Robert Reinecke , Sebastian Gnann, Lina Stein , Marc Bierkens, Inge de Graaf, Tom Gleeson, Gualbert Oude Essink, Edwin Sutanudjaja, Claudia Ruz-Vargas, Jarno Verkaik, Thorsten Wagener 

At what depth groundwater can be found below the land surface is key to understanding whether it is
potentially accessible to ecosystems and humans, or what role it plays in the water cycle. Knowledge of
ground-water table depth (WTD) exists at regional scales in many places, but a bottom-up knowledge
aggregation to obtain a coherent global picture is exceptionally challenging. Uncertainty in global-scale
WTD knowledge severely affects our ability to assess groundwater’s future role in a water cycle altered by
changes in climate, land use, and human water use. Global groundwater models offer a top-down pathway
to gain this knowledge. However, we find them highly uncertain: four models investigated show WTD
disagreements of more than 100 m for one-third of the global land area. Averaged across the models, we
estimate that 23% [most deviating model: 71%] of the land area contains shallow groundwater potentially
accessible to ecosystems and humans, <10m depth, 57% [29%] is potentially accessible to humans through
pumping, 10-100m, while 20% [0.01%] is potentially too costly to access or inaccessible, >100m.
Depending on the model, +-63% of global forest coverage and +-54% of irrigated land is inside areas of
potentially ecosystem-accessible water, and +-33% of the global population lives in areas with potentially
human-accessible groundwater. These results add significant uncertainties to any global-scale analysis,
which will not significantly reduce without dedicated efforts. We outline three pathways to reduce this
uncertainty through better global datasets, alternative strategies for model evaluation, and greater
cooperation with experts.