Geoscience Reference
In-Depth Information
Stemflow and interception loss from the trunks
In several past analyses of interception the water balance of the trunks and stems has been
treated separately from that of the leaves (Rutter
et al
., 1975; Gash
et al
., 1995). Since
the evaporation from the trunks is usually very small compared with the evaporation from
the canopy, the resulting losses consist mainly of the evaporation of the water remaining
on the trunks after the end of stemflow. Thus, for precipitation events long enough to
saturate the trunk storage, the total loss is equal to the maximal trunk storage
S
tic
. When the
precipitation does not quite saturate the trunk storage, by analogy with (3.14) or (3.19), the
loss may be taken as (
p
t
PD
), where
p
t
is the proportion of the precipitation that is diverted
to stemflow.
In most situations, however, these losses are considerably smaller than those from the
leaves of the canopy. For example, for pine forests in Great Britain (Gash, 1979; Gash
et al
.,
1980), the trunk losses were found to be about 2% to 9% of the total interception loss; for
the Amazonian rain forest a value of about 9% was observed (Lloyd
et al
., 1988).
3.4.3 Experimental determination of the vegetation structure parameters
The main surface parameters controlling the interception loss are
S
ic
and
c
; the surface
roughness
z
0
probably plays only a minor role through its effect on the evaporation rate.
The storage capacity
S
ic
is usually estimated with a procedure proposed by Leyton
et al
.
(1967; see also Gash and Morton, 1978). The method is based on the observation that, as
indicated in Equation (3.6), the loss is equal to the canopy storage when evaporation is equal
to zero. As before, let
P
represent the average rainfall rate during an event of duration
D
.
Then, in a plot of the gross, i.e. total precipitation (
PD
), versus the net precipitation, i.e.
throughfall [(1
−
c
)
PD
], for a number of observed precipitation events,
S
ic
can be taken
as the intercept of the lower envelope with a slope of unity; the lower envelope repre-
sents the events with minimal
E
i
, so that
PD
=
(1
−
c
)
PD
+
S
ic
. The data points must be
taken from events of sufficiently long duration, to ensure that the canopy is fully saturated.
This is illustrated in Figure 3.19. Observe, however, that the vertical axis should represent
(1
−
p
t
)
PD
, instead of
PD
, to account for stemflow, but the difference is usually small and
can be neglected. The free throughfall coefficient
p
can be determined from throughfall
measurements for small storms insufficient to saturate the canopy (Gash and Morton, 1978)
and the canopy cover can then be obtained by assuming
c
=
−
p
).
Typical values of the specific canopy storage capacity (
S
ic
/
c
) (i.e. the storage capacity
per unit area of cover) and of
c
are respectively, 0.8-1.2 mm and 0.68-1.00 for dense pine
forest (Gash and Morton, 1978; Gash
et al
., 1980), around 0.56 mm and 0.45 for sparse
pine forest (Gash
et al
., 1995), 0.8 mm and 0.92 for Amazonian rain forest (Lloyd
et al
.,
1988), and 0.64 mm and 0.64 for sparse pine forest and 0.35 mm and 0.60 for eucalyptus
forest (Valente
et al.
, 1997). For grasses ranging in height between 0.1 m and 0.5 m, (
S
ic
/
c
)
values ranging between 0.43 and 2.8 mm have been reported (Merriam, 1961).
The stemflow parameters
S
tic
and
p
t
can be determined as, respectively, the mean slope
and the intercept of the regression of stemflow versus precipitation for each tree on which
measurements are made (see, for example, Gash and Morton, 1978). Typical values for
S
tic
and
p
t
are, respectively, 0.014-0.74 mm and 0.016-0.29 for dense pine forest, 0.17 mm and
0.0275 for sparse pine forest, 0.15 mm and 0.036 for Amazonian forest, 0.019 mm and
0.0038 for Mediterranean sparse pine forest and 0.027 mm and 0.017 for eucalyptus forest.
(1
