Data for: Why Hydrological Memory Dominates in Low-latitude Highlands: A Mechanistic Shift in Ecosystem Response to Extremes

Version 2.0 (October 13, 2025):This version includes a significant update to ensure the completeness and reproducibility of the research. Key improvements are:1. Data Integrity: A previously missing data file has been recalculated and is now included in the dataset.2. Script Validation: All analysis scripts have been thoroughly tested to confirm correct execution and ensure full reproducibility of the results presented in the manuscript.Understanding how compound extremes affect terrestrial ecosystems is a major challenge in Earth system science. Although the combined effects of stressors are recognized, the manner in which the prestress state determines the basic response mechanism remains unclear. In this study, we used the "natural experiment" methodology to compare two major extreme events within a monsoon-influenced, low-latitude highlands setting to explain mechanistic changes in land-atmosphere interactions. By analyzing an extensive set of remote sensing and reanalysis data with nonlinear structural equation modelling, we show that the ecosystem response shifted from a traditional water-limited paradigm during the 2010 drought to an energy-governed paradigm during the 2019 heatwave. Our results suggest that this transition is governed by the antecedent root zone soil moisture status, which acts as a tipping point fundamentally shifting the impact of atmospheric factors on canopy evapotranspiration screens, such as temperature and vapor pressure deficit. This study highlights a possible threshold-type state dependence non-linearity, which is lacking in major Earth System Models. Incorporating this “hydrological memory” is crucial for minimizing uncertainties in climate projections and for correctly assessing the vulnerability of ecosystems in a warming world. This repository contains NCL scripts and documentation for analyzing climate data in the Low Latitude Highland region. The scripts generate various figures showing the relationship between vegetation productivity (GPP), root-zone available water content, and terrain characteristics.