Evaluating uncertainty to improve a common monitoring method and guide management decisions for diamond-backed terrapins

Designing monitoring with adequate certainty to evaluate management actions can be challenging, especially for elusive species in relatively inaccessible habitats. The diamond-backed terrapin (Malaclemys terrapin) is considered a high priority species for management in all states within its range. Among key threats to terrapin populations is bycatch mortality in crab pots, prompting states throughout the species' range to consider crabbing regulations and creating a need for a monitoring approach able to evaluate the success of any new crabbing regulations to benefit terrapins. Because terrapins occupy extensive and often difficult to access estuarine habitat, it is hard to collect sufficient data to estimate population responses accurately and precisely for management. To assist state managers with decisions regarding monitoring bycatch reduction regulations, we modeled 12 years of terrapin capture-recapture data. We used estimates from those models to simulate capture-recapture data pre- and post-implementation of bycatch reduction that resulted in a 0.20 absolute increase in mean apparent survival probability. Results indicated weak reliability of a monitoring approach using only manual seining of tidal creeks to detect a real management effect, with a positive management effect detected at the 95% certainty level only 34% of the time. When we considered 85% and 75% certainty thresholds, we detected a positive effect on survival among 61% and 75% of simulations, respectively. Low within-year recapture probability and the effect of tide amplitude on terrapin availability indicate there is low feasibility of improving monitoring precision at a single site, requiring monitoring of more sites to improve confidence in the detection of the management effect. The number of sites monitored depends on the acceptable level of certainty. We recommend that researchers and management entities that use seining assess the level of certainty they require to evaluate management actions and increase the number of sites sampled to meet that level of certainty. Additionally, the use of multiple monitoring methods and integrated models should be explored to reduce uncertainty and to allow for easier monitoring of more populations over broader spatial scales.