The long-term increase of satellite-based proxies of vegetation cover is a well-documented response of seasonally snow-covered ecosystems to climate warming. However, observed greening trends are far from being uniform and substantial uncertainty remains concerning the underlying causes of this spatial variability. Here, we processed surface reflectance of the moderate resolution imaging spectroradiometer (MODIS) to investigate trends and drivers of changes in the annual peak values of the Normalized Difference Vegetation Index (NDVI). Our study focuses on the above treeline ecosystems in the European Alps. The NDVI changes of these ecosystems are highly sensitive to land cover and biomass changes and are marginally affected by anthropogenic disturbances. We found a widespread greening for the period 2000-2020, a pattern that is consistent with the overall increase of summer temperature. At the local scale, the spatial variability of greening was mainly due to the preferential response of north-facing slopes between 1900 m and 2400 m. Using high resolution imagery, we noticed that the presence of screes and outcrops locally magnified this response. At the regional scale, we identified hotspots of greening where vegetation cover is sparser than expected given the elevation and exposure. Most of these hotspots experienced delayed snowmelt and green-up dates in recent years. We conclude that the ongoing greening in the Alps primarily reflects the high responsiveness of sparsely vegetated ecosystems that are able to benefit the most from temperature and water-related habitat amelioration above treeline.

We exploit available time series of the moderate resolution imaging spectroradiometer (MODIS) to provide a comprehensive picture of recent greening and its spatial variability in the European Alps and to improve our understanding of its drivers. Our study utilizes annual peak values of the Normalized Difference Vegetation Index (NDVI) as a proxy of land surface greenness. We quantified the significance and magnitude of NDVI trends for high elevation ecosystems that are located between the treeline and the permanent snowline.

We first selected a set of 250-m resolution MODIS pixels having non-forested land cover classes with a tree cover density below 5%, an elevation above 1400 m and an average NDVImax above 0.15. We discarded pixels exhibiting more than 10% of settlements (ski resorts) or permanent water using data layers of European settlements (https://land.copernicus.eu/pan-european/GHSL/european-settlement-map/) and of permanent water (https://land.copernicus.eu/pan-european/high-resolution-layers/water-wetness/status-maps/water-wetness-2018). We also removed pixels for which we found significant abrupt changes of NDVImax within the period 2000-2020, as this might be indicative of a physical or anthropogenic disturbance unrelated to climate trend. We ended up with 511,375 pixels - see the file data_GCB.csv - of which 284,346 (55.6%) exhibited an average NDVImax below 0.65. Our main findings are based on this data subset to avoid saturation effect of NDVImax for closed canopies.