Content uploaded by L.A. Bruijnzeel
Author content
All content in this area was uploaded by L.A. Bruijnzeel on Jan 30, 2016
Content may be subject to copyright.
FOREST HYDROLOGY
Impacts of Forest Plantations
on Streamflow
D F Scott, Okanagan University College, Kelowna,
Canada
L A Bruijnzeel, Vrije Universiteit, Amsterdam, The
Netherlands
R A Vertessy, CSIRO Land and Water, Canberra,
Australia
I R Calder, University of Newcastle upon Tyne, UK
&2004, Elsevier Ltd. All Rights Reserved.
Scott, D F
Okanagan University College
Kelowna, British Columbia VIV 1V7, Canada
Bruijnzeel, L A
Faculty of Earth and Life Sciences
Vrije Universiteit
1081 HV Amsterdam, The Netherlands
Vertessy, R A
Cooperative Research Centre for Catchment Hydrology
CSIRO Land and Water
Canberra, ACT 2601, Australia
Calder, I R
Centre for Land Use and Water Resources Research
University of Newcastle upon Tyne
Newcastle upon Tyne, NE1 7RU, UK
Characteristics of Plantation Forests
Tree plantations for the production of timber have
been established for more than a century (tropical
teak and mahogany plantations, for instance, date
back to the mid-nineteenth century), but it is mainly
in the last few decades that an exponential expansion
of this form of land use has occurred. Taking one of
the most popular tree types for plantations as an
example, eucalyptus have been planted on an
estimated 17 million ha worldwide, of which more
than 90% have been established since 1955 and
roughly 50% during the last decade. The total
plantation area around the globe is 187 million ha
of which over % are in Asia, with Europe having the
next largest share (17%). Eucalyptus and Pinus are
the dominant genera within the broadleaved and
coniferous plantations, respectively. Although forest
plantations only occupy about 5% of the world’s
forest area, they are estimated, as of 2000, to supply
35% of all roundwood, a figure that is expected to
rise to 44% by 2020 as natural forests continue to
decline and demands keep rising.
Plantations are typically established at a regular
spacing (1000–2000 stems ha
1
), and individual
stands (compartments) have the same age and are
composed of a single species or clone. Often
plantations are particularly productive because the
tree species being grown are exotic to the area and
thus free of their native pests and diseases. Generally,
a distinction is made between industrial plantations
(aimed at producing wood for commercial purposes,
including construction timber, panel products, furni-
ture timber, and paper pulpwood) and nonindustrial
plantations (aimed at fuelwood production, protec-
tion of catchment areas for soil and water conserva-
tion, provision of wind- or fog breaks, etc.).
Scope
There appears to be a significant disparity between
public and scientific perceptions of the hydrological
role of forests in general and of plantations in
particular. Arguably, the contrasts in views are
especially pronounced in tropical regions where calls
for massive reforestation programs to restore dry
season flows as well as to suppress flooding and
stream sedimentation are heard the most frequently.
Often, however, these expected hydrological benefits
are not realized and in a number of cases forest
plantations have even been observed to aggravate the
situation. This article recapitulates the current
understanding of how forest plantations affect the
hydrological functioning of catchments. Other arti-
cles outline the principles of the forest hydrological
cycle (see Hydrology: Forest Hydrological Cycle
(00206)) and indicate the hydrological effects of
various forest management activities (see Forest
Hydrology: Impacts of Forest Management on
Streamflow (00269)) and forest conversion to other
land uses. Strictly speaking, the term ‘‘reforestation’’
should be used to describe the planting of trees in
areas that were once covered by natural forest
whereas ‘‘afforestation’’ applies to plantation estab-
lishment in areas that are too dry to support natural
forest vegetation. To avoid semantic problems, the
term ‘‘forestation’’ is used mostly in the following to
denote either type of planting.
The Forest Plantation Water Budget
The hydrology of tree plantations is most easily
discussed with the aid of a simple water budget
3B2v7:51c
GML4:3:1EFORS 00272 Prod:Type:
pp:1211ðcol:fig::NILÞ
ED:Mahashankar
PAGN: umadevi SCAN:
A0005
S0005
P0005
P0010
S0010
P0015
S0015
P0020
equation, most simply expressed in equivalent units
of water depth (mm per unit of time):
P¼ET QþDS
where Pis total precipitation (mostly rainfall, some-
times also fog or snow), ET the sum of various
evaporation components (often referred to as evapo-
transpiration), Qthe surface runoff or streamflow,
and DSthe change in (subsurface) storage of water in
the catchment (soil water and groundwater reserves).
Evapotranspiration ET dominates the water bal-
ance of all but the most humid forest plantations.
