1

Interception: Its Importance, its Isotopic Impact,

and How to Model and Measure it
Hubert H.G. Savenije and Miriam Coenders-Gerrits
Delft University of Technology, Delft, The Netherlands

1 Introduction where I is the interception process (mm d−1 ), SI is the

interception storage (mm), and EI is the evaporation
To hydrologists who focus on the terrestrial part of the from interception (mm d−1 ). The interception I is the
hydrological cycle, precipitation (rainfall, snowfall, hail, part of the rainfall that is intercepted and, after a short
and dew) is generally seen as the starting point of the period of time, becomes the evaporative flux EI . For a
water cycle. Although one could equally well start with time scale larger than approximately one to two days, it
any other point in the cycle, e.g. advection of moisture to is safe to assume that the intercepted water has turned
the land or evaporation from the oceans, it is common to into an evaporation flux: I = EI .
begin the terrestrial water cycle with the precipitation, Next, it is important to define the location of the
probably because precipitation is more easy to observe interception process in the hydrological cycle. The most
than other incoming fluxes. The terrestrial water cycle logical place is between the atmosphere and the first sep-
involves many interconnected processes, separation aration point where the rainfall splits into interception,
points, feedback loops, and stocks where the water surface runoff, and infiltration (Figure 1). The intercep-
resides over longer or shorter periods of time. Figure 1 tion stock (SI ) is located at the first separation point.
illustrates these processes with their separation points In this definition, the interception process includes
and the evaporation fluxes that feedback moisture to evaporation from wet leaves, wet land cover (included
the atmosphere, which in turn sustain precipitation. man-made structures and roads), wet mulch, wet forest
Starting from the rainfall, the first process that comes floor, and even wet soil. In short, it is the fast evaporation
into view is interception. It retains the water before it mechanism that dries a moist surface during and directly
can continue its path in the water cycle and it allows for after a rainfall event.
a direct feedback loop to the atmosphere. A more narrow definition of interception, often used,
Depending on the precise definition of this process, is the difference between rainfall and “throughfall,” the
interception can be understood to mean different things. rain that falls through the leaves of, for instance, a tree.
First of all, it is important to decide whether we con- In this definition, interception is merely the amount of
sider interception as a stock, a flux, or a process. If we water retained by the leaves of a tree. This is not a correct
consider it a stock, then it is the amount of rainfall that definition of interception because it does not account for
can be temporarily stored on the land (with its natural the intercepted water by the ground cover, which can be
and man-made cover) to be evaporated shortly after (or substantial [1]. If interception by the ground is included,
during) the rainfall event. Here, we actually mean the then the total interception is larger than what is found
interception capacity, generally expressed as a volume in the literature based on the difference between rainfall
per unit area, or depth (L). If we consider interception and throughfall [2].
as a flux, then it is the evaporation from intercepted The total evaporation E in a catchment is the sum of a
water, which we express in depth per unit of time (L/T). number of different processes: interception EI , transpira-
However, it is more appropriate to consider interception tion ET , soil evaporation ES , and open water evaporation
as a process. The interception process is the sum of the EO [3]. In the broader definition of interception, we
rate of change of intercepted water and its evaporation: combine evaporation from leaf interception with wet
dSI surface evaporation. The wet surface interception, or
I= + EI (1) ground interception, is not the same as soil evaporation,
dt
Encyclopedia of Water: Science, Technology, and Society, edited by Patricia A. Maurice.
Copyright © 2019 John Wiley & Sons, Inc.
DOI: 10.1002/9781119300762.wsts0061
2 Atmosphere and Precipitation, Ice and Glaciers, Oceans and Coasts, Soils and Mineral–Water Interface

