Hydrological change is one of the most critical driving forces that affect surface erosion, weathering and civilizational evolution. However, reconstructing past hydrology remains a major challenge, as reliable hydrological proxies are scarce. Traditional paleoclimate proxies, such as δ¹⁸O, often incorporate temperature effects, making it difficult to isolate signals of hydrological change.

Recent studies have proposed that river water δ⁷Li is primarily controlled by hydrological processes, showing a negative correlation with runoff (Zhang et al., 2022; Pogge von Strandmann et al., 2023), implying that δ⁷Li preserved in lake sedimentary records may record past hydrological changes. However, whether δ⁷Li can be reliably used to reconstruct past hydrology has not yet been systematically assessed, as constraints on the relationship between dissolved δ⁷Li and hydrology fingerprints from sedimentary archives (e.g., lake level) are still required. This has hindered the development of Li isotopes as a novel tool for reconstructing past hydrological changes.

Now, a new study led by researchers from the Institute of Earth Environment of the Chinese Academy of Sciences (IEECAS) focused on the Muztag Glacier catchment in the tectonically active Pamir Plateau—known as the “Father of Glaciers”. Using high-resolution weekly sampling, the team systematically investigated seasonal behaviors in riverine δ⁷Li and their relations with hydrological changes.

The results show remarkable seasonal variations in riverine δ⁷Li, with values decreasing by up to ~15.4‰ from the dry to wet seasons. Such a large seasonal difference has not been observed previously.

More importantly, the riverine δ⁷Li unexpectedly exhibit a clear hydrological control, showing a significant relationship between δ⁷Li and lake water levels (r²=0.61). Specifically, higher δ⁷Li values correspond to lower lake levels, whereas lower δ⁷Li values occur during periods of higher lake levels. In contrast, no clear correlation is observed between δ⁷Li and temperature.

Significant correlations between δ⁷Li and lake levels, riverine ⁸⁷Sr/⁸⁶Sr and Al/Si ratios suggest that hydrologically driven variations in water–rock interaction time affect secondary mineral formation and seasonal riverine δ⁷Li (up to ~15‰). In detail, fast surficial runoff with short water–rock interaction time in the wet season facilitates more congruent weathering, resulting in rock-like riverine δ⁷Li. In contrast, slow flow during the dry season prolongs water–rock interaction time, promotes secondary mineral formation and higher δ⁷Li values.

Furthermore, the study compared the potential of classic δ¹⁸O and new δ⁷Li as hydrological proxies, and found that riverine δ¹⁸O reflects temperature in westerly regions and precipitation in monsoon regions, while δ⁷Li consistently reflects hydrological changes across both climate regions, as well as in other global climate settings.

These findings provide new insights into the controls on riverine δ⁷Li and its potential applications: (1) Seasonal variations in riverine δ⁷Li can reach up to ~15‰, representing the largest seasonal amplitude reported so far. (2) Even in tectonically active and highly erosive glacial environments, δ⁷Li can still respond rapidly and sensitively to hydrological changes. (3) The study quantitatively establishes the relationship between dissolved δ⁷Li and actual lake-level variations. This provides the first field-based evidence for using δ⁷Li (e.g., from lacustrine carbonates) in sedimentary archives to trace past hydrologic state in the future. (4) Compared with the traditional proxy δ¹⁸O, lithium isotopes emerge as a more sensitive indicator of hydrological variability and can effectively exclude the influence of temperature.

Therefore, δ⁷Li exhibits its potential to become a new and reliable proxy for reconstructing past hydrological changes, opening a new window for hydrological reconstruction. Given the close links between hydrological variability and fields such as paleoclimate reconstruction, extreme drought and flood events, the rise and fall of civilizations, vegetation succession, and the weathering–carbon cycle, this new proxy is expected to have broad applications in future research.

This work, published in Geochimica et Cosmochimica Acta, was supported by the National Natural Science Foundation of China, the International Partnership Program of Chinese Academy of Sciences for Future Network, and other funding sources.