Rephotography, permanent plots and remote sensing data provide varying insights on vegetation change on subantarctic Macquarie Island, 1980–2015

Abstract

AbstractDespite its remoteness, the vegetation of Subantarctic Macquarie Island is amongst the most thoroughly surveyed and monitored in Australia. Eradication of invasive European rabbits was achieved in 2011, after a decline in introduced rabbit populations in the eighties and nineties and a resurgence in the early twenty‐first century. The impacts of introduced rabbits and climate change on the vegetation of Macquarie Island are well‐documented from long‐term plots situated in some, but not all, of the vegetation types. Satellite data are sporadic, with substantial could‐free coverage rare. The island has been well‐photographed over many years, and several series of long‐term photo‐monitoring sites have been established over the past few decades. We compared insights on vegetation change derived from repeat landscape photography three years after myxoma virus was established, rabbit population resurgence and rabbits eradication, with plant species cover data from a set of long‐term monitoring sites, and with spectral vegetation indices derived from satellite imagery covering the transition from high rabbit numbers to no rabbits. The repeat photography confirmed many well‐documented vegetation trends, as well as providing an indication of previously unreported trends: stability of Acaena herbfield despite reduced disturbance; increase in cover of Pleurophyllum hookeri; increase in short grassland vegetation; and constancy in cover of Azorella macquariensis. Although subject to errors related to differences in sample size and resolution, the rephotography data gave the best information on spatial and temporal patterns of overall vegetation change. The plot data, where comparable, were not fully consistent with the rephotography data and are expensive to collect, requiring scientific expertise in the field and much time, while the sporadically available remote sensing data were effective in monitoring the revegetation/devegetation of ground and the recovery of Poa foliosa but we were unable to detect other changes.