EI P E upland forest catchments of Britain, evaporation from

s
interception amounts to 35% in areas with an annual
ET
rainfall of more than 1000 mm yr−1 , but that this per-
First separation point Qo
F Q centage is higher in areas with lower rainfall, amounting
Qs
to about 40–50% in areas with 500–600 mm yr−1 . In
Second separation point Qg
Calder’s definition, however, interception is merely the
C R
difference between rainfall and throughfall. But if we use
the broader definition presented above, interception is
P = rainfall R = percolation considerably larger. After a rainfall event, not only the
EI = interception C = capillary rise leaves of vegetation are wet, also the surface underneath
Qo = overland flow Es = soil evaporation (rock, soil, mulch layer, roads, build-up area, etc.), which
F = infiltration Qg = seepage flow becomes dry within the same day, particularly in warm
ET = transpiration Qs = subsurface flow climates. Moreover, there are many stagnant pools that
continue to evaporate until a day after the event. The
Figure 1 Separation points and evaporative fluxes. The additional wet surface evaporation can be as much as the
interception origins from the first separation point on the land interception by leaves, particularly in dry climates. All
surface, the transpiration, and soil evaporation origins from the these interception processes are fast, having an average
second separation point within the root zone. Interception is
constrained by the interception capacity of the surface; residence time in the order of one day.
transpiration and soil evaporation are constrained by the moisture Another ecological function of interception is that
content of the root zone. canopies distribute rainfall to preferential dripping
points, where the dripping intensity is larger than the
free precipitation [10]. Anybody who has tried to shelter
which is a different process with different timing. Soil from the rain under a tree knows that after some time
evaporation is the evaporation from soil moisture that preferential dripping points appear that need to be
reaches the surface by capillary rise. Soil evaporation avoided, because one is likely to get more wet under such
only comes into play after wet surface interception has a dripping point than out in the open. These preferential
evaporated. Similarly, it is reasonable to assume that dripping points trigger the occurrence of preferential
transpiration only starts after leaf interception has fully infiltration pathways that recharge the soil moisture in
evaporated. So there is a logical sequence in how the the root zone, and eventually facilitate recharge and sub-
energy available for evaporation is allocated to the differ- surface drainage when there is excess of water. Brooks
ent processes. After or during rainfall, evaporation from et al. [11] demonstrated that this self-organization leads
the canopy precedes canopy transpiration. Likewise, to the existence of “two water worlds” [12]. Apparently,
ground interception precedes soil evaporation [4]. ecosystems evolve in a way that they optimize their
water use and water storage to overcome dry spells [13],
but also their drainage. The interception process is an
2 The Importance of Interception important part of this.
Note that the wetted soil surface should not be consid-
Interception is one of the most underrated, and underes- ered part of the soil moisture that feeds the transpiration
timated, processes in rainfall–runoff analysis [5]. Some process. The wet surface (extending to several millime-
models even disregard interception completely (particu- ters of soil depth) feeds back the intercepted water
larly event-based models, based on the simplified catch- through direct evaporation and not via a delayed tran-
ment model of Dooge [6, 7], the argument being that it is spiration process. Even dry sand, without vegetation, can
generally a small proportion of the total evaporation. This intercept water. Also, in sandy soils, recharge is prefer-
may be acceptable for short duration events, but neglect- ential, leading to wet “fingers” of infiltration, surrounded
ing interception can lead to serious mistake in hydrologi- by dry pockets. These dry pockets only get moist after
cal modeling or water resources assessment. In Section 4, prolonged rainfall. After a rainfall event on dry sand, a
it will be shown how interception can be modeled. But wet “crust” of soil is formed, underlain by dry sand, which
first, we explain why interception is such an important subsequently dries out. While walking over this crust,
process in the hydrological cycle. one often breaks it and sees the dry sand underneath.
Firstly, the amount of interception is not small. Beven This sandy surface can intercept several millimeters of
[8] stated that evaporation from intercepted water on rainfall.
leaf surfaces in rough canopies can be very efficient Transpiration works completely differently. First of all,
and a significant component of the total water balance it is a physiological process intimately tied to CO2 assim-
in some environments. Calder [9] showed that in the ilation. The time scale of transpiration is determined by
 Interception: Its Importance, its Isotopic Impact, and How to Model and Measure it 3

