Global Recharge Data Set Indicates Strengthened Groundwater Connection to Surface Fluxes

Abstract

Groundwater is an invaluable global resource, but its long-term viability as a resource for consumption, agriculture, and ecosystems depends on precipitation recharging aquifers. How much precipitation recharges groundwaters varies enormously across Earth's surface, yet recharge rates often remain uncertain. Here we use a global synthesis of field-estimated recharge across six continents to show that globally recharge first-order follows a simple function of climatic aridity. We use this relationship to estimate long-term recharge outside of permafrost regions. Our aridity-based recharge estimates are consistent with the global field data but, on average, double previous estimates of global models. The spatial and process-averaging conveyed by the recharge rates generated by hydrological models results in predictions that are underestimations at both high and low recharge rates. Our recharge estimates suggest that more groundwater must contribute to evapotranspiration and streamflow than previously represented by global hydrological models and global water cycle diagrams.

Full text

1. Introduction
Groundwater constitutes almost all of Earth's liquid fresh water (Abbott etal.,2019; Gleeson etal.,2016) and is
extensively extracted, with global withdrawals of hundreds of cubic kilometers per year (Döll etal.,2014; Margat
& Van der Gun,2013; Sutanudjaja et al., 2018). Groundwater provides approximately 2 billion people with
drinking water (Morris etal.,2003) and supplies almost 40% of irrigated lands worldwide (Siebert etal.,2010).
Groundwater also shapes ecosystems and landscapes as rivers and vegetation can source their waters from aqui-
fers (Berghuijs & Kirchner,2017; Evaristo & McDonnell,2017; Fan etal.,2017; Jasechko etal.,2016).
The dynamic roles of groundwater are not always apparent, but aquifers must be sufficiently recharged for
groundwater to sustain ecosystems and water resources into the future (Alley etal.,2002; Gleeson etal.,2012).
Earth's diversity of landscapes and climates results in groundwater recharge rates that vary by orders of magni-
tudes globally (MacDonald etal.,2021; Moeck etal.,2020; Scanlon etal.,2006). Yet, for most of Earth's surface,
groundwater recharge rates remain uncertain because measurements are sparse (Moeck etal., 2020; Scanlon
etal.,2006), and large-scale modeled recharge remains mostly unvalidated (de Graaf etal.,2015,2019; Döll &
Fiedler,2008; Li etal.,2021; Müller Schmied etal.,2021; Reinecke etal.,2021). In addition, upscaling recharge
estimates derived from extensively studied sites to other locations is challenging because many landscape, vege-
tation, and surface properties can affect recharge (Crosbie etal., 2018; De Vries & Simmers, 2002; Moeck
etal.,2020). These issues are problematic because accurate recharge estimates are needed to assess the sustain-
ability of groundwater use and the role of groundwater in supporting ecosystems and surface waters (Gleeson
etal.,2020).
Abstract Groundwater is an invaluable global resource, but its long-term viability as a resource
for consumption, agriculture, and ecosystems depends on precipitation recharging aquifers. How much
precipitation recharges groundwaters varies enormously across Earth's surface, yet recharge rates often remain
uncertain. Here we use a global synthesis of field-estimated recharge across six continents to show that globally
recharge first-order follows a simple function of climatic aridity. We use this relationship to estimate long-term
recharge in energy-limited systems outside of permafrost regions. Our aridity-based recharge estimates are
consistent with the global field data but, on average, double previous estimates of global models. Our higher
recharge estimates are likely caused by preferential groundwater recharge and discharge occurring at grid
scales finer than global models. The higher recharge estimates suggest that more groundwater contributes to
evapotranspiration and streamflow than previously represented by global hydrological models and global water
cycle diagrams.
Plain Language Summary Groundwater is an essential resource for societies and ecosystems.
The rate at which rainfall and snow replenish groundwater storage is important as it dictates the upper limit
of sustainable groundwater use. Here we use measurements of groundwater recharge to show how climate
determines groundwater recharge rates. Measured recharge rates, on average, strongly exceed those of
models. This suggests there is more recharge globally than currently acknowledged. Consequently, also more
groundwater recharge must get back to Earth's surface via river flow or water use of vegetation.
BERGHUIJS ETAL.
© 2022 The Authors.
This is an open access article under
the terms of the Creative Commons
Attribution-NonCommercial License,
which permits use, distribution and
reproduction in any medium, provided the
original work is properly cited and is not
used for commercial purposes.
Global Recharge Data Set Indicates Strengthened
Groundwater Connection to Surface Fluxes
Wouter R. Berghuijs1 , Elco Luijendijk2,3 , Christian Moeck4 , Ype van der Velde1 , and
Scott T. Allen5
1Department of Earth Sciences, Free University Amsterdam, Amsterdam, The Netherlands, 2Bundesgesellschaft Für
Endlagerung, Peine, Germany, 3Department of Earth Science, University of Bergen, Bergen, Norway, 4Department Water
Resources and Drinking Water, Eawag - Swiss Federal Institute of Aquatic Science and Technology, Dübendorf, Switzerland,
5Department of Natural Resources and Environmental Science, University of Nevada, Reno, NV, USA
Key Points:
• A global recharge data set indicates
that climate strongly shapes the
fraction of precipitation that will
recharge groundwaters
• This recharge data set indicates more
recharge globally than existing global
hydrological models suggest
• Thus, more groundwater must
contribute to evaporation and
streamflow than represented by
current global models and water cycle
diagrams
Supporting Information:
Supporting Information may be found in
the online version of this article.
Correspondence to:
W. R. Berghuijs,
w.r.berghuijs@vu.nl
Citation:
Berghuijs, W. R., Luijendijk, E., Moeck,
C., van der Velde, Y., & Allen, S. T.
(2022). Global recharge data set indicates
strengthened groundwater connection
to surface fluxes. Geophysical Research
Letters, 49, e2022GL099010. https://doi.
org/10.1029/2022GL099010
Received 6 APR 2022
Accepted 18 NOV 2022
Author Contributions:
Conceptualization: Wouter R. Berghuijs,
Christian Moeck, Ype van der Velde,
Scott T. Allen
Data curation: Wouter R. Berghuijs
Formal analysis: Wouter R. Berghuijs,
Elco Luijendijk
Investigation: Wouter R. Berghuijs
Methodology: Wouter R. Berghuijs
Resources: Wouter R. Berghuijs
Software: Wouter R. Berghuijs
Validation: Wouter R. Berghuijs
Visualization: Wouter R. Berghuijs
Writing – original draft: Wouter R.
Berghuijs
10.1029/2022GL099010
RESEARCH LETTER
1 of 9

