Directory UMM :Data Elmu:jurnal:T:Tree Physiology:Vol16.1996:
Tree Physiology 16, 809--815
© 1996 Heron Publishing----Victoria, Canada
Measuring stem water content in four deciduous hardwoods with a
time-domain reflectometer
STAN D. WULLSCHLEGER, PAUL J. HANSON and DONALD E. TODD
Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6422, USA
Received February 28, 1996
Summary New technologies in time-domain reflectometry
offer a reliable means of measuring soil water content. Whether
these same technologies can be used or adapted to estimate the
water content of other porous media, such as the woody tissue
of forest trees, has not been thoroughly addressed. Therefore,
curves relating the apparent dielectric constant (Ka) to volumetric water content (g cm −3) were constructed for large-diameter
stems of red maple (Acer rubrum L.), white oak (Quercus
alba L.), chestnut oak (Q. prinus L.), and black gum (Nyssa
sylvatica Marsh.). This information was combined with previously published data and a proposed ‘‘universal’’ calibration
equation for wood was derived. Stainless-steel rods (15-cm
wave guides) were inserted into 160 trees (30 to 49 per species)
growing in an upland oak--hickory forest and stem water contents estimated monthly during 1994 and 1995 with a time-domain reflectometer (TDR). Volumetric water contents in April
ranged from 0.28 g cm −3 for red maple to 0.43 g cm −3 for black
gum, with no evidence that water content changed as a function
of stem diameter. Stem water contents estimated during 1994
(a wet year) increased from May to July, reached a maximum
in midsummer (0.41 to 0.50 g cm −3), and then decreased in
November. During 1995 (a dry year), stem water contents for
red maple and black gum (two diffuse-porous species) decreased from May to August, reached a minimum in September
(0.29 to 0.37 g cm −3), slightly increased in October and November, and then decreased in December. A different trend was
observed during 1995 for white oak and chestnut oak (two
ring-porous species), with water contents remaining fairly
stable from May to August, but decreasing abruptly in September and again in December. Stem water contents estimated with
a TDR broadly agreed with gravimetric analyses of excised
stem segments and increment cores, although there was evidence that overestimation of water content was possible with
TDR as a result of wounding following wave guide installation.
Nonetheless our results hold promise for the application of
TDR to the study of stem water content and to the study of
whole-plant water storage.
Keywords: Acer rubrum, apparent dielectric constant, capacitance, Nyssa sylvatica, Quercus alba, Quercus prunus, stem
water storage, TDR.
Introduction
Water stored within the woody tissues of forest trees has been
viewed as a reservoir from which water could be withdrawn to
buffer the evaporative demands of a transpiring plant canopy
(Reynolds 1965, Turner and Waggoner 1968). This pool of
available water, whether it is drawn from intracellular or extracellular storage (Ewers and Cruiziat 1991, Holbrook 1995),
may be of sufficient volume to influence the whole-plant water
balance of some species. Studies in Scots pine (Pinus
sylvestris L.) showed that 30 to 50% of the water transpired by
a stand could be supplied over short periods of time from water
stored within the sapwood (Waring et al. 1979). This agrees
closely with the results of Waring and Running (1978) who
reported that water stored in the sapwood of old-growth
Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) was capable of supplying up to 50% of the tree’s short-term water
requirements. These estimates, however, far exceed those obtained in an earlier study with Scots pine (Roberts 1976) and a
more recent modeling exercise with Thuja occidentalis L.
(Tyree and Yang 1990), where stem water storage was estimated to contribute little to daily water use. Water stored
within stems of deciduous hardwoods and the use of these
supplies to offset whole-plant water requirements have been
less well studied, perhaps because there is some expectation
that stem water storage in hardwoods is less important than in
conifers (Hinckley et al. 1978, Chaney 1981).
Despite the need to quantify better the contribution of stem
water storage to whole-plant water balance, Holbrook (1995)
points out that this requires more than simply establishing the
presence of water in the stem. Attention must be directed
toward improved estimates of whole-tree water uptake from
soils and water loss by the entire canopy, and improved methods for in situ monitoring of stem water content. Attempts to
visualize better the spatial distribution of water in woody
tissues have led to the promising use of gamma-ray attenuation
(Edwards and Jarvis 1983), nuclear magnetic resonance
(Byrne et al. 1986), and computer tomography (Raschi et al.
