Data & Scripts for: Hydraulic resonance and adaptive decoupling dictate ecosystem resilience to compound climate extremes

Earth System Models (ESMs) rarely account for non-linear ecosystem tipping points, relying instead on linear and additive logic to project carbon sink responses to compound climate extremes. Here, we explore an extreme "resilience paradox" in the South Chinese monsoon forests (SCMFs): gross primary production (GPP) underwent catastrophic collapse during a moderate 2003 heat-dry wave but was highly resilient to a 2013 compound extreme with roughly 75% more thermodynamic energy surplus. To reconcile this disparity, we present the Ecological Amplitude Modulation (EAM) framework. By implementing spatiotemporal fast-MEEMD and Hilbert-Huang Transform (HHT) analyses, we frame the ecosystem response as a multiplicative process. We detect a quasi-40-day spectral gap, which functions as an intrinsic biophysical firewall separating high-frequency physiological carrier waves from low-frequency environmental modulators. We show that the 2003 crash was excited by “hydraulic resonance”—a deleterious frequency-locking amongst these cross-scale signals, which left an enduring spectral scar and determined an asymmetric, sluggish recovery. In contrast, the 2013 resistance was derived from "adaptive decoupling," which insulated the physiological engine effectively from damaging sub-seasonal desiccation. These results reveal that ecosystem resilience is a frequency-dependent topological property rather than a static capacity based on magnitudes, providing a critical diagnostic measure of carbon sink dynamics under an increasingly severe Anthropocene climate.