Quantifying drought-induced ecosystem vulnerability, e.g. in terms of plant productivity, based on vulnerability curves and coupled processes is a frontier issue and a challenge in the field of climate change risk. Primary productivity vulnerability to drought varies among and within ecosystem types, obscuring generalized patterns of ecological stability. Thus, we constructed drought vulnerability curves between aboveground net primary productivity (ANPP) and drought intensity for forest and grassland, through global data collection, bias checking, and systematic integration. Based on the improvement of theoretical analysis of ecosystem vulnerability and quantitative evaluation methods, we investigated the processes of sensitivity and adaptation to reveal vulnerability mechanisms. The ANPP was nonlinearly reduced along with drought intensity gradients, with an increasing trend for forests and a decreasing trend for grasslands. Under the same drought duration, both the forest and grassland ANPP decreased with increasing intensities (from light to severe drought). However, the forest and grassland ANPP exhibited divergent responses with durations under the same drought intensity. The combinations of drought duration and intensities also affected the ANPP. For example, negative responses of the grassland ANPP to increasing intensities differed among the durations (p = 0.001) and maximum occurred under the severe drought. Furthermore, we quantified the vulnerability by analyzing sensitivity and adaptation to short- (≤ 3years) and long-term (> 3 years) drought. Grasslands showed higher sensitivity to short-term drought than to long-term drought, and the contrary for forests. Forests exhibited a certain adaptation to long-term drought, but to a lesser extent than grasslands. Comprehensively, grasslands presented lower vulnerability to growing-season drought with higher adaptation and lower sensitivity than forests. These findings suggest that quantitative assessment on ecosystem response to drought from the viewpoints of vulnerability curves and processes should be promoted in and coupled with future climate change risks and sustainable management of different ecosystems.

基于脆弱性曲线和耦合过程来量化干旱引起的生态系统脆弱性，例如在植物生产力方面，是气候变化风险领域的前沿问题和挑战。初级生产力对干旱的脆弱性在生态系统类型之间和内部各不相同，掩盖了生态稳定的普遍模式。因此，我们通过全球数据收集，偏差检查和系统整合，在地上净初级生产力（ANPP）与森林和草原的干旱强度之间建立了干旱脆弱性曲线。在对生态系统脆弱性理论分析和定量评估方法进行改进的基础上，我们研究了敏感性和适应性过程以揭示脆弱性机制。随着干旱强度梯度的增加，ANPP呈非线性下降趋势，其中森林呈上升趋势，草地呈下降趋势。在相同的干旱持续时间下，森林和草原的ANPP都随着强度的增加而减少（从轻度干旱到严重干旱）。然而，在相同干旱强度下，森林和草原ANPP在持续时间上表现出不同的响应。干旱持续时间和强度的组合也影响了ANPP。例如，草地ANPP对强度增加的负面反应在持续时间之间有所不同（在相同的干旱强度下，森林和草原ANPP在持续时间上表现出不同的响应。干旱持续时间和强度的组合也影响了ANPP。例如，草地ANPP对强度增加的负面反应在持续时间之间有所不同（在相同的干旱强度下，森林和草原ANPP在持续时间上表现出不同的响应。干旱持续时间和强度的组合也影响了ANPP。例如，草地ANPP对强度增加的负面反应在持续时间之间有所不同（p  = 0.001），在严重干旱下最大。此外，我们通过分析对短期（≤3年）和长期（> 3年）干旱的敏感性和适应性来量化脆弱性。草原对短期干旱的敏感性高于对长期干旱的敏感性，而对森林则相反。森林对长期干旱表现出一定的适应性，但程度不及草原。总体而言，与森林相比，草原对生长季节干旱的脆弱性较低，适应性和敏感性较低。这些发现表明，应从脆弱性曲线和过程的角度出发，对生态系统对干旱的反应进行定量评估，并将其与未来的气候变化风险和不同生态系统的可持续管理相结合。