Beyond an annual precipitation of c. 2000 mm and
under conditions of lowered evaporative demand
(e.g., montane or coastal fog belts) the balance
between evaporation and streamflow tips toward
streamflow. There are two main components to
forest ET: transpiration (the water which is taken
up from the soil by roots and passes through the trees
to be transpired from the stomata of the leaves, E
t
)
and interception (the water that is caught in the
canopy and evaporates directly back into the atmo-
sphere without reaching the ground, E
i
). Under
closed canopy conditions, usually rather minor
additional components of evaporation are evapora-
tion from the soil surface (E
s
), which in a forest
includes interception by the litter layer, and evapora-
tion from understory vegetation. The presence or
absence of a forest cover has a profound influence on
the magnitude of ET, and by implication, also on
streamflow Q.
Rainfall Interception
Compared to short, simple vegetation canopies
(grassland, agricultural crops), tree plantations in-
crease evaporation losses by intercepting a larger
portion of incident rainfall. Generally, annual inter-
ception totals associated with the dense canopies
typical of evergreen coniferous plantations are higher
than those of deciduous broadleaved forests. Inter-
ception is also particularly high (expressed as a
fraction of total precipitation) where rainfalls are
frequent but of low intensity, especially where the
evaporation process is aided by the influx of
relatively warm air striking over a cooler vegetation
surface as is often observed in near-coastal areas. An
example of this effect comes from the UK where
conifer plantations have been established in upland
heath and grasslands in Scotland and Wales. Here the
nature of the precipitation, proximity to the ocean,
and the change in canopy density may increase
interception losses to as much as 35–40% of annual
precipitation.
At the other end of the interception spectrum (E
i
c.
6%) are the Eucalyptus plantations of the humid,
subtropical eastern escarpment in South Africa. This
is an area of high seasonal rainfall (1200–1500 mm),
much of which falls in the form of infrequent storms
of short duration but high intensity. Interception
losses from pine plantations in the same area are
somewhat higher (13%), reflecting their denser
canopies compared to the more open canopies of
the eucalypt stands. In the cooler southwestern part
of South Africa, an area of winter rainfall of lower
intensity, interception losses from pine plantations
are higher again (18%) than in the pine plantations
in the subtropical areas. In the case of both pine and
eucalypt plantations, though, there is a net increase
in interception over the grasslands and scrub vegeta-
tion they replace because of the higher leaf area,
greater depth of canopy, and aerodynamic roughness
associated with timber plantations.
Rainfall interception in tropical tree plantations
ranges from relatively low values in eucalypt stands
(c. 12%) (Figure 1a), to c. 20% for broadleaved
hardwood species such as teak and mahogany
(Figure 1b), and 20–25% for pines (Figure 1c)and
other conifers (Araucaria, Cupressus), with the
higher values usually found in upland situations
where rainfall intensities are generally lower. Well-
developed dense stands of the particularly fast-
growing Acacia mangium, on the other hand, may
intercept as much as 30–40% of incident rain.
Typical interception values for the rainforests re-
placed by these plantations range from 10–20% in
most lowland situations to 20–35% in montane
areas.
Transpiration
Transpiration (soil water uptake) is the second large
component in the evaporation budget of forest
plantations. Usually, plantation water uptake rates
are similar to those of natural forest occurring in the
area of planting but under certain conditions water
use of the (usually exotic) newcomers may be higher,
particularly under subhumid conditions where the
natural vegetation consists of more open woodland
or scrub. Examples include the replacement of dry
forest/scrub by fast-growing plantations of Eucalyp-
tus camaldulensis and E. tereticornis in South India,
and by E. grandis in southeastern Brazil and South
Africa. Likewise, water uptake rates reported for
(vigorously growing and densely stocked) stands of
Acacia mangium in Malaysia and for various species
planted in the lowland rainforest zone of Costa Rica
are such that they must exceed the water use of the
old-growth rain forests they are replacing, possibly
P0025
S0020
P0030
P0035
P0040
S0025
P0045
2
FOREST HYDROLOGY /Impacts of Forest Plantations on Streamflow
EFORS : 00272
by 100–250 mm year
1
. Even greater differences in
transpiration can be expected where plantations are
established in areas with (natural) grassland or
degraded cropland. For example, whilst forest water
uptake under humid tropical conditions typically
exceeds that of pasture by about 200 mm year
1
, the
difference may increase to as much as 500–
700 mm year
1
under more seasonal conditions.
Such differences reflect the contrasting rooting
depths of trees and grassland as well as the tendency
for natural grasslands to go dormant during ex-
tended dry periods while the (exotic) trees continue
to take up water.
