Unveiling the ~three-year oscillation in terrestrial water storage change: detection and attribution analysis

Unveiling the ~three-year oscillation in terrestrial water storage change: detection and attribution analysisXiaolong Li1,2, Taoyong Jin1,3,*, Jürgen Kusche2, Benjamin Daniel Gutknecht2, Mingyu Xiao1, Charlotte Hacker2, Fupeng Li21 School of Geodesy and Geomatics, Wuhan University, Wuhan 430079, China2 Institute of Geodesy and Geoinformation, University of Bonn, 53115 Bonn, Germany3 Key Laboratory of Geospace Environment and Geodesy, Ministry of Education, Wuhan 430079, ChinaCorresponding author: Taoyong Jin ([email protected])Key Points:Significant ~3-year oscillation is identified in terrestrial water storage change~3-year oscillation is driven by Earth’s surface envelope, not interior dynamicsLinear combination of 4 major climate modes explains the ~3-year oscillation wellAbstractAnalyzing the interannual variability in terrestrial water storage change (TWSC) is crucial for managing water resources and addressing climate change impacts. Previous studies have examined the response of water storage to climate variability, and discussed a six-year oscillation that distributed in water storage in many regions. However, this study identifies a significant ~three-year oscillation (TYO) in TWSC at global and regional scale, revealing its unique driving factors. We identify the TYO in four independent TWSC datasets. Analysis of the length of day and geomagnetic fields indicates that the TYO is unrelated to Earth’s interior dynamics and is not a harmonic of the six-year oscillation. Further analysis confirms the presence of TYO in precipitation and runoff, and also in five ocean-atmosphere coupling system variables and four major climate modes, indicating its manifestation across a wide range of Earth’s external envelope variables. Using multiple linear regression to isolate climate mode effects, we find that climate variability plays a pivotal role in driving the TYO in TWSC, precipitation, and runoff. Basin-scale analysis consistently supports these findings, with South American basins exhibiting the strongest TYO influence.1 IntroductionSince the launch of the Gravity Recovery and Climate Experiment and its follow-on mission (GRACE-/FO), terrestrial water storage anomaly (TWSA) has provided unprecedented insights into variations in total water storage on and beneath the land surface, integrating components such as soil moisture, groundwater, snow, and surface water (Tapley et al., 2004; Rodell et al., 2018; Gerdener et al., 2023). Derived from the temporal differences in TWSA, terrestrial water storage change (TWSC) reflects the net balance among precipitation, evapotranspiration, and runoff, which are the major fluxes within the terrestrial water cycle, and is thus more suitable than TWSA for diagnosing water balance components and their responses to climatic forcing (Syed et al., 2008; Chen et al., 2018). Analyzing the interannual variability of TWSC is particularly crucial for advancing our understanding of long-term water cycle variability, assessing the impacts of climate change on regional and global water availability, and uncovering linkages between hydrology and the Earth system processes (WCRP Global Sea Level Budget Group, 2018; Tapley et al., 2019; Wu et al., 2025).Earth’s climate variability on interannual and decadal scales is primarily influenced by large-scale, periodic shifts in the coupled ocean-atmosphere system, such as the El Niño-Southern Oscillation (ENSO), and these climate modes are expected to play a role in shaping the dynamics of water storage (Martens et al., 2018). Phillips et al., (2012) found that storage anomalies in some regions showed high correlation with ENSO index, and tropical regions show strong negative correlation and arid regions exhibit positive correlation. Ni et al., (2018) suggested that the interannual water storage changes are strongly correlated with ENSO over much of the globe, with maximum cross-correlated coefficients up to about 0.70. Guo et al., (2021) discussed the link between 14 main climate modes and water storage anomalies in detail, and found that the four global important patterns of coupled atmosphere-ocean variability, i.e. ENSO, the Pacific Decadal Oscillation (PDO), the Atlantic Multidecadal Oscillation (AMO), and the Indian Ocean Dipole (IOD) affect variation in water storage in 76.5%, 74.6%, 59.7%, and 46.4% of the global land surface, respectively. These studies typically examine the entire interannual frequency band, while in recent years, some studies have identified significant oscillations within TWSC. For example, Pfeffer et al. (2023) detected a prominent ~six-year oscillation in TWSC from GRACE/-FO with global prevalence and distinct regional variations. This signal is consistent with oscillations in Earth system components such as the length-of-day (Abarca Del Rio et al., 2000), geomagnetic field (Gillet et al., 2010), global mean sea level, precipitation, and surface temperature (Moreira et al., 2021). However, the origins of this ~six-year oscillation and its implications for deep Earth-climate coupling remain debated (Cazenave et al., 2023).In contrast, much less attention has been paid to relatively higher-frequency oscillations, such as the ~three-year oscillation (TYO), despite emerging evidence of its widespread presence in the terrestrial water cycle and related climate modes. For instance, Bruun et al. (2017) detected significant 2.5- and 3.8-year oscillations in the Southern Oscillation Index from 1951 to 2016, while Li et al. (2013) identified similar oscillations in the Niño 3.4 index over a long historical period (since early 14th century). Additionally, as shown by Pfeffer et al. (2022), precipitation and TWSC in regions such as the Central Amazon Corridor, North Australia, and the Indian Northern Plains exhibit significant TYOs. Global datasets, including surface temperature (Pfeffer et al., 2023) and GMSL (Moreira et al., 2021), also show consistent TYOs. Schmidt et al. (2008) also identified significant periods in the range of 2.1 to 2.5 years (beyond annual and semiannual) in GRACE and WaterGAP data, though the shorter GRACE record at the time precluded identifying longer periods. These findings suggest that the TYO is pervasive across Earth’s surface envelope and may be closely linked to ocean–atmosphere coupling processes within it.In summary, exploring the TYO of TWSC and conducting attribution analysis are important steps toward improve our understanding of the interannual variability in terrestrial water cycle and reveals the potential impact of climate variability. In Section 2, we outlined the Data sets and methods. Results, discussions, and conclusions are shown in Section 3, Section 4, and Section 5.2 Data and Methods2.1 Terrestrial water storage change