"Recent vegetation phenology variability and wild reindeer migration in Hardangervidda plateau (Norway)"
Résumé
More than others, arctic and subarctic ecosystems are affected by rapid consequences of global climate changes. The effects on Scandinavian reindeer migrations and interactions with environment are studied here. The herbivorous plays numerous roles both in Scandinavian natural and cultural landscapes (Forbes et al., 2007). In particular, reindeer grazing history is suspected to strongly interact with balance of carbon in tundra ecosystems (Väisanen et al., 2014). Moreover, the sentinel-specie allow us a better understanding of interactions between arctic ungulates and cold ecotones in the context of increasing extent and biomass of vascular plants (shrubs, forbes) for Eurasiatic subpolar latitudes (Macias-Fauria et al., 2012). The IUCN red list of endangered species has just classified reindeer Rangifer tarandus as vulnerable (Gunn, IUCN, 2016). Wild reindeer (Rangifer tarandus tarandus L.) herds in Hardangervidda mountainous plateau (Norway) constitute one of the three isolated populations along Fennoscandia mountain range. Population is declining (above 30 000 individuals in 1960, above 5000 in 2010, Uboni et al., 2016) threatening it to bottle-neck effects and then serious difficulties to adapt to potential effects of climate change (directs: summer and weather conditions; indirect: forage resources availability and accessibility, insect harassment and parasitism). The study aims firstly to understand temporal and spatial variability of seasonal habitat choices for all individuals. We then characterize phenological variability with Corine Land Cover ecological habitat assessment and bi-monthly NDVI index (MODIS 13Q1, 2000-2015, 250 m.). Thirdly, we test relationships between 2000-2015 averaged onset and monthly median center location for population and individuals.
The GPS-tracks dataset comes from Cagnacci et al., 2015, monitoring 7 individuals with GPS collars. The temporal coverage is 2007-2010, with a bi-monthly accuracy. The biogeophysical descriptors are represented by altitudes, past monthly temperatures (1950-1990), Corine Land Cover ecological habitats (European Environment Agency, 2006) and temporal series of MODIS 13Q1 NDVI index (2000-2015, bi-monthly). The study area is delimited by convex hull GIS-tool, taking account of each reindeer location. We developed a Python toolbox automating regional sampling from numerous raster datasets. The extent of the polygonal grid is represented by the convex hull calculation process. The script automatically generates a continuous grid (250 meters grid size) and associated centroid of polygons. Such script includes the possibility to export the entire dataset (with geographical IDs), allowing us to compute statistical analysis using regular spreadsheets software.
First analysis of GPS-locations median centers shows a strong differentiation between habitat choices during hot and cold seasons with a swing between sparsely vegetated areas in winter and mainly moors and heathlands in summer. Nevertheless, all along the year, results showed a large preference for moors and heathland with a relatively small inter-annual variability (from 57 to 63% of GPS locations). Sparsely vegetated areas represent the second most frequent habitat for both population and individuals (min: 18%). The summer GPS-locations also described a significant diversity and equality in habitat choices (birch forests or mixed birch-coniferous forest). Comparing monthly median centers reveal regularity of inter-annual location and habitat selection the calving period, in June. Concerning the phenological peak by type of ecological habitats (up to 0.7 NDVI threshold), dates and intensities are very dependent of altitudes and monthly temperatures distribution among ecological habitats. Depending on contextual conditions, wild reindeer could benefit of landscape heterogeneity (at regional/annual and at local/seasonal scales) to adapt to phenological variability during the growth period. These first results could be reinforced by a stronger quantification of relationships between reindeer relative position, landscape heterogeneity and bioclimatological variability. It could be in particular interesting to examine the links between phenological seasonal gradient, GPS-locations and slope aspects (Måtell et al., 2006). The automation of the vectorial grid creation and the attributes extraction will allow us to facilitate the measurement of reindeer densities as well as the collection and preprocessing of geophysical descriptors.
Loading...