F0005 Figure 1 The contrasting canopies of (a) Eucalyptus spp., (b) teak (Tectona grandis), and (c) pines (Pinus caribaea) lead to
differences in amounts of rainfall interception and in the drop size spectra (and thus eroding power) of water dripping from the canopy.
(Photographs by LA Bruijnzeel.)
FOREST HYDROLOGY /Impacts of Forest Plantations on Streamflow
3
EFORS : 00272
Total Evapotranspiration (ET)
It follows from the above increases in rainfall
interception and transpiration that are typically
associated with the establishment of tree plantations
in areas of (natural) grasslands or (degraded) crop-
land that overall ET totals can be much increased
after forestation. As shown in Figure 2, total ET
values for actively growing plantations may ap-
proach 1500 mm year
1
and, occasionally, as much
as 1700–1900 mm year
1
. Such very high values
must be considered the exception rather than the
rule, however, and probably reflect the advection of
warm, dry air flowing in from adjacent grassland
areas which tends to greatly enhance evaporation
rates. Nevertheless, the fear is justified that the much
increased water use of tree plantations compared to
the pastures and crops they replace will lead to
substantial reductions in catchment water yields,
particularly during the dry season, if entire catch-
ments are planted.
Effects of Tree Plantations on Streamflow
Effects of Associated Land Management
In discussing the hydrological effects of establishing
timber plantations, it is important to be clear about
the site-specific conditions and management prac-
tices associated with the land-use change and their
contribution to the effect of a change in land use.
Such background information may be very important
in assessing the overall hydrological effect of the
plantations. The following examples illustrate the
need to specify more than simply the change in
vegetation cover itself.
Firstly, a forest or plantation would be expected,
normally, to have a continuous groundcover of leaf
or needle litter and some shade-tolerant shrubs. In
parts of southern China and adjacent countries,
however, all litter and understory plants may be
collected for fuel, a practice that has a profound
influence on the occurrence of surface runoff and
erosion in the plantation (Figure 3). Elsewhere (as on
the Indian subcontinent) forests and plantations are
used to graze cattle, a practice that requires regular
burning to stimulate the growth of fresh grass shoots.
The combined effect of burning and trampling by
livestock may promote massive surface erosion,
sometimes to the extent of initiating gullies.
Contrary to popular belief, it is not the intercep-
tion of rainfall by the main tree canopy that protects
the soil underneath against the erosive impact of the
rains. Rather it is the combined protection afforded
by the understory vegetation and a well-developed
litter layer that prevents the soil from being eroded.
In fact, the erosive power of rain dripping from the
canopies of tall trees is often greater than that in the
open, because of the associated increases in drop
sizes. The largest increases in drop diameters are
observed for drip from large-leaved trees such as teak
or Gmelina, whereas those falling from eucalyptus or
pines are more modest in size (Figure 4). Such
findings underscore the importance of maintaining a
good groundcover in plantations if runoff and
erosion problems are to be avoided (cf. Figure 3).
A final example of the importance of management
comes from the wet and peaty hill country of
Scotland and the English borderlands, where surface
drains are usually excavated prior to the planting of
400
600
800
1000
1200
1400
1600
1800
2000
1000 1500 2000 2500 3000 3500 4000 4500 5000
Mean annual
p
reci
p
itation
(
mm
)
Annual evaporation (mm)
Zhang Grass Zhang Forest Pine plantation Eucalyptus plantation
Other hardwood
plantation
Fynbos Grassland n. Forest
F0010 Figure 2 Total evaporation (ET) from forest plantations and other vegetation types as a function of precipitation. Data mostly from
humid tropical (Bruijnzeel 1997) and South African plantations (courtesy of D Le Maitre, CSIR, South Africa, unpublished compilation).
The curves define average forest and grassland water use in southeastern Australia (adapted from Zhang, et al. 1999), and have been
extrapolated for rainfall 42000 mm.
S0030
P0050
S0035
S0040
P0055
P0060
P0065
P0070
4
FOREST HYDROLOGY /Impacts of Forest Plantations on Streamflow
EFORS : 00272
coniferous trees (mostly Sitka spruce, Picea sitch-
ensis), to improve the success of tree crop establish-
ment. However, the influence of the drainage ditches
on streamflow has proved more important than the
vegetation change from heath and moorland to tree
plantation itself.
Forestation and Water Yield
The increased evaporation from timber plantations
replacing shorter vegetation types (Figure 2), not
unexpectedly, translates into decreases in annual
streamflow totals after plantation establishment.
Although there are no stringent (paired) catchment
experiments in the humid tropics proper, there is
overwhelming evidence to this effect from the
F0015 Figure 3 The practice of repeated removal of needle litter from coniferous forests in parts of mainland southeast Asia often leads to
dramatic increases in surface runoff and erosion. (Photographs by courtesy of C Cossalter.)
