IMPERFECT DETECTION SUBSTANTIALLY UNDERESTIMATES PATHOGEN PREVALENCE IN FREE-RANGING EASTERN BOX TURTLES (TERRAPENE CAROLINA CAROLINA)

Wildlife epidemiologists aim to understand how pathogens spread across landscapes; however, imperfect detection due to host-pathogen biology, inappropriate sample collection or handling, or diagnostic assay performance limitations may lead to inaccurate prevalence estimates and ineffective management actions. In chelonian disease surveillance, pathogen prevalence is typically determined using quantitative real-time PCR (qPCR), but if an animal is infected but not shedding during surveillance, or the pathogen burden is too low to detect, the animal will appear qPCR negative. Occupancy modeling can provide a more accurate estimate of the true prevalence within a population by accounting for false-negative tests. We used single-season and dynamic occupancy models to generate unbiased prevalence estimates for three pathogens in free-ranging eastern box turtles (Terrapene carolina carolina). We analyzed Terrapene adenovirus, Terrapene herpesvirus 1, and box turtle Mycoplasmopsis sp. detection data from 778 turtles captured across five sites in central Illinois, USA, from 2016–22. During a single surveillance year in 2022, the reported raw prevalence values ranged from 5.6–7.7%; however, when controlling for imperfect detection, we estimated the prevalence for these three pathogens to range from 18.3–56.8%. This is higher than the prevalence previously thought to occur in Illinois box turtle populations. However, the detection probability for these pathogens in 2022 was relatively low and ranged from 24.1–48.3%, emphasizing the need for repeated sampling to determine true pathogen status. Continued investigation of pathogen occurrence, prevalence, and latent disease periods will aid veterinary management and promote effective conservation strategies for this species.