Vegetation dynamics in northern south America on different time scales

The overarching goal of this doctoral thesis was to understand the dynamics of vegetation activity occurring across time scales globally and in a regional context. To achieve this, I took advantage of open data sets, novel mathematical approaches for time series analyses, and state-of-the-art technology to effectively manipulate and analyze time series data. Specifically, I disentangled the longest records of vegetation greenness (>30 years) in tandem with climate variables at 0.05° for a global scale analysis (Chapter 3). Later, I focused my analysis on a particular region, northern South America (NSA), to evaluate vegetation activity at seasonal (Chapter 4) and interannual scales (Chapter 5) using moderate spatial resolution (0.0083°). Two main approaches were used in this research; time series decomposition through the Fast Fourier Transformation (FFT), and dimensionality reduction analysis through Principal Component Analysis (PCA). Overall, assessing vegetation-climate dynamics at different temporal scales facilitates the observation and understanding of processes that are often obscured by one or few dominant processes. On the one hand, the global analysis showed the dominant seasonality of vegetation and temperature in northern latitudes in comparison with the heterogeneous patterns of the tropics, and the remarkable longer-term oscillations in the southern hemisphere. On the other hand, the regional analysis showed the complex and diverse land-atmosphere interactions in NSA when assessing seasonality and interannual variability of vegetation activity associated with ENSO. In conclusion, disentangling these processes and assessing them separately allows one to formulate new hypotheses of mechanisms in ecosystem functioning, reveal hidden patterns of climate-vegetation interactions, and inform about vegetation dynamics relevant for ecosystem conservation and management.