1995), and to the application of methods such as stem capacitance (Holbrook et al. 1992), electric resistance (Borchert
1994), and time-domain reflectometry (Constantz and Murphy
810
WULLSCHLEGER, HANSON AND TODD
1990, Holbrook et al. 1992) for the study of stem water content. Time-domain reflectometry (TDR), has been used for
years to estimate soil water content (Topp et al. 1980) and is
based on the speed at which an electromagnetic wave is propagated through a water-bearing material and the dielectric constant of that material (Fellner-Feldegg 1969). Constantz and
Murphy (1990) used the concepts of TDR to follow changes in
stem water content over time for various forest tree species and
noted that the technology provided a rapid and convenient
means of measuring stem water content and potentially estimating water storage. Holbrook and Sinclair (1992) built on
these studies and used TDR to monitor both stem water content
and water storage in an arborescent palm, concluding that the
technology readily detected changes in stem water content and
that up to 60% of the total water transpired by Sabal palmetto
((Walt.) Lodd, ex J.A. and J.H. Schult.) during an imposed
period of drought was drawn from stem storage.
Given the initial success with which TDR has been applied
to trees, our objective was to use this technology to monitor
seasonal changes in woody tissue water content of four deciduous hardwoods. Curves relating the apparent dielectric constant (Ka) to volumetric water content (g cm −3) were developed
for both ring-porous and diffuse-porous species, and a proposed ‘‘universal’’ calibration equation was derived. This equation was then used to estimate monthly stem water content for
160 trees growing in an upland oak--hickory forest during 1994
and 1995. Stem water contents were compared among species,
across a range of stem diameters, and against estimates of
water content determined on excised stem segments and on
increment cores. We conclude that TDR is a reliable technique
for measuring stem water content, but note that as applied in
our study it has some shortcomings.
Materials and methods
Site description
This study was conducted on a south-facing slope within the
Walker Branch Watershed, a part of the U.S. Department of
Energy’s Oak Ridge Reservation in Anderson County, Tennessee (35°58′ N and 84°17′ W). The vegetation is typical of an
upland oak--hickory (Quercus--Carya spp.) forest, consisting
primarily of chestnut oak (Q. prinus L.), white oak
(Q. alba L.), black gum (Nyssa sylvatica Marsh.) and red maple (Acer rubrum L.). Although 17 species occupy the study
site, these four comprise almost 75% (15.2 out of 21.1 m2 ha −1
total) of the site’s basal area. Mean annual rainfall (30-year
average) is 134 cm and median temperature is 14.4 °C. Compared with the 30-year average, 1994 was a wet year and 1995
was a dry year (Table 1). Soils at the study site are primarily
Typic paleudults. A comprehensive description of the climate,
vegetation, soils, and land use history of the Walker Branch
Watershed can be found in Johnson (1989).
Calibration procedures
Although a universal calibration curve has been proposed for
use in determining soil water content (Topp et al. 1980), it is
doubtful that such a curve developed for soils could be used to
Table 1. Monthly air temperatures and rainfall for 1994 and 1995 at
the Walker Branch Watershed study site. Mean air temperatures for
each month were calculated from hourly observations.
Month
January
February
March
April
May
June
July
August
September
October
November
December
Mean air temperature Precipitation
30-year
1994
o
C
1995
o
C
1994
cm
1995
cm
mean
−0.8
5.7
9.6
16.5
17.5
23.6
23.9
23.4
19.6
15.3
11.6
7.1
3.4
4.0
11.8
16.0
19.1
22.4
25.5
25.7
20.6
14.7
6.2
3.1
14.2
24.9
25.4
22.6
5.9
19.4
12.8
10.6
8.2
6.4
9.8
7.3
13.5
9.7
9.4
6.4
17.5
5.5
4.7
1.5
10.2
12.8
13.2
9.2
11.6
10.9
13.9
10.2
11.8
10.5
12.8
9.4
9.7
8.2
11.4
13.5
cm
estimate accurately the water content of woody tissues (Constantz and Murphy 1990, Holbrook and Sinclair 1992). Therefore, curves relating apparent dielectric constant (Ka) to
volumetric water content (g cm −3) were specifically constructed for large-diameter stems of red maple, white oak,
chestnut oak, and black gum. One tree of each species was
felled early on December 2, 1993 and the boles sectioned into
segments of about 30 to 40 cm in length. Water loss from
fresh-cut surfaces was minimized with a coating of paraffin
(Gulf Lite and Wizard Inc., Memphis, TN). Four segments
ranging in diameter from 16 to 22 cm were selected for each
species and two stainless-steel wave guides (15 cm long, 2.5
cm spacing) were inserted radially into pre-drilled holes. Wave
guides were also inserted into two additional segments of each
species and estimates of stem water contents for these samples
were used to verify the accuracy of the calibration curve.
Wave-guide length was considered 13 cm with 2 cm of the
stainless-steel rods exposed for sensor head attachment.