Geophysical Research Letters
BERGHUIJS ETAL.
10.1029/2022GL099010
2 of 9
Regional analysis across carbonate rock landscapes indicates that many widely used hydrological models seem
to underestimate recharge (Hartmann etal.,2017). However, it remains unclear how widespread this model bias
is, as the enhanced recharge rates were attributed to the strong preferential flows in karst landscapes (Hartmann
etal., 2017). Yet, other evidence also suggests that global models overestimate the sensitivity of recharge to
climate change across arid regions in Africa because recharge induced by intense rainfall can lead to focused
recharge through losses from ephemeral overland flows, which are often not captured by large-scale models
(Cuthbert etal.,2019). Such discrepancies between models and observations are based on recharge and ground-
water observations across specific landscapes and climate conditions. Thus, it remains unclear how widespread
such issues are across other parts of Earth.
A recent global synthesis of recharge measurements from 5237 sites globally (Moeck etal.,2020) may alleviate
this issue. This synthesis data set provides a basis to investigate how observation-based recharge values vary
globally and to what extent there may be a widespread recharge bias in existing models. However, such investi-
gations are hampered by the large unquantified uncertainty associated with observation-based recharge estimates
(Crosbie etal.,2010,2018; Moeck etal.,2020). In addition, the exact spatial scale and period these measure-
ments represent remain uncertain and will never exactly overlap with those of models. However, the large number
of sites in the data set still allows to investigate the primary controls on global patterns of recharge and quantify
to what extent there could be a systematic recharge bias in existing models.
Here we show that climate aridity (Trabucco & Zomer,2009)—the ratio of potential evapotranspiration to
precipitation—strongly controls the fraction of precipitation that becomes groundwater. We parameterize a func-
tion that captures this relationship using the synthesis of groundwater recharge estimates (Moeck etal.,2020). We
show that the synthesis of groundwater recharge estimates (Moeck etal.,2020) indicates that existing hydrologi-
cal models, that have been previously used to predict recharge across the globe, underestimate recharge.
2. Methods and Data
2.1. Recharge Data
We obtain recharge rates from a recent global synthesis of groundwater recharge rates of 5237 sites located across
all continents but Antarctica (Moeck etal., 2020). The compiled data primarily originate from tracer methods
(∼80%) but are also derived from water table fluctuations, water balance methods, lysimeters, heat tracers, and
geophysical methods. This large variety of methods can affect estimated recharge rates at individual sites. The
recharge estimation studies cover the period from 1968 to 2018. The mean recharge rate is 234mmyr
−1, but over
40% of data points have rates between 0 and 25mmyr
−1 (median 51.3mmyr
−1). The data set contains recharge
rate estimates based on datasets that exceed at least 1year to avoid bias in the rates due to seasonal effects and
incomplete annual recharge values. The 5237 sites are assumed to represent naturally occurring recharge, as
recharge rates presumed to be affected by irrigation or managed aquifer recharge were already omitted by Moeck
etal.(2020). Study sites where rivers and streams dominate the estimated recharge were also omitted by Moeck
etal.(2020). Almost all these measurements will fall on recharge zones of the landscape, which in surface area
strongly dominate over the discharge zones near rivers (e.g., O'loughlin,1981). The global data has no quality
flags or uncertainties on recharge estimates because these estimates are also typically absent in many of the past
reports. For more information on the data, see Moeck etal.(2020) and the references therein. Most of the obser-
vations (n=4,386) originate from Australia (Crosbie etal.,2010) but these data have a similar relationship of
recharge fractions with aridity as the other data in the data set (Figure S1 in Supporting InformationS1).
2.2. Climate Data
We use temperature, aridity, precipitation, and FAO Penman-Monteith potential evapotranspiration data from
WorldClim (Fick & Hijmans,2017) and the global aridity and potential evapotranspiration database (Trabucco &
Zomer,2009). We define the aridity index as the ratio of mean potential evapotranspiration to mean precipitation.
Accordingly, high aridity index values reflect drier climates, whereas low values reflect humid climates. Regions
are classified as likely to have permafrost conditions when the mean annual temperature is below −2°C.
Writing – review & editing: Wouter
R. Berghuijs, Elco Luijendijk, Christian
Moeck, Ype van der Velde, Scott T. Allen

