高寒生态系统脆弱性及其对气候变化和人类活动的响应

了解陆地生态系统的脆弱性和基本机制是适应和减轻全球气候变化影响的决策基础。生态系统的脆弱性可以通过生产力对气候变化的敏感性和适应性进行量化。采用1982-2018年青海省境内基于遥感的现实净初级生产力（NPP_R）和气候驱动的潜在净初级生产力（NPP_C），量化了高寒生态系统的敏感性（Sensitivity）、适应性（Adaptability）和脆弱性（Vulnerability）。然后探讨了生态系统脆弱性的时空变化，并分别从人类活动和气候变化的影响方面分析了其基本机制。结果表明：（1）基于NPP_R和NPP_C的生态系统脆弱性在空间上呈现出中度脆弱的模式，脆弱性从东南向西北由不脆弱依次递增到极度脆弱等级。（2）耕地的脆弱性较低，基于NPP_R和NPP_C的指数分别为-1.31和-0.93，这是由于其适应水平较高而敏感性较低；森林次之，指数为-1.18（NPP_R）和-1.06（NPP_C）；草原的指数为-0.17（NPP_R）和-0.17（NPP_C）；而荒漠的脆弱性较高，指数为0.77（NPP_R）和0.78（NPP_C），这是由于其敏感性较高，适应性较低。（3）基于NPP_R的高寒草地的脆弱性有两个温度阈值（-2.2±0.8）℃和（5.5±0.8）℃，一个降水阈值（387±45.6）mm，两个干旱指数阈值为（14.2±20.2）和（78.2±20.2）。而基于NPP_C的脆弱性也发现了同样的阈值，并且数值相似。阈值表明最佳气候条件下，生态系统将具有较高的适应性和较低的敏感性，即较低的脆弱性。但如果气温较低或较高，或者降水较低，生态系统的脆弱性将会更高。（4）人类活动对东部地区生态系统的脆弱性产生了强烈的影响，但就整个青海省的生态系统而言，这些影响在区域平均水平上较小。这项研究表明，在高寒脆弱的生态系统中，气候条件决定了脆弱性在空间上的分布情况，这应该被视为生态保护决策的理论基础。此外，本研究发现的阈值将为生态系统生态学提供一个案例研究，并应在世界各地的脆弱生态系统中广泛探索。

Vulnerability of alpine ecosystems and its response to climate change and human activities

Understanding vulnerability of terrestrial ecosystems and underlying mechanisms are fundaments for decision-making on the adaptation and mitigation of global climate change impacts. Vulnerability of ecosystem could be quantified by the sensitivity and adaptability of productivity to climatic changes. In this study, a method was developed to quantify the sensitivity, adaptability and vulnerability for alpine ecosystems. The method respectively quantified the vulnerability, sensitivity and adaptability of ecosystems through the remote sensing-based actual net primary productivity (NPP_R) and the climate-driven potential net primary productivity (NPP_C) over Qinghai province from 1982 to 2018. Then the spatial and temporal changes were explored in ecosystem vulnerability and the underlying mechanisms were analyzed in the terms of the effects from human activities and climate change respectively. The results showed that:(1) the NPP_R- and NPP_C-based vulnerability of the ecosystems showed a spatial pattern as moderately vulnerable, vulnerability increases from southeast to northwest in descending order from not vulnerable to extremely vulnerable. (2) Cropland had lower vulnerability with the NPP_R- and NPP_C- based index of -1.31 and -0.93, due to its higher adaptation level and lower sensitivity level, followed by forest with the index of -1.18 (NPP_R) and -1.06 (NPP_C), and grassland with the index of -0.17 (NPP_R) and -0.17 (NPP_C), while desert had higher vulnerability level with the index of 0.77 (NPP_R) and 0.78 (NPP_C), due to higher sensitivity level and lower adaptation level. (3) The NPP_R-based vulnerability of alpine grassland was found having the two thresholds (-2.2±0.8)℃ and (5.5±0.8)℃ of temperature, the one precipitation threshold (387±45.6) mm, therefore the two aridity index thresholds of (78.2±20.2) and (14.2 ±20.2). And the same thresholds were found for the NPP_C-based vulnerability, which has the similar value as the NPP_R-based vulnerability. The thresholds indicate that under optimal climate conditions, ecosystems will have higher adaptation and lower sensitivity, that is, lower vulnerability. But it will be higher vulnerability if temperatures are very cold or very warm, or the precipitation are very low. (4) The human activities exert strong impact on the vulnerability of the ecosystems over the eastern regions though the impacts were smaller in the term of the regional average for the ecosystems in whole province. This study demonstrated that vulnerability is spatially vary but dependent on climate condition in alpine fragile ecosystems, which should be considered as a theoretical fundament in the decision-making on ecological protection. Further the thresholds found in this study would provide a case study in ecosystem ecology and should be widely explored over the fragile ecosystems worldwide.