Measuring Amazon rainforest resilience from remotely sensed data

The Amazon rainforest has a major contribution to the bio-geochemical functioning of the Earth system and has been projected to be at risk of large-scale, potentially irreversible, dieback to a savanna state. Measuring the resilience of the Amazon rainforest empirically is critical to helping us understand the magnitude and frequency of disturbances that the rainforest can still recover from. Different means to quantify resilience in practice have been proposed. Here we determine the Amazon rainforest resilience based on a space-for-time replacement, and then estimating the average residence time in the forest state. This 'global' notion of resilience is different from local measures based on variance or autocorrelation and thus provides complementary information. We study the dependence of the exit-time-base resilience on total rainfall and, in order to study the evolution of the Amazon rainforest, we also estimate changes in their resilience over the years.