COS 64-6
Vegetation greening in the US: Trends and drivers over 25 years

Wednesday, August 13, 2014: 9:50 AM
Regency Blrm F, Hyatt Regency Hotel
Brody Sandel, Department of Biological Sciences, Aarhus University, Aarhus C, Denmark
Jens-Christian Svenning, Department of Bioscience, Aarhus University, Aarhus C, Denmark

Terrestrial vegetation influences atmospheric composition, global climate and biodiversity. Human activities are leading to broad-scale changes in vegetation, due to direct exploitation, land use, and climate change. These influences are likely to persist and intensify in the near future, and it is crucial to understand how vegetation will respond. Many regions around the world are greening, though the precise mechanism for these patterns is not well understood. We examined spatiotemporal variation in vegetation greenness across the continental United States from 1983-2006, utilizing satellite-based measurements of the normalized difference vegetation index (NDVI) and asked whether NDVI changes could be attributed to three widely invoked explanations: climate change, land use changes and rising atmospheric CO2concentration.


There was an overall greening trend over this period, with an average rate of increase of about 0.1%/year. NDVI increased most in the winter and decreased slightly during the summer, leading to reduced seasonality. Precipitation primarily determined the spatial pattern of long-term NDVI, while year-to-year variations were jointly determined by temperature, precipitation and CO2. Within-year NDVI was driven primarily by temperature, except for an expanding region of the southwest that is limited by precipitation. While there was a warming trend across this period, it was insufficient to explain the magnitude of NDVI increase. Only by including CO2 could the trend be satisfactorily explained. Land use change had a minimal impact at this scale. Thus, we provide evidence for a crucial broad-scale CO2 fertilization effect on NDVI. The central importance of CO2 is problematic for making predictions because the functional relationship between CO2 and plant productivity over large scales is not well understood beyond the range of historically observed CO2 concentrations.