Assessing the Hydrological Effects of Forest Plantations in Brazil - River Conservation and Management - page 59

'[(P EC -Q)-AE]/AE' ratio are 7%, 13% and 26%
respectively. These patterns are consistent with
conclusions made by Farley et al . (2005) that
absolute reduction in annual runoff may be greater
in wetter catchments, but proportionately the
impact of reduction is likely to be more severe in
drier areas.
The analysis in Table 5.1 also shows the increase
in evapotranspiration [(P EC -Q)-AE] associated
with forest plantations on a catchment scale,
based on a comparison of the mass water
balance with the estimated climatic actual
evapotranspiration (AE). This is equivalent to
an increase in the 'green water flux' caused by
plantations ( sensu Falkenmark and Folk, 2002).
Consequently, an important parameter in Table
5.1 is catchment water yield ('Q'). Calculated Q
values indicate that flow was maintained in all
experimental catchments, suggesting that despite
increasing evapotranspiration, eucalyptus and pine
plantations in the climatic and soil conditions
of the study locations do not affect base-flow
water

actual evapotranspiration (AE) (Thornthwaite and
Matter, 1955). This confirms that models of
estimated potential evapotranspiration based on
temperature and solar radiation provide the best
stream-flow simulations for use in catchment
rainfall-runoff relationships (Oudin et al ., 2005).
The (P EC -Q) value of catchment F in Table 5.1
is also similar to the annual value of evaporation
measured by the eddy covariance method in the
same catchment (Cabral et al ., 2010).
Stape et al . (2008) found a similar relationship
between water availability and wood production of
a Eucalyptus grandis plantation in the State of Bahia,
north-east Brazil. That region has an annual rainfall
of 1040 mm, high potential evapotranspiration and
a low average water surplus. Adding irrigation
water increased wood yield by 92 m 3 ha − 1 yr − 1 ,
which is among the highest values ever reported for
forests. In rain-fed eucalyptus plots, transpiration
was 1261 mm in a year with above average annual
precipitation, but only 962 mm in an average
rainfall year. In comparison, annual transpiration
of the irrigated plots was about 1250 mm in both
years, demonstrating that the supply of water is
a key factor determining the potential growth of
eucalyptus plantations.
Analysing these and other experimental results
from eucalyptus plantations in Hawaii, Binkley
et al . (2004) concluded that the increase in wood
production from 1.4 to 3.5 kg m − 2 yr − 1 caused
by irrigation also increased water use efficiency
from 6.9 to 9.0 kg m − 3 . Therefore, the hydrological
impact of forest plantations depends in part on both
water consumption and water use efficiency - the
same amount of water can produce twice the wood
yield on sites with higher water supply.
As a result of analysing data from over 250
experimental catchments worldwide, Zhang et al .
(2001) developed a simple mathematical model
relating catchment-scale evapotranspiration with
annual precipitation. The so-called 'Zhang curves'
predict that increased water availability increases
water consumption and wood production. The
model also indicates that afforestation of former
grassland will increase water consumption, but
the

conditions, described as 'blue water' by

Falkenmark and Folk (2002).

Discussion

The effects of forest plantation on water balance
vary in Brazil, as elsewhere in the world, according
to local climatic, soil and management factors. By
analysing over 500 observations worldwide, Farley
et al . (2005) summarized the effects of afforestation
as follows: for regions where natural runoff is
less than 10% of the mean annual precipitation,
afforestation would result in a complete loss of
stream flow; where natural runoff is 30% of the
mean annual precipitation, water yield would be
cut by half or more as a result of plantation forests.
These conclusions have major implications for
catchment management and aquatic ecosystems.
Comparing the mass water balance calculated for
experimental catchments with the climatic water
balance of that particular region and taking account
of differences in local and regional precipitation,
it is clear that catchment-scale evapotranspiration
(P EC -Q)

difference in water consumption between

matches closely the climatic values of forest and grass is only evident when annual