the soil moisture stock, which makes the time scale of the because where the canopy dries out quickly and com-
process much longer (average residence times varying pletely, the ground surface remains wet much longer.
between weeks and months, depending on the root zone If the surface is wetted again by rainfall, the mixture of
storage). Moreover, the process of transpiration does enriched and new water can contribute enriched water to
not change the isotope composition of the evaporated deeper soil layers. Open water evaporation from ponds
moisture, whereas evaporation of a water surface does functions much like ground interception. It also enriches
[14, 15]. One can distinguish fast transpiration from the remaining water, which partly infiltrates. As a result,
delayed transpiration. Fast transpiration is from shallow ground interception, soil evaporation, and pond infil-
rooted plants (typically grass and annual crops) with a tration are responsible for the higher concentration of
time scale of less than a month; delayed transpiration is heavy isotopes in the soil, while the preferential recharge,
from deeply rooted plants (trees, shrubs, and perennial facilitated by concentrated dripping from the canopy,
crops), which have a time scale longer than a month. Fast transports the rainwater directly to the groundwater
transpiration only draws on the upper soil layer, whereas and causes the deeper groundwater to be close to the
delayed transpiration draws on deeper soil layers. Open meteoric line.
water evaporation can be considered separately, when
necessary, and is identifiable relatively simply.

4 How to Model Interception

3 Effect of Interception on the Isotopic Evaporation
Composition of Moisture For time steps shorter than a day, the computation of
interception should be done with a Rutter model [17],
Depending on the location on Earth, precipitation con-
using Eq. (1), whereby the stock of interception SI is
tains a fixed proportion of heavy isotopes of oxygen (18 O)
accounted for. For practical purposes, however, and in
and hydrogen (2 H). This is called the local meteoric water
view of the typical time scale of the process where the
line. Soil moisture generally deviates from the meteoric
intercepted water evaporates within a time span of a
line, while groundwater and streamflow generally lie
day, interception can be computed by a simple daily
close to the meteoric line, leading to the two different
model. At this time scale EI = I, an interception can be
“water worlds” described earlier. Unlike what one might
conveniently described as a threshold process acting on
expect, it is not the transpiration that is responsible
the daily precipitation P (mm d−1 ) using a daily threshold
for this deviation. Transpiration does not fractionate
D (mm d−1 ). For the daily threshold, one can take the
between heavy and lighter isotopes; it merely transports
amount of interception storage that can evaporate during
moisture with a certain isotope composition from the
a day, or the potential evaporation on a day, whichever is
root zone to the air. Canopy interception contributes
smallest. Daily interception (mm d−1 ) is then compute as:
to fractionation, as rainwater on the leaves evaporates
during rainfall events, or between events before leaves I = min(P, D) (2)
become dry [16]. But equally, or even more important
mechanisms of fractionation, are ground interception implying that if the rainfall is less than the threshold D,
(including temporary ponding), open water evaporation, the interception I is equal to the rainfall P, and otherwise
and soil evaporation, which enrich the remaining mois- it is equal to the threshold D. De Groen and Savenije
ture with heavy isotopes and cause deviation from the [18] estimated the daily threshold D to have a value of
meteoric line. 1–5 mm d−1 , depending on the land use and the poten-
Soil evaporation is different from ground intercep- tial evaporation. Pitman [19] argued that in Southern
tion in the sense that it does not directly evaporate the Africa it could be as much as 8 mm d−1 in forests (maybe
wetness left behind by a recent rainfall event, but rather because forest floor interception can spread over more
draws on older soil moisture that comes to the surface than one subsequent dry days).
by capillary rise. However, in terms of isotopic fraction- Monthly evaporation from interception can be
ation, soil evaporation functions similarly to ground determined very efficiently, even for poorly gauged
interception. Where transpiration does not fractionate catchments, using the Markov property of rainfall [18].
the water that is transported from the root zone to the The Markov property implies that the probability that
stomata from which it evaporates, ground interception a certain day is a rainday depends purely on whether
and soil evaporation do, because lighter isotopes evapo- the previous day was a wet or a dry day. Combination
rate more easily than their heavier counterparts. Ground of Eq. (2) with the logarithmic probability distribution
interception enriches more than canopy interception, of monthly rainfall leads to the following equation for
4 Atmosphere and Precipitation, Ice and Glaciers, Oceans and Coasts, Soils and Mineral–Water Interface

