Integrating conventional tagging and acoustic telemetry improves estimates of post-release survival in a highly targeted reef fish

By integrating large-scale conventional tagging data with direct acoustic telemetry observations within a statistical framework, this study demonstrates that post-release survival of gag reef fish declines with capture depth due to barotrauma, providing refined annual and seasonal survival estimates to improve fisheries management.

The Gist

Imagine you are running a massive fishing tournament. Every time a fisherman catches a fish, they have to decide: keep it for dinner, or let it go back into the ocean. The big question for scientists and managers is: When we let a fish go, does it actually survive?

This paper is about solving a mystery: How many "Gag" fish (a type of grouper) survive after being caught and released, and why do some die while others live?

To solve this, the researchers had to combine two different detective tools, because neither tool worked perfectly on its own. Here is the story of how they did it, explained simply.

The Two Detective Tools

1. The "Wanted Poster" Method (Conventional Tagging)
Imagine you catch a fish, put a sticker with a phone number on it, and throw it back. You tell the fisherman, "If you catch this fish again, call us and we'll give you a free T-shirt!"

• The Good: You can tag thousands of fish over a huge area. It's like having a massive army of spies.
• The Bad: You don't know what happens to the fish that don't get caught again. Did they die? Did they just swim away and never get caught? Or did someone catch them, eat them, and throw the sticker in the trash without calling? It's like trying to guess the weather by only looking at the days it rains; you miss all the sunny days.

2. The "Smart Watch" Method (Acoustic Telemetry)
Now, imagine you give a fish a tiny, high-tech smart watch that talks to underwater microphones. You can see exactly where the fish goes every second.

• The Good: You know for a fact if the fish is alive or dead. If the watch stops moving, you know the fish died. It's like having a live video feed of the fish's life.
• The Bad: These watches are expensive and heavy. You can only tag a few dozen fish, and they usually stay in one small neighborhood. It's like having a crystal-clear video of one house, but no idea what's happening in the rest of the city.

The Big Idea: Mixing the Tools

The researchers realized they needed to mix these two methods. They took the massive data from the "Wanted Posters" (thousands of fish) and used the crystal-clear truth from the "Smart Watches" (dozens of fish) to calibrate the whole system.

Think of it like this:

• The Smart Watches told them the true survival rate for fish caught in specific spots.
• They used that truth to "teach" the computer model how to interpret the Wanted Posters.
• Suddenly, the model could look at the thousands of "Wanted Posters" and say, "Ah, I know that fish didn't get caught again because it died, not because it got lucky."

The Big Discovery: The "Elevator" Effect

The study found a clear pattern, and it has to do with depth.

Imagine the ocean is a giant elevator.

• Shallow Water (The Lobby): When fish are caught in shallow water (like the lobby of a building), they are fine. They swim away, and 97% of them survive. They are happy campers.
• Deep Water (The Penthouse): When fish are caught deep down (like the penthouse), they suffer from "barotrauma." This is like the "bends" divers get. Their swim bladders (a balloon inside them that helps them float) expand like a balloon popped out of a soda can. It hurts, they can't swim right, and they float to the surface, making them easy targets for sharks or birds.
• The Result: The deeper the fish is caught, the less likely it is to survive. At the deepest point they studied, only about 32% of the fish survived.

Why Does This Matter?

This isn't just about fish; it's about rules and regulations.

1. Seasonal Patterns: The researchers found that in the summer, fishermen go deeper to catch fish. This means more fish die after being released in the summer. In the winter, they fish in shallower water, so more fish survive.
2. Better Management: Because they now know exactly how depth affects survival, managers can make smarter rules. For example, they might close the fishing season during the summer months when the fish are most vulnerable, or encourage fishermen to use special "descender devices" (like a weighted line) to pull deep-water fish back down slowly so their swim bladders don't explode.

The Takeaway

By combining the "big picture" data with the "close-up" truth, the scientists built a super-accurate map of fish survival. They proved that while most fish survive shallow catches, deep-water fishing is a deadly game of chance for them.

This new method is a game-changer. It allows scientists to predict exactly how many fish will survive in any given month or year, helping to keep the fish populations healthy for everyone to enjoy in the future. It's like upgrading from guessing the weather to having a perfect, 100% accurate forecast.

Technical Summary

1. Problem Statement

Accurate estimation of post-release survival is critical for fisheries stock assessments and management, particularly for recreational fisheries where regulatory discards often exceed retained harvest. Two primary methods exist for estimating this survival:

• Conventional Tag-Return Studies: Offer large sample sizes and broad environmental coverage but suffer from unknown reporting rates. Consequently, they can only estimate relative survival (recapture ratios) rather than absolute survival.
• Acoustic Telemetry: Provides direct observation of individual fate (absolute survival) with high temporal resolution but is limited by small sample sizes, high costs, and restricted spatial extent.

The Gap: No existing framework has formally integrated these two data types within a unified statistical model to estimate absolute post-release survival across broad environmental gradients while leveraging the strengths of both methods.