All calibration segments (16 total) and segments used to
verify the calibration curve (8 total) were weighed and then
allowed to dry on a greenhouse bench for 2 to 4 months. At
3-week intervals, each segment was weighed and the Ka measured with a time-domain reflectometer (Model 6050X1, Soil
Moisture Equipment Corp., Santa Barbara, CA). At completion, all segments were placed in an oven and dried at 60 °C to
constant weight. Segment volumes were determined by the
water displacement technique and then, knowing the wet
weight of each segment at any sampling date and the final dry
mass of each segment, wood densities (g cm −3) and volumetric
stem water contents (g cm −3) were calculated. A single calibration curve was developed from these data and those published
previously for sapwood blocks of Pinus radiata D. Don (P. insignis Dougl. ex Loud.) (Constantz and Murphy 1990). The
accuracy of this calibration equation was verified by comparing stem water contents estimated by TDR with those determined by gravimetric analyses on additional stem segments
MEASURING STEM WATER CONTENT BY TIME-DOMAIN REFLECTOMETRY
and on increment cores taken from randomly chosen trees at
the time of wave guide installation in the field.
Estimates of stem and soil water content
Seasonal and species-specific variations in stem water content
were examined for red maple, white oak, chestnut oak, and
black gum trees into which stainless-steel wave guides had
been installed at about breast height during mid-March 1994.
A total of 160 trees were examined, 35 red maples, 46 white
oaks, 49 chestnut oaks, and 30 black gum. These trees were
measured monthly during both 1994 and 1995 beginning in
early April and continuing until early December. Multiple sets
of wave guides were installed into a few trees (four of each
species) to investigate whether the time since wave guide
installation influenced estimates of stem water content.
Variation in soil water content (0 to 35 cm depth) was
determined for the study site with a TDR and 310 pairs of
stainless-steel wave guides. These were measured at least once
a month during 1994 and 1995. Soil water contents were
adjusted for percent coarse fragment (Drungil et al. 1987) and
soil matrix potentials (MPa) were calculated based on soilmoisture release curves generated for the A and B horizons of
these Typic paleudult soils (Peters et al. 1970).
Statistical analyses
Species-specific differences in stem water content throughout
the season were identified by a repeated measures analysis of
variance (Moser et al. 1990), with species as the between-subjects factor and sampling date as the within-subjects factor. A
one-way analysis of variance was used to test whether the
length of time since wave guide installation influenced estimates of stem water content, whereas a two-way analysis of
811
variance was used to test whether species differences in stem
water content were consistent across stem diameter classes.
Duncan’s multiple range test was used for mean separation
when differences between species were significant.
Results
Segments of stem used to construct calibration curves for the
four species had midwinter volumetric water contents that
ranged from 0.46 g cm −3 for white oak to 0.56 g cm −3 for red
maple (Figure 1). Over the 2- to 4-month period of drying,
these values decreased to about 0.12 g cm −3 for all species.
Apparent dielectric constants over this period ranged from
20.1 in red maple to 7.4 in black gum, and there was a
consistent, positive relationship between Ka and stem water
content for each of the four species. When these data were
combined with those of Constantz and Murphy (1990), a
curvilinear relationship was observed (Figure 2). A second-order quadratic was fitted to the pooled data, yielding the expression,
θ = −0.251 + 4.66 × 10 −2Ka − 4.93 × 10−4K2a ,
(1)
where θ is stem water content (g cm −3).
Stem water contents estimated with Equation 1 were in close
agreement with gravimetric analyses on independent stem
segments and on increment cores from the four species (Figure 3). There was, however, evidence early in the study (1994)
that overestimation of stem water content was possible with
TDR as a result of wounding following wave guide installation
(Table 2). In the case of chestnut oak, stem water contents
measured by TDR within 1 week of wave guide installation
Figure 1. Relationships between apparent dielectric constant and volumetric
stem water content for (A) red maple,
(B) white oak, (C) chestnut oak and
(D) black gum. The dashed line is the
calibration curve derived for soils by
Topp et al. (1980).
812
WULLSCHLEGER, HANSON AND TODD
Figure 2. Calibration curve of stem water content and apparent dielectric constant obtained from data for the four hardwood species used in
this study and from data of Constantz and Murphy (1990).
Little additional effect was observed on stem water content
when wave guides were left in place for up to 30 weeks.
Stem water contents estimated within 2 weeks of installing
the stainless-steel wave guides ranged from 0.28 g cm −3 for red
maple to 0.43 g cm −3 for black gum, with no evidence that
water content changed as a function of stem diameter (Table 3). There was an indication, however, that water content
differed (P < 0.01) among species. Stem water content averaged across stem diameters was 0.28 g cm −3 for red maple,
0.39 g cm −3 for white oak, whereas chestnut oak and black
gum values averaged 0.42 g cm −3 and above (Table 3).