monthly interception: have a very small spatial correlation (a few kilometers),

( ( )) but apparently the Markov property has a high one,
nr D
Im = Pm 1 − exp − (3) of the same order of magnitude as monthly rainfall
Pm (100–1000 km).
where I m is the monthly evaporation from interception There is some debate on the relative contributions of
(mm mo−1 ), Pm is the monthly rainfall (mm mo−1 ), and interception and transpiration to total terrestrial evapo-
nr is the number of raindays per month (d mo−1 ). Gerrits ration [22]. Wang-Erlandsson et al. [4] calculated the dis-
et al. [20] followed a similar approach to upscale inter- tribution of the different evaporative fluxes globally for
ception to annual amounts. different land covers (Figure 2). They concluded that, on
As a result of the Markov property of daily rainfall, average, transpiration amounts to 59% of the total terres-
the number of raindays in a month is a function of the trial evaporation, interception to 31%, soil evaporation to
monthly rainfall: 6%, and open water evaporation to 4%. Depending on the
land cover, these percentages vary, although not substan-
30p01
nr = (4) tially. For instance, in evergreen broadleaf (i.e. tropical)
1 − p11 + p01
forest, the percentages of transpiration and interception
where p01 is the probability of a rainday after a dry day are 54% and 45%, respectively; in evergreen needle forest,
and p11 is the probability of a rainday after a rainday. it is 50% and 46%; and in deciduous broadleaf (i.e. tem-
These probabilities appear to be power functions of perate), it is 64% and 31%. In general, one can conclude
monthly rainfall. The coefficients of the power functions that the total interception (from vegetation and ground
are location dependent, depending on the climate and floor) is substantial and in the same order of magnitude
the topography, and can be derived from daily time series as the transpiration.
of selected raingauges. Fischer et al. [21] demonstrated Besides interception forming a considerable part
that, in East Africa, these coefficients can be interpolated of monthly evaporation, it is crucial for determining
in space as a function of elevation and distance to the antecedent conditions in rainfall–runoff modeling. It is
ocean. This is surprising, since daily rainfall is known to often said that during storm events, evaporation from

Transpiration evaporation Et Soil moisture evaporation Esm 1500

(a) (b)
45° N 45° N
1000

mm yr–1
0° 0°
500

45° S 45° S
0

Vegetation interception evaporation Ev Floor interception evaporation Ef

(c) (d)
45° N 45° N

0° 0°

45° S 45° S

Inland water evaporation Ew Total evaporation E

(e) (f)
45° N 45° N

0° 0°

45° S 45° S

Figure 2 Mean annual evaporation as estimated by Wang-Erlandsson et al. [4]. Gray indicates areas where the evaporative flux is 0.
 Interception: Its Importance, its Isotopic Impact, and How to Model and Measure it 5

1
(a)

0.8
45° N

0.6

0° 0.4

0.2

45° S
0

0.4+
(b)
0.35
45° N 0.3

0.25
0.2
0° 0.15

0.1

0.05
45° S
0

+
(c)
0.5
45° N
0.4

0.3

0°
0.2

0.1
45° S
0

Figure 3 Continental recycling of precipitation computed by Van der Ent et al. [23]: (a) Continental precipitation recycling ratio
𝜌C = 𝜌C, i + 𝜌C, t , (b) continental precipitation recycling ratio for interception 𝜌C, i = 𝜌C − 𝜌C, t , and (c) continental precipitation recycling ratio
for transpiration 𝜌C, t = 𝜌C − 𝜌C, i . The recycling ratio is the percentage of the precipitation that relies on terrestrial evaporation. Red color
indicates the regions where the recycling ratio is high. Comparison of (b) and (c) shows that interception and transpiration are equally
relevant to sustain continental precipitation.