S0045
P0075
FOREST HYDROLOGY /Impacts of Forest Plantations on Streamflow
5
EFORS : 00272
subhumid tropics (notably India), the subtropics
(mostly South Africa), and the temperate zone
(including southeast Australia and New Zealand).
Considerable differences have been observed be-
tween species but these are not necessarily the same
in different areas. For example, in southeast Aus-
tralia and New Zealand greater reductions in flow
were observed after planting pines (Pinus radiata)on
grassland than in the case of planting eucalyptus
(Figure 5). Conversely, in South Africa, other
variables being equal, the effect of planting Euca-
lyptus grandis was more pronounced than that of P.
radiata or P. patula (see Figure 8 below). Such
contrasts mainly reflect differences in growth perfor-
mance between regions and to a lesser extent
differences in rainfall interception dynamics.
Published experimental results often represent the
maximum possible impacts on streamflow. In the real
world, variations in site characteristics and planta-
tion management may exert a moderating influence
on the hydrological impacts of forestation. Moderat-
ing factors include the fraction of the catchment
planted, planting position within the catchment
(upstream or downstream parts, close to or away
from the streams, blocks vs. strips, etc.), and
variations in stand age and productivity between
species. These factors are elaborated upon briefly
below.
Catchments are rarely completely planted with
trees because some land is usually reserved for other
uses or it may be inaccessible or otherwise unsui-
table. The classical forest hydrology literature
suggests that the magnitude of the change in
catchment water yield is linearly proportional to
the percentage of catchment planted or cleared, with
increases in flow after forest removal and reductions
after forestation (Figure 6). Hence, in the case of
plantations, one could assume that if only half of a
grassland catchment would be forested then the
estimated reduction in mean annual runoff would
also be about half of the maximum reduction
predicted by Figure 5 for a given annual rainfall
total (assuming that plantation position in the
catchment does not influence the result).
Few experimental data are available on the
influence of plantation position on catchment water
balance changes. Under humid conditions in the
eastern USA, the reverse operation (i.e., forest clear-
cutting) did not show a significant difference in
streamflow response after cutting the upper half of
the catchment or the lower half. Also, elimination of
the vegetation around streams in one experiment in
the summer-rainfall zone of South Africa did not lead
to greater increases in streamflow than when remov-
ing an equal area of forest away from the stream.
However, several other experiments in South Africa
700
600
500
400
300
200
100
0
0 10 20 30 40 50 60 70 80 90 100
Annual streamflow change (mm)
Percentage of area reforested or deforested
SEE = 89 mm
y = 3.26x
F = 66.4
S = 0.01
r2 = 0.50
F0030Figure 6 Changes in annual water yield vs. percentage forest
cover change (solid circles denote experimental data of Bosch
and Hewlett (1982); open circles those of Trimble, et al. (1987).
SEE, standard error of estimate. (After Trimble, et al. (1987).)
0
0.2
0.4
0.6
0.8
1
01 5
Drop diameter (mm)
Eucalyptus camaldulensis Tectona
g
randi
s
Pinus caribaea
Normalized cumulative
volume
234 6
F0020 Figure 4 Characteristic drop size spectra for rain dripping from
pine trees (Pinus caribaea), teak (Tectona grandis), and
eucalyptus (Eucalyptus camaldulensis) as measured in South
India. (After Calder (1999).)
0
100
200
300
400
500
600 800 1000 1200 1400 1600
Eucalypt forest
Pine forest
Mean annual runoff reduction (mm)
Mean annual rainfall
(
mm
)
Tumut
Glendhu
Waiwhiu
Lidsdale
F0025 Figure 5 Potential reduction in mean annual streamflow
estimated to result from forestation of grasslands with eucalyptus
and pines in southeast Australia. Shown (as symbols) are field
data from four pine forestation experiments in Australia and New
Zealand. (After Vertessy, et al. (2003).)
P0080
P0085
P0090
6
FOREST HYDROLOGY /Impacts of Forest Plantations on Streamflow
EFORS : 00272
showed that an area of plantation near streams had
roughly double the effect of the same area of mid-
slope planting.
Such contrasting results may be explained in terms
of average soil water surplus or deficit, depth to the
groundwater table and slope morphology. All these
factors influence hillslope hydrological behavior.
Where rainfall is plenty, slopes steep and convex,
and the groundwater table rather deep (say, more
than 3 m), no major spatial effect is expected. This is
because rainwater infiltrating into the soil percolates
more or less vertically to the water table, then moves
laterally as groundwater to the nearest stream
without being taken up again by the roots of the
trees. Conversely, where soil water is scarcer, slopes
gentle and concave, and depth to the water table
shallow, a more pronounced effect is possible
because trees located closer to the stream will have
more ready access to the groundwater table. As such,
they are likely to consume more water than trees
further away from the stream that have less direct
access to groundwater to supplement diminished soil
water reserves.
