Simulations reveal the power and peril of artificial breeding sites for monitoring and managing animals.

Despite common use, the efficacy of artificial breeding sites (e.g., nest boxes, bat houses, artificial burrows) as tools for monitoring and managing animals depends on the demography of target populations and availability of natural sites. Yet, the conditions enabling artificial breeding sites to be useful or informative have yet to be articulated. We use a stochastic simulation model to determine situations where artificial breeding sites are either useful or disadvantageous for monitoring and managing animals. Artificial breeding sites are a convenient tool for monitoring animals and therefore occupancy of artificial breeding sites is often used as an index of population levels. However, systematic changes in availability of sites that are not monitored might induce trends in occupancy of monitored sites, a situation rarely considered by monitoring programs. We therefore examine how systematic changes in unmonitored sites could bias inference from trends in the occupancy of monitored sites. Our model also allows us to examine effects on population levels if artificial breeding sites either increase or decrease population vital rates (survival and fecundity). We demonstrate that trends in occupancy of monitored sites are misleading if the number of unmonitored sites changes over time. Further, breeding site fidelity can cause an initial lag in occupancy of newly installed sites that could be misinterpreted as an increasing population, even when the population has been continuously declining. Importantly, provisioning of artificial breeding sites only benefits populations if breeding sites are limiting or if artificial sites increase vital rates. There are many situations where installation of artificial breeding sites, and their use in monitoring, can have unintended consequences. Managers should therefore not assume that provision of artificial breeding sites will necessarily benefit populations. Further, trends in occupancy of artificial breeding sites should be interpreted in light of potential changes in the availability of unmonitored sites and the potential of lags in occupancy owing to site fidelity.