Modelling global pyrogeography using data derived from satellite imagery

Doutoramento em Engenharia Florestal e dos Recursos Naturais - Instituto Superior de Agronomia - UL

Vegetation burning has an important impact on the global atmosphere and vegetated land surface. Deforestation fires, peatland fires, and ecosystems with shortening fire return interval contribute substantially to the build-up of atmospheric greenhouse gases affecting environmental quality and the climate system at local and regional scales. Recognition of the role of fire in the Earth system led to its designation as an Essential Climate Variable (ECV), a physical, chemical, or biological variable that has a crucial contribution towards characterization of Earth’s climate. The central task of this thesis was the development of a new global classification and map of fire regimes, using multiple correspondence analysis and hierarchical clustering, and relying on active fire data from the Moderate Resolution Imaging Spectroradiometer (MODIS) MCD14ML product. That work was preceded by study dedicated to a thorough screening and exploratory spatial analysis of the dataset, and led to the development of an improved algorithm for identifying individual active fire clusters, and to global analysis of size inequality in their statistical distributions. In addition to this core research, other continental-global pyrogeography studies were developed, and are presented, dealing with: the time lag between the timing of optimal fire weather conditions and peak fire season dates as a diagnostic of anthropogenic vegetation burning; the spatial non-stationarity in the parameters of the relationship between population density and area burned; and the modulation of weekly cycles of vegetation burning in African croplands by regionally dominant religious affiliation. We hope that this set of studies may constitute a useful contribution to the burgeoning topic of global pyrogeography