Furthermore, there is the intuitive notion that the
further away one gets from a stream, the smaller the
probability that water infiltrating into the soil will
actually contribute to streamflow. These ideas have
been tested in modeling experiments in the context of
southeast Australia, the results of which lend support
to the notion that plantation position could affect
catchment water yield under conditions of low
rainfall (700 mm), gentle slopes, and high water-
tables (Figure 7). Indeed, the predicted effect on
streamflow of tree planting differed strongly depend-
ing whether forestation started at the top of the
hillsides and progressively moved downslope or vice
versa. The curves of Figure 7 also suggest that under
the prevailing conditions planting of the lower 30%
of the catchment would have a much greater impact
than planting the uppermost 30%. Similarly, a
related modeling study indicated that planting trees
in strips about 40 m wide parallel to the contour with
bands of pasture in between leads to greater tree
water use and better growth than when the same
10
80
70
60
50
40
30
20
0 10080604020
Proportion with trees (%)
Annual runoff (mm)
Bottom−up
Top−down
F0035Figure 7 Results from a numerical modeling experiment
showing two sets of predictions of annual streamflow after
planting trees on a catchment under pasture in central New South
Wales, Australia (mean annual rainfall 700 mm). The upper curve
(solid line) shows changes in annual flow with forestation starting
at the top of the catchment and progressing downslope. The
lower curve (dashed line) shows the comparative response when
forestation starts at the bottom of the catchment and progresses
upslope. (After Vertessy, et al. (2003).)
0
100
200
300
400
500
600
02468101214161820
Years after
p
lantin
g
Flow reduction (mm)
Westfalia, E. grandis Mok-A, E. grandis
Mok-B, P. patula CP3, P. patula
Lb-B, P. radiata
F0040 Figure 8 Reductions in streamflow as measured in five catchment afforestation experiments in South Africa. The curves are scaled
for 100% planting of the catchment and smoothed to the mean annual runoff (MAR) prior to planting. (Based on Scott and Smith
(1997).)
P0095
P0100
FOREST HYDROLOGY /Impacts of Forest Plantations on Streamflow
7
EFORS : 00272
number of trees are planted in a single block at mid-
slope position.
More work is needed to ascertain optimal planta-
tion positions to minimize the hydrologic impacts of
forestation under contrasting climatic and topo-
graphic conditions. Process-based, spatially distrib-
uted hydrological models can be used to assess how
different planting strategies would impact on catch-
ment flow regimes. Whilst such models are difficult
to set up and apply, the effort is surely worthwhile
given the level of investment that goes into planning
any significant forestation initiative.
Much can be learned on the effects of species,
plantation age, and vigor from a particularly
comprehensive series of long-term paired catchment
studies of the hydrological effects of afforesting
natural grasslands and scrublands in subtropical
South Africa. Ten paired catchment experiments
have studied the effects of afforestation with Pinus
radiata,P. patula, and Eucalyptus grandis within
catchments. The research sites are all in the high
rainfall zone of South Africa (mean annual precipita-
tion 1100–1600 mm). Experimental control was
provided by catchments kept under native vegeta-
tion. Although generally steep, the catchments have
deep, well-drained soils and show very low storm-
flow response to rainfall. The catchments are all in
good hydrological condition (i.e., no significant
surface erosion); thus, the experimental comparison
is between the two vegetation covers, reflecting,
ultimately, the differences in total evaporation.
The resulting streamflow reductions over time
after planting follow a sigmoidal pattern comparable
to a growth curve (Figure 8). There are clear
differences between the effects of eucalyptus and
pines, but there is also a large amount of variation
from year to year within a single experiment and
between different experiments, even in comparable
catchments in one locality. The highest flow reduc-
tions occur once the tree crop is mature, and range,
for a 10% level of planting, from 17.3 mm or 10%
year
1
in a drier catchment to 67.1 mm or 6.6%
year
1
in wetter catchments (Figure 8). As such,
relative streamflow reductions (%), for a set age, are
greater in drier catchments but absolute reductions
(mm) are greater in wetter catchments. In other
words, the reductions are positively related to water
availability. The lower of these reductions in stream-
flow are similar to results obtained after planting E.
globulus in high elevation grassland areas in the
subhumid South of India (c. 20 mm per 10% forest
year
1
) whereas the highest reductions in South
Africa rather resemble the changes observed after
planting P. caribaea on seasonal grasslands in Fiji
(50–60 mm per 10% year
1
). Similar effects on
streamflow have been recorded under the more
temperate conditions of New Zealand (see also
Figure 5), where conversion of pastures and tussock
grassland to P. radiata plantations, over a range of
climates, led to streamflow reductions of 20–45 mm
year
1
per 10% of catchment planted, the amount
again being dependent on water availability.
