Changes in the ground surface temperature in permafrost regions along the Qinghai–Tibet engineering corridor from 1900 to 2014: A modified assessment of CMIP6

• Published in: Advances in Climate Change Research
• Main Authors: Zan-Pin Xing, Lin Zhao, Lei Fan, Guo-Jie Hu, De-Fu Zou, Chong Wang, Shu-Ci Liu, Er-Ji Du, Yao Xiao, Ren Li, Guang-Yue Liu, Yong-Ping Qiao, Jian-Zong Shi
• Format: Article in Journal/Newspaper
• Language: English
• Published: KeAi Communications Co., Ltd. 2023
• Subjects: Ground surface temperature; CMIP6; Cold bias; Permafrost; Qinghai–Tibet engineering corridor; Meteorology. Climatology; QC851-999; Social sciences (General); H1-99; permafrost
• Online Access: https://doi.org/10.1016/j.accre.2023.01.007; https://doaj.org/article/b8cd96f77c9b40cf80a1661750cc0fe2

Numerous studies were published in the last two decades to evaluate and project the permafrost changes in its thermal state, mainly based on the soil temperature datasets from the Coupled Model Intercomparison Project (CMIP), and discuss the impacts of permafrost changes on regional hydrological, ecological and climatic systems and even carbon cycles. However, limited monitored soil temperature data are available to validate the CMIP outputs, resulting in the over-projection of future permafrost changes in CMIP3 and CMIP5. Moreover, future permafrost changes in CMIP6, particularly over the Qinghai–Tibet Plateau (QTP), where permafrost covers more than 40% of its territory, are still unknown. To address this gap, we evaluated and calibrated the monthly ground surface temperature (GST; 5 cm below the ground surface), which was often used as the upper boundary to simulate and project permafrost changes derived from 19 CMIP6 Earth System Models (ESMs) against in situ measurements over the QTP. We generated the monthly GST series from 1900 to 2014 for five sites based on the linear calibration models and validated them through the three other sites using the same calibration methods. Results showed that all of the ESMs could capture the dynamics of in situ GST with high correlations (r > 0.90). However, large errors were detected with a broad range of centred root-mean-square errors (1.14–4.98 °C). The Top 5 model ensembles (MME5) performed better than most individual ESMs and averaged multi-model ensembles (MME19). The calibrated GST performed better than the GST obtained from MME5. Both annual and seasonal GSTs exhibited warming trends with an average annual rate of 0.04 °C per decade in the annual GST. The average seasonal warming rate was highest in winter and spring and lowest in summer. This reconstructed GST data series could be used to simulate the long-term permafrost temperature over the QTP.