Geophysical Research Letters
BERGHUIJS ETAL.
10.1029/2022GL099010
3 of 9
2.3. Relationship Between Climate and Recharge
We use a mathematical expression that describes the global relationship between climate aridity and groundwater
recharge fractions:


=

1− ln

+1

1+ln(+1
)

(1)
where

is groundwater recharge (mmyr
−1), P is precipitation (mmyr
−1),

is aridity (dimensionless), defined as
the ratio of potential evapotranspiration to precipitation (EP/P), and

(dimensionless) is a constant equating to the
fraction of precipitation that becomes recharge for

→
0
(i.e., humid conditions).

is the characteristic exponent
(dimensionless) of the aridity index. We calibrate the

and

using a least absolute residuals fit. The sigmoidal
equation was selected because it is among the simplest equations that enforce physically realistic upper and lower
limits for recharge fractions. It closely follows the exponential decrease of recharge fraction with increasing arid-
ity visible in global recharge data. We reorganize the equation to estimate total recharge (mm yr
−1) using global
precipitation and aridity data:

=⋅

1− ln

+1

1+ln(+1
)

(2)
2.4. Groundwater Recharge Estimates From Global Models
We obtained simulated diffuse recharge estimates from the PCR-GLOB hydrological model (de Graaf
etal.,2015,2019) and the WaterGAP Global Hydrology Model (versions v2.1f and v2.2d) (Döll & Fiedler,2008;
Müller Schmied etal.,2021), and machine learning (Mohan etal.,2018). Also considering recharge from surface
water bodies did not change the overall results significantly. For the 5237 stations with recharge data, we compare
the observed recharge with the simulated recharge (Figure3). The simulated recharge values from the global
hydrological models represent mean annual recharge over a period that ranges from the year 1960 to 2001 (Döll
& Fiedler,2008), 1957 to 2002 (De Graaf etal.,2015), 1960 to 2010 (de Graaf etal.,2019) and 1901 to 2016
(Müller Schmied etal.,2021), respectively.
3. Results and Discussion
3.1. The Relationship Between Aridity and Recharge
The observation-based recharge estimates from sites spanning most regions of the globe show that recharge
fractions are strongly controlled by climate aridity (Figure1), despite many other factors also affecting ground-
water recharge globally (e.g., Moeck etal.,2020). In humid climates, typically, larger fractions of precipitation
recharge groundwater. This recharge fraction shrinks with increasing aridity, often approaching almost zero in
very arid sites. This relationship is nonlinear, and the empirical data show substantial variation for a given aridity,
reflecting an influence of other environmental conditions. However, the pattern is sufficiently monotonic to yield
a highly significant correlation between climate aridity and the fraction of precipitation that recharges groundwa-
ters (Spearman ρ=−0.674; p<0.001). This relationship is consistent with past work, which indicated that both
precipitation and potential evapotranspiration can strongly affect groundwater recharge (e.g., Moeck etal.,2020;
Scanlon etal., 2006), but goes beyond these past works by also showing how the partitioning of precipitation
changes. This quantified partitioning pattern is important as it substantiates the relative importance of recharge
across different climates.
The vast majority (i.e., 99%) of the observation-based recharge values are from regions with climate aridi-
ties exceeding 0.75. These aridities cover most of Earth's surface aside from several of Earth's wettest regions
(e.g., Congo Basin, Amazonia, Southeastern Asia), which largely fall outside the observational range. The
observation-based recharge values (Figure1b) suggest recharge fractions can shrink again at very low aridities,
but this remains uncertain because only few of the recharge-measurement sites fall in energy-limited systems with
aridity below one. In addition, the observation-based recharge sites fall outside regions underlain by permafrost

