Terrestrial water storage (TWS) is generally defined as the sum of all water resources stored on the land surface and in the subsurface (Figure 1). It is an essential component of the hydrological cycle that determines freshwater availability on Earth[1]. In regions experiencing frequent water scarcity, TWS often provides the last critical buffer against extreme droughts through natural water storage systems such as underground aquifers, glaciers, and wetlands.

Figure 1. Illustration of terrestrial water storage (TWS).

Graphic by Andreas Hendrich, GFZ: https://www.globalwaterstorage.info/en/anwendungen/terrestrial-water-storage

The global TWS is, however, being threatened by multi-dimensional natural and human factors[2],[3]. It is affected by climate change which accelerates the hydrological cycle through enhancing evapotranspiration and altering precipitation patterns across the globe. The changed hydrological cycle is exacerbated by accumulated daily human activities at various levels. For instance, the impoundment of fresh water in reservoirs, land use and land cover change, and irrigation for food production exacerbate the TWS variations in time and space contributing to changes in water and energy budgets. These natural and human-induced variations in TWS exert additional pressures on freshwater resources worldwide, which in turn affect the health of ecosystems and the daily livelihood of society at large, especially in those water-stressed regions[4].

It is interesting if climate change signals can be detected in the TWS, which may help differentiate human contributions from natural variability in climate change. However, the challenge remains, because it is difficult to obtain long-term trends in global TWS due to limited global observation records, decadal climate variability, highly uneven hydrological responses to climate change, and complicated interactions between climate, hydrological, and social systems[5]. Although the long-term trends are normally required to detect a signal, detecting climate change signals at multiple temporal (from daily to century) and spatial scales (from local to global), and from different perspectives, will contribute to fully understanding the evolving patterns of climate change over time and across space that would further help identify critical drivers and effective adaptation to climate change. Likewise, it is interesting to apply such idea to the TWS domain and to even the shorter timescales at the global spatial pattern. 

A recent study, building upon the “fingerprints” – the conception of human effects on climate that is based on detection and attribution techniques – was conducted to identify the relationship between annual global mean TWS and the daily surface air temperature and humidity fields using multi-model hydrological simulations. The study followed a previously established machine learning approach that makes use of statistical learning techniques[6], but extended the focus to extract externally forced “fingerprints” from global weather patterns of daily surface air temperature and specific humidity.

The overall finding highlights that approximately 50% of days for most years since 2016 have climate change signals emerged above the noise of internal variability globally. Climate change signals in global mean TWS have been consistently increasing over the last few decades, and in the future, are expected to emerge from the natural climate variability.

The temperature fingerprint reveals the dominance of large negative regression coefficients over the tropical lands, suggesting a stronger signal of the annual global average of terrestrial water storage (AGTWS) change in these regions embedded daily temperature fluctuation compared to westerlies dominated regions. On the other hand, the spatial patterns of humidity fingerprint show more evenly distributed magnitudes compared to the temperature fingerprint, indicating a more homogeneous distribution (Figure 2). Negative coefficients are observed over most of the land areas in the Northern Hemisphere and along the Equator. The results also show a broad congruency between the two sets of AGTWS estimates over different intervals, especially from the early 1960s onwards when decreasing trends are perceptible in both models (Figure 3a) and the average of reanalyses (Figure 3b).

Figure 2. The annual average fingerprint of external forcing and its variations at latitudinal bands

Figure 3. AGTWS estimates from models and reanalysis datasets

Additionally, strong weak signals of human-induced climate change in AGTWS do not emerge from natural variability at global and continental scales in the historical climate (HIST, 1861-2005) simulations. However, the detected fraction of days started to climb in 2015. Climate change signals in global TWS have accumulated over the last few decades, and the results showed that these signals will possibly be uniformly detected at daily timescales in the next few decades. The signal of forced climate change under RCP6.0 emerges later than that under RCP2.6 due to higher climate variability under higher emission scenarios. The AGTWS is expected to decrease by about 40 kg m⁻² under RCP6.0, compared to about 20 kg m⁻² under RCP2.6.

This study has provided multiple implementations, both academically and practically. It implies the possibility of detecting climate change signals at the earliest possible moment using daily global data. It also reveals that the already limited freshwater availability will likely become scarcer around the globe as a significant decline in the annual global average of TWS in response to global warming. The declining freshwater storage under climate change will jeopardize water ecosystem services worldwide, exacerbating water-related risks on human and natural systems. All these findings are calling for the immediate action for sustainable water futures, battling against the climate change.

[1] Rodell, M., & Famiglietti, J. S. (2001). An analysis of terrestrial water storage variations in Illinois with implications for the Gravity Recovery and Climate Experiment (GRACE). Water Resources Research, 37(5), 1327-1339.

[2] IPCC. Climate Change 2022: Impacts, Adaptation and Vulnerability. Cambridge University Press, 609 Cambridge, UK and New York, USA, 2022.

[3] Rodell, M., Famiglietti, J. S., Wiese, D. N., Reager, J. T., Beaudoing, H. K., Landerer, F. W., & Lo, M. H. (2018). Emerging trends in global freshwater availability. Nature, 557(7707), 651-659.

[4] Xu, L., & Famiglietti, J. S. (2023). Global patterns of water‐driven human migration. Wiley Interdisciplinary Reviews: Water, 10(4), e1647.

[5] Tapley, B. D., Watkins, M. M., Flechtner, F., Reigber, C., Bettadpur, S., Rodell, M., ... & Velicogna, I. (2019). Contributions of GRACE to understanding climate change. Nature climate change, 9(5), 358-369.

[6] Sippel, S., Meinshausen, N., Fischer, E. M., Székely, E., & Knutti, R. (2020). Climate change now detectable from any single day of weather at global scale. Nature climate change, 10(1), 35-41.