Detections of vegetation greening trends at the global scale are often attributed to increases in atmospheric carbon dioxide (fertilization effect) and a lengthening of the growing season. Yet, regional contributions to this greening trend are not well understood. Satellite-derived observations of Normalized Difference Vegetation Index (NDVI) are widely used as a proxy for observing temporal and spatial changes in plant productivity. We analyzed the USGS EROS NDVI dataset from 1989 to 2014 for trends in maximum annual NDVI for the Greater Yellowstone Ecosystem (GYE) to address the following questions:
- Do we detect change in vegetation greening in the GYE during the past three decades?
- How do vegetation responses vary spatially across the GYE?
- What are the regional- and local-scale drivers of trends in vegetation greening?
To detect changes in vegetation greening we calculated the annual change in maximum NDVI for each pixel within the GYE at 1 km resolution. To isolate a climate effect we masked disturbed areas to calculate NDVI trends unaffected by disturbance during the study period. We examined the drivers of detected vegetation changes in two ways, by fitting an empirical model and through simulation modeling using the dynamic vegetation model LPJ-GUESS.
Results/Conclusions
Our analysis indicated positive (greening) and negative (browning) NDVI trends in 20.4% and 6.2% respectively of the area within the Greater Yellowstone Ecosystem. The highest concentration of pixels with a statistically significant (p<.05) greening trend in NDVI was in the southwestern region of the GYE, mostly within National Forest designations, especially Caribou-Targhee National Forest. Browning trends in NDVI were detected sporadically on the eastern half of the GYE. We filtered out pixels that had burned since 1984 (using Monitoring Trends in Burn Severity data), land cover layers that were not natural vegetation (using National Land Cover Database), logged stands (using North American Forest Dynamics data), and areas affected by bark beetles (using Aerial Detection Surveys data). Of the remaining pixels within the GYE, only 11.1% had a greening trend and 3.6% had a browning trend. This suggests local-scale disturbance and management history are primary drivers of detected greening and browning vegetation trends in the GYE. Our results suggest that at the local scale, the direct effects of increases in atmospheric carbon dioxide and elongated growing season due to climate change have yet to be strongly influential on vegetation greening, despite other indications of climate change impacts on the GYE.