The Vegetation Greenness Trend in the Arctic and Subarctic of North America from Landsat Data Record

Abstract:

Circum-Arctic warming has been linked to a range of vegetation changes, including increased shrub-tundra biomass, shrub encroachment, and altered forest succession. Coarse spatial resolution AVHRR satellite data have been commonly used to study these vegetation dynamics at continental scales, usually via the Normalized Difference Vegetation Index (NDVI) metric. However, AVHRR cannot resolve local, patch-scale trends in vegetation, or provide sufficient spatial detail to relate observed trends in NDVI to specific land cover types. We have assembled a record of atmospherically-corrected 30-meter Landsat data acquired during peak greenness months over three decades (1990-1992, 2000-2002, and 2010-2012) to study the vegetation dynamics of the entire Arctic and Subarctic of North America, and compare with the results derived from 1/12-degree AVHRR NDVI3g data for 1990-2012. Landsat NDVI values were screened for abrupt changes indicative of disturbance, and then assessed for statistically significant temporal trends. Landsat NDVI, at both 30-meter resolution and the aggregated 1/12-degree NDVI3g resolution, showed a strong greening trend continuously along the Arctic coast and sporadic browning trends inland (e.g. southern NWT). Approximately 29% and 3% of the study area shows greening and browning trend respectively. The spatial pattern and summary statistics are generally in good agreement with those for AVHRR results and are consistent with previous results indicating increased shrub-tundra biomass. Additional efforts are focusing on separating long-term, climate-induced vegetation change from disturbance recovery, and in relating observed changes to specific vegetation types and topographic/edaphic conditions. These results also illustrate the importance of long-term, calibrated satellite records for monitoring gradual changes in terrestrial ecosystems.