Satellite remote sensing of vegetation dynamics in the context of climate change
Vegetation is a key component of the Earth's climate system. Understanding vegetation dynamics in a changing climate requires both in situ and remote sensing data. Satellite remote sensing is especially indispensible for continuous monitoring of vegetation over large areas. This dissertation is focused on investigation of vegetation dynamics in the broader context of climate change using satellite data over two critical regions: the arctic-boreal area in the northern high latitudes and Amazonia in South America. The northern high latitudes have experienced amplified warming. We found the response of the arctic-boreal vegetation to this warming to be different between North America and Eurasia during a 30-year period since 1982: the relationship between vegetation green-up and temperature rise was stable over Eurasia, but in North America, the amount of vegetation green-up per unit amount of warming has decreased since the beginning of 21st century. This could partly be explained by the unmatched northward movements of temperature and precipitation patterns in North America. The Amazonian rainforests have highly dense canopies of green leaves. In such dense media, reflection of solar radiation tends to saturate. Thus, the satellite measurements are weakly sensitive to vegetation changes. At the same time, the data are strongly influenced by changing sun-sensor geometry. This makes it difficult to discriminate between vegetation changes and sun-sensor geometry effects. We developed a new physically based approach to detect changes in dense forests. Analyses of several years of data from three sensors on two satellites under a range of sun-sensor geometries provide robust evidence for a sunlight driven seasonal cycle in structure and greenness of Amazonian rainforests. The 2005 and 2010 dry-season droughts decreased the photosynthetic activity of Amazonian rainforests. We demonstrate that satellite data capture such decreases. Furthermore, we show that in 2004 and 2007, when there was lower wet-season water abundance compared to normal years, the photosynthetic activity of Amazonian forests also decreased. Potentially frequent water deficits over Amazon in the future, irrespective of whether they occur in the dry or wet season, will decrease the photosynthetic activity of Amazonian forests, and provide a positive feedback to global warming.