The timing of the first significant reductions in
flow after planting varies quite widely depending on
the rate at which catchments are dominated by the
plantation crop. The pine plantations in the high
altitude grasslands at Cathedral Peak in South Africa
(CP in Figure 8) usually took several years to have a
clear impact on streamflow. However, the same
species of pine had an earlier effect on streamflow
(within 3 years) under the drier conditions prevailing
in the Mokobulaan B catchment in Mpumalanga
Province (Mok-B in Figure 8). Other conditions
remaining the same, eucalyptus have a slightly earlier
impact on streamflows than pines in South Africa,
normally within 2–3 years.
A key factor influencing the degree of streamflow
reduction after forestation is the vigor of the trees.
Usually, there is a close link between the growth rate
of a plantation and its overall water uptake. A new
finding from the South African afforestation experi-
ments is that the flow reductions are diminishing
again during the postmaturation phase of the
plantations, both in the case of pines (after about
30 years) and in at least one of the two eucalypt
experiments (after 15 years). This undoubtedly
mirrors the gradually decreased vigor of older trees
as has also been observed in old-growth native
eucalypt forest in southeast Australia and tropical
rainforest in Amazonia. In industrial plantation
forestry, short- and medium-length tree rotations
will tend to keep the trees in their peak water use
phase, but longer rotation crops, such as those aimed
at producing good quality saw timber, are more
likely to have a smaller effect on water yield later on
in the rotation.
Forestation and Low Flows
Declines in streamflow following the establishment
of plantations are recorded in all components of the
annual hydrograph (i.e., stormflows and baseflows).
In South Africa, effects on total and low flows follow
the same pattern, but low flows are decreased more
than are total flows at the same age (Figure 9).
Similar effects have been found in the temperate zone
as well as in Fiji, India (even more so after coppicing
and resprouting), and Malawi. The effect of foresta-
tion on low flows in subhumid areas has two
supposed sources. First, exotic plantations, in con-
P0105
P0110
P0115
P0120
P0125
S0050
P0130
8
FOREST HYDROLOGY /Impacts of Forest Plantations on Streamflow
EFORS : 00272
trast to the native grasses or scrub vegetation they
replace, do not go dormant in the dry season. The
second cause, though less easily quantified, is that of
steadily reducing soil water reserves as the trees
mature. Low flows are a reflection of the amounts of
soil water and groundwater stored in the catchment
and as these are steadily depleted by tree water
uptake so will low flows diminish accordingly. It is
clear from the South African experiments that total
water use by the tree crop can exceed annual rainfall
in many years and that, once dry season flow has
ceased altogether, the occurrence of rainstorms may
not easily cause the streams to flow again.
Strongly reduced baseflows after forestation of
(nondegraded) grassland or scrubland can thus be
expected to be a generally occurring phenomenon.
The magnitude of this effect is probably related to
the capacity of the soils to store water and to the
extent that this water can be accessed by the roots of
the tree crop. Thus, where the new trees are able to
occupy a much greater volume of soil through their
deeper roots, reductions in baseflows following
forestation can be expected to be proportionately
larger than in situations where rooting volume is
restricted.
Finally, it is important to bear in mind that the
above examples concern situations where the soil is
not degraded and rainfall infiltration generally
proceeds unimpaired. Under such conditions, stream-
flow amounts will simply reflect the change in
vegetation water use and low flows will be thus
(much) reduced (see Figures 2 and 6). However, in
areas with degraded, compacted soils where much of
the rain may run off along the surface as overland
flow (and therefore not contribute to soil water
reserves), the planting of trees can be expected to
ultimately have a positive effect on infiltration.
Theoretically, the extra water entering the soil
through improved infiltration after forestation may
moderate or, in extreme cases, even reverse the
adverse effect of the larger water use of the trees on
streamflow. In all cases, the net effect of tree planting
on the baseflow from degraded areas will reflect a
trade-off between these two effects. Where infiltra-
tion is already sufficient to accommodate most of the
rainfall, any further improvements by forestation
will not tip the balance. Rather, water yield will be
reduced even further (Figure 6). However, where
soils are deep but overland flow during rainfall is
rampant and much is to be gained from improved
infiltration (Figure 10a), it cannot be excluded that a
net positive effect on low flows may occur. The
experimental evidence for this contention is only
indirect, however, and based on a comparison of
observed reductions in stormflow response (having a
positive effect on soil water reserves) (Figure 10a) vs.
increases in vegetation water use (having a negative
effect on soil water reserves) (Figure 2).
