Local variation in mountain birch spring phenology along an altitudinal gradient in northern coastal Fennoscandia

Abstract

Currently there is a lack of spring phenology studies covering small-scale altitudinal gradients of mountain birch in coastal northern Fennoscandia, a region characterized by exceptionally high precipitation of snow, which gives reason to investigate the influence of snow cover on birch budburst in this area. Furthermore, ground phenology studies are today increasingly replaced by satellite data studies, but only too rarely is the latter approach properly validated. In order to provide a better basis for studies of local variation in spring phenology of birch in coastal altitudinal gradients, this study aimed at exploring how well (1) thermal sum models could predict budburst in individual birch trees with the inclusion of local conditions such as snow depth, soil temperature and altitude as predictors, and (2) NDVI values from high-resolution satellite images reflect leaf phenology in birch. The result for (1) showed that a simple thermal time model with spring air temperatures provided the best explanation for budburst, and that snow depth, soil temperature and altitude did not improve the predictions. The results for (2) showed that the correspondence between NDVI values and leaf phenology was generally high, but there were still some variation between in situ measurements and the satellite-derived NDVI, probably reflecting the limited capacity of satellite imagery to capture the phenology of merely one species.