Geophysical Research Letters
BERGHUIJS ETAL.
10.1029/2022GL099010
4 of 9
(Obu,2021), where recharge processes often differ from non-permafrost regions (Walvoord & Kurylyk,2016). It
is good to note that a large part of the observation-based global data set (n=4,386) originates from the Australian
continent, mostly synthesized by another study (Crosbie etal.,2010). These Australian data have higher recharge
rates (mean=244mmyr
−1) compared to the remainder of the global data set (mean=188mmyr
−1) (Figure S1 in
Supporting InformationS1), but both parts of this global data set have a similar pattern of recharge fractions that
shrink with aridity according to the worldwide trend (Figure1; Figure S1 in Supporting InformationS1). There-
fore, the aridity-recharge relationship is likely representative for large parts of Earth's surface, but how recharge
in very humid and permafrost regions evolves with aridity cannot be directly constrained by the existing data.
Much of the variations in recharge can be described by a sigmoidal function of climate aridity (Equation1;
Figure1b). Calibrated on all data, this function describes how recharge exceeds 50% of precipitation (α=0.72,
with 95% confidence bounds 0.71, 0.73) when aridity approaches one (i.e., precipitation equals potential evapo-
transpiration), and decreases with increasing aridity (β=15.11, with 95% confidence bounds 14.91, 15.30). The
relationship seems inaccurate at low aridities, where both high and low recharge rates can occur.
Large but unquantified uncertainties associated with recharge measurements can limit the correspondence
between our model and the observations. However, although this parameterization is simple, it captures the
observed global trend in the fraction of precipitation that becomes groundwater recharge more accurately than
widely used global hydrological models (Figure3), which underestimate recharge in both arid and humid regions
(Extended Data Figure S2 in Supporting InformationS1). The parsimony of our aridity-recharge relationship
(Figure1b) may limit its predictive power but using more predictor variables does not substantially improve its
predictive capacity (Figure S3 in Supporting InformationS1). A split-sample test using 80% of the data for cali-
bration and the remaining 20% for validation still yields relatively narrow confidence bounds of the fitted param-
eters (95% confidence intervals α=0.69–0.75, β=14.0–16.2, not displayed), thus also subsets of the empirical
data effectively constrain the relationship (Figure S3 in Supporting InformationS1). Thus, the seemingly overly
simple predictions of groundwater recharge based on only climate aridity appear surprisingly effective compared
to the status quo, despite excluding many other factors that may also affect groundwater recharge.
3.2. Global Recharge Pattern
The parameterized relationship between climate aridity and recharge fraction (Equation1; Figure1b) enables
estimating the global distributions of groundwater recharge fractions (Figure2a) and total groundwater recharge
(Figure2b) using global aridity and precipitation data (see Methods). The estimated global pattern of groundwa-
ter recharge fractions shows large regional differences in how much precipitation recharges groundwater, broadly
Figure 1. Groundwater recharge fractions vary with aridity. Recharge fractions (the ratio between long-term recharge and
long-term precipitation) at the 5237 sites and the global pattern of climate aridity (a), whereby recharge negatively correlates
with aridity (b). The gray markers indicate the recharge fractions of individual groundwater recharge sites, whereas dark
markers average across 2% of the sites, removing most local site-to-site variability. The pink shading indicates a 25–75th
percentile over 100 data points. The red line depicts the calibrated sigmoid function Equation1. These data show how a
distinct trend of groundwater recharge fractions decreasing with aridity. The relationship is least constrained at low aridities,
where both high and low recharge rates can occur. There is substantial site-to-site recharge variability that is not explained by
aridity which is caused by other conditions that affect the physical recharge rates, and the large uncertainties associated with
recharge measurements.

Geophysical Research Letters
BERGHUIJS ETAL.
10.1029/2022GL099010
5 of 9
Figure 2. Estimated global patterns of groundwater recharge outside of permafrost regions. Estimates of groundwater recharge fractions vary regionally (a) and are
based on global climate data and Equation1. The absolute groundwater recharge values show high spatial variation because both the precipitation amount and the
fraction of precipitation that becomes recharge are correlated with aridity (b) (note the logarithmic color scales). Markers indicate the observations at the 5237 sites (b).
Permafrost regions are classified by having a mean annual temperature below −2°C. The estimates exclude regions with mean temperatures below −2°C because these
regions lack observations, whereas regions with aridity below 1 are excluded indicating that the data in these very humid regions is limited.

