Conservation beyond boundaries: using animal movement networks in protected area assessment.

Protected areas (PAs) are a core component of conservation policy and practice. However, many species for which they are designated are highly mobile, and may move among sites within and beyond PA boundaries. Environmental impacts on sites beyond those boundaries could thus impact the PA's protected populations, with the risk of adversely affecting its conservation objectives. Conservationists therefore urgently need tools to assess impacts on PAs and their populations of developments beyond their boundaries. We present a framework for using network analysis of observations of marked individuals to assess the protection footprint of PAs in the wider landscape and the impact footprint of developments within or neighboring PAs. We illustrate the use of this framework by assessing the impact of a current airport development proposal on a partially protected wetland, the Tagus estuary in Portugal, specifically by evaluating the extent of noise disturbance on the PA's population of Black‐tailed Godwits, a protected migratory wader species. By analyzing individual movements between sites and across seasons, we find disturbance impacts on up to 68.3% of individual Black‐tailed Godwits overall, greatly exceeding the estimates of 0.46–5.5% in the airport's Environmental Impact Assessments which derived from count data. We then compared the pre‐development network with simulated networks that represented two levels of site avoidance by removing sites predicted to receive noise in excess of two thresholds, 55 or 65 dB(A). Avoidance by Godwits of the potentially‐impacted sites is predicted to have a larger effect on network structure than removing the same number of sites at random, suggesting that an airport in the proposed location would operate as a barrier to connectivity in this wetland. We also highlight the role that network analyses of locations of marked individuals can play in assessing environmental impacts on protected sites and populations. [ABSTRACT FROM AUTHOR]