Seasonal changes in stem water content were observed during 1994 and 1995, although the pattern of change was qualitatively different in the two years. Stem water contents
measured during 1994 (a wet year) increased from May to July,
reached maximum values in midsummer (0.41 to 0.50 g cm −3),
and then decreased gradually to a minimum in November
(Figure 4). Over the year, stem water contents averaged 0.36
g cm −3 for red maple, 0.47 g cm −3 for white oak, 0.49 g cm −3
for chestnut oak, and 0.46 g cm −3 for black gum. During 1995
(a dry year), stem water contents for red maple and black gum
(two diffuse-porous species) decreased from May to August,
reached a minimum value in September (0.29 to 0.37 g cm −3),
increased slightly in October and November, and then decreased in December (Figure 5). A somewhat different trend
was observed in white oak and chestnut oak (two ring-porous
species), with stem water contents being stable from May to
August, but then decreasing abruptly in September and again
in December. During 1995, water contents averaged 0.38
g cm −3 for red maple, 0.50 g cm −3 for white oak, 0.53 g cm −3
for chestnut oak, and 0.48 g cm −3 for black gum.
Discussion
Figure 3. Stem water contents estimated by time-domain reflectometry and by gravimetric analysis of increment cores. The 1/1 line
represents an essential agreement between the techniques for monitoring stem water content. Data were collected during the 1994 growing
season.
were higher than those determined gravimetrically on increment cores. For red maple, white oak, and black gum, differences due to the method of estimating stem water content were
apparent 12 weeks after wave guide installation (Table 2).
Techniques for measuring stem water content in trees range
from simple correlations with electrical resistance (Dixon et al.
1978, Davis et al. 1979) to rather complex associations between the absorption of ionizing radiation and the chemical
composition of the material being studied (Raschi et al. 1995).
In most cases, however, it appears that reliable spatial and
temporal estimates of water content are difficult to obtain on
woody tissues. Invasive sampling of stem water content, such
as the analysis of stem cores, is quick and repeatable, but has
been criticized because of the possible bias introduced as water
is forced out of the sample during extraction (Whitehead and
Jarvis 1981). More complicated methods, such as estimating
water content with the attenuation of gamma radiation (Ed-
Table 2. Stem water contents (g cm--3) (mean ± SD) of four deciduous hardwoods as affected by the time since wave guide installation. Values
within a row followed by the same letter are not significantly different.
Species
Red maple
White oak
Chestnut oak
Black gum
Time since wave guide installation (weeks)
Gravimetric
30
12
1
0.44 ± 0.09 a
0.57 ± 0.01 a
0.51 ± 0.04 a
0.60 ± 0.03 a
0.45 ± 0.08 a
0.58 ± 0.06 a
0.51 ± 0.03 a
0.58 ± 0.04 a
0.27 ± 0.03 b
0.52 ± 0.01 b
0.51 ± 0.01 a
0.47 ± 0.08 b
0.29 ± 0.04 b
0.47 ± 0.01 b
0.41 ± 0.02 b
0.42 ± 0.05 b
MEASURING STEM WATER CONTENT BY TIME-DOMAIN REFLECTOMETRY
813
Table 3. Stem water contents (g cm--3) (mean ± SD) of four deciduous hardwoods as related to stem diameter. A diameter tape was used to measure
stem diameters during late March, whereas water contents were estimated by time-domain reflectometry in early April. Average values of stem
water content for the four species followed by the same letter are not significantly different.
Species
Red maple (n = 35)
White oak (n = 46)
Chestnut oak (n = 49)
Black gum (n = 30)
1
Diameter class (cm)
Average
20--30
30--40
40--50
> 50
0.26 ± 0.06
0.39 ± 0.03
0.42 ± 0.05
0.43 ± 0.06
0.30 ± 0.06
0.40 ± 0.05
0.43 ± 0.04
0.42 ± 0.07
0.31 ± 0.01
0.37 ± 0.05
0.41 ± 0.05
0.47 ± 0.04
0.32 ± 0.05
0.39 ± 0.02
0.45 ± 0.03
nd 1
0.28 ± 0.06 c
0.39 ± 0.04 b
0.42 ± 0.05 a
0.43 ± 0.06 a
No black gum trees of this diameter class were measured.
Figure 4. Seasonal estimates of stem water content during 1994 for (A)
red maple, (B) white oak, (C) chestnut oak and (D) black gum. Each
point represents the mean ± SD for 30 to 49 trees.
Figure 5. Seasonal estimates of stem water content during 1995 for (A)
red maple, (B) white oak, (C) chestnut oak and (D) black gum. Each
point represents the mean ± SD for 30 to 49 trees.
wards and Jarvis 1983) and nuclear magnetic resonance
(Byrne et al. 1986), may provide improved spatial and temporal estimates of stem water content, but these techniques suffer
from problems of expense, portability, and calibration.
