Speaker
When
Where
Role of precipitation variability on the distribution of tropical forests and savannas
Tropical forests and savannas are the most extensive ecosystems in the tropics, accounting for ~60% of global terrestrial productivity and providing essential ecosystem services. Traditionally, their distribution has been linked to climatic factors, particularly total annual precipitation and disturbance regimes such as fire. However, the influence of daily precipitation variability remains less explored, despite large precipitation events and prolonged dry spells becoming more frequent with climate change, independent of trends in annual precipitation changes.
In this talk, I present an analysis that integrates remote sensing data with machine learning models to examine how soil properties, fire occurrence, and precipitation at multiple temporal scales influence the distribution of forests and savannas. Our results show that the length of the dry spell during the dry season can be a more significant predictor of forest and savanna distribution than annual precipitation (MAP), fire occurrence, and soil properties. Furthermore, tropical forests are more common in regions with short dry spells during both dry and wet seasons, indicating that forests depend not only on high MAP but also on short dry periods.
Our results have the potential to improve our understanding of the risk of transitions between dryland and non-dryland ecosystems across the tropics, associated with less explored tipping points in precipitation variability, such as dry spell length during the dry season. This is crucial for reducing uncertainty in future projections of land-atmosphere feedbacks, particularly those related to the global water and carbon cycles.
Can't make it in person? Join Virtually