarily released by the angler upon noticing the mrPAT tag.

Thirty-four cobia were also PAT-tagged and ranged in size from 699 mm FL (27.5 in) to 1112 mm FL (43.8 in), with a mean of 938 ± 106 mm FL (Table 1). Due to the size requirements for PAT tag deployments, 65% (22 of 34) of these individuals exceeded the legal minimum size limit. Fight times for PAT-tagged individuals ranged from 0.5 to 6.0 min (mean = 2.8 ± 1.6 min), and handling times ranged from 2 to 8 min (mean = 3.8 ± 1.4 min). All PAT-tagged cobia exhibited shallow hooking locations (i.e., visibly hooked in the mouth or jaw).

Of the 34 PAT tags deployed (6 MiniPATs, 11 sPATs, and 17 mrPATs), 28 successfully transmitted data, while six mrPATs failed to report (DNR; Table 1). Five PAT tags (4 MiniPATs and 1 sPAT) were physically recovered, providing the entire archived time series. Among the 28 tags that transmitted data, 15 released on schedule, whereas 13 detached prematurely due to predation (Fish 12), angler harvest (Fish 15), or tag attachment failure (TAF; n = 11). PAT tags experiencing TAF reported specific release reasons, including “Floater” or “Pin Broke,” leading to early transmissions and reduced deployment durations (mean = 41% ± 32% of the scheduled deployment period). Floater indicated the tag and tether assembly detached from the fish and surfaced, triggering the tag’s pre-configured conditional release parameters. Pin Broke indicated that the tag detected a broken release pin and surfaced, likely due to external forces exerted on the tether (e.g., entanglement with underwater structures or human interference). Two TAFs occurred within the first 10 d after tagging, including one mrPAT (Fish 21) that detached on the same day the fish was recaptured and released by an angler after one day at liberty. Although the angler released the fish with the mrPAT tag still attached, the tag transmitted later that day, suggesting that the recapture event may have triggered its premature detachment, either as a potential PRM or TAF. Another cobia (Fish 24) tagged with a mrPAT on 21 October 2021 was recaptured on 8 October 2022, also nearly one year after being tagged. The mrPAT completed its 103-d deployment, releasing on schedule on 1 February 2022, and its tether remained attached when the fish was recaptured in October 2022.

Table 1
Summary of pop-up archival transmitting (PAT) tag deployments on cobia in the northwestern Gulf of Mexico.
Fish ID	FL (mm)	Tag Model	Tagging Location	Tagging Date	Pop-up Date	Pop-up Location	DAL	Release Reason
1	945	sPAT	27.64°N, 97.01°W	17 Jun 2020	10 Aug 2020	27.73°N, 96.89°W	54	TAF – Pin Broke
2^a	1035	MiniPAT	27.39°N, 97.08°W	04 Aug 2020	12 Oct 2020	27.67°N, 96.92°W	69	TAF – Floater
3	895	mrPAT	27.75°N, 96.99°W	06 Aug 2020	15 Sep 2020	27.52°N, 97.23°W	40	TAF – Floater
4	1092	mrPAT	27.92°N, 96.65°W	07 Sep 2020				DNR
5	838	mrPAT	27.79°N, 96.96°W	16 Sep 2020				DNR
6	865	mrPAT	27.66°N, 97.17°W	25 Sep 2020	15 Oct 2020	26.23°N, 97.17°W	20	TAF – Floater
7^b	898	MiniPAT	27.79°N, 96.96°W	01 Oct 2020	30 Mar 2021		180	Scheduled
8	916	mrPAT	28.07°N, 96.77°W	01 Oct 2020	27 Nov 2020	28.14°N, 96.59°W	57	Scheduled
9	925	mrPAT	27.83°N, 97.02°W	01 Oct 2020	18 Dec 2020	26.64°N, 96.82°W	78	Scheduled
10	923	mrPAT	27.83°N, 97.02°W	01 Oct 2020	04 Dec 2020	27.78°N, 96.53°W	64	Scheduled
11^a	925	MiniPAT	27.65°N, 97.13°W	13 Oct 2020	11 Apr 2021	27.60°N, 97.19°W	180	Scheduled
12^a	990	MiniPAT	27.64°N, 97.13°W	13 Oct 2020	20 Oct 2020	27.64°N, 97.13°W	1	M – Predation
13^a	1067	MiniPAT	27.62°N, 97.13°W	13 Oct 2020	20 Nov 2020	26.89°N, 97.35°W	38	TAF – Pin Broke
14	995	sPAT	27.75°N, 96.99°W	14 Oct 2020	13 Dec 2020	26.58°N, 97.17°W	60	Scheduled
15	1090	mrPAT	27.71°N, 97.03°W	14 Oct 2020	19 Oct 2020	27.60°N, 97.08°W	5	M – Harvested
16	867	mrPAT	27.79°N, 96.96°W	14 Oct 2020	12 Feb 2021	25.98°N, 96.71°W	121	Scheduled
17	865	mrPAT	27.82°N, 97.01°W	14 Oct 2020				DNR
18	860	mrPAT	27.64°N, 97.15°W	04 Oct 2021				DNR
19	1035	mrPAT	27.64°N, 97.15°W	04 Oct 2021				DNR
20	927	mrPAT	27.62°N, 97.13°W	04 Oct 2021				DNR
21	918	mrPAT	27.70°N, 97.09°W	07 Oct 2021	08 Oct 2021	27.91°N, 96.92°W	1	TAF – Floater
22^c	1086	mrPAT	27.70°N, 97.09°W	07 Oct 2021	15 Feb 2022		131	Scheduled
23	972	mrPAT	27.62°N, 97.13°W	21 Oct 2021	01 Feb 2022	26.46°N, 96.81°W	103	Scheduled
24	965	mrPAT	27.62°N, 97.13°W	21 Oct 2021	01 Feb 2022	26.17°N, 96.83°W	103	Scheduled
25^d	1054	MiniPAT	27.71°N, 97.09°W	21 Oct 2021	16 Nov 2021		26	TAF
26	967	sPAT	27.82°N, 96.85°W	06 Aug 2024	05 Oct 2024	28.08°N, 96.38°W	60	Scheduled

