Instruction manual

11
since the mass of air is negligible. The core sample is then immediately submerged in a
beaker of water which has been placed on an accurate mass balance. The balance reading
will indicate an immediate increase in mass, which equals the displacement of water, D
T
[kg]. The total volume,
V
T
[m
3
], of the sample is then equal to
ρ
L
times D
T
, where the
density of water,
ρ
L
, is assumed to be 1000 kg m
3
. The core sample is then oven-dried to
determine the mass of dry wood,
M
D
[kg]. The difference between the fresh weight and
the dry weight, (
M
F
-
M
D
) is equal to the mass of water,
M
L
[kg], contained in the fresh core
sample. Thus, the volume fraction of water is
F
L
=
M
L
/(
ρ
L
V
T
). Similarly, the volume
fraction of wood
F
M
=
M
D
/(
ρ
M
V
T
) where the density of dry wood,
ρ
M
equals 1530 kg m
3
.
1.6. Estimating Volumetric Sapflow
Equation (4) provides an estimate of the values of
J
at any point in the conducting
sapwood. It is widely recognized that sap flux density is not uniform throughout the
sapwood, but rather peaks at a depth of 10 - 20 mm in from the cambium. Consequently,
sampling at several depths in the sapwood is necessary to characterize the sapflow
velocity profile (Cohen et al., 1981; Edwards and Warrick, 1984; Green and Clothier,
1988). A volumetric measure of total sap flux can be obtained by the integration of these
point estimates over the sapwood conducting area. The most common approach is to fit a
least-squares polynomial to the depthwise estimates of sap flux density, and then to
integrate the fitted function over the sapwood cross section. This is the approach that we
favour. An alternative, simpler integration method, as presented by Hatton et al. (1990), is
based on a weighted average approach. According to Hatton et al. this simpler approach is
a more robust estimator of the volume flux when the velocity profiles exhibit large
curvature.
In the ANALYSIS program described later we use the most common approach to
calculate volume sap flow. Since our probes measure
J
at four radial depths, we fit a
second-order regression of the form
(
)
J r
r
r
=
+
+
α
β
γ
2
(5)