in the stem cylinder between 4 and 51 m (
Q
t_med
; kg
stem
h
– 1
)
was derived as the difference between the inflow (Q
t_4m
) into,
and outflow (
Q
t_51m
) from, this section:
Q
t_med
=
Q
t_4m
–
Q
t_51m
(2)
where Q
t_4m
is assumed daily total tree water use. Similar cal-
culations were made for branches:
Q
crown_med
= Q
crown_tot
– Q
crown_top
(3)
where Q
crown_med
is sap flow for the mid-crown, Q
crown_tot
is sap
flow for the whole crown (equal to that for the whole tree) and
Q
crown_top
is sap flow measured just below the top.
Estimation of water storage and tissue volume
We estimated wood and foliage water storage gravimetrically
(biometric samples), hydrometrically (based on sap flow) and
volumetrically (with dendrometers). For some of the water
content estimates, it was necessary to determine bole, branch
and foliage volumes. Total volumes of aboveground tree tis-
sues including the stem and foliage were estimated biometri-
cally. Volumes of stem xylem, phloem and bark, and branch
xylem, phloem and bark, were estimated from diameters and
cores taken at stem heights of 4, 46 and 51 m.
Foliage volume was calculated from (1) measurements of
height above ground, diameter, length and foliage volume of
all live branches (Ishii et al. 2002) and (2) estimated foliage
quantity based on sapwood basal area and branch size and po-
sition. Sapwood cross-sectional area at any height on the bole
of a Douglas-fir is related linearly to the amount of foliage
above that point (Long et al. 1981). In addition, our two esti-
mates of foliage quantity were compared with a third estimate
derived from sapwood cross-sectional area at 4 m (McDowell
et al. 2002).
The longitudinal or vertical section of the stem in Figure 1
was reconstructed from the sapwood cross-sectional area us-
ing a tree stem form factor (Korf et al. 1972, Philip 1994) with
total stem volume converted to sapwood and phloem volumes.
Free water volume in the sapwood (V
w_free
/
V, expressed as a
percentage of total sapwood fresh volume, V) was calculated
by subtracting the volume of water in the heartwood (V
w_htrw
,
taken as mostly physically bound) from that in the sapwood
(V
w_sapw
) (Zimmermann and Brown 1971, Kravka and Èermák
1995):
V
V
V
V
V
V
w_ free
w_ sapw
w_ htrw
=
–
(4)
Changes in stem radius of Psme 1373 at heights of 4 and
46 m were measured with a temperature compensated elec-
tronic radial dendrometer (DR-01, EMS Brno, Czech Repub-
TREE PHYSIOLOGY ONLINE at http://heronpublishing.com
DYNAMICS OF TREE WATER STORAGE AND STEM DIAMETER CHANGE
183
Figure 1. Sample tree (right)
showing the positions of sap
flow sensors at heights of 4, 46
(B), 51 (M) and 56 m (T).
Vertical pattern of stem form
(A; delimiting sapwood and
heartwood) and free water
content in the sapwood,
phloem and needles (B) in the
old-growth Douglas-fir sample
tree (Psme 1373). Horizontal
bars represent 1-m thick layers
above ground, including tree
stem, branches and needles.
Downloaded from https://academic.oup.com/treephys/article-abstract/27/2/181/1664618
by guest
on 25 July 2018
lic). A steel radial rod was inserted through a 7-mm diameter
hole (which extended 80 mm through the sapwood and was a
little wider than the rod so that the rod was not touching the
sapwood) and screwed tightly into the heartwood. A magnetic
sensor (Diana Inc., U.K.), whose sensitive point was in direct
contact with a smooth bark surface (located 5 cm below the
rod), was fastened to the rod; its temperature was measured
with an attached platinum thermometer. The bark was re-
moved and smoothed to a distance about 1 mm from the bark
cambium and phloem. The dendrometers were insulated and
shielded as described for the sap flow sensors.
Gravimetric water storage estimates
Stem tissue water content was estimated by classical methods.
Bark, phloem, and xylem radial water contents (% of volume)
were measured gravimetrically on 5.2-mm diameter cores
taken with an increment corer (Suunto, Finland). Immediately
after sampling, each core was protected by tightly wrapping it
in aluminum foil and stored in a shielded polyethylene bag.