Table 1
continued.
Fish ID	FL (mm)	Tag Model	Tagging Location	Tagging Date	Pop-up Date	Pop-up Location	DAL	Release Reason
27^a	1029	sPAT	28.04°N, 96.34°W	06 Aug 2024	19 Sep 2024	26.43°N, 97.02°W	44	TAF – Floater
28	1003	sPAT	27.64°N, 97.14°W	16 Sep 2024	15 Nov 2024	27.35°N, 97.27°W	60	Scheduled
29	756	sPAT	28.18°N, 96.60°W	01 Oct 2024	23 Nov 2024	25.81°N, 97.15°W	53	TAF – Floater
30	699	sPAT	28.20°N, 96.55°W	01 Oct 2024	30 Nov 2024	28.30°N, 95.99°W	60	Scheduled
31	756	sPAT	28.08°N, 96.72°W	01 Oct 2024	26 Oct 2024	21.61°N, 97.36°W	26	TAF – Floater
32	902	sPAT	28.07°N, 96.76°W	01 Oct 2024	30 Nov 2024	26.58°N, 96.99°W	60	Scheduled
33	731	sPAT	28.07°N, 96.76°W	01 Oct 2024	06 Oct 2024	26.74°N, 97.34°W	5	TAF – Floater
34	1112	sPAT	28.07°N, 96.76°W	01 Oct 2024	30 Nov 2024	28.02°N, 96.40°W	60	Scheduled

FL, fork length; DAL, days at liberty; TAF, tag attachment failure; DNR, did not report; M, mortality

^aRecovered tag

^bThe first Argos location was reported on the beach five days after the tag released on schedule

^cNo reported Argos locations

^dTag transmitted on schedule following a 180-d deployment period but experienced tag attachment failure for unknown reasons after

26 days at liberty

Post-Release Mortality

Post-release fate was inferred using data from the 28 reporting PAT tags. Fish 15 was harvested 5 d after tagging, as indicated by Argos-derived locations tracking toward land and later confirmed by the angler. This individual was not classified as a PRM, as it had resumed feeding prior to capture (i.e., it actively took a baited hook or lure). Fish 12’s MiniPAT tag was ingested within 1.5 h of release, as inferred from recovered archival depth, temperature, and light-level time series. Given the short time interval between release and predation, this event was classified as a PRM. The remaining 26 tagged cobia exhibited horizontal movements, depth, temperature, or tilt data consistent with survival.

PRM estimates varied depending on the calculation method. Using the first estimation approach (Eq. 1), which considered only the single predated individual (Fish 12), the PRM rate was 3.6% (95% CI: 0.0–10.7%) among the 28 transmitting PAT tags. The second, more conservative approach (Eq. 2) incorporated two TAFs (Fish 21 and 33) that occurred within the first 10 d after tagging and six tags that DNR classified as mortalities. Under this method, nine mortalities (1 PRM + 2 TAF + 6 DNR) were assumed from the 34 total PAT tag deployments, yielding a PRM rate of 26.5% (95% CI: 11.8–41.2%).

Horizontal Movements

Dispersal from the tagging area was evaluated for 21 of the 28 reporting PAT tags using quality pop-up locations from deployments > 10 d (Table 2). One mrPAT (Fish 22) did not report locations, while two MiniPATs (Fish 7 and 25) reported locations 5 and 154 d after tag detachment, respectively, and were thus excluded from dispersal calculations. Notably, Fish 25’s MiniPAT tag transmitted on schedule following a 180-d deployment period over the outer shelf off Georgia and South Carolina. However, a review of the transmitted depth time series revealed that this tag experienced TAF and started floating over the mid-shelf off Packery Channel, Texas, after 26 d based on maximum likelihood locations. Despite sensor data indicating a TAF event, technical support from the tag manufacturer (Wildlife Computers) could not determine the cause of the delayed transmission.

PAT tags released between mid-August and mid-April, with deployment durations ranging from 20–180 d (mean = 73 ± 43 d; Table 2). Seventeen individuals reported pop-up locations with mean dispersal distances of 173.4 ± 172.7 km (range = 8.6–719.6 km) south-southwest (n = 14) or 46.9 ± 10.3 km (range = 35.8–56.2 km) east (offshore; n = 3) from their original tagging location between September and April. Four individuals exhibited northeastward movements, with mean dispersal distances of 30.7 ± 17.8 km (15.2–54.4 km) between August and November.

Three PAT tags (Fish 16, 29, and 31) reported pop-up locations in Mexican waters, including one mrPAT (Fish 31) releasing 719.6 km south of its tagging location off Cabo Rojo on the Tamaulipas-Veracruz shelf. One PAT tag (Fish 16) released only 2.7 km into Mexican waters, though it is likely that the fish was still in U.S. waters at the time of release, as the first successful Argos transmission occurred approximately 11 h later, exhibiting a strong southward drift trajectory based on subsequent Argos locations. Seven additional PAT tags (Fish 6, 9, 14, 23, 24, 27, and 32) reported 18.3–70.3 km north of the U.S.-Mexico maritime boundary (Fig. 1 & Figure S1). Except for three individuals (Fish 2, 27, and 31), all tagged cobia reported pop-up and maximum likelihood locations in continental shelf waters (< 200 m depth) between El Mezquital, Tamaulipas, Mexico, and Galveston, Texas, USA.

