Abstract
Ecological traps, poor‐quality habitat that nonetheless attract individuals, have been observed in both natural and human‐altered settings. Until recently, ecological traps were considered a kind of source–sink system, but source–sink theory does not model maladaptive habitat choice, and therefore cannot accurately represent ecological traps or predict their population‐level consequences. Although recent models of ecological traps addressed this problem, they used patch‐based models containing only two habitats that were very different from one another, but were internally homogeneous. These sorts of patch models may not apply to many real populations, and using them for populations in landscapes with mosaic or gradient habitat structures may be misleading. I developed models that treat source–sink dynamics and ecological traps as special cases of a single process, in which the attractiveness and quality of the habitat are separate variables that can be either positively or negatively related, and in which habitat quality varies continuously throughout the landscape. As expected, sinks are less detrimental to populations than ecological traps, in which preferential use of poor habitat elevates extinction risk. Furthermore, ecological traps may be undetected, and may even appear to be sources, when population sizes are large, but may still prevent recovery in spite of the availability of high‐quality habitat when populations drop below threshold levels. Conservation biologists do not routinely consider the possibility that apparent sinks are actually traps, but since traps should be associated with the rapidly changing and novel habitat characteristics primarily produced by human activities, ecological traps should be considered an important and potentially widespread conservation concern.