A new 2010 permafrost distribution map over the Qinghai–Tibet Plateau based on subregion survey maps: a benchmark for regional permafrost modeling

Abstract

Permafrost over the Qinghai–Tibet Plateau (QTP) has received increasing attention due to its high sensitivity to climate change. Numerous spatial modeling studies have been conducted on the QTP to assess the status of permafrost, project future changes in permafrost, and diagnose contributors to permafrost degradation. Due to the scarcity of ground stations on the QTP, these modeling studies are often hampered by the lack of validation references, calibration targets, and model constraints; however, a high-quality permafrost distribution map would be a good option as a benchmark for spatial simulations. Existing permafrost distribution maps for the QTP can poorly serve this purpose. An ideal benchmark map for spatial modeling should be methodologically sound, of sufficient accuracy, and based on observations from mapping years rather than all historical data spanning several decades. Therefore, in this study, we created a new permafrost distribution map for the QTP in 2010 using a novel permafrost mapping approach with satellite-derived ground surface thawing and freezing indices as inputs and survey-based subregion permafrost maps as constraints. This approach accounted for the effects of local factors by incorporating (into the model) an empirical soil parameter whose values were optimally estimated through spatial clustering and parameter optimization constrained by survey-based subregion permafrost maps, and the approach was also improved to reduce parametric equifinality. This new map showed a total permafrost area of about 1.086×106 km2 (41.2 % of the QTP area) and seasonally frozen ground of about 1.447×106 km2 (54.9 %) in 2010, excluding glaciers and lakes. Validations using survey-based subregion permafrost maps (κ=0.74) and borehole records (overall accuracy =0.85 and κ=0.43) showed a higher accuracy of this map compared with two other recent maps. Inspection of regions with obvious distinctions between the maps affirms that the permafrost distribution on this map is more realistic than that on the Zou et al. (2017) map. Given the demonstrated excellent accuracy, this map can serve as a benchmark map for constraining/validating land surface simulations on the QTP and as a historical reference for projecting future permafrost changes on the QTP in the context of global warming. The dataset is available from the repository hosted on Figshare (Cao et al., 2022): https://doi.org/10.6084/m9.figshare.19642362.