Warming permafrost on a global scale is projected to have significant impacts on engineering, hydrology and environmental quality. Greater warming trends are predicted on the Qinghai–Tibetan Plateau (QTP), but most models for mountain permafrost have not considered the effects of water phase change and the state of deep permafrost due to a lack of detailed information. To better understand historical and future permafrost change based on in situ monitoring and field investigations, a numerical heat conduction permafrost model was introduced which differentiated the frozen and thawed state of soil, and considered unfrozen water content in frozen soil, distribution of ground ice and geothermal heat flow. Simulations were conducted at two sites with validation by long‐term monitoring of ground temperature data. After forcing with reconstructed historical ground surface temperature series starting from 1966, the model predicted permafrost changes until 2100 under different RCP scenarios. The results indicate a slow thermal response of permafrost to climate warming at the two investigated sites. Even under the most radical warming scenario (RCP8.5), deepening of the permafrost table is not obvious before 2040. At both sites, the model indicates that shallow permafrost may disappear but deep permafrost may persist by 2100. Moreover, the simulation shows that the degradation modes may differ between zones of discontinuous and continuous permafrost. The main degradation mode of the site in the discontinuous zone appears to be upward thawing from the permafrost base, while that of the site in the continuous zone is downward thawing at the permafrost table with little change at the permafrost base.

中文翻译：

---

1966-2100年青藏高原多年冻土变化模拟——以青藏工程走廊沿线两个钻孔为例

预计全球范围内的永久冻土变暖将对工程、水文和环境质量产生重大影响。青藏高原（QTP）预计会有更大的变暖趋势，但由于缺乏详细信息，大多数山地永久冻土模型没有考虑水相变化和深层永久冻土状态的影响。为了在原位监测和实地调查的基础上更好地了解历史和未来多年冻土的变化，引入了数值热传导多年冻土模型，该模型区分了土壤的冻融状态，并考虑了冻土中未冻水含量、地冰分布和地热热流。在两个地点进行了模拟，并通过对地面温度数据的长期监测进行了验证。在对 1966 年开始的重建历史地表温度序列进行强制后，该模型预测了不同 RCP 情景下到 2100 年的永久冻土变化。结果表明，两个调查地点的永久冻土对气候变暖的热响应缓慢。即使在最激进的变暖情景（RCP8.5）下，2040年之前多年冻土表的加深并不明显。在两个地点，模型表明浅层多年冻土可能会消失，但深层多年冻土可能会持续到2100年。此外，模拟表明，不连续和连续多年冻土带的退化模式可能不同。不连续带场地的主要退化模式似乎是从多年冻土基底向上解冻，