Abstract
Arctic ecosystems are particularly vulnerable to impacts of climate change; however, the complex relationships between climate and ecosystems make incorporating effects of climate change into population management difficult. This study used structural equation modelling (SEM) and a 24-year multifaceted monitoring data series collected at Zackenberg, North-East Greenland, to untangle the network of climatic and local abiotic and biotic drivers, determining their direct and indirect effects on two herbivores: musk ox ( Ovibos moschatus ) and collared lemming ( Dicrostonyx groenlandicus ). Snow conditions were determined to be the central driver within the system, mediating the effects of climate on herbivore abundance. Under current climate change projections, snow is expected to decrease in the region. Snow had an indirect negative effect on musk ox, as decreased snow depth led to an earlier start to the Arctic willow growing season, shown to increase fecundity and decrease mortality. Musk ox are therefore expected to be more successful under future conditions, within a certain threshold. Snow had both positive and negative effects on lemming, with lemming expected to ultimately be less successful under climate change, as reduction in snow increases their vulnerability to predation. Through their capacity to determine effects of climatic and local drivers within a hierarchy, and the relative strength and direction of these effects, SEMs were demonstrated to have the potential to be valuable in guiding population management.