Table 2
Dispersal of cobia classified as survivors (> 10 days at liberty) in the northwestern Gulf of Mexico.
Fish ID	Tagging Date	Pop-up Date	DAL	Net Displacement (km)	Geodesic Bearing (°)	Azimuthal Direction
1	17 Jun 2020	10 Aug 2020	54	15.20	54.98	NE
2	04 Aug 2020	12 Oct 2020	69	34.30	29.82	NE
3	06 Aug 2020	15 Sep 2020	40	34.68	225.51	SW
6	25 Sep 2020	15 Oct 2020	20	157.42	180.09	S
8	01 Oct 2020	27 Nov 2020	57	18.92	67.84	NE
9	01 Oct 2020	18 Dec 2020	78	132.47	170.43	S
10	01 Oct 2020	04 Dec 2020	64	48.78	95.82	E
11	13 Oct 2020	11 Apr 2021	180	8.61	229.86	SW
13	13 Oct 2020	20 Nov 2020	38	84.19	196.54	SW
14	14 Oct 2020	13 Dec 2020	60	130.85	188.58	S
16	14 Oct 2020	12 Feb 2021	121	201.86	172.29	S
23	21 Oct 2021	01 Feb 2022	103	133.01	164.71	S
24	21 Oct 2021	01 Feb 2022	103	163.36	168	S
26	06 Aug 2024	05 Oct 2024	60	54.39	61.98	NE
27	06 Aug 2024	19 Sep 2024	44	191.62	202.93	SW
28	16 Sep 2024	15 Nov 2024	60	34.72	205.27	SW
29	01 Oct 2024	23 Nov 2024	53	268.09	193.01	SW
30	01 Oct 2024	30 Nov 2024	60	56.16	80.26	E
31	01 Oct 2024	26 Oct 2024	26	719.64	185.72	S
32	01 Oct 2024	30 Nov 2024	60	166.81	188.96	S
34	01 Oct 2024	30 Nov 2024	60	35.75	97.21	E

Dispersal is calculated as the net displacement (minimum haversine distance), geodesic bearing, and azimuthal direction between tagging and pop-up locations. DAL, days at liberty.

Unlike dispersal estimates based solely on tagging and pop-up locations, geolocated maximum likelihood tracks revealed more complex movement patterns, with total distances traveled ranging from 697.5–5256.7 km (mean = 2284.0 ± 1860.2 km; Fig. 2). The shortest net dispersal distance (8.6 km) was recorded for Fish 11, which also had the longest maximum likelihood track (5256.7 km). Tagged on 13 October 2020, this individual traveled 300.2 km northeast to offshore Galveston, Texas by 7 December 2020, then moved 141.3 km south toward Port Mansfield, Texas, by 18 January 2021, before returning 8.6 km southwest of its original tagging site by 11 April 2021, completing a 180-d scheduled deployment. Fish 2, tagged on 4 August 2020, displayed an extensive 1781.9-km track, initially moving south 217.4 km into Mexican waters by 1 September 2020, before returning north to shelf slope and off-shelf waters, ultimately reporting a pop-up location 34.4 km northeast of its tagging site on 12 October 2020. Fish 7, tagged off Port Aransas, Texas, on 10 October 2020, traveled 256.6 km north to offshore Galveston by 18 November 2020, overwintering primarily over the mid-shelf between Port Aransas and Galveston.

The remaining three maximum likelihood tracks were shorter deployments (26–44 d), covering distances between 697.5 and 1230.6 km. Fish 13, tagged on 13 October 2020, moved to the shelf edge by 22 October 2020 before returning to nearshore waters on 28 October 2020, eventually heading south with a pop-up location 84.2 km from its tagging site by 20 November 2020. Fish 25, tagged on 21 October 2021, initially moved 99.2 km east-southeast to the shelf edge by 24 October 2021 before returning to nearshore waters by 29 October 2021. It then traveled 161.4 km north of its tagging site, ultimately in mid-shelf waters on 15 November 2021. Finally, Fish 27 tagged on 6 August 2024 moved 115.7 km south-southeast to off-shelf waters by 10 August 2024, returned to mid-shelf waters, then exhibited a second off-shelf excursion on 31 August 2024 leaving from mid-shelf waters off Mexico before returning to U.S. mid-shelf waters by 12 September 2024.

Core Use Areas

Overlapping core use areas (50% RD) identified a high-use region along the central and South Texas Shelf, frequented by multiple individuals (n = 1–4) throughout all observed months (Fig. 3). This high-use region extended southward toward Mexico from August to September, overlapping shelf slope and off-shelf waters, before shifting north toward Galveston from October to December, then returning southward from January to February, representing overwintering habitat. The high-use region followed a north-south orientation mirroring the Texas Coastal Bend and the 200 m isobath along the shelf edge. The number of days and individuals observed each month reflects tag deployment durations, with most cobia tagged in the fall and individuals tracked for < 10 days within a given month excluded.

Several cobia demonstrated interannual site fidelity to the tagging region and specific structures. For example, Fish 11, tagged in October 2020, had a pop-up location just 8.6 km southwest of its original tagging location in April 2021. Additional examples of site fidelity were obtained from recaptured individuals. One cobia tagged on 21 October 2021 was recaptured on 8 October 2022 at the same standing oil and gas platform where it was initially tagged off Packery Channel. Another cobia, tagged on 17 June 2020 at the artificial reef structure M/V Kinta S, a 47.2-m steel cargo ship submerged at the Corpus Christi Nearshore Reef located off Packery Channel and Port Aransas, was recaptured at the same structure nearly one year later, on 14 June 2021.

