HYDROLOGICAL VARIABILITY ON VEGETATION SEASONALITY, PRODUCTIVITY AND COMPOSITION IN TROPICAL ECOSYSTEMS OF AFRICA

Abstract

As the world's second largest and most populous continent behind Asia, the African continent harbors vast area of tropical ecosystems spanning great gradients from savannas to rainforests, which provide essential ecosystem services at regional and global scales. In the context of global climate change, many of these drought-sensitive ecosystems are expected to be modified or endangered. This dissertation aims to advance fundamental understandings on response and vulnerability of African tropical ecosystems to various possible hydroclimate changes at continental scales by addressing two major questions: (1) What climatic variables do control vegetation seasonality and phenology in tropical ecosystems and how (Chapter 2-5)? (2) How do intra-seasonal and inter-annual hydrological variabilities influence function and structure of tropical ecosystems (Chapter 6-8)? Chapter 2 provides an advanced algorithm to derive vegetation phenological and trajectory information over Africa. Using the derived decadal-long phenological record, I explore the hydrological controls on the phenology of African savannas and woodlands in Chapter 3 by examining phase correspondences and inter-annual variations between rainy season and growing season, revealing the possible role of groundwater in African woodlands. Chapter 3 also provides the non-linear impacts of rainy season length on ecosystem structures (e.g. tree fraction). Chapter 4 reveals a possible seasonal decoupling between canopy functioning and canopy structure from optical and microwave remote sensing respectively in African tropical evergreen forests. Chapter 5 discovers a diverse seasonal pattern of canopy photosynthesis in tropical evergreen forests at global scale, and a threshold of annual total rainfall (~2000mm/year) governs the supply-demand relationship between seasonal subsurface water storage and dry season water deficit, which determines these global patterns. Chapter 6 studies the inter-annual sensitivity of different biomes to precipitation variability using multi-sensor data, and generates an improved vegetation fractional cover products. Chapter 7 and Chapter 8 respectively study the impact of individual and interactive changes in rainfall characteristics (i.e. frequency, intensity and rainy season length) on ecosystem productivity and biome distributions. The results strongly demonstrate that only considering changes in mean state of climate is not enough to capture ecosystem responses and explicitly accounting for variability at intra-seasonal scale is essential to predict ecosystem changes.