Recent advancements in the use of TDR for measuring in
situ water content of soils have led some to consider whether
such methods can be applied to other porous media, such as the
woody tissue of forest trees. Time-domain reflectometry was
used by Constantz and Murphy (1990) to monitor daily and
seasonal changes in water content for a range of tree species
and it was concluded that TDR offered a rapid and convenient
technique for measuring sapwood water content. Short-term
(days) and long-term (months) studies indicated that stem
water content was responsive to flood irrigation in the genus
Juglans, and that water content in the genera Aesculus, Eucalyptus, Pinus, Quercus, and Sequoia exhibited annual variations that ranged from 0.61 to 0.52 g cm −3 (a 15% change) in
the ring-porous species Quercus argifolia Née (Q. oxyadenia
814
WULLSCHLEGER, HANSON AND TODD
Torr.) and 0.71 to 0.39 g cm −3 (a 45% change) in the diffuseporous species Aesculus californica (Spach) Nutt. (Constantz
and Murphy 1990). From the perspective of whole-plant water
balance, the magnitude of this annual variation reflects the
degree to which water is moved to and from storage compartments in the stem, or more specifically in the sapwood (Holbrook 1995).
We found little evidence that the magnitude of annual variation in stem water content differed among the four species
during 1994 (a wet year). Variation in water content ((maxmin)/max) ranged from 15 to 20% for all trees and the direction of change indicated a net movement of water into stem
storage. It was clear, however, that during 1995 (a dry year)
annual variation in water content was greater for stems of red
maple and black gum (two diffuse-porous species) than for
stems of white oak and chestnut oak (two ring-porous species).
Variation in stem water content was 39% in red maple, 35% in
black gum, 16% in white oak, and 19% in chestnut oak. Much
of this variation occurred between the early spring (May) when
soil water contents were high and late summer (September)
when soil water contents were low. In the ring-porous species,
there was evidence that stem water content decreased throughout the summer in parallel with soil water availability (data not
shown). The direction of change during 1995 suggested a net
movement of water out of storage, presumably to offset some
portion of the seasonal water requirements of a transpiring
plant canopy. Our data are not sufficiently resolved to address
changes in stem water content over shorter periods of time,
although we suspect that these cycles do occur (Constantz and
Murphy 1990).
Previous studies have shown that TDR can be used with
success to estimate stem water content in forest trees (Constantz and Murphy 1990) and an arborescent palm (Holbrook
and Sinclair 1992). Like these authors, we recognize that
various shortcomings of the technique must be addressed before fully understanding how data on stem water content are to
be interpreted. Perhaps most serious is that estimates of stem
water contents obtained with the TDR are averaged or integrated throughout the length of the wave guides and for an area
of wood roughly twice that of the wave guide separation (2.5
cm in our study). For diffuse-porous species this may not be a
problem given the preponderance of sapwood versus heartwood, but in ring-porous species such averaging of stem water
content will likely bias estimates because of the varying contributions of sapwood and heartwood. Constantz and Murphy
(1990) cautioned that when wave guides are installed radially
into a stem, as they were in both our study and theirs, the
resulting data must be interpreted with an appreciation for the
cross-sectional morphology of wood. Clearly, given the current wave-guide configuration, stem water contents estimated
for white oak and chestnut oak reflect changes not only in
sapwood water content, but changes (or lack thereof) in heartwood water content as well. If stem water storage is confined
to sapwood, then the annual variation in stem water content as
calculated in our study for the oaks in particular may show less
seasonal fluctuation than otherwise suspected. This partially
explains why annual variation in stem water content was lower
for the ring-porous species than for the two diffuse-porous
species.
Another concern related to the use of TDR to estimate
woody-tissue water content is that the relationship between
volumetric water content and Ka for trees is quite different from
that observed for soils. Samples of sapwood from Pinus radiata showed a much smaller change in Ka for a given change
in water content than is typical of most organic soils (Constantz and Murphy 1990). We also observed that the relationship between water content and Ka was apparently unique for
wood and agree that the calibration curve for soil is not appropriate for use in estimating woody-tissue water content. Constantz and Murphy (1990) suggested that species-specific
calibrations may provide more accurate estimates of stem
water content for a given species, but we note that their calibration curve was developed based solely on data from Pinus
radiata and the resulting equation applied to conifers and
hardwoods alike. Based on our analysis, however, which included data from deciduous hardwoods and the conifer data of
Constantz and Murphy (1990), it appears that a single ‘‘universal’’ calibration curve relating stem water content and Ka may
be warranted.