interception is negligible as compared to other fluxes. the rainfall into the effective part that contributes to soil
This may be true during the event itself, which is of short moisture (and hence the rainfall–runoff process) and the
duration, but it is untrue in relation to the build-up of ineffective part that does not, i.e. interception.
the antecedent conditions. It is widely recognized that Finally, evaporation from interception is an important
the success of event-based modeling depends to a very process in moisture recycling to support continental
large extent on the antecedent conditions, particularly rainfall. Figure 3 (from Van der Ent et al. [23]) shows
the distribution of the soil moisture in the unsaturated that interception is only slightly less important than
soil. If we want to get that right, it is important to split transpiration in sustaining global precipitation. In
6 Atmosphere and Precipitation, Ice and Glaciers, Oceans and Coasts, Soils and Mineral–Water Interface

months with high rainfall (i.e. the wet season), inter- the plant/tree (see Levia and Germer [27] for a review).
ception is an important mechanism; in dry months, It is a very efficient way of a plant to direct the water
transpiration is often dominant, but then (typically after to the roots. Some trees have a branch structure that
the wet season) there are hardly any rainfall generating actually facilitates stemflow, whereas others, such as firs,
mechanisms to benefit from this feedback. Shuttleworth direct the water away from the stem. Stemflow can be
[3] observed that half the evaporation from intercep- measured by winding a half-tube around a tree stem
tion occurs during the storm itself, providing instant and connecting it to tipping bucket recorder or simply a
moisture feedback. Hence, the moisture feedback to the container (Figure 4).
atmosphere, which is such an important mechanism Because throughfall is very heterogeneous, even to the
to support continental rainfall in the Sahel, and the extent that it concentrates in preferred dripping points,
Amazon [24, 25], relies to a large extent on interception. it is very hard to measure. In Figure 5, we see a set-up
The reason why transpiration is relatively small in wet with gutters connected to tipping bucket recorders and
months is because the energy available for evaporation a dense network of individual accumulating raingauges.
is first consumed by the interception process, preceding The water from the tipping buckets is conserved in con-
transpiration, and because solar radiation is inhibited by tainers as a backup and for chemical and isotope analysis.
clouds in wet months. Although this is a very dense set-up, there is still the risk
that the preferential dripping points are not well sampled.
Some researchers advocate to reallocate the throughfall
5 Measuring Interception collectors regularly and randomly.
An interesting effort to measure the water intercepted
Classically, canopy interception is determined by sub- by the canopy directly was undertaken by Van Emmerik
tracting throughfall and stemflow from precipitation et al. [28], who used accelerometers to detect the change
in open air (see Zimmermann and Zimmermann [26] in the frequency spectrum of a tree as it swings with more
for a review). Stemflow is the part of the precipitation or less mass in the canopy. Another interesting approach
that flows along the branches and stem to the base of to measure canopy interception was followed by Friesen
et al. [29], who tried to determine the weight of the tree
by measuring stem compression.
Obviously field observations of canopy interception
are not simple, and it is hard to make accurate estimates
on the basis of field trials. Another way to determine
interception is by hydrological modeling, whereby the
interception storage capacity is the only unknown. The
physical range of this storage capacity is limited. For
grass and cropland, a feasible range is 1–2 mm. For

Figure 4 A winding half-open hose glued to the tree and

connected to a tipping bucket recorder to measure stemflow in a Figure 5 Experimental set-up to measure canopy interception in a
beech forest in Luxembourg. Source: Photograph courtesy of beech forest in Luxembourg. Source: Photograph courtesy of
Miriam Coenders. Miriam Coenders.
 Interception: Its Importance, its Isotopic Impact, and How to Model and Measure it 7

