Global warming is predicted to change ecosystem functioning and structure in Arctic ecosystems by strengthening top-down species interactions, i.e. predation pressure on small herbivores and interference between predators. Yet, previous research is biased towards the summer season. Due to greater abiotic constraints, Arctic ecosystem characteristics might be more pronounced in winter. Here we test the hypothesis that top-down species interactions prevail over bottom-up effects in Scandinavian mountain tundra (Northern Sweden) where effects of climate warming have been observed and top-down interactions are expected to strengthen. But we test this 'a priori' hypothesis in winter and throughout the 3-4 yr rodent cycle, which imposes additional pulsed resource constraints. We used snowtracking data recorded in 12 winters (2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015) to analyse the spatial patterns of a tundra predator guild (arctic fox Vulpes lagopus, red fox Vulpes vulpes, wolverine Gulo gulo) and small prey (ptarmigan, Lagopus spp). The a priori top-down hypothesis was then tested through structural equation modelling, for each phase of the rodent cycle. There was weak support for this hypothesis, with top-down effects only discerned on arctic fox (weakly, by wolverine) and ptarmigan (by arctic fox) at intermediate and high rodent availability respectively. Overall, bottom-up constraints appeared more influential on the winter community structure. Cold specialist predators (arctic fox and wolverine) showed variable landscape associations, while the boreal predator (red fox) appeared strongly dependent on productive habitats and ptarmigan abundance. Thus, we suggest that the unpredictability of food resources determines the winter ecology of the cold specialist predators, while the boreal predator relies on resource-rich habitats. The constraints imposed by winters and temporary resource lows should therefore counteract productivity-driven ecosystem change and have a stabilising effect on community structure. Hence, the interplay between summer and winter conditions should determine the rate of Arctic ecosystem change in the context of global warming.

Traditional grazing areas in Europe have declined substantially over the last century. Specifically, in northern Fennoscandia, the grazing land is disturbed by cumulative land-use pressures. Here we analysed the configuration of the grazing land for reindeer and sheep in northern Fennoscandia in relation to the concurrent land-use pressures from tourism, road and railway networks, forestry, industrial and wind energy facilities, together with predator presence and climate change. Our results show that 85% of the region is affected by at least one land-use pressure and 60% is affected by multiple land-use pressures, co-occurring with predator presence and rising temperatures. As such, a majority of the grazing land is exposed to cumulative pressures in northern Fennoscandia. We stress that, if the expansion of cumulative pressures leads to grazing abandonment of disturbed areas and grazing intensification in other areas, it could irreversibly change northern vegetation and the Fennoscandian mountain landscape.

Studies of ecological processes should focus on a relevant spatial scale, as crude spatial resolution will fail to detect small scale variation which is of potentially critical importance. Remote sensing methods based on multispectral satellite images are used to assess primary productivity and aerial photos to map vegetation structure. Both methods are based on the principle that photosynthetically active vegetation has a characteristic spectral signature. Yet they are applied differently due to technical differences. Satellite images are suitable for calculations of vegetation indices, for example Normalized Difference Vegetation Index (NDVI). Colour infrared aerial photography was developed for visual interpretation and never regarded for calculation of indices since the spectrum recorded and post processing differ from satellite images. With digital cameras and improved techniques for generating colour infrared orthophotos, the implications of these differences are uncertain and should be explored. We tested if plant productivity can be assessed using colour infrared aerial orthophotos (0.5 m resolution) by applying the standard NDVI equation. With 112 vegetation samples as ground truth, we evaluated an index that we denote rel‐NDVIortho in two areas of the Fennoscandian mountain tundra. We compared the results with conventional SPOT5 satellite‐based NDVI (10 m resolution). rel‐NDVIortho was related to plant productivity (Northern area: P = <0.001, R2 = 0.73; Southern area: P = <0.001, R2 = 0.39), performed similar to SPOT5 satellite NDVI (Northern area: P = <0.001, R2 = 0.76; Southern area: P = <0.001, R2 = 0.40) and the two methods were highly correlated (cor = 0.95 and cor = 0.84). Despite different plant composition, the results were consistent between areas. Our results suggest that vegetation indices based on colour infrared aerial orthophotos can be a valuable tool in the remote sensing toolbox, offering a high‐spatial resolution proxy for plant productivity with less signal degradation due to atmospheric interference and clouds, compared to satellite images. Further research should aim to investigate if the method is applicable to other ecosystems.

No abstract