Inconsistent relationships between annual tree ring-widths and satellite- measured NDVI in a mountainous subarctic environment

Recent climate warming has intensified efforts to understand patterns of vegetation productivity at northern high latitudes. Remotely-sensed and dendrochronological measurements have indicated significant changes in aboveground vegetation production in these regions, but the relationship between the two indicators remains unclear. We examined the relationship between remotely-sensed data, using the normalized difference vegetation index (NDVI), and annual ring-width indices obtained from trees and shrubs in the boreal mountains of southwest Yukon. A 24-year (1990-2013) trend in annual maximum NDVI was generated from two Advanced Very High Resolution Radiometer (AVHRR) datasets: the Alaska Composite NDVI (1 km), and the third generation Global Inventory Modeling and Mapping Studies (GIMMS(3g)) NDVI (8 km). Twenty sites were systematically selected for dendrochronological sampling and chronologies were developed for the dominant coniferous (Picea glauca and Picea mariana) and deciduous (Salix glauca, Alnus crispa, and Populus tremuloides) species at each site. NDVI trends indicated low-level greening across much of the region during the 24-year period. However, ring-width trends varied in direction and magnitude and the agreement with annual maximum NDVI was low. Correlations between annual ring-width indices and biweekly NDVI values were also low, although eight coniferous chronologies did show significant positive correlations with GIMMS(3g) biweekly data, suggesting that radial growth may be reflected in NDVI at some point during the growing season. Reasons for the inconsistent relationships between ring-width and NDVI remain unclear. Future studies should examine the role of climatic variables on the ring-width-NDVI relationship, especially when multiple datasets and various resolutions are being used. Investigating the relative contributions of plant growth, establishment of new plants, and changes in plant resource allocation to NDVI trends are critical to achieve a more functional understanding of the changes occurring in mountainous boreal ecosystems.