Reducing Uncertainties in Conservation Decision-Making for American Alligators

Effective conservation decision-making necessitates monitoring programs that are designed to collect unbiased and precise measurements of relevant attributes deemed to reduce structural uncertainty of the managed resource state. American alligators (Alligator mississippiensis; hereafter alligator) are a keystone species within the southeastern United States that have cascading effects on ecosystem structure and function, and are managed under consumptive use management programs throughout their range. Management of alligator populations in South Carolina is challenging due to pervasive uncertainties regarding the size class distribution, which is only partially observable using the primary monitoring tool (nightlight surveys), a lack of demographic parameter estimates, and identification of measurable attributes that could pose conservation threats (e.g., drought, contaminants). My objective was to develop analytical tools to reduce partial observability in alligator monitoring and identify potential drivers of alligator population dynamics to reduce structural uncertainty. I developed a Bayesian integrated population model (IPM) that produced among the first demographic parameter estimates for alligators in South Carolina and determined that survival probabilities increased greatly among immature size classes, but are relatively similar among adults (\u3e0.90); a pattern that has been previously reported for American crocodiles (Crocodylus acutus). The IPM produced size-class specific abundance estimates for alligators from count data with prolific state uncertainty (\u3e60% unknown size observations). In general, alligator abundance trends were uncertain and appeared to vary spatially, though the mean population growth (λ) estimates for all sites, IPM versions, and the Lefkovtich matrix were \u3c1, indicating a population decline. However, the 95% Bayesian credible intervals for λ at one survey site included 1, indicating some uncertainty. I then used the demographic parameter estimates to simulate virtual alligator populations under varying gradations of initial population density, harvest rate to determine an optimal level of spatiotemporal replication for a monitoring programs. To evaluate the need to obtain size class-specific abundance estimates, the simulated count data from the underlying virtual population was total individuals (of all size classes). Based on fundamental objectives to maximize financial effectiveness and minimize management and ecological uncertainty, all of the harvest and density scenarios (except low density and maximum harvest) selected a monitoring program with six temporal replicates (the maximum) and 320 spatial replicates (1 spatial replicate = 0.5 km river segment). In general, data reliability (precision and accuracy) was more sensitive to increasing temporal, compared to spatial, replication, which has been previously reported in other simulation based studies in which detection probabilities are low (p\u3c 0.10). Moreover, all scenarios and monitoring programs induced changes in alligator size class structure, though the effects were minimized with reduced harvest rate, increase survey effort and population density. In synthesis, the demographic parameter estimates produced by the IPM can and are being used to improve monitoring methodology for alligators in South Carolina, and provide a mechanism to increase the demographic resolution of monitoring data, inform optimal monitoring decisions, and explore further uncertainties associate with harvest decisions. Finally, to better elucidate potential drivers of alligator population status, I evaluated total mercury (THg) concentrations in adult alligator whole blood from a longitudinal mark-recapture study. I determined that THg in whole blood was best described by an interactive effect of sex and predicted age, as calculated by predicted age at first capture using a recently developed growth model for alligators in South Carolina. THg concentrations averaged 0.16 ± 0.05 mg kg-1 ww and were slightly higher in males than female, though the overall average is significantly lower than other estimates reported in the Florida Everglades and the Savannah River Site in South Carolina. The quadratic effect of THg with predicted age, in which older individuals had lower levels than younger individuals is novel, and contrasts with previous assumptions that THg bioaccumulates with age (i.e., does not decrease). We posit that determinate (asymptotic) growth, which could accompany age-related changes in foraging patters and metabolism, could potentially explain the lower THg we detected in the oldest individuals. The results from our study could highlight the need for long-term longitudinal monitoring of sentinel species to further evaluate our hypotheses