Geophysical Research Letters
BERGHUIJS ETAL.
10.1029/2022GL099010
6 of 9
consistent with the data set comprising observations from the 5237 sites
(Figure2a; Figure 1b). We exclude regions that can have permafrost (i.e.,
mean temperature below −2°C). Energy-limited regions with aridity below
1 are excluded from Figure1 to indicate that the empirical climate aridity
function poorly fits observations in these regions.
Estimated groundwater recharge fractions are low (<0.1) across roughly half
of Earth's surface (excluding permafrost regions) (Figure2a), as drylands
are very prevalent across all continents but Europe (Figure 1a) (Berg &
McColl,2021). Recharge fractions increase across more humid parts of Earth
such as most of Europe, eastern North America, central Africa, Southern
Asia, and most of South America. These regional patterns are both pres-
ent in observations and the estimated global pattern. Absolute recharge rates
show largely similar regional patterns (Figure2b), but the differences in esti-
mated recharge are even greater between humid and arid regions. Estimated
recharge would be highest in the equatorial wet regions and coastal regions
of Central and North America, Europe, and Oceania (consistent with earlier
global estimates), but large parts of these areas have aridities below 1 which
means recharge estimates are hard to constrain because few data exist at these
locations and the function starts to poorly fit the data. Nevertheless, even
when recharge fractions are low, the potential of high absolute recharge rates
will remain substantial in these regions as they experience high precipitation
rates.
Observation-based recharge values more than double those of several previous
global model estimates (Figure3; Figure S4 in Supporting InformationS1).
Such model estimates have not been systematically evaluated with observed
recharge data but rather with proxies such as streamflow measurements and
groundwater levels, or only with a small amount of field data. If we compare
the recharge rates from the widely used PCR-GLOB and WaterGAP global
hydrological models with the recharge observations at the 5237 sites, we find
that these models on average have 50% less recharge than the empirical data
(Figure3; Figures S4a–S4d in Supporting InformationS1). A similar but even more substantial difference is pres-
ent in another global recharge estimate based on 715 sites with recharge data (Figure3; Figure S4e in Supporting
Information S1). Split-sample tests do not show any such biases resulting from our aridity-recharge fraction
relationship (Figure S2 in Supporting InformationS1). An example realization of our relationship (Figure S4f in
Supporting InformationS1) shows how much better it explains observed recharge than other global hydrologi-
cal models (Figures S4a–S4e in Supporting InformationS1). The biases of the hydrological models arise from
underestimations at both high and low recharge rates. The difference in modeled and field-estimated recharge
may partly arise from the difference in the scales they represent. Global hydrological models simulate hydrologi-
cal behavior at multiple km
2 per grid-cell scale, thereby covering both recharge and discharge zones. In contrast,
most observations will be in recharge zones (Moeck etal.,2020), but discharge zones tend to cover only a small
part of the Earth's surface (e.g., O'loughlin,1981).
4. Implications and Conclusions
Aquifer storages are governed by the balance between recharge and discharge of groundwater to surface waters
and vegetation, in addition to human abstractions (Alley etal.,2002). Where observations are available, field
observations of recharge more than double most previous model estimates (Figure3; Figure S4 in Supporting
InformationS1). This enhanced recharge does not counter the current understanding of regional groundwater
overuse and its threats to global water security (Famiglietti,2014) because groundwater overuse results in storage
depletion and declining water levels that have been robustly documented in many more arid areas across the globe
(e.g., Rodell etal.,2018).
Most recharge will resurface as evapotranspiration or river flow (Alley etal.,2002). Thus, higher recharge rates
imply that groundwater's role in evapotranspiration and surface water fluxes is larger than previously modeled.
Figure 3. Comparison of observed versus predicted recharge for several
global recharge predictions. Moving averages of recharge predicted by
global models such as PCR-GLOB, WATER-GAP, and machine learning are
systematically lower than recharge of the 5237 observation sites (as indicated
by lines above the 1:1-line). The predictions by global models underestimate
recharge by more than 50% compared to the recharge measurement. Using
the sigmoid function (Equation1) largely removes this bias and produces
an overall average recharge of a very similar magnitude as global recharge
estimates. The presented recharge rates are moving averages over 10% of
the data. More detailed comparisons of modeled and observed recharge are
presented in Figure S4 in Supporting InformationS1.

Geophysical Research Letters
BERGHUIJS ETAL.
10.1029/2022GL099010
7 of 9
This also suggest that global hydrological models have overestimated (near) surface fluxes, such as overland flow,
shallow subsurface flows through the unsaturated zone, and soil-moisture-fed evapotranspiration.
The implied greater role of groundwater in supplying streamflow and evapotranspiration is consistent with
global observations that have shown that vegetation can source substantial parts of their water from groundwa-
ter, and vegetation disproportionally occurs near zones where it can access groundwater as a water source (Fan
etal., 2017; Koirala etal.,2017). It is also consistent with the observation that most precipitation is stored in
landscapes for at least several months before being observed in rivers (Jasechko etal.,2016), but note that older
water can also have other sources as water also can reside in soils and reservoirs for months before being meas-
ured as streamflow (Messager etal.,2016; Sprenger etal.,2019). These dynamic connections with vegetation and
streams likely predominantly occur in the upper layers of groundwater as deeper groundwaters mostly exchange
slowly with the Earth's surface (Berghuijs & Kirchner,2017; Jasechko etal.,2016,2017).
Recharge and its main potential fates (i.e., streamflow vs. evapotranspiration) depend strongly on climate arid-
ity (Budyko,1974). How much precipitation becomes streamflow shrinks with increasing aridity, whereas the
evaporative fraction grows with increasing aridity (Budyko,1974) (Figure S5 in Supporting InformationS1).
In humid areas, which typically have substantial recharge, both streamflow and evapotranspiration can have
groundwater contributions as streams typically have water levels below adjacent groundwater levels (Jasechko
et a.,2021). Losing rivers are more common in drier climates (Jasechko et a.,2021), suggesting a smaller role
for recharge in their streamflow and probably more recharge ultimately going to evapotranspiration. The relative
contribution of groundwater for transpiration is also reported to grow with aridity (Evaristo & McDonnell,2017)
though conservation of mass dictates that groundwater will typically only be a small component of total evap-
otranspiration across arid landscapes (i.e., recharge<< evapotranspiration). In mesic regions, the fraction of
precipitation that recharges groundwater derived from the synthesis recharge data set tends to exceed the fraction
that typically becomes streamflow (Figure S5 in Supporting InformationS1), which suggests that also a part of
evapotranspiration is supplied by groundwater. The gradients of recharge fraction with climate aridity may also
help to assess the impacts of climate changes on groundwater recharge. The effects of climate change on recharge
are currently highly uncertain and mostly unquantified (IPCC,2021).
A strong connection of groundwater with surface water and plant transpiration remains absent from most
diagrams of the global water cycle (Abbott etal.,2019; Dorigo etal.,2021; Oki & Kanae,2006). Although such
water cycle diagrams may not be intended as complete representations of the hydrological cycle, they often play
an important role in teaching, research, communication, and policymaking (Abbott etal.,2019). Therefore, we
need to consider revising those diagrams by increasing the rate at which groundwater is being replenished and
discharged and strengthening the link of groundwater with incoming precipitation, surface waters, and vegetation
(e.g., Miguez-Macho & Fan,2021).
The underrepresentation of groundwater as a key contributor to evapotranspiration and river flows may be
pervasive in hydrological models. Recharge is an internal flux that accumulates uncertainties and errors of
other components of the budget (Reinecke etal.,2021), and models are often not designed to treat groundwa-
ter recharge as a main source of streamflow and evapotranspiration. Preferential flow paths that can recharge
groundwaters are important in virtually any landscape (Beven & Germann,2013; Nimmo,2012) and contribute
disproportionally to fluxes such as recharge (Berghuijs & Kirchner,2017). Many of these pathways are absent in
global hydrological models. Connections of groundwaters with streamflow and evapotranspiration could also be
strengthened by including lateral groundwater flows (Maxwell & Condon,2016). Many of these lateral connec-
tions between surface water and groundwater likely occur at scales smaller than the grid-cells of most models
and thus require implicit sub-grid parameterizations (Fan etal.,2019). Strengthening the groundwater connection
to surface fluxes in these models is essential, given that models are the foundation of our understanding of our
planet, and underpin present-day environmental science and policymaking.
Data Availability Statement
All data used in this study are available via the cited sources. Precipitation data are available at https://www.
worldclim.org/data/v1.4/worldclim14.html. Potential evapotranspiration and aridity data are available at https://
cgiarcsi.community/data/global-aridity-and-pet-database/. Recharge data are available at https://opendata.eawag.
ch/dataset/globalscale_groundwater_moeck.

