Abstract Groundwater, Earth’s largest source of liquid freshwater, is essential for sustaining ecosystems and meeting societal demands. However, quantifying global groundwater withdrawals remains a significant challenge due to inherent uncertainties in input data, sectoral allocation assumptions, and model parameterization. In this study, we analyze global groundwater withdrawals from 2001 to 2020 using a newly developed data-driven Global Groundwater Withdrawal (GGW) model and quantify uncertainties through Monte Carlo simulations. The GGW model integrates reported country-level data with global grid-based datasets to estimate annual withdrawals across domestic, industrial, and agricultural sectors at a 0.1° resolution (≈ 10 km). Our results indicate an average global groundwater withdrawal of 648 km³ a⁻¹, with an uncertainty range of 465-881 km³ a⁻¹. Agriculture accounts for 50% of total withdrawals, followed by domestic use at 34.5% and industrial use at 15.5%. Temporal analysis shows increasing groundwater withdrawal in 66% of the 44 IPCC WGI reference regions over the 20 years, with a global average annual increase of 0.5% (varying regionally from 6.5% annual increase to 9% annual decrease). Comparison with previous studies highlights the impact of methodological choices and assumptions about groundwater withdrawal on the resulting global estimates. Our findings underscore the need for comprehensive uncertainty assessments and improved datasets. Expanding spatial coverage in underrepresented regions and enhancing temporal resolution, particularly for dynamic variables like irrigated areas, are crucial for more accurate groundwater withdrawal assessments. These improvements will enable better management and conservation of this vital resource in the face of growing global demands and climate change impacts.
Nazari et al. (2025) studied this question.