Forestation and Stormflows
Forest hydrological research has shown that the
influence of vegetation cover or type on catchment
runoff response to rainfall (‘‘stormflows’’) is inversely
related to the size of the rainfall event that generates
the flows. This can be explained as follows: in small
to medium storm events the combined water storage
capacity of vegetation layers, litter, surface depres-
sions, and the soil mantle will be considerable
relative to the amount of rain delivered by the storm.
As a result, the associated catchment response will be
much reduced in the case of a good forest cover. The
soil mantle is potentially the largest water store, but
its capacity to accommodate additional rain varies as
a function of soil wetness. Where previous uptake by
the vegetation has depleted soil water reserves (as is
often the case during summer), storage capacities,
and thus stormflow reduction, will be relatively high
(Figure 10a). However, once the soil has become
thoroughly wetted by previous rains (typically during
winter or the main rainy season), very little
opportunity to store additional water will remain,
regardless of vegetation type (Figure 10b). In
addition, as rainfall events increase in size, so does
the relatively fixed maximum storage capacity of the
soil become less important in determining the size of
the stormflows. In other words, the presence or
absence of a well-developed forest cover has a
significant effect in the case of small events but this
typically makes very little difference (less than 10%)
0
10
20
30
40
50
60
70
80
90
100
-1 1 3 5 7 9 11 13 15
Years after afforestation
Reduction in streamflow (%)
Total flow
Low flow
F0045 Figure 9 Pooled results from two catchment eucalyptus
afforestation experiments in South Africa, showing the pattern
of flow reductions as a function of plantation age, and illustrating
the greater and earlier effects on the low flow component. (After
Scott and Smith (1997).)
P0135
P0140
S0055
P0145
FOREST HYDROLOGY /Impacts of Forest Plantations on Streamflow
9
EFORS : 00272
in the case of truly large events (floods) generated by
extreme and prolonged rainfall (Figure 10b). Under
such conditions, runoff response is governed almost
entirely by the capacity of the soil to accommodate
and transfer the rain.
However, where degradation of a catchment’s soils
has produced strong reductions in canopy and
groundcover (including litter), and above all in
infiltration capacity and soil depth through contin-
ued erosion (and thus overall soil water storage
opportunity), reforestation could clearly lead to an
improvement of most or all these factors over time.
These ideas are conceptualized in Figure 11.
Concluding Remarks
Catchment experiments all over the world have
demonstrated convincingly that total amounts of
streamflow emanating from catchments where forest
plantations have replaced (natural) grassland or
scrubland, or (degraded) cropland, are invariably
much reduced. In addition, the reductions in base-
flows during the dry season are relatively greater
than during the wetter season. Small to medium-
sized stormflows are also reduced significantly by
forestation but the effect on occurrence and size of
flood peaks associated with truly large rainfall events
is very limited.
These observations differ strongly from the pop-
ular view held by many foresters, policy-makers, and
the public at large that forestation will lead to (more
or less rapidly) increased streamflows and the
High
Low
Shallow Deep
Soil depth: Available storage
Storm response (QS/P)
Degraded grassland
Grassland
Old forest
Immature forest
F0055Figure 11 Postulated generalized relationship between catch-
ment storage capacity and stormflow response to rainfall, as
affected by vegetation cover. (After Scott, et al. (2004).)
180
160
140
120
100
20
40
60
80
Peak discharge in cubic feet per second
Annual frequency-equaled or exceeded
024 681012
1958
1945
1937
1935
180
160
140
120
100
20
40
60
80
Peak discharge in cubic feet per second
Five-
y
ear frequenc
y
-equaled or exceeded
0 5 10 15 20 25 30
(a) (b)
1935−36 to 1939−40
1951−52 to 1955−56
F0050 Figure 10 Frequency distributions for peak discharges during (a) summer and (b) winter in the While Hollow catchment, Tennessee,
USA before (1935) and after (1937–1958) reforestation. (Modified from Tennessee Valley Authority (1961) Forest Cover Improvement
Influences upon Hydrologic Characteristics of White Hollow Watershed,1935–1958. Report no. 0-5163A. Knoxville, TN: Tennessee
Valley Authority.)
P0150
S0060
P0155
P0160
10
FOREST HYDROLOGY /Impacts of Forest Plantations on Streamflow
EFORS : 00272
elimination of flooding. Although the establishment
of forest plantations on degraded land will improve
the soil’s capacity to absorb rainfall, this is likely to
take at least several decades. However, because water
use by the trees is much increased within a few years
already compared to that of the former vegetation,
the balance of probability is that low flows will also
be reduced in this case. Establishing the precise
hydrological effects of reforesting areas in various
stages of soil degradation constitutes a prime
research need.