Geophysical Research Letters
BERGHUIJS ETAL.
10.1029/2022GL099010
8 of 9
References
Abbott, B. W., Bishop, K., Zarnetske, J. P., Minaudo, C., Chapin, F. S., Krause, S., et al. (2019). Human domination of the global water cycle
absent from depictions and perceptions. Nature Geoscience, 12(7), 533–540. https://doi.org/10.1038/s41561-019-0374-y
Alley, W. M., Healy, R. W., LaBaugh, J. W., & Reilly, T. E. (2002). Flow and storage in groundwater systems. Science, 296(5575), 1985–1990.
https://doi.org/10.1126/science.1067123
Berg, A., & McColl, K. A. (2021). No projected global drylands expansion under greenhouse warming. Nature Climate Change, 11(4), 331–337.
https://doi.org/10.1038/s41558-021-01007-8
Berghuijs, W. R., & Kirchner, J. W. (2017). The relationship between contrasting ages of groundwater and streamflow. Geophysical Research
Letters, 44(17), 8925–8935. https://doi.org/10.1002/2017gl074962
Beven, K., & Germann, P. (2013). Macropores and water flow in soils revisited. Water Resources Research, 49(6), 3071–3092. https://doi.
org/10.1002/wrcr.20156
Budyko, M. I. (1974). Climate and life (p.508). Academic Press.
Crosbie, R. S., Jolly, I. D., Leaney, F. W., & Petheram, C. (2010). Can the dataset of field based recharge estimates in Australia be used to predict
recharge in data-poor areas? Hydrology and Earth System Sciences, 14(10), 2023–2038. https://doi.org/10.5194/hess-14-2023-2010
Crosbie, R. S., Peeters, L. J., Herron, N., McVicar, T. R., & Herr, A. (2018). Estimating groundwater recharge and its associated uncer-
tainty: Use of regression kriging and the chloride mass balance method. Journal of Hydrology, 561, 1063–1080. https://doi.org/10.1016/j.
jhydrol.2017.08.003
Cuthbert, M. O., Taylor, R. G., Favreau, G., Todd, M. C., Shamsudduha, M., Villholth, K. G., etal. (2019). Observed controls on resilience of
groundwater to climate variability in sub-Saharan Africa. Nature, 572(7768), 230–234. https://doi.org/10.1038/s41586-019-1441-7
de Graaf, I. E. M., Gleeson, T., van Beek, L. R., Sutanudjaja, E. H., & Bierkens, M. F. (2019). Environmental flow limits to global groundwater
pumping. Nature, 574(7776), 90–94. https://doi.org/10.1038/s41586-019-1594-4
de Graaf, I. E. M., Sutanudjaja, E. H., van Beek, L.P.H., & Bierkens, M. F. P. (2015). A high-resolution global-scale groundwater model. Hydrol-
ogy and Earth System Sciences, 19(2), 823–837. https://doi.org/10.5194/hess-19-823-2015
De Vries, J. J., & Simmers, I. (2002). Groundwater recharge: An overview of processes and challenges. Hydrogeology Journal, 10(1), 5–17.
https://doi.org/10.1007/s10040-001-0171-7
Döll, P., & Fiedler, K. (2008). Global-scale modeling of groundwater recharge. Hydrology and Earth System Sciences, 12(3), 863–885. https://
doi.org/10.5194/hess-12-863-2008
Döll, P., Mueller Schmied, H., Schuh, C., Portmann, F. T., & Eicker, A. (2014). Global-scale assessment of groundwater depletion and related
groundwater abstractions: Combining hydrological modeling with information from well observations and GRACE satellites. Water Resources
Research, 50(7), 5698–5720. https://doi.org/10.1002/2014wr015595
Dorigo, W., Dietrich, S., Aires, F., Brocca, L., Carter, S., Cretaux, J. F., etal. (2021). Closing the water cycle from observations across scales:
Where do we stand? Bulletin of the American Meteorological Society, 102(10), 1–95. https://doi.org/10.1175/bams-d-19-0316.1
Evaristo, J., & McDonnell, J. J. (2017). Prevalence and magnitude of groundwater use by vegetation: A global stable isotope meta-analysis.
Scientific Reports, 7(1), 1–12. https://doi.org/10.1038/srep44110
Famiglietti, J. S. (2014). The global groundwater crisis. Nature Climate Change, 4(11), 945–948. https://doi.org/10.1038/nclimate2425
Fan, Y., Clark, M., Lawrence, D. M., Swenson, S., Band, L. E., Brantley, S. L., etal. (2019). Hillslope hydrology in global change research and
