Environmental DNA concentrations vary greatly across productive and degradative conditions, with implications for the precision of population estimates.

Population size is an important metric to inform the conservation and management of species. For aquatic species, environmental DNA (eDNA) concentration has been suggested for non-invasively estimating population size. However, many biotic and abiotic factors simultaneously influence the production and degradation of eDNA which can alter the relationship between population size and eDNA concentration. We investigated the influence of temperature, salinity, and ranavirus infection on eDNA concentrations using tadpole mesocosms. Using linear regression models, we tested the influence of each experimental treatment on eDNA concentrations at three time points before and during epidemics. Prior to infection, elevated temperatures lowered eDNA concentrations, indicating that degradation was the driving force influencing eDNA concentrations. During early epidemics, no treatments strongly influenced eDNA concentrations and in late epidemics, productive forces dominated as ranavirus intensity and dead organisms increased eDNA concentrations. Finally, population size was only an important predictor of eDNA concentration in late epidemics and we observed high levels of variation between samples of replicate mesocosms. We demonstrate the complexities of several interacting factors influencing productive and degradative forces, variation in influences on eDNA concentration over short time spans, and examine the limitations of estimating population sizes from eDNA with precision in semi-natural conditions.