2. Methodology

The study utilized Gag (Mycteroperca microlepis) as a case study, integrating two distinct datasets collected off Tampa Bay, Florida:

1. Conventional Tagging Data: 3,363 gag released between 2009–2012 (Observer Program), with 356 recaptures.
2. Acoustic Telemetry Data: 58 gag released between 2013–2017, equipped with acoustic transmitters and conventional tags, monitored by 30 receiver stations.

Statistical Framework

The authors developed a discrete-time Bayesian statistical model based on a modified Cormack-Jolly-Seber (CJS) framework. The model partitions the probability of recapture ($R$) into four components:

1. $\psi$ (Immediate Post-Release Survival): The probability a fish survives the release event.
2. $\phi$ (Remaining at Risk): The probability the fish survives natural mortality and retains its tag until the next time step.
3. $\eta$ (Encounter Probability): The probability an angler encounters the fish.
4. $\xi$ (Reporting Probability): The probability an angler reports the tag if encountered.

Key Innovation:

• Shared Parameters: The model links the two datasets by assuming the post-release survival ($\psi$) is driven by the same biological mechanism (capture depth) for both tagged groups.
• Anchoring: The acoustic telemetry data provides direct observations of fate ($y_j \sim \text{Bernoulli}(\psi^*_j)$), acting as an "anchor" to estimate absolute survival. This allows the model to separate the unknown reporting rate ($\xi$) from the survival rate in the conventional tagging data.
• Predictors: Capture depth was identified as the primary predictor for survival via logistic regression ($\text{logit}(\psi) = \beta_0 + \beta_1 \times \text{Depth}$). Other variables (length, hook trauma, temperature) were excluded via exploratory analysis as they did not significantly explain variation.
• Implementation: The model was implemented in Stan (via R) using Markov Chain Monte Carlo (MCMC) with 4 chains, 1,000 warm-up, and 1,000 sampling iterations.

3. Key Contributions

• Methodological Integration: First application of a unified Bayesian framework to simultaneously estimate absolute post-release survival by combining large-scale relative risk data (conventional tags) with direct fate data (acoustic telemetry).
• Scalability: The approach allows for the extrapolation of absolute survival estimates across broad environmental gradients (specifically depth) that exceed the spatial limits of telemetry studies alone.
• Management Relevance: The model produces monthly and annual survival probabilities tailored to specific fishing conditions, moving beyond static assumptions used in previous stock assessments.

4. Key Results

• Depth-Dependent Survival: There is a strong negative correlation between capture depth and post-release survival, consistent with barotrauma effects.
  • Shallow Waters (<25m): Survival was extremely high (approx. 97%). No mortality was observed in acoustic telemetry releases shallower than 25m.
  • Deep Waters: Survival declined significantly with depth. At the maximum observed capture depth (93m), predicted survival dropped to 32% (80% CI: 0.12–0.64).
• Model Parameters:
  • Reporting Rate ($\xi$): Estimated at 14% (80% CI: 0.13–0.15).
  • Natural Mortality ($\phi$): Estimated at 98% monthly survival probability.
• Temporal Patterns: Applying the model to recreational fishery data (2023–2025) revealed seasonal patterns:
  • Fall/Winter: Shallow fishing depths (10–20m) corresponded to high survival (~93%).
  • Summer: Fishing moved to deeper waters (peaking at 38m in July), resulting in lower survival (~83%).
• Annual Estimates: Annual post-release survival estimates for 2023, 2024, and 2025 were 86%, 92%, and 87%, respectively. These align closely with the 88% rate assumed in the most recent Gulf of Mexico Gag stock assessment (SEDAR72).

5. Significance and Implications

• Validation of Stock Assessments: The study provides independent, data-driven validation for the survival rates currently used in stock assessments, increasing confidence in management decisions.
• Targeted Management: The identification of seasonal survival dips (summer) due to deeper fishing offers a scientific basis for temporal management strategies. For instance, the study supports recent regulatory shifts (moving the season start from June to September) that likely reduced dead discards by avoiding peak deep-water fishing periods.
• Mitigation Strategies: The results highlight the efficacy of descending devices. Since survival drops sharply with depth, management efforts should focus on education regarding venting and recompression devices specifically during deep-water fishing seasons.
• Transferability: This integrated framework is applicable to other highly targeted reef fisheries where regulatory discards are high, offering a robust path to refine mortality estimates without requiring prohibitively expensive telemetry studies for every scenario.

Limitations

The authors note two primary limitations:

1. Assumption of Removal: The model assumes unreported fish are removed from the risk pool (harvested or lost). If unreported fish are re-released with intact tags, mortality risk may compound, potentially leading to an underestimation of overall survival.
2. Representativeness: The study relies on Observer Program data (for-hire fleet) as a proxy for the broader private recreational fleet. Differences in fishing depth or practices between these fleets could introduce bias.