Earth system modeling. Water Resources Research, 55(2), 1737–1772. https://doi.org/10.1029/2018wr023903
Fan, Y., Miguez-Macho, G., Jobbágy, E. G., Jackson, R. B., & Otero-Casal, C. (2017). Hydrologic regulation of plant rooting depth. Proceedings
of the National Academy of Sciences, 114(40), 10572–10577. https://doi.org/10.1073/pnas.1712381114
Fick, S. E., & Hijmans, R. J. (2017). WorldClim 2: New 1-km spatial resolution climate surfaces for global land areas. International Journal of
Climatology, 37(12), 4302–4315. https://doi.org/10.1002/joc.5086
Gleeson, T., Befus, K. M., Jasechko, S., Luijendijk, E., & Cardenas, M. B. (2016). The global volume and distribution of modern groundwater.
Nature Geoscience, 9(2), 161–167. https://doi.org/10.1038/ngeo2590
Gleeson, T., Cuthbert, M., Ferguson, G., & Perrone, D. (2020). Global groundwater sustainability, resources, and systems in the anthropocene.
Annual Review of Earth and Planetary Sciences, 48(1), 431–463. https://doi.org/10.1146/annurev-earth-071719-055251
Gleeson, T., Wada, Y., Bierkens, M. F., & Van Beek, L. P. (2012). Water balance of global aquifers revealed by groundwater footprint. Nature,
488(7410), 197–200. https://doi.org/10.1038/nature11295
Hartmann, A., Gleeson, T., Wada, Y., & Wagener, T. (2017). Enhanced groundwater recharge rates and altered recharge sensitivity to climate
variability through subsurface heterogeneity. Proceedings of the National Academy of Sciences, 114(11), 2842–2847. https://doi.org/10.1073/
pnas.1614941114
IPCC. (2021). Climate change 2021: The physical science basis. Contribution of working group I to the sixth assessment report of the inter-
governmental panel on climate change [Masson-Delmotte In V. Masson-Delmotte, P. Zhai, A. Pirani, S. L. Connors, C. Péan, etal. (Eds.),
Cambridge University Press.
Jasechko, S., Kirchner, J. W., Welker, J. M., & McDonnell, J. J. (2016). Substantial proportion of global streamflow less than three months old.
Nature Geoscience, 9(2), 126–129. https://doi.org/10.1038/ngeo2636
Jasechko, S., Perrone, D., Befus, K. M., Cardenas, M. B., Ferguson, G., Gleeson, T., etal. (2017). Global aquifers dominated by fossil groundwa-
ters but wells vulnerable to modern contamination. Nature Geoscience, 10(6), 425–429. https://doi.org/10.1038/ngeo2943
Jasechko, S., Seybold, H., Perrone, D., Fan, Y., & Kirchner, J. W. (2021). Widespread potential loss of streamflow into underlying aquifers across
the USA. Nature, 591(7850), 391–395. https://doi.org/10.1038/s41586-021-03311-x
Koirala, S., Jung, M., Reichstein, M., de Graaf, I. E. M., Camps-Valls, G., Ichii, K., etal. (2017). Global distribution of groundwater-vegetation
spatial covariation. Geophysical Research Letters, 44(9), 4134–4142. https://doi.org/10.1002/2017gl072885
Li, B., Rodell, M., Peters-Lidard, C., Erlingis, J., Kumar, S., & Mocko, D. (2021). Groundwater recharge estimated by land surface models: An
evaluation in the conterminous United States. Journal of Hydrometeorology, 22(2), 499–522. https://doi.org/10.1175/jhm-d-20-0130.1
MacDonald, A. M., Lark, R. M., Taylor, R. G., Abiye, T., Fallas, H. C., Favreau, G., etal. (2021). Mapping groundwater recharge in Africa from
ground observations and implications for water security. Environmental Research Letters, 16(3), 034012. https://doi.org/10.1088/1748-9326/
abd661
Margat, J., & Van der Gun, J. (2013). Groundwater around the world: A geographic synopsis. Crc Press.
Maxwell, R. M., & Condon, L. E. (2016). Connections between groundwater flow and transpiration partitioning. Science, 353(6297), 377–380.
https://doi.org/10.1126/science.aaf7891
Messager, M. L., Lehner, B., Grill, G., Nedeva, I., & Schmitt, O. (2016). Estimating the volume and age of water stored in global lakes using a
geo-statistical approach. Nature Communications, 7(1), 1–11. https://doi.org/10.1038/ncomms13603
Acknowledgments
We thank Dr. Ying Fan and Andrew
Western for their constructive reviews that
helped to improve this paper.