Vertical and Thermal Habitat Utilization

Vertical and thermal habitat use was analyzed for 24 reporting PAT tags with deployments > 10 d. The mean deployment length for these tags, including four recovered archival records, was 73 ± 43 d (1747 total days). Incorporating 24-h archival data (mean deployment = 83 ± 66 d; 331 total days) with aggregated 24-h transmitted datasets resulted in 380 d of TAD and TAT histograms from 6 tags, 848 d of minimum and maximum depth data from 15 tags, and 1322 d of minimum and maximum temperature data from 24 tags. Differences in the number of tags and days across datasets were due to variations in PAT tag models and the success of transmission to the Argos satellite system following tag release.

Cobia occupied depths from the surface to a maximum of 121 m (Fig. 4). The median daily maximum depth remained relatively stable across most months, ranging between 27 m (IQR 25–28.3 m) and 34 m (IQR 30–37 m), with exceptions recorded in February (42 m; IQR 37–48 m) and April (17 m; IQR 16–26 m), both based on a single tagged cobia (Fish 11). The maximum and minimum recorded temperatures occupied by cobia were 34.5°C and 13.3°C, respectively (Fig. 4). The median daily maximum and minimum temperatures declined from June to October, ranging from 27.5°C (IQR 26.9–28.8°C) to 30.6°C (IQR 29.5–31.6°C) and 24.2°C (IQR 24.1–24.4°C) to 28.8°C (IQR 28.4–29.4°C), respectively. During the overwintering period from November to February, temperatures further declined, ranging from 18.1°C (IQR 17.8–18.5°C) to 24.8°C (IQR 23.5–26.1°C) for median daily maximum temperatures and from 16.9°C (IQR 15.4–17.1°C) to 23.5°C (IQR 22.1–25.2°C) for minimum temperatures.

Cobia exhibited a strong seasonal preference for surface waters, spending 36.4% and 45.2% of their time within the top 3 m of the water column during August and September, respectively, when temperatures were highest (87.0%–99.2% of time spent in 28–33°C waters; Fig. 5). As temperatures cooled, cobia progressively reduced their time near the surface, from 13.5% in November (84.3% at 22–24°C), decreasing further to 5.7%–8.1% from December (75.3% at 19–21°C) through March (54.3% at 16–18°C; Fig. 5).

Diel Behavior

Diel patterns in depth use became more apparent from September through March (Fig. 6). In September, cobia exhibited reverse diel vertical behavior, spending more time in surface waters (0–5 m) during the day and descending to depths of 5–15 m at night while remaining in consistently warm waters (≥29°C). From October through March, cobia transitioned to normal diel vertical behavior, with increased daytime depth use relative to nighttime. This diel pattern became more pronounced from December through March, with Fish 11 primarily occupying depths of 25–35 m during the day, extending to 45 m in February. At night, Fish 11 either moved into shallower waters or exhibited increased vertical activity across depths ranging from 5–30 m, again reaching 45 m in February. While this bimodal depth distribution was evident, thermal habitat use remained less distinct, as cobia generally occupied similar temperatures throughout the diel cycle (Figure S2). However, a gradual decline in temperature use was observed, from ~ 23°C in November to ~ 17°C February.

Transmitted time-series data from the two longest MiniPAT deployments (Fish 7 and 11), spanning October to March, were highly sparse, containing only 0 and 1 complete 24-h day(s) of data within their respective 180-d deployment periods. Therefore, diel depth patterns were analyzed using the recovered archival time series from Fish 11 in relation to seasonal water column thermal structure (Fig. 7). Normal diel vertical behavior was evident from October through March, but distinct seasonal behavioral differences emerged. In October, oscillatory swimming behavior was observed throughout both day and night in warm (≥23°C) nearshore to mid-shelf waters, with Fish 11 frequently moving between the seafloor and surface at night while remaining at deeper depths during the day (Fig. 7). Deeper habitat use increased as temperatures declined from November to March. By mid-February to early March, the water column had become largely well-mixed and cold, with the formation of inversion layers—where warmer temperatures occurred at depth compared to near the surface [33]—in mid-shelf waters offshore of Port Aransas. During this period, Fish 11 exhibited submergence behavior, occupying the relatively warmer waters at depth (Fig. 7). By March, Fish 11 exhibited more pronounced ‘U-shaped’ dive profiles in cooler (16–20°C) nearshore to mid-shelf waters, shifting between shallow waters (0–10 m) at night and deeper focal depths (20–40 m) during the day (Fig. 7).

This study provides novel information on the seasonal movements, habitat use, and post-release mortality of Gulf cobia. Notably, our results address high priority data gaps identified by the most recent stock identification workshops, stock assessments, and management profiles [6, 17, 34] by providing new information on overwintering habitat and stock distribution in the western Gulf. Despite a relatively limited sample size, dispersal estimates based on pop-up locations occurring between November and February suggest that most cobia moved south after October to overwinter on the South Texas Shelf north of the U.S.-Mexico maritime boundary. A smaller proportion of fish appeared to undertake nearshore-offshore migrations to mid-shelf depths adjacent to the tagging region, which has also been observed or suggested for Atlantic cobia off North Carolina and Virginia [10, 12, 13] and Gulf cobia off Mississippi ([35]; Eric Gigli, Mississippi Department of Marine Resources, personal communication). Cobia movement into Mexican waters was documented in this study; however, this appeared to be a less common occurrence rather than the norm as only two tagged fish were in Mexican waters when their PAT tags released. Moreover, maximum likelihood tracks spanning 180 days for two individuals (Fish 7 and 11) indicate that cobia have access to suitable habitat and can remain off Texas through the winter. Outside of the overwintering period, brief incursions into Mexican waters were observed for two cobia during the late summer. Though long-distance movements deep into Mexican waters can occur based on rare conventional tag recaptures and one pop-up location in this study, collectively, our findings do not support the notion that a large fraction of the Gulf cobia stock overwinters in the southern Gulf.

