Rethinking Global Groundwater Models to Feed Crops and Protect Ecosystems

Groundwater is the invisible workhorse of global agriculture. Aquifers supply roughly half of the water used for irrigation worldwide, underpinning harvests that feed billions of people, yet the same hidden reserves sustain springs, wetlands, rivers and the ecosystems that depend on them. A new perspective published in Nature Water by Inge de Graaf of Wageningen University & Research and colleagues argues that the scientific community has become remarkably good at documenting how much groundwater the world is losing, but remains surprisingly poor at answering the question that actually matters for policy: how can this finite resource be used sustainably, at the same time, for both food production and a healthy environment?

The authors’ central claim is that global groundwater modelling has delivered an impressive diagnostic picture of depletion but a limited understanding of sustainable use. Over the past decade, large-scale hydrological models have mapped where water tables are falling fastest, from the High Plains aquifer in the United States to the Indo-Gangetic plain and the North China Plain. Studies such as the 2019 analysis led by de Graaf in Nature demonstrated that groundwater depletion is not a local curiosity but a global phenomenon, with falling water tables affecting streamflow and ecosystems across entire river basins. Complementary work by researchers including Bridget Scanlon and Carole Dalin has quantified the dependence of food systems on non-renewable aquifer storage, linking over-pumping to national food security and international food trade. Yet, as the Nature Water perspective points out, these efforts largely stop at the diagnosis.

The gap, according to the authors, lies in how the models are built. Most global assessments treat groundwater as a reservoir to be withdrawn, tracking the volume of water removed from storage. What they rarely capture is the full set of processes that connect pumping to the things society actually cares about: the yield of a wheat field under a given irrigation schedule, the baseflow that keeps a river alive through a dry summer, the depth at which a wetland’s vegetation can no longer reach the water table, or the feedback between soil moisture, evapotranspiration and crop physiology. Without those process-based links, a model can say that an aquifer is being depleted, but it cannot say what will happen to food production or to ecosystems if pumping is reduced by ten, twenty or fifty percent, nor where a given reduction would do the most good or the most harm.

To close this gap, de Graaf and her co-authors, including Anton Urfels of the International Rice Research Institute and Cornell University, Fatemeh Karandish, Sida Liu and Iwan Supit of Wageningen, propose a shift towards integrated, process-based frameworks that support decision-making across scales. In practical terms, this means coupling global groundwater models with crop growth models, ecohydrological representations of surface-water dependency, and the socio-economic realities of farming decisions. Crop models of the kind represented in the literature by work such as that of Joost Kroes on agrohydrology and Joseph Amthor on crop physiology simulate how plants respond to water stress, allowing researchers to translate a change in water availability directly into a change in yield. Ecohydrological theory, advanced by researchers such as Tom Gleeson, Mark Cuthbert and Donald Perrone, describes how aquifers exchange water with rivers and groundwater-dependent ecosystems, and how slowly those connections recover once disturbed.

The perspective also stresses that scale matters in ways current approaches handle poorly. A pumping decision is made by an individual farmer, in a specific field, in a particular season, often in response to rainfall that arrived or failed to arrive weeks earlier. Global models, by contrast, operate on grid cells of tens to hundreds of kilometres and on annual or monthly time steps. Between those two extremes lies the entire chain of cause and effect that determines whether groundwater use is sustainable. The authors argue that frameworks capable of moving information across these scales, from field-level agronomy up to basin-level water allocation and back down again in the form of policy-relevant guidance, are essential if groundwater governance is to be based on evidence rather than on crisis response.

Climate change sharpens the urgency of this argument. The Intergovernmental Panel on Climate Change has documented intensifying droughts and shifting precipitation patterns, and groundwater is increasingly treated as the buffer that will carry agriculture through a more volatile climate. But the perspective warns that treating aquifers as a climate insurance policy without understanding the environmental consequences risks simply transferring one crisis into another. When water tables fall, streams that depend on groundwater discharge lose flow, groundwater-dependent ecosystems degrade, and in coastal regions the void left by fresh water can be invaded by salt. Recent work by de Graaf and colleagues published in Environmental Research Letters in 2024, and by Cuthbert and colleagues in Nature Climate Change in 2019, has shown how slowly these systems respond to changes in pumping, with time lags that can span decades or centuries. Decisions made today may lock in environmental damage long after the agricultural benefits have been realised.

The food security dimension is equally stark. The Food and Agriculture Organization’s State of Food and Agriculture reports, along with analyses by Matti Kummu and colleagues on global food security, highlight how much of the world’s food production depends on water supplies that are under stress. In regions such as Iran, where Karandish and colleagues have published detailed analyses in the Journal of Hydrology in 2025, agriculture draws heavily on groundwater in semi-arid conditions where surface water is unreliable. In such settings, the trade-off between pumping for crops and preserving environmental flows is not an abstraction; it plays out every growing season in the choices farmers make about how much to irrigate and where. The authors argue that integrated frameworks can make these trade-offs explicit, quantifying the yield gained per unit of environmental harm and identifying management options, such as adjusted cropping patterns, deficit irrigation or targeted aquifer recharge, that shift the balance.

What distinguishes this perspective from a simple call for more modelling is its emphasis on decision support. The authors are explicit that the goal is not a bigger or more detailed simulation, but a framework that produces information usable by the people who allocate water: basin authorities designing pumping quotas, agricultural agencies advising on crop choices, and international bodies assessing food security risks. That requires models that represent adaptation itself, the ways farmers and water managers respond to scarcity, as part of the system rather than as an afterthought. It also requires honest treatment of uncertainty, since projections of future recharge, crop demand and ecosystem response all carry substantial error bars that decision-makers need to understand rather than ignore.

The work is grounded in a broader research programme. De Graaf leads a project funded by the European Research Council under the European Union’s Horizon Europe programme, an ERC Starting Grant that supports the development of the integrated ideas presented in the paper. The perspective draws on a body of literature spanning hydrology, agronomy, environmental science and sustainability studies, from Konikow’s foundational estimates of global groundwater depletion to the recent syntheses of the field’s state of the art. In pulling these threads together, the authors are effectively issuing a roadmap for the next generation of groundwater science: less emphasis on measuring the hole we are digging, more emphasis on understanding the machinery that connects that hole to our food supply and our ecosystems.

The stakes could hardly be higher. Groundwater depletion is already reducing crop yields in some of the world’s most productive agricultural regions, and the ecosystems that depend on aquifers are among the fastest-disappearing on Earth. If the authors are right, the tools that revealed the problem are not, in their current form, the tools that will solve it. Building integrated, process-based frameworks that link the water table to the wheat field and the wetland, and that work at the scales where decisions are actually made, is presented as the necessary next step. Whether the research community and the policy world can build and apply such frameworks fast enough may help determine how much of the world’s hidden water, and the food and ecosystems it sustains, survives the coming decades.

Subject of Research: Integrated process-based groundwater modelling for balancing agricultural irrigation with environmental sustainability

Article Title: Balancing groundwater for crop production and a healthy environment

Article References: de Graaf, I., Urfels, A., Karandish, F., Liu, S., & Supit, I. (2026). Balancing groundwater for crop production and a healthy environment. Nature Water. https://doi.org/10.1038/s44221-026-00720-y

Image Credits: AI Generated

DOI: 10.1038/s44221-026-00720-y

Keywords: groundwater, depletion, irrigation, crop production, food security, hydrology, ecosystems, water management, global modelling, climate change, aquifers, sustainability