Geophysical Research Letters
BERGHUIJS ETAL.
10.1029/2022GL099010
9 of 9
Miguez-Macho, G., & Fan, Y. (2021). Spatiotemporal origin of soil water taken up by vegetation. Nature, 598(7882), 624–628. https://doi.
org/10.1038/s41586-021-03958-6
Moeck, C., Grech-Cumbo, N., Podgorski, J., Bretzler, A., Gurdak, J. J., Berg, M., & Schirmer, M. (2020). A global-scale dataset of direct natural
groundwater recharge rates: A review of variables, processes and relationships. Science of the Total Environment, 717, 137042. https://doi.
org/10.1016/j.scitotenv.2020.137042
Mohan, C., Western, A. W., Wei, Y., & Saft, M. (2018). Predicting groundwater recharge for varying land cover and climate conditions–a global
meta-study. Hydrology and Earth System Sciences, 22(5), 2689–2703. https://doi.org/10.5194/hess-22-2689-2018
Morris, B. L., Lawrence, A. R., Chilton, P.J., Adams, B., Cay-low, R. C., & Klinck, B. A. (2003). Groundwater and its susceptibility to degra-
dation: A global assessment of the problems and options for management (Vol. 03–3). UNEP Early Warning & Assessment Report Series RS.
Müller Schmied, H., Cáceres, D., Eisner, S., Flörke, M., Herbert, C., Niemann, C., etal. (2021). The global water resources and use model
WaterGAP v2. 2d: Model description and evaluation. Geoscientific Model Development, 14(2), 1037–1079. https://doi.org/10.5194/
gmd-14-1037-2021
Nimmo, J. R. (2012). Preferential flow occurs in unsaturated conditions. Hydrological Processes, 26(5), 786–789. https://doi.org/10.1002/
hyp.8380
Obu, J. (2021). How much of the Earth's surface is underlain by permafrost? Journal of Geophysical Research: Earth Surface, 126(5),
e2021JF006123. https://doi.org/10.1029/2021jf006123
Oki, T., & Kanae, S. (2006). Global hydrological cycles and world water resources. Science, 313(5790), 1068–1072. https://doi.org/10.1126/
science.1128845
O'loughlin, E. M. (1981). Saturation regions in catchments and their relations to soil and topographic properties. Journal of Hydrology, 53(3–4),
229–246. https://doi.org/10.1016/0022-1694(81)90003-2
Reinecke, R., Müller Schmied, H., Trautmann, T., Andersen, L. S., Burek, P., Flörke, M., etal. (2021). Uncertainty of simulated groundwa-
ter recharge at different global warming levels: A global-scale multi-model ensemble study. Hydrology and Earth System Sciences, 25(2),
787–810. https://doi.org/10.5194/hess-25-787-2021
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. https://doi.org/10.1038/s41586-018-0123-1
Scanlon, B. R., Keese, K. E., Flint, A. L., Flint, L. E., Gaye, C. B., Edmunds, W. M., & Simmers, I. (2006). Global synthesis of groundwater
recharge in semiarid and arid regions. Hydrological Processes: International Journal, 20(15), 3335–3370. https://doi.org/10.1002/hyp.6335
Siebert, S., Burke, J., Faures, J. M., Frenken, K., Hoogeveen, J., Döll, P., & Portmann, F. T. (2010). Groundwater use for irrigation–a global
inventory. Hydrology and Earth System Sciences, 14(10), 1863–1880. https://doi.org/10.5194/hess-14-1863-2010
Sprenger, M., Stumpp, C., Weiler, M., Aeschbach, W., Allen, S. T., Benettin, P., etal. (2019). The demographics of water: A review of water ages
in the critical zone. Reviews of Geophysics, 57(3), 800–834. https://doi.org/10.1029/2018rg000633
Sutanudjaja, E. H., Beek, R. V., Wanders, N., Wada, Y., Bosmans, J. H., Drost, N., etal. (2018). PCR-GLOBWB 2: A 5 arcmin global hydrological
and water resources model. Geoscientific Model Development, 11(6), 2429–2453. https://doi.org/10.5194/gmd-11-2429-2018
Trabucco, A., & Zomer, R. J. (2009). Global aridity and PET database (CGIAR consortium for spatial information. Retrieved from https://cgia-
rcsi.community/data/global-aridity-and-pet-database/
Walvoord, M. A., & Kurylyk, B. L. (2016). Hydrologic impacts of thawing permafrost—A review. Vadose Zone Journal, 15(6), vzj2016.01.0010.
https://doi.org/10.2136/vzj2016.01.0010