Acknowledgments
We thank R. Borchert, J. Constantz, and N.T. Edwards for their helpful
reviews of the draft manuscript. This research was sponsored by the
Program for Ecosystem Research, Environmental Sciences Division,
Office of Health and Environmental Research, U.S. Department of
Energy under contract No. DE-AC05-96OR22464 with Lockheed
Martin Energy Research Corp. Publication No. 4577, Environmental
Sciences Division, Oak Ridge National Laboratory.
References
Borchert, R. 1994. Electrical resistance as a measure of tree water
status during seasonal drought in a tropical dry forest in Costa Rica.
Tree Physiol. 14: 299--312.
Byrne, G.F., M.D. Fenn and M.I. Burgar. 1986. Nuclear magnetic
resonance studies of water in tree sections. Agric. For. Meteorol.
38:307--317.
Chaney, W.R. 1981. Sources of water. In Water Deficits and Plant
Growth, Vol. VI. Woody Plant Communities. Ed. T.T. Kozlowski.
Academic Press, New York, NY, pp 1--47.
Constantz, J. and F. Murphy. 1990. Monitoring moisture storage in
trees using time domain reflectometry. J. Hydrol. 119:31--42.
Davis, W., A. Shigo and R. Weyrick. 1979. Seasonal changes in
electrical resistance of inner bark in red oak, red maple, and eastern
white pine. For. Sci. 25:282--286.
Dixon, M.A., R.G. Thompson and D.S. Fensom. 1978. Electrical
resistance measurements of water potential in avocado and white
spruce. Can. J. For. Res. 8:73--80.
Drungil, C.E.C., T.J. Gish and K. Abt. 1987. Soil moisture determinations in gravelly soils with time domain reflectometry. Am. Soc.
Agric. Engineers Paper No. 87--2568.
Edwards, W.R.N. and P.G. Jarvis. 1983. A method for measuring
radial differences in water content of intact tree stems by attenuation of gamma radiation. Plant Cell Environ. 6:255--260.
MEASURING STEM WATER CONTENT BY TIME-DOMAIN REFLECTOMETRY
Ewers, F.W. and P. Cruiziat. 1991. Measuring water transport and
storage. In Techniques and Approaches in Forest Tree Ecophysiology. Eds. J.P. Lassoie and T.M. Hinckley. CRC Press, Boca Raton,
FL, pp 91--115.
Fellner-Feldegg, J. 1969. The measurement of dielectrics in the time
domain. J. Phys. Chem. 73:616--623.
Hinckley, T.M., J.P. Lassoie and S.W. Running. 1978. Temporal and
spatial variations in the water status of forest trees. For. Sci.
Monogr. 20.
Holbrook, N.M. and T.R. Sinclair. 1992. Water balance in the arborescent palm, Sabal palmetto. II. Transpiration and stem water storage.
Plant Cell Environ. 15:401--409.
Holbrook, N.M., M.J. Burns and T.R. Sinclair. 1992. Frequency and
time-domain dielectric measurements of stem water content in the
arborescent palm, Sabal palmetto. J. Exp. Bot. 43:111--119.
Holbrook, H.M. 1995. Stem water storage. In Plant Stems: Physiology
and Functional Morphology. Ed. B.L. Gartner. Academic Press,
New York, pp 151--174.
Johnson, D.W. 1989. Site description. In Analysis of Biogeochemical
Cycling Processes in Walker Branch Watershed. Eds. D.W. Johnson
and R.I. Van Hook. Springer-Verlag, New York, pp 6--20.
Moser, E.B., A.M. Saxton and S.R. Pezeshki. 1990. Repeated measures analysis of variance: application to tree research. Can. J. For.
Res. 20:524--535.
Peters, L.N., D.F. Grigal, J.W. Curlin and W.J. Selvidge. 1970. Walker
Branch Watershed Project: chemical, physical and morphological
properties of the soils on Walker Branch Watershed. Oak Ridge
National Laboratory, Technical Manual, ORNL-TM-2968, 96 p.
815
Raschi, A., R. Tognetti, H.-W. Ridder and C. Béres. 1995. Water in the
stems of sessile oak (Quercus petraea) assessed by computer tomography with concurrent measurements of sap velocity and ultrasound emission. Plant Cell Environ. 18:545--554.
Reynolds, E.R.C. 1965. Transpiration as related to internal water
content. Nature 208:1001--1002.
Roberts, J. 1976. An examination of the quantity of water stored in
mature Pinus sylvestris trees. J. Exp. Bot. 27:473--479.
Topp G.C., J.L. Davis and A.P. Anan. 1980. Electromagnetic determination of soil water content: measurements in coaxial transmission
lines. Water Resour. Res. 16:574--582.
