Impact of precipitation on the resilience of tropical forests to non-Gaussian Lévy fluctuations

Abstract

Understanding the resilience of tropical vegetation to perturbations and disturbances is crucial for predicting ecosystem responses to climate change. Here we investigate the stability of tropical forest ecosystems across varying precipitation levels and the influence of extreme events, which are modeled as burst-like pulses following a heavy-tailed distribution, using an α-stable Lévy process. The non-Gaussian index α and noise intensity ε of α-stable Lévy processes characterizes the frequency and the intensity of these extreme events. We propose a novel global resilience measure based on the stationary density to quantify the probability of the system to remain within its basin of attraction despite extreme perturbations. Our findings reveal that higher precipitation levels inherently provide greater stability to the forest state, even in the presence of larger noise intensities and higher frequencies of small jumps in extreme events. In contrast, at a low precipitation level, forest resilience is markedly reduced and declines rapidly with rising noise intensity, indicating a higher susceptibility to perturbations. Our study highlights the critical role of precipitation in modulating the resilience of tropical forests to disturbances, realistically modelled as non-Gaussian Lévy fluctuations.