Instagram Occurrence Data Sampling for Species Distribution Modelling

Worldwide, Species Distribution Models (SDMs) are applied to model treeline species under current, past and future climate conditions. SDM approaches could also provide knowledge about treeline species of the southern Andes, which have received little attention in treeline research. Recently published studies addressing the southern Andes focus on local treeline sites, but large-scale modelling studies remain absent due to the limited accessibility of treeline sites. The main reason for the absence of modelling studies can be contributed to the lack of high quality and quantity occurrence data.

Against this background, social media sources are increasingly recognized as suitable databases for ground truthing in remote sensing classification approaches. We developed a novel social-media-based Instagram Ground Truthing method to determine valid occurrence data of the treeline species Nothofagus pumilio in the southern Andes.

Instagram postings of landscape images providing clear evidence of the species and its location were identified and transferred to a map. Sentinel 2 data provided the basis for supervised classification using the random forest algorithm. The resulting occurrence data, covering the entire distribution range of Nothofagus pumilio, were evaluated using a total of 1317 georeferenced points obtained by the Instagram Ground Truthing method.

Finally, the validated presence-absence data set of Nothofagus pumilio can then be used for further modelling studies under different climate conditions.