Turner, N.C. and P.E. Waggoner. 1968. Effects of changing stomatal
width in a red pine forest on soil water content, leaf water potential,
bole diameter and growth. Plant Physiol. 43:973--978.
Tyree M.T. and S. Yang. 1990. Water-storage capacity of Thuja, Tsuga
and Acer stems measured by dehydration isotherms. Planta
182:420--426.
Waring, R.H. and S.W. Running. 1978. Sapwood water storage: its
contribution to transpiration and effect upon water conductance
through the stems of old-growth Douglas-fir. Plant Cell Environ.
1:131--140.
Waring, R.H., D. Whitehead and P.G. Jarvis. 1979. The contribution of
stored water to transpiration in Scots pine. Plant Cell Environ.
2:309--317.
Whitehead, D. and P.G. Jarvis. 1981. Coniferous forests and plantations. In Water Deficits and Plant Growth. Vol. VI. Woody Plant
Communities. Ed. T.T. Kozlowski. Academic Press, New York, NY,
pp 49--152.
© 1996 Heron Publishing----Victoria, Canada
Measuring stem water content in four deciduous hardwoods with a
time-domain reflectometer
STAN D. WULLSCHLEGER, PAUL J. HANSON and DONALD E. TODD
Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6422, USA
Received February 28, 1996
Summary New technologies in time-domain reflectometry
offer a reliable means of measuring soil water content. Whether
these same technologies can be used or adapted to estimate the
water content of other porous media, such as the woody tissue
of forest trees, has not been thoroughly addressed. Therefore,
curves relating the apparent dielectric constant (Ka) to volumetric water content (g cm −3) were constructed for large-diameter
stems of red maple (Acer rubrum L.), white oak (Quercus
alba L.), chestnut oak (Q. prinus L.), and black gum (Nyssa
sylvatica Marsh.). This information was combined with previously published data and a proposed ‘‘universal’’ calibration
equation for wood was derived. Stainless-steel rods (15-cm
wave guides) were inserted into 160 trees (30 to 49 per species)
growing in an upland oak--hickory forest and stem water contents estimated monthly during 1994 and 1995 with a time-domain reflectometer (TDR). Volumetric water contents in April
ranged from 0.28 g cm −3 for red maple to 0.43 g cm −3 for black
gum, with no evidence that water content changed as a function
of stem diameter. Stem water contents estimated during 1994
(a wet year) increased from May to July, reached a maximum
in midsummer (0.41 to 0.50 g cm −3), and then decreased in
November. During 1995 (a dry year), stem water contents for
red maple and black gum (two diffuse-porous species) decreased from May to August, reached a minimum in September
(0.29 to 0.37 g cm −3), slightly increased in October and November, and then decreased in December. A different trend was
observed during 1995 for white oak and chestnut oak (two
ring-porous species), with water contents remaining fairly
stable from May to August, but decreasing abruptly in September and again in December. Stem water contents estimated with
a TDR broadly agreed with gravimetric analyses of excised
stem segments and increment cores, although there was evidence that overestimation of water content was possible with
TDR as a result of wounding following wave guide installation.
Nonetheless our results hold promise for the application of
TDR to the study of stem water content and to the study of
whole-plant water storage.
Keywords: Acer rubrum, apparent dielectric constant, capacitance, Nyssa sylvatica, Quercus alba, Quercus prunus, stem
water storage, TDR.
Introduction
Water stored within the woody tissues of forest trees has been
viewed as a reservoir from which water could be withdrawn to
buffer the evaporative demands of a transpiring plant canopy
(Reynolds 1965, Turner and Waggoner 1968). This pool of
available water, whether it is drawn from intracellular or extracellular storage (Ewers and Cruiziat 1991, Holbrook 1995),
may be of sufficient volume to influence the whole-plant water
balance of some species. Studies in Scots pine (Pinus
sylvestris L.) showed that 30 to 50% of the water transpired by
a stand could be supplied over short periods of time from water
stored within the sapwood (Waring et al. 1979). This agrees
closely with the results of Waring and Running (1978) who
reported that water stored in the sapwood of old-growth
Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) was capable of supplying up to 50% of the tree’s short-term water
requirements. These estimates, however, far exceed those obtained in an earlier study with Scots pine (Roberts 1976) and a
more recent modeling exercise with Thuja occidentalis L.
(Tyree and Yang 1990), where stem water storage was estimated to contribute little to daily water use. Water stored
within stems of deciduous hardwoods and the use of these
supplies to offset whole-plant water requirements have been
less well studied, perhaps because there is some expectation
that stem water storage in hardwoods is less important than in
conifers (Hinckley et al. 1978, Chaney 1981).
