Permafrost Degradation Risk Evaluation in the Qinghai‐Tibet Plateau Under Climate Change Based on Machine Learning Models

Abstract

AbstractPermafrost in the Qinghai‐Tibet Plateau (QTP) is sensitive to climate warming, but the associated degradation risk still lacks accurate evaluation. To address this issue, machine learning (ML) models are established to simulate the mean annual ground temperature (MAGT) and active layer thickness (ALT), and climate data from shared socioeconomic pathways (SSPs) are prepared for evaluation in the future period. Based on the projections, permafrost is expected to remain relatively stable under the SSP1‐2.6 scenario, and large‐scale permafrost degradation will occur after the 2050s, resulting in area losses of 30.15% (SSP2‐4.5), 58.96% (SSP3‐7.0), and 65.97% (SSP5‐8.5) in the 2090s relative to the modeling period (2006–2018). The average permafrost MAGT (ALT) is predicted to increase by 0.50°C (59 cm), 0.67°C (89 cm), and 0.79°C (97 cm) in the 2090s with respect to the modeling period under the SSP2‐4.5, SSP3‐7.0, and SSP5‐8.5 scenarios, respectively. Permafrost in the Qilian Mountains and Three Rivers Source region are fragile and vulnerable to degradation. In the future period, permafrost on the sunny slopes is more prone to degradation and the sunny‐shade slope effect of permafrost distribution will be further enhanced under climate warming. The lower limit of permafrost distribution is expected to rise by about 100 m in the 2050s under the SSP2‐4.5 scenario. These findings can provide valuable insights about future permafrost changes in the QTP.