Eint Precipitation

Litter

Su
Geotextile
EI
Infiltration Weighing
device

SI

Valve

Figure 6 Ground interception weighing device in a beech forest in Luxembourg. Source: Photograph and sketch courtesy of Miriam
Coenders.

dense forest, it may be bound between 2 and 5 mm. Of floor interception capacity is the main unknown. When
course, it also depends on the season. Therefore, the a reasonable range of interception capacity is assumed
interception capacity of vegetation is often connected (1–3 mm, depending on the season), the quantity of
to the Leaf Area Index (LAI), which is readily available forest floor interception then largely depends on the
from remote sensing products. temporal distribution of the rainfall and the energy
Ground interception, or forest floor interception, is available for evaporation.
often disregarded, but can be a substantial part of the
interception. Even during winter, the fallen leaves in a
deciduous forest have a substantial ground interception 6 Consequences of Underestimating
capacity (as can be seen in Figure 5). The transition from Interception
leaf cover to humus to soil may be gradual, and it may
be hard to separate the free water from capillary water, Figures 2 and 3 clearly show how important interception
but in any case, the evaporation from the ground, which is for the global water balance and for terrestrial moisture
is different from transpiration, can be a substantial part recycling. Also, interception is mainly responsible for the
of the total evaporation. Moreover, as stated before, it is fractionation of isotopes in the hydrological cycle. Inter-
this part of the evaporation that is largely responsible for ception cannot be simply neglected or lumped with other
the enrichment of the soil moisture with heavy isotopes. evaporation mechanisms. Disregarding or underestimat-
Gerrits et al. [1] were the first to try to measure ground ing interception can lead to serious modeling mistakes,
interception continuously. They used a weighing device particularly when one uses automated calibration tech-
using strain gauges as shown in Figure 6 to measure the niques. If interception is modeled incorrectly, the error
ground interception in different ground covers (beech will be compensated by other parameters, to satisfy the
leaves, moss with grass, and needles). The two containers
goodness-of-fit criterion.
were weighed independently. Ground interception was
The most common mistake of lumping interception
than computed by:
( ) with transpiration leads to an overdimensioning of the
dSu dSl soil moisture stock. This can be seen easily. If the inter-
EI,ground = P − +
dx dx ception is forced through the transpiration process, a
where Su [L] is the storage in the upper container and Sl correct representation of the total flux and time scale in
[L] is the storage in the lower container. the model can only be achieved if the soil moisture stock
Similar to canopy interception, ground interception is is overdimensioned. If transpiration and interception are
not easy to observe at larger scale. As a result, a modeling of the same order of magnitude, then the modeled soil
approach is a more practical route, whereby the forest moisture stock becomes double its “real” value.
8 Atmosphere and Precipitation, Ice and Glaciers, Oceans and Coasts, Soils and Mineral–Water Interface

Further Reading

Carlyle-Moses, D.E. and Gash, J.H.C. (2011). In: Forest Gerrits, A.M.J. and Savenije, H.H.G. (2011). In: Treatise on
Hydrology and Biogeochemistry (eds. D.F. Levia, D. Water Science, vol. 2 (ed. P. Wilderer), 89–101. Oxford
Carlyle-Moses and T. Tanaka), 407–423. Dordrecht: Academic Press.
Springer. Llorens, P. and Domingo, F. (2007). J. Hydrol. 335 (1–2):
37–54.