List of Technical Nomenclature
See also:Forests, Tree Physiology and Climate
(00114). Hydrology: Forest Hydrological Cycle (00206);
Forests and Streamflow (00208); Forests and Water
Quality (00209); Impacts of Forest Management on
Streamflow (00269). Silviculture: Forest Plantations
(00212); Short Rotation Forestry for Biomass Production
(00236). Soils and Site: Forests and Soil Development
(00239); Soil and its Relationship to Forest Productivity
and Health (00240); Water Storage and Movement
(00249). Tree Physiology: A Whole Tree Perspective
(00098); Root System Physiology (00102).
Further Reading
Bridges EM, Hannam ID, Oldeman LR, et al. (2001)
Response to Land Degradation. Enfield, NH: Science
Publishers Inc.
Bruijnzeel LA (1997) Hydrology of forest plantations in the
tropics. In: Nambiar EKS and Brown AG (eds) Manage-
ment of Soil, Nutrients and Water in Tropical Plantation
Forests, pp. 125–167. Canberra, Australia: ACIAR/
CSIRO/CIFOR.
Bruijnzeel LA (2004) Tropical forests and environmental
services: not seeing the soil for the trees? Agriculture,
Ecosystems and Environment (in press).
Calder IR (1999) The Blue Revolution: Land Use and
Integrated Water Resources Management. London:
Earthscan Publications.
Calder IR, Rosier PTW, Prasanna KT, and Parameswar-
appa S (1997) Eucalyptus water use greater than rainfall
input: a possible explanation from southern India.
Hydrology and Earth System Science 1: 249–256.
Dye PJ (1996) Climate, forest and streamflow relationships
in South African afforested catchments. Commonwealth
Forestry Review 75: 31–38.
Fahey B and Jackson RJ (1997) Hydrological impacts of
converting native forests and grasslands to pine planta-
tions, South Island, New Zealand. Agricultural and
Forest Meteorology 84: 69–82.
Gilmour DA, Bonell M, and Cassells DS (1987) The effects
of forestation on soil hydraulic properties in the Middle
Hills of Nepal: a preliminary assessment. Mountain
Research and Development 7: 239–249.
Johnson R (1998) The forest cycle and low river flows: a
review of UK and international studies. Forest Ecology
and Management 109: 1–7.
McJannet DL, Silberstein RP, and Vertessy RA (2001)
Predicting the water use and growth of plantations on
hillslopes: the impact of planting design. In Proceedings
of MODSIM2001, International Congress on Modelling
and Simulation, December 2001, Canberra, vol. 1, pp.
455–460.
Scott DF (1999) Managing riparian zone vegetation to
sustain streamflow: results of paired catchment experi-
ments in South Africa. Canadian Journal of Forest
Research 29: 1149–1157.
Scott DF and Smith RE (1997) Preliminary empirical
models to predict reductions in total and low flows
resulting from afforestation. Water South Africa 23:
135–140.
Scott DF, Bruijnzeel LA, and Mackensen J (2004) The
hydrological and soil impacts of forestation in the
tropics. In: Bonell M and Bruijnzeel LA (eds) Forests–
Water–People in the Humid Tropics, pp. 00–00. Cam-
bridge, UK: Cambridge University Press.
Sikka AK, Samra JS, Sharda VN, Samraj P, and Lakshma-
nan V (2003) Low flow and high flow responses to
converting natural grassland into bluegum (Eucalyptus
globulus) in Nilgiris watersheds of South India. Journal
of Hydrology 270: 12–26.
Trimble SW, Weirich FH, and Hoag BL (1987) Reforesta-
tion and the reduction of water yield on the southern
Piedmont since circa 1940. Water Resources Research
23: 425–437.
Vertessy RA, Zhang L, and Dawes WR (2003) Plantations,
river flows and river salinity. Australian Forestry 66: 55–
61.
Waterloo MJ, Bruijnzeel LA, Vugts HF, and Rawaqa TT
(1999) Evaporation from Pinus caribaea plantations on
former grassland soils under maritime tropical condi-
tions. Water Resources Research 35: 2133–2144.
Zhang L, Dawes WR, and Walker GR (2001) Response of
mean annual evapotranspiration to vegetation changes at
catchment scale. Water Resources Research 37: 701–
708.
Zhou GY, Morris JD, Yan JH, Yu ZY, and Peng SL (2001)
Hydrological impacts of reafforestation with eucalyptus
and indigenous species: a case study in southern china.
Forest Ecology and Management 167: 209–222.
S0065
FOREST HYDROLOGY /Impacts of Forest Plantations on Streamflow
11
EFORS : 00272