The observation of cobia exhibiting both resident and long-distance dispersal behaviors in this study suggests the presence of multiple contingents [36] and evidence of a partial migration strategy within the Gulf stock. While some long-distance dispersal into Mexico was observed, most fish displayed restricted movements, aligning with conventional tagging indications of a more resident contingent in the western Gulf [18]. Partial migration is likely the most common migration strategy in fish populations, being widely reported across numerous orders, habitat types, and spatial scales [37, 38, 39]. While partial migration may enhance stock resiliency in the face of spatially varying environmental conditions and stabilize population dynamics across time, overfishing of one contingent could lead to adverse effects on recruitment potential and stock recovery, especially if this overfishing is associated with loss of genetic diversity [40]. For Gulf cobia, this may be particularly applicable to migrations along the Florida Panhandle and the associated recreational fishing tournaments targeting this seasonal migration [16, 34]. While genetically distinct contingents of Atlantic cobia have been identified [10], studies with a similar level of genetic resolution are lacking for the Gulf stock. The dynamics of partial migration in Gulf cobia have significant implications for assessment and management; therefore, studies to determine whether genetically distinct contingents exist and whether migration patterns have changed over time are warranted.

Dispersal and maximum likelihood tracks estimated from PAT tag data indicated cobia did not move north of Galveston, Texas. While this finding apparently suggests low connectivity with the northern Gulf (e.g., Louisiana), several factors including the timing of tagging in this study (June through October) and tag retention may have influenced the extent of northward movement we observed. For example, most PAT tags were programmed to release during the winter months, given the interest in overwintering habitat identification. Thus, only two fish, with PAT tag releases in March and April, provided evidence of a return migration north in the spring, albeit to the tagging region. The timing of tagging may have also limited our inferences on northern Gulf connectivity as northward migrating fish in the spring may have already passed the tagging region by June when tagging commenced each year; however, supplemental tagging of 22 cobia between 2020 and 2021 with acoustic transmitters possessing a four-year battery life also failed to indicate movements north of the tagging region despite significant acoustic receiver infrastructure off Galveston, Texas (Streich, unpublished). Nevertheless, the potential for tag timing bias underscores the need for expanded cooperative tagging and genetic sampling with seasonal considerations to better understand whether observed movement patterns represent stock substructure (e.g., metapopulation dynamics).

Our study provides the first quantitative estimate of post-release mortality for recreationally caught Gulf cobia. In the most recent stock assessment (SEDAR 28 Update), a post-release mortality rate of 5% was assumed for the recreational and commercial hook-and-line fisheries [14]. This estimate was derived from limited commercial logbook data, suggesting that a high percentage of discarded cobia were released alive in bandit (98%) and longline (92%) fisheries and from South Carolina Department of Natural Resources observations that only 1 of 60 cobia collected for broodstock died within one week of collection and transport [6]. This study’s estimated PRM rate ranged from 3.6% to 26.5%, depending on how TAF and non-reporting PAT tags were considered. The PRM rate derived from only definitively inferred mortalities (3.6%; 95% CI: 0.0-10.7%) is similar to the 5% rate used in the most recent stock assessment and a recent estimate of 0% for PAT-tagged cobia in Virginia [13]. One limitation of our PRM estimates is that fish handling practices may not fully represent the Gulf recreational fishery. For example, because identifying seasonal movements and overwintering habitat was the primary goal of this study, fish were netted, tagged with gills submerged in a cooler of water, and fully revived on the side of the tagging vessel, potentially resulting in a PRM rate estimate that is biased low. In this case, our more conservative estimate may be more realistic and precautionary until such uncertainty around handling practices is reduced. While cobia may be hardy, accurate and precise estimates of post-release mortality are critical because small changes in fishing mortality can result in large changes in estimated spawning stock biomass and ultimately population growth and sustainability [20]. This is especially true considering the Gulf recreational fishery where, like many other recreational fisheries, cobia discards have and continue to exceed total landings [41]. For these reasons, we recommend a PRM study encompassing a wider regional extent to better capture variation in Gulf-wide fishing and handling practices and an increased sample size to ensure the most accurate estimate of PRM.

PAT-tagged cobia exhibited clear seasonal shifts in vertical and thermal habitat use, with individuals occupying shallower depths and warmer surface waters during summer months and shifting to deeper, cooler waters in winter, consistent with patterns reported for cobia in Virginia [13]. In this study, cobia spent 36.4% and 45.2% of their time in the top 3 m of the water column in August and September, respectively, corresponding to the warmest water temperatures available (31°-33°C). Surface water use decreased to 5.7%-8.1% from December through March. The preference for warmer waters in summer may reflect both enhanced foraging opportunities and the species’ previously observed association with productive coastal shelf habitats. During the overwintering period (i.e., November to February), cobia median daily maximum temperatures ranged from 18.1°C (IQR 17.8–18.5°C) to 24.8°C (IQR 23.5–26.1°C). Cobia also spent significantly more time at temperatures below 18°C during the coldest months (i.e., January and February) than cobia tagged off Virginia (25–50% in this study vs. less than 12% in [13]), which could be a function of limited sample size and spatiotemporal differences in oceanographic conditions.