References
1 Gerrits, A.M.J., Savenije, H.H.G., Hoffmann, L., and 16 Allen, S.T., Keim, R.F., Barnard, H.R. et al. (2016).
Pfister, L. (2007). Hydrol. Earth Syst. Sci. 11: 695–701. WIREs Water doi: 10.1002/wat2.1187.
2 Gerrits, A.M.J., Pfister, L., and Savenije, H.H.G. 17 Rutter, A.J., Kershaw, K.A., Robins, P.C., and Morton,
(2010). Hydrol. Process. 24 (21): 3011–3025. A.J. (1971). Agric. Meteorol. 9: 267–384.
3 Shuttleworth, W.J. (1993). In: Handbook of Hydrology 18 De Groen, M.M. and Savenije, H.H.G. (2006). Water
(ed. D.R. Maidment), 4.1–4.53. McGraw-Hill. Resour. Res. 42, W12417: 1–10.
4 Wang-Erlandsson, L., van der Ent, R.J., Gordon, L.J., 19 Pitman, W.V. (1973). A mathematical model for gen-
and Savenije, H.H.G. (2014). Earth Syst. Dynam. 5: erating monthly river flows from meteorological data
441–469. doi: 10.5194/esd-5-441-2014. in Southern Africa. 2/73, University of Witwatersrand,
5 Savenije, H.H.G. (2004). Hydrol. Process. 18 (8): Department of Civil Engineering, Hydrological Unit,
1507–1511. South Africa.
6 Dooge, J.C.I. (1973). Linear theory of hydrologic sys- 20 Gerrits, A.M.J., Savenije, H.H.G., Veling, E.J.M., and
tems. Agriculture Research Service Technical Bulletin Pfister, L. (2009). Water Resour. Res. 45 (4, W04403):
No. 1468. US Department of Agriculture. 1–15.
7 Dooge, J.C.I. (2003). Linear theory of hydrologic sys- 21 Fischer, B.M.C., Mul, M.L., and Savenije, H.H.G.
tems. EGU Reprint Series, 1, European Geosciences (2013). Hydrol. Earth Syst. Sci. 17: 2161–2170.
Union, Katlenburg-Lindau, Germany. 22 Coenders-Gerrits, A.M.J., Van der Ent, R.J.,
8 Beven, K.J. (2001). Rainfall-Runoff Modelling, the Bogaard, T.A. et al. (2014). Nature 506: E1–E2. doi:
Primer. John Wiley & Sons. 10.1038/nature12925.
9 Calder, I.R. (1990). Evaporation in the Uplands. John 23 Van der Ent, R.J., Wang-Erlandsson, L., Keys, P.W.,
Wiley & Sons. and Savenije, H.H.G. (2014). Earth Syst. Dynam. 5:
10 Coenders-Gerrits, A.M.J., Hopp, L., Savenije, H.H.G., 471–489. doi: 10.5194/esd-5-471-2014.
and Pfister, L. (2013). Hydrol. Earth Syst. Sci. 17: 24 Eltahir, E.A.B. and Bras, R.L. (1994). Q. J. R. Meteorol.
1749–1763. Soc. 120: 861–880.
11 Brooks, J.R., Barnard, H.R., Coulombe, R., and 25 Savenije, H.H.G. (1995). J. Hydrol. 167: 57–78.
McDonnell, J.J. (2010). Nat. Geosci. 3 (2): 100–104. 26 Zimmermann, A. and Zimmermann, B. (2014). Agric.
12 McDonnell, J.J. (2014). Wiley Interdiscip. Rev. Water 1 For. Meteorol. 189–190: 125–139.
(4): 323–329. 27 Levia, D.F. and Germer, S. (2015). Rev. Geophys. 53
13 Gao, H., Hrachowitz, M., Schymanski, S.J. et al. (3): 673–714.
(2014). Geophys. Res. Lett. 41: 7916–7923. doi: 28 Van Emmerik, T., Steele-Dunne, S., Hut, R. et al.
10.1002/2014GL061668. (2017). Sensors 17 (5): 1098.
14 Ehleringer, J.R. and Dawson, T.E. (1992). 29 Friesen, J., van Beek, C., Selker, J. et al. (2008).
Plant Cell Environ. 15 (9): 1073–1082. doi: Water Resour. Res. 44, W00D15: 1–5. doi:
10.1111/j.1365-3040.1992.tb01657.x. 10.1029/2008WR007074.
15 Kendall, C. and McDonnell, J.J. (1998). Isotope Tracers
in Catchment Hydrology. Elsevier.