Climate extremes like devastating droughts and floods are reshaping our world, making it more crucial than ever to understand how water resources respond. This is where GRAiCE, presented in our recent paper in Scientific Data, steps in. Our newly developed dataset is a groundbreaking resource for monitoring global freshwater storage dynamics. By offering a fresh perspective on changes in water availability, it advances global efforts in understanding and managing the challenges posed by climate change on freshwater systems.

What is GRAiCE and why does it matter?

GRAiCE is a dataset designed to measure changes in freshwater storage worldwide. It represents a game-changer for freshwater research helping track how much water is “lost” or “gained” in response to climate dynamics, especially during extreme events like droughts and floods.

While satellite observations from the GRACE and GRACE-FO missions provide invaluable insights into terrestrial water storage anomalies (TWSA), they come with limitations. Temporal gaps and the absence of data before 2002 hinder the ability to assess long-term changes. GRAiCE bridges this gap by reconstructing global freshwater storage changes from 1984 to 2021 using state-of-the-art machine learning models (see Video below).

TWSA evolution (1984–2021) generated by BiLSTM predictions from GRAiCE.

The science behind GRAiCE

At the heart of GRAiCE are advanced machine learning models trained on GRACE/GRACE-FO satellite data and supplemented with climate and vegetation dynamics. Using Long Short-Term Memory (LSTM) and Bidirectional LSTM (BiLSTM) neural networks, we captured the complex dependencies within time-series data, enabling accurate reconstructions of TWSA. These models estimate changes at a spatial resolution of 0.5° globally, revealing how water systems respond to climatic drivers over time (Figure 1).

What GRAiCE can do

GRAiCE accurately reproduces observed freshwater storage changes at both global and river basin scales. A key strength of the dataset is its ability to highlight the influence of extreme climatic events, such as El Niño and La Niña phases, on freshwater resources. During El Niño years, the dataset reveals characteristic patterns like increased rainfall over the tropical Pacific and severe precipitation deficits in tropical continental regions. For example, South America often faces intense flooding in the south and severe droughts in the north, while the 2015–2016 El Niño caused a significant decline in water storage across Southern Africa due to extreme drought. On the other hand, La Niña events generally produce opposite effects, such as the substantial rise in water storage observed in eastern Australia during the 2010–2011 La Niña (Figure 2). Understanding these patterns is crucial for helping communities prepare for water scarcity or flooding before they strike.

GRAiCE goes beyond advancing scientific knowledge: it is a powerful resource with real-world applications. By detecting climate-driven changes in freshwater storage and uncovering long-term shifts in the global hydrological cycle, GRAiCE offers critical insights into how water resources have responded to natural climate variability.

This information supports not only researchers, but also policymakers, urban planners, and water managers with actionable insights to address pressing challenges such as:

• Sustainable water resource management: in regions facing severe droughts or floods, GRAiCE can guide strategies to allocate water resources efficiently, ensuring access for communities, agriculture, and ecosystems.
• Risk assessment and adaptation planning: the dataset helps identify regions that are most vulnerable to water scarcity or flooding, enabling targeted interventions to strengthen resilience to future climate extremes.
• Climate and water governance: by providing a reliable record of past freshwater changes, GRAiCE supports informed decision-making to balance climate mitigation efforts with equitable water use, particularly in areas under stress from rapid urbanization or industrial growth.

As a result, GRAiCE empowers global efforts to adapt to and mitigate the effects of climate change on freshwater systems.

GRAiCE is essential for our planet's future

Developing GRAiCE was not without its challenges, from fine-tuning parameters and designing machine learning models to ensuring robust predictions. But the payoff was worth it. As the climate crisis intensifies, tools like GRAiCE are indispensable for building resilience. By better understanding freshwater dynamics, we can safeguard water resources for future generations.

Looking ahead

The GRAiCE dataset opens new avenues for advancing water-related research and management. The potential applications of this dataset range from improving water management strategies in drought-prone regions to supporting equitable and sustainable freshwater distribution worldwide.

By making GRAiCE openly available, we aim to encourage collaboration among scientists, policymakers, and water managers to address the growing challenges of climate extremes and human pressures on global water systems.

We hope GRAiCE will inspire further research and action toward a more sustainable management of freshwater resources, building resilience against future water challenges.

The GRAiCE dataset is publicly available at: https://doi.org/10.5281/zenodo.10953658