Archival records of temperature and depth time series and proportional time at depth and temperature data permitted the first analysis of diel habitat use for Gulf cobia. These data revealed oscillatory, submergence, regular- and reverse diel vertical migration behaviors depending on the season. Submergence behavior was particularly pronounced during the overwintering period and may suggest the use of thermal refugia, a strategy documented in other pelagic predators such as tunas and swordfish, which may adjust depth selection to maintain favorable body temperatures and enhance metabolic performance (e.g., [42, 43]. Increased use of deeper depths during the winter months could also reduce metabolic costs by limiting exposure to fluctuations in surface water temperature that are amplified by extreme weather events like Winter Storm Uri (2021), which resulted in air temperatures at or below freezing for nearly a week in the study region. These seasonal differences in habitat use have implications for the Gulf recreational fishery, namely that cobia may be more available to surface-oriented recreational methods (e.g., hook-and-line, sight fishing) during the warmer months, but reduced surface water use during the winter months could lower encounter rates and bias abundance perceptions during the winter. Maximum likelihood movement tracks from this study demonstrate at least some fish display resident behaviors in the tagging region during the overwintering period; however, anecdotal evidence suggests cobia are not targeted during the winter months as sea state conditions worsen and many anglers shift their focus to other outdoor activities such as hunting or may not be aware that cobia are still present. These dynamics underscore the importance of accounting for seasonal effects of depth- and temperature-mediated catchability and subsequent shifts in vulnerability and fishing effort, as failure to consider these seasonal changes in habitat use could bias stock assessments.

This study provides the first comprehensive characterization of Gulf cobia movements, habitat use, and post-release mortality in the western Gulf. While limited conventional tag recaptures and PAT tag dispersal observations support the potential for extensive migrations to overwintering habitats in the southern Gulf, our findings do not support the notion that a large fraction of the Gulf cobia stock overwinters off Mexico. Instead, most fish displayed more restricted movement behavior, overwintering on the South Texas Shelf. This finding bodes well for U.S. fishery managers tasked with ending overfishing of the Gulf cobia stock and suggests the probability of significant removals of Gulf cobia in Mexico during the overwintering period is minimal. Accordingly, the Gulf stock should be more directly affected by regulatory actions designed to manage fishing mortality in U.S. waters. Our estimated PRM rate (3.6%) is close to the 5% rate used in cobia stock assessments, but it may be biased low, as preferential handling in this study may not represent true recreational fishing practices. Given the observed sensitivity of our estimate depending on the treatment of TAF and non-reporting tags, a higher estimate could be incorporated into the next Gulf cobia assessment as a precautionary measure. The seasonal variation in thermal and vertical habitat utilization exhibited by PAT-tagged cobia has direct effects on catchability and vulnerability to recreational fishing gear. Considering changing climate patterns may further alter intra- and interannual migratory behaviors and distributions [44, 45], continued monitoring to determine if current regulations are effective in conserving each stock contingent is key. Finally, observed movement patterns provide evidence for a partial migration strategy; however, how the fishery interacts with each contingent is uncertain at this time, highlighting the need for increased cooperative tagging and genetic sampling efforts across the Gulf during all seasons to better understand these dynamics and promote sustainable management of this popular recreational fishery.

PAT: pop-up archival transmitting

Gulf: Gulf of Mexico

SEDAR: Southeast Data, Assessment, and Review

ASMFC: Atlantic States Marine Fisheries Commission

SAMFC: South Atlantic Fishery Management Council

GMFMC: Gulf of Mexico Fishery Management Council

PRM: post-release mortality

DNR: did not report

TAF: tag attachment failure

BL: body lengths

RD: residency distribution

TAD: time-at-depth

TAT: time-at-temperature

DAL: days at liberty

IQR: interquartile range

Ethics approval and consent to participate

Fish capture and tagging followed procedures approved by the Texas A&M University-Corpus Christi Institutional Animal Care and Use Committee (Animal Use Protocol #2020-04-01 & #2023-0007) and the Texas Parks and Wildlife Department (Scientific Research Permit #SPR-0303-279)

Consent for publication

Not applicable

Availability of data and materials

The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests

The authors declare that they have no competing interests.

Funding

This work was funded by the Gulf Fishery Management Council (Council) under award NA15NMF4410011 from the National Oceanic and Atmospheric Administration (NOAA), U.S. Department of Commerce and by U.S. Fish and Wildlife Service grant TX-T-221-R-1, F20AF11348 awarded under the authority of the State Wildlife Grant Program (CFDA# 15.634) issued to the Texas Parks and Wildlife Department. This work was also funded by the NOAA Marine Fisheries Initiative (MARFIN) under award NA23NMF4330382 to Texas A&M University-Corpus Christi. The statements, findings, conclusions, and recommendations are those of the author(s) and do not necessarily reflect the views of the Council or U.S. Department of Commerce.

Author’s contributions

MS: Conceptualization, funding acquisition, data collection and curation, data analysis, writing – original draft, writing – review and editing. DC: Conceptualization, data curation, data analysis, writing – original draft, writing – review and editing. GS: Conceptualization, funding acquisition, writing – review and editing.

Acknowledgements

We thank the staff and students of the Center for Sportfish Science and Conservation for their help in the field. In particular, we thank Jeff Kaiser for offering his personal boat as a research platform for many of the tagging trips. We sincerely thank Captain Patrick Somerville for his contributions to satellite tagging efforts and the numerous private anglers who conventionally tagged cobia during their fishing trips. Staff of the Harte Research Institute provided significant logistical support to make this work possible.

