Quantifying permafrost ground ice contents in the Tien Shan and Pamir (Central Asia): a petrophysical joint inversion approach using a geometric mean model

Abstract

Abstract. In the Central Asian Tien Shan and Pamir mountain ranges, permafrost is extensive, but in situ data on permafrost remain scarce. Quantitative analysis of permafrost's subsurface components – ice, water, air, and rock – is vital for not only discerning the impact of climate change on increased slope instability due to permafrost degradation, but also understanding its role as a potential water resource in high-altitude environments. Recent studies have employed a petrophysical joint inversion (PJI) approach combining geoelectrical and seismic refraction data to model the subsurface's four phases (fractions of air, water, ice, and rock). However, most of these studies primarily rely on Archie’s law, which has limitations in coarse blocky substrates typical of mountainous terrains. Recognizing this limitation, the electrical geometric mean (PJI-GM) model may be used as an alternative implementation within PJI. In this study, we assess the suitability of using the PJI-GM model across an extensive geophysical dataset comprising 22 profiles in Central Asia (Kyrgyzstan and Tajikistan). Our goals are to (i) address the existing data gap concerning mountain permafrost and ground ice contents in the Tien Shan and Pamir of Central Asia and (ii) evaluate the performance of the PJI-GM model in comparison to Archie's law within the PJI framework across the different landforms at remote sites. The findings reveal that the ground ice content is more specific to landform types than to the different geographic regions surveyed, with rock glaciers exhibiting the highest mean ice contents (38 %–60 %), followed by moraines (18 %–40 %), talus slopes (20 %–40 %), and fine-grained sediments (0 %–20 %). The PJI-GM model performed especially well for ice-rich landforms such as rock glaciers, accurately reflecting high ice contents with minimal variability between different model runs. The quality of a model result was here assessed by comparing a multitude of different model runs with different sets of inversion parameters and petrophysical variables using a clustering approach. This research provides one of the first comprehensive (geophysical) in situ datasets on permafrost on various landforms and sites in Central Asia, highlighting the potential of the PJI-GM model as a more suitable alternative to Archie’s law, particularly for rock glaciers and other ice-rich landforms. These findings significantly advance our understanding of permafrost in the Tien Shan and Pamir and serve as a baseline dataset for future modeling studies.