Against the backdrop of global change, the water cycle is being reshaped. How does precipitation enter the soil? How is soil water absorbed by plants? Does deep soil water contribute to groundwater recharge? These questions may seem disparate, yet they all point to the same critical medium—the soil water.
Recently, the team led by Professor Zhu Guofeng from Northwest Normal University published a data paper in Scientific Data, presenting and releasing a global-scale dataset of stable isotopes in soil water. This dataset systematically integrates observations of stable hydrogen and oxygen isotopes in soil water collected from 1975 to 2024, comprising a total of 27,455 records. It covers six continents, 37 countries, and 463 observation sites, providing an essential data foundation for understanding global soil water cycling processes, calibrating hydrological models, and studying vegetation strategies for water use.
Figure 1. (a) Spatial distribution of sampling points in the global soil water stable isotope dataset; (b) Temporal distribution of records; and (c) Distribution classified by category.
Where does the data come from?
Data sources are categorized into three paths:
Literature extraction (15,051 records): A systematic search was conducted across Web of Science, Scopus, and Google Scholar, and data were extracted from 262 peer-reviewed papers spanning the period from 1975 to 2024.
Open Database (973 records): From the Water Isotopes Open Repository;
In-situ measured data (11,431 records): This dataset was continuously monitored by the Shiyang River Basin Observatory at Northwest Normal University starting from 2015 and extending through 2024, accounting for approximately 41.6% of the global dataset.
Technical validation
To ensure cross-regional comparability, the team has uniformly standardized and quality-controlled historical data, and implemented standardized management across the entire process—from sampling to analysis—for the measured data.
The study employed a commonly used hierarchical scheme in the field of eco-hydrology to categorize all records into four standard profile intervals: shallow soil water (0–10 cm), root-active layer (10–40 cm), middle soil layer (40–100 cm), and deep soil layer (>100 cm).
To address the issue of inconsistent sampling depths, the research team employed median-depth matching and a “shallow-first” rule to enhance the comparability of data from different soil layers. Most soil water samples were obtained using vacuum low-temperature condensation extraction. Isotope measurements were primarily conducted using laser spectrometers and isotope ratio mass spectrometers, thereby ensuring data quality at the methodological level.
In the in-situ field measurement and laboratory analysis, soil water extraction was performed using the LI-2100 Automated Vacuum Water Extraction System (Beijing LICA United Technology Ltd.). The extraction conditions were as follows: a high-vacuum environment of ≤10 Pa, a constant temperature of 95℃, and continuous extraction for 2–3 hours to ensure thorough recovery of water. Subsequently, the samples were analyzed using a liquid water isotope analyzer to determine δ²H and δ¹⁸O. Each sample was injected six times; the first two injections were discarded to eliminate memory effects, and the average of the remaining four injections was taken as the final result.
Figure 2. Global average spatial distribution map of soil–water stable isotope profiles. This figure shows the δD and δ¹⁸O values (‰) measured at global observation sites. The color gradient of the circles represents the arithmetic mean isotope composition across the entire vertical soil profile at each observation site.
Figure 3. The relationship between δ²D and δ¹⁸O at different soil depths.
What results does this dataset present?
In terms of spatial coverage, this dataset encompasses a variety of typical climate zones and ecosystems worldwide, and it fills data gaps in traditionally under-observed regions such as northern Russia and Alaska.
In terms of time, the data cover the period from 1975 to 2024, and the number of records has grown rapidly since 1990, reflecting that soil water isotope research is gradually shifting from localized observations toward comprehensive global-scale analyses.
In terms of isotopic composition, the global range of δ²H in soil water is approximately from -249.7‰ to 4.62‰, while the range of δ¹⁸O is roughly from -25‰ to 15‰, reflecting significant spatial variations shaped jointly by diverse climatic, topographic, soil, and hydrological processes.
In the three different soil layers, both δ²H and δ¹⁸O showed a significant positive correlation; however, the strength of the correlation and the slope varied with depth, indicating that soil water at different depths is influenced differently by processes such as evaporation, precipitation input, root water uptake, and groundwater recharge.
After integrating multi-source data, this dataset boasts advantages such as broad spatial coverage, long time series, large sample size, and standardized metrics, providing a solid empirical foundation for global research on the soil water cycle.
Conclusion
This study not only constructs a global dataset of stable isotopes in soil water that boasts broad coverage, long temporal span, and a large sample size, but also transforms long-term, scattered observational data with varying formats into standardized, traceable, and reusable data resources. This dataset can provide empirical support for source apportionment of soil water, identification of vertical migration processes, studies on plant water use, and assessment of groundwater recharge. It also lays a solid data foundation for hydrological model calibration, cross-regional comparisons, and global change research. Although soil water remains hidden underground, it is deeply interconnected with precipitation, vegetation, groundwater, and terrestrial ecosystems. Stable isotopes serve as crucial “fingerprints” that reveal this elusive hydrological cycle.
Data acquisition channels:
Global Soil Water Stable Isotope Dataset: https://data.mendeley.com/datasets/9vzhz97m5j/332;
Journal of Publication: Scientific Data [Impact Factor: 7.2]
Research institutions: Northwest Normal University, Gansu Province Key Laboratory of Oasis Resources, Environment and Sustainable Development, and others.
Study location: Global scale; in-situ measured data primarily come from the Shiyang River Basin Observatory at Northwest Normal University.
Equipment used:LI-2100 Automated Vacuum Water Extraction System
DOI: https://doi.org/10.1038/s41597-026-07262-8