Costa DP, Breed GA, Robinson PW. New insights into pelagic migrations: implications for ecology and conservation. Annu Rev Ecol Evol Syst. 2012;43:73-96.
Chin A, Simpfendorfer CA, White WT, Johnson GJ, McAuley RB, Heupel MR. Crossing lines: a multidisciplinary framework for assessing connectivity of hammerhead sharks across jurisdictional boundaries. Sci Rep. 2017;7:46061.
Rooker JR, Dance MA, Wells RJD, Ajemian MJ, Block BA, Castelton MR, Drymon JM, Falterman BJ, Franks JS, Hoffmaer ER, Kraus RT, McKinney JA, Secor DH, Stunz GW, Walter JF. Population connectivity of pelagic megafauna in the Cuba-Mexico-United States triangle. Sci Rep. 2019;9:1663.
Grimes CB, Johnson AG, Fable WA Jr. Delineation of king mackerel (Scomberomorus cavalla) stocks along the U.S. east coast and in the Gulf of Mexico. In: Kumpf HE, Vaught RN, Grimes CB, Johnson AG, Nakamura EL, editors. Proceedings of the Stock Identification Workshop; 1985 Nov 5–7; Panama City Beach, FL. NOAA Tech Memo NMFS-SEFC-199. 1987.
Goethel DR, Berger AM, Cadrin SX. Spatial awareness: good practices and pragmatic recommendations for developing spatially structured stock assessments. Fish Res. 2023;264:106703.
SEDAR (Southeast Data, Assessment, and Review). SEDAR 28 – Gulf of Mexico cobia stock assessment report. North Charleston (SC): SEDAR; 2013. 616 p. Available from: http://www.sefsc.noaa.gov/sedar/Sedar_Workshops.jsp?WorkshopNum=28.
SEDAR (Southeast Data, Assessment, and Review). SEDAR 58 – Atlantic cobia stock assessment report. North Charleston (SC): SEDAR; 2020. 500 p. Available from: http://sedarweb.org/sedar-58.
SAFMC (South Atlantic Fishery Management Council). Amendment 31 to the fishery management plan for coastal migratory pelagic resources in the Gulf of Mexico and Atlantic region – Atlantic migratory group cobia management. North Charleston (SC): SAFMC; 2018.
GMFMC (Gulf of Mexico Fishery Management Council). Framework amendment 7 to the fishery management plan for coastal migratory pelagic resources in the Gulf of Mexico and Atlantic region – modifications to Gulf of Mexico migratory group cobia size and possession limits. Tampa (FL): GMFMC; 2019.
Perkinson M, Darden T, Jamison M, Walker MJ, Denson MR, Franks J, Hendon R, Musick S, Orbesen ES. Evaluation of the stock structure of cobia (Rachycentron canadum) in the southeastern United States by using dart-tag and genetics data. Fish Bull. 2019;117(3):220-33.
Darden TL, Walker MJ, Brenkert K, Yost JR, Denson MR. Population genetics of cobia (Rachycentron canadum): implications for fishery management along the coast of the southeastern United States. Fish Bull. 2014;112:24-35.
Gallagher R. Using acoustic telemetry and population genetics to investigate cobia stock structure in the southeast U.S. [master’s thesis]. Raleigh (NC): North Carolina State University; 2020.
Jensen DR, Graves JE. Movements, habitat utilization, and post-release survival of cobia (Rachycentron canadum) that summer in Virginia waters assessed using pop-up satellite archival tags. Anim Biotelemetry. 2020;8:1-3.
SEDAR (Southeast Data, Assessment, and Review). SEDAR 28 – Gulf of Mexico cobia update assessment report. North Charleston (SC): SEDAR; 2020. 147 p. Available from: http://www.sefsc.noaa.gov/sedar/Sedar_Workshops.jsp?WorkshopNum=28.
GMFMC (Gulf of Mexico Fishery Management Council). Final amendment 32 to the fishery management plan for coastal migratory pelagic resources in the Gulf of Mexico and Atlantic region – modifications to the Gulf of Mexico migratory group cobia catch limits, possession limits, size limits, and framework procedure. Tampa (FL): GMFMC; 2022.
GMFMC (Gulf of Mexico Fishery Management Council). Something’s fishy with cobia response summary. 2019-S28Update-WP-03. North Charleston (SC): SEDAR; 2020.
SEDAR. SEDAR 58 – Cobia stock ID process report compilation. North Charleston (SC): SEDAR; 2018. 116 p. Available from: http://sedarweb.org/sedar-58-stock-id-process.
Dippold DA, Leaf RT, Franks JS, Hendon JR. Growth, mortality, and movement of cobia (Rachycentron canadum). Fish Bull. 2017;115:460-72.
Muoneke MI, Childress WM. Hooking mortality: a review for recreational fisheries. Rev Fish Sci. 1994;2:123-56.
Musyl MK, Moyes CD, Brill RW, Mourato BL, West A, McNaughton LM, Chiang WC, Sun CL. Postrelease mortality in istiophorid billfish. Can J Fish Aquat Sci. 2015;72:538-56.
Burns KM, Froeschke JT. Survival of red grouper (Epinephelus morio) and red snapper (Lutjanus campechanus) caught on J-hooks and circle hooks in the Florida recreational and recreational-for-hire fisheries. Bull Mar Sci. 2012;88(3):633-46.
Curtis JM, Johnson MW, Diamond SL, Stunz GW. Quantifying delayed mortality from barotrauma impairment in discarded red snapper using acoustic telemetry. Mar Coast Fish. 2015;7:434-49.
François P, Sidonie C, Caroline C, Jean-Marc G. The effect of hook type and trailing gear on hook shedding and fate of pelagic stingray (Pteroplatytrygon violacea): new insights to develop effective mitigation approaches. Mar Policy. 2019;107:103594.
Schlenker LS, Latour RJ, Brill RW, Graves JE. Physiological stress and post-release mortality of white marlin (Kajikia albida) caught in the United States recreational fishery. Conserv Physiol. 2016;4:cov066.
Orbesen ES, Brown CA, Snodgrass D, Serafy JE, Walter JF 3rd. At-vessel and postrelease mortality rates of bluefin tuna (Thunnus thynnus) associated with pelagic longline gear in the northern Gulf of Mexico. Fish Bull. 2019;117:15-23.
Musyl MK, Domeier ML, Nasby-Lucas N, Brill RW, McNaughton LM, Swimmer JY, Lutcavage MS, Wilson SG, Galuardi B, Liddle JB. Performance of pop-up satellite archival tags. Mar Ecol Prog Ser. 2011;433:1-28.
Graves JE, Luckhurst BE, Prince ED. An evaluation of pop-up satellite tags for estimating postrelease survival of blue marlin (Makaira nigricans) from a recreational fishery. Fish Bull. 2002;100:134-42.
Goodyear CP. Factors affecting robust estimates of the catch-and-release mortality using pop-off tag technology. Am Fish Soc Symp. 2002;30:172-9.
Pedersen MW, Patterson TA, Thygesen UH, Madsen H. Estimating animal behavior and residency from movement data. Oikos. 2011;120(9):1281-90.
Nelson D, Stieglitz JD, Cox GK, Heuer RM, Benetti DD, Grosell M, Crossley DA 2nd. Cardio-respiratory function during exercise in the cobia, Rachycentron canadum: the impact of crude oil exposure. Comp Biochem Physiol C Toxicol Pharmacol. 2017;201:58-65.
Hijmans RJ, van Etten J, Sumner M, Cheng J, Baston D, Bevan A, et al. raster: geographic data analysis and modeling. R package version 3.6-31. Vienna (Austria): R Foundation for Statistical Computing; 2025. Available from: https://CRAN.R-project.org/package=raster.
R Core Team. R: a language and environment for statistical computing. Vienna (Austria): R Foundation for Statistical Computing; 2025. Available from: https://www.R-project.org/.
Weatherly G, Wienders N, Harkema R. Temperature inversions in the open Gulf of Mexico. J Geophys Res Oceans. 2003;108(C6).
GSMFC (Gulf States Marine Fisheries Commission). Management profile for Gulf of Mexico Cobia. Ocean Springs (MS): GSMFC; 2019. Publication no. 287. Available from: https://www.gsmfc.org/publications/GSMFC%20Number%20287.pdf.
Hendon JR, Franks JS. Sport fish tag and release in Mississippi coastal waters and the adjacent Gulf of Mexico, Project F-132, segments 7–9. Final report. Ocean Springs (MS): Mississippi Department of Marine Resources; 2010.
Secor DH. Specifying divergent migrations in the concept of stock: the contingent hypothesis. Fish Res. 1999;43(1–3):13–34.
Robichaud D, Rose GA. Migratory behaviour and range in Atlantic cod: inference from a century of tagging. Fish Fish. 2004;5(3):185-214.
Secor DH, Piccoli PM. Determination of frequency of anadromous migrations by Chesapeake Bay striped bass based upon otolith microchemical analysis. Fish Bull. 2007;105:62-73.
Chapman BB, Skov C, Hulthén K, Brodersen J, Nilsson PA, Hansson L-A, Brönmark C. Partial migration in fishes: definitions, methodologies and taxonomic distribution. J Fish Biol. 2012;81(2):479-99.
Ryman N, Utter F, Laikre L. Protection of intraspecific biodiversity of exploited fishes. Rev Fish Biol Fish. 1995;5:417-46.
Matter VM, Nuttall MA. Recreational survey data for cobia in the Gulf of Mexico. 2019-S28Update-WP-02. North Charleston (SC): SEDAR; 2020.
Brill RW, Lutcavage ME. Morphophysiological adaptations for pelagic life. In: Block BA, Stevens ED, editors. Tuna: Physiology, Ecology, and Evolution. San Diego: Academic Press; 2001. p. 31-62.
Hoolihan JP, Luo J, Abascal FJ, Campana SE, De Metrio G, Dewar H, et al. Vertical and horizontal movements of the broadbill swordfish (Xiphias gladius) in the Atlantic Ocean. Fish Oceanogr. 2011;20(6):479-92.
Crear DP, Watkins BE, Friedrichs MAM, St-Laurent P, Weng KC. Estimating shifts in phenology and habitat use of cobia in Chesapeake Bay under climate change. Front Mar Sci. 2020a;7:579135.
Crear DP, Watkins BE, Saba VS, Graves JE, Jensen DR, Hobday AJ, Weng KC. Contemporary and future distributions of cobia, Rachycentron canadum. Divers Distrib. 2020b;26(8):1002-15.

