NDVI
Bi-weekly normalized difference vegetation index (NDVI) data for the approx. 8 × 8 km2 pixel incorporating the research area was extracted from the third generation of the Global Inventory Modeling and Mapping Studies (GIMMS3g) dataset for the period July 1981 - 2015. The GIMMS3g-NDVI dataset is assembled from different sensors of the Advanced Very High-Resolution Radiometers (AVHRR) satellites. It was tested whether there were positive linear trends in maximum NDVI (NDVImax) and mean late summer (July-August) NDVI over time with one-sided Pearson’s correlation tests. Furthermore, it was tested whether there were positive relationships between the mean standardized shoot length chronology / branch initiation on the one hand and NDVImax / late summer NDVI on the other. In addition, negative relationships between standardized branch mortality frequencies in the previous year and NDVImax / late summer NDVI were hypothesized and tested. Lastly, positive relationships between seasonal climate variables (mean temperatures and precipitation sums; the same as in the climate-growth models) were hypothesized and tested for significance with one-sided Pearson’s correlation tests.