Within 24 h, each core was cut into short pieces of known
length; these were individually marked, weighed, oven-dried
at 90 °C for 24 h and re-weighed. The specific mass of xylem
dry matter was assumed to be 1.54 g cm
– 1
and volume was at-
tributed to three fractions: dry matter, water and air (Kravka
and Èermák 1995). The phloem was assumed to contain the
same fraction of free water as the xylem sapwood. The amount
of free stored water was calculated by multiplying the volume
of a particular tissue by the volumetric percentage of free wa-
ter. We defined free water as the amount of water measured in
the sapwood after subtracting the amount of water measured in
the heartwood.
Needle water content was measured at 55.9, 51.1, 44.2, 39.1
and 26.4 m in current-year, 1- and 2-year old foliage sampled
from the south side of the crown. Needle free water was esti-
mated based on percent water content values obtained from
small samples multiplied by the quantity of foliage in 1-m
zones. Samples were taken in late October when tissues were
well hydrated. Needles were oven-dried at 65 °C for 72 h. The
foliage area, mass and volume in these zones were estimated
knowing total foliage parameters (from sapwood area) and its
distribution along the stem (derived from the distribution of
branch foliage volumes).
Hydrodynamic water storage estimates (sap flow)
For Psme 1373, we had estimates of the total foliage and its
vertical distribution and the amount of water lost from six
branches (q
br
). When q
br_mean
(kg m
leaf
– 2
h
– 1
) for each crown sec-
tion was multiplied by the leaf area for that section and then
summed to give Q
crown
(kg h
– 1
),
Q
crown
overestimated
Q
t
, be-
cause branch sap flows were measured in branches at the outer,
more exposed edges of the crown, which overestimated water
loss for that section. To correct this error, q
br_mean
was con-
verted to total sap flow for each section of the crown by using
apparent leaf area A
app
(m
2
part
– 1
) and the formula:
Q
crown
=
q
br_mean
A
app
(5)
where A
app
was determined in an iterative process so that Q
crown
equaled Q
t
for each day (assuming that there were no water
losses from a stem without foliage). Daily total
Q
crown
was first
calculated by multiplying q
br_mean
by leaf area (A
actual
) and this
value was compared with the daily summed Q
t
derived from
sap flow data. Actual leaf area was reduced and
Q
crown
was re-
calculated. The process was continued until Q
crown
matched Q
t
for that day. At that point, A
app
for the tree crown was known.
The change in stored water (
∆Q; dm
3
) at any time in the
whole tree (or in a specified part of the tree) became discern-
ible when the difference in sap flow between the stem and
small branches was calculated:
Cum
(
branch
stem
t
t
∆
∆
Q
Q
Q
t
=
+
∑
–
)
1
(6)
where
∆t is the time step and can range from the length of time
between data logging to the entire day. Negative values of
∆Q
occur between sunrise and early to mid-afternoon and repre-
sent times when water stores are being depleted. Positive val-
ues of
∆Q indicate refilling of depleted water storage tissue,
which occurs from mid-afternoon to well into the evening or
until the next dawn.
Because the daily totals of flow, Q
crown
and
Q
t
, were equal
(only small differences can be expected between consecutive
days), data collected over short periods within a day can be
compared to estimate the amount of water extracted from tree
storage during a particular day (W
stor
):
W
stor
= (±)
∆
Q =
Q
t
– Q
crown
(7)
where, for each recorded time step (1 or 15 min) during a diur-
nal course, a series of differences, +
∆Q and –∆Q resulted.
Their summation for the morning hours provided an estimate
of the use of stored water (–
∆Q), whereas their summation in
the late afternoon and evening hours (+
∆Q) gave an estimate
of the refilling of stored reserves.
Volumetric water storage estimates (dendrometers)
The diurnal curve of cumulated
∆Q should reflect changes in
tissue water content and thus should be comparable with diur-
nal changes in the volume of extensible tissue (e.g., sapwood,
phloem and a negligible part of cork bark (see Molz and
Klepper 1973, Hinckley and Bruckerhoff 1975)), measured as
∆R with dendrometers in addition to water changes in inelastic
tissue (e.g., cavitation). This comparison can be made only
when
∆
Q and ∆
R are expressed in comparable units. First, re-
corded data changes in stem radius (
∆R based on an initial ra-
dius, R
orig
) were converted to changes in stem cross-sectional
area (
∆A). These changes were expressed on a volume basis,
∆V, by multiplying the length (L) of the corresponding stem
segment, L
i
(or lengths of upper, middle and lower part of the
stem) by a stem form parameter (
f ):
184
ÈERMÁK, KUÈERA, BAUERLE, PHILLIPS AND HINCKLEY
TREE PHYSIOLOGY VOLUME 27, 2007
Downloaded from https://academic.oup.com/treephys/article-abstract/27/2/181/1664618
by guest
on 25 July 2018