No competing interests reported.

FigureS1.png
Figure S1. Geographic dispersal (black arrows) of twenty cobia classified as survivors (>10 days at liberty) along the Texas shelf in the northwestern Gulf of Mexico (inset). Deployment (squares) and release locations (triangles) of PAT tags are colored by month. The 200 m isobath (gray lines) represents the shelf edge contour; solid black lines indicate the exclusive economic zone, and the dotted line marks the U.S. federal-state seaward boundary.
FigureS2.pdf
Figure S2. Mean proportion of time spent at temperature by time of day for all recovered archival records (n = 4), aggregated by month. Plots are colored by the proportion of time, with values greater than 0.5 truncated for clearer illustration. The total number of cobia (n) and histogram data days (d) are provided for each month. Individuals with <10 days of histogram data within a given month were excluded.

Download PDF

Reviewers agreed at journal
22 Dec, 2025
Reviewers agreed at journal
22 Dec, 2025
Reviewers invited by journal
05 Nov, 2025
Editor assigned by journal
31 Oct, 2025
Submission checks completed at journal
29 Oct, 2025
First submitted to journal
27 Oct, 2025

You are reading this latest preprint version

Seasonal movements and habitat use of Cobia (Rachycentron canadum) in the Gulf of Mexico

Status:

Version 1

Abstract

Background

Results

Conclusions

Figures

Background

Methods

Fish Tagging

Post-release Mortality

Geolocation Estimation

Vertical and Thermal Habitat Utilization

Results

Data Recovery

Post-Release Mortality

Horizontal Movements

Core Use Areas

Vertical and Thermal Habitat Utilization

Diel Behavior

Discussion

Conclusions

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1