Despite the need to quantify better the contribution of stem
water storage to whole-plant water balance, Holbrook (1995)
points out that this requires more than simply establishing the
presence of water in the stem. Attention must be directed
toward improved estimates of whole-tree water uptake from
soils and water loss by the entire canopy, and improved methods for in situ monitoring of stem water content. Attempts to
visualize better the spatial distribution of water in woody
tissues have led to the promising use of gamma-ray attenuation
(Edwards and Jarvis 1983), nuclear magnetic resonance
(Byrne et al. 1986), and computer tomography (Raschi et al.
1995), and to the application of methods such as stem capacitance (Holbrook et al. 1992), electric resistance (Borchert
1994), and time-domain reflectometry (Constantz and Murphy
810
WULLSCHLEGER, HANSON AND TODD
1990, Holbrook et al. 1992) for the study of stem water content. Time-domain reflectometry (TDR), has been used for
years to estimate soil water content (Topp et al. 1980) and is
based on the speed at which an electromagnetic wave is propagated through a water-bearing material and the dielectric constant of that material (Fellner-Feldegg 1969). Constantz and
Murphy (1990) used the concepts of TDR to follow changes in
stem water content over time for various forest tree species and
noted that the technology provided a rapid and convenient
means of measuring stem water content and potentially estimating water storage. Holbrook and Sinclair (1992) built on
these studies and used TDR to monitor both stem water content
and water storage in an arborescent palm, concluding that the
technology readily detected changes in stem water content and
that up to 60% of the total water transpired by Sabal palmetto
((Walt.) Lodd, ex J.A. and J.H. Schult.) during an imposed
period of drought was drawn from stem storage.
Given the initial success with which TDR has been applied
to trees, our objective was to use this technology to monitor
seasonal changes in woody tissue water content of four deciduous hardwoods. Curves relating the apparent dielectric constant (Ka) to volumetric water content (g cm −3) were developed
for both ring-porous and diffuse-porous species, and a proposed ‘‘universal’’ calibration equation was derived. This equation was then used to estimate monthly stem water content for
160 trees growing in an upland oak--hickory forest during 1994
and 1995. Stem water contents were compared among species,
across a range of stem diameters, and against estimates of
water content determined on excised stem segments and on
increment cores. We conclude that TDR is a reliable technique
for measuring stem water content, but note that as applied in
our study it has some shortcomings.
Materials and methods
Site description
This study was conducted on a south-facing slope within the
Walker Branch Watershed, a part of the U.S. Department of
Energy’s Oak Ridge Reservation in Anderson County, Tennessee (35°58′ N and 84°17′ W). The vegetation is typical of an
upland oak--hickory (Quercus--Carya spp.) forest, consisting
primarily of chestnut oak (Q. prinus L.), white oak
(Q. alba L.), black gum (Nyssa sylvatica Marsh.) and red maple (Acer rubrum L.). Although 17 species occupy the study
site, these four comprise almost 75% (15.2 out of 21.1 m2 ha −1
total) of the site’s basal area. Mean annual rainfall (30-year
average) is 134 cm and median temperature is 14.4 °C. Compared with the 30-year average, 1994 was a wet year and 1995
was a dry year (Table 1). Soils at the study site are primarily
Typic paleudults. A comprehensive description of the climate,
vegetation, soils, and land use history of the Walker Branch
Watershed can be found in Johnson (1989).
Calibration procedures
Although a universal calibration curve has been proposed for
use in determining soil water content (Topp et al. 1980), it is
doubtful that such a curve developed for soils could be used to
Table 1. Monthly air temperatures and rainfall for 1994 and 1995 at
the Walker Branch Watershed study site. Mean air temperatures for
each month were calculated from hourly observations.
Month
January
February
March
April
May
June
July
August
September
October
November
December
Mean air temperature Precipitation
30-year
1994
o
C
1995
o
C
1994
cm
1995
cm
mean
−0.8
5.7
9.6
16.5
17.5
23.6
23.9
23.4
19.6
15.3
11.6
7.1
3.4
4.0
11.8
16.0
19.1
22.4
25.5
25.7
20.6
14.7
6.2
3.1
14.2
24.9
25.4
22.6
5.9
19.4
12.8
10.6
8.2
6.4
9.8
7.3
13.5
9.7
9.4
6.4
17.5
5.5
4.7
1.5
10.2
12.8
13.2
9.2
11.6
10.9
13.9
10.2
11.8
10.5
12.8
9.4
9.7
8.2
11.4
13.5
cm
estimate accurately the water content of woody tissues (Constantz and Murphy 1990, Holbrook and Sinclair 1992). Therefore, curves relating apparent dielectric constant (Ka) to
volumetric water content (g cm −3) were specifically constructed for large-diameter stems of red maple, white oak,
chestnut