Modelling movement and landscape connectivity of New Zealand native birds in highly structured agroecosystem networks

Understanding how spatial heterogeneity affects movement and dispersal is critical for maintaining functional connectivity in agroecosystems. Least-cost path models are popular conservation tools to quantify the cost of a species dispersing though the landscapes. However, the variability of species in life history traits and landscape configurations can affect their space-use patterns and should be considered in agroecosystem management aiming to improve functional biodiversity. In this study, we modelled the connectivity properties of native species on a real agroecosystem landscape dominated by sheep and beef farming in north Canterbury, New Zealand, where the recovery of native bird population is desired. We chose two species to act as case studies that were contrasting in their mobility: New Zealand pigeon/kererū (Hemiphaga novaeseelandiae; highly mobile) and southern brown kiwi/tokoeka (Apteryx australis; flightless). Networks of the least-cost paths of kererū and tokoeka were constructed based on their habitat preferences and movement capacities, and we compared and contrasted the connectivity properties and network topographies of their networks. We then compared the network metrics of western side (higher density of forest) with the eastern side (dominated by grazed grassland) of the study area where the vegetation composition was vastly different for both species. The results shown three variables were the most important contributors to the structure of the dispersal networks: the nature of the matrix, spatial structure of vegetation patches, and the gap-crossing ability of the study species. Tokoeka were able to utilise smaller habitat patches as stepping-stones for dispersal, while kererū can select more preferred habitat patches due to their high movement capacity. In contrast to the eastern side, we observed the western/forested side to have more, and stronger, links among habitat patches for both species, due to the presence of several large patches of native forest. Our work suggested that one size does not fit all, rather, conservation strategies that account for species’ life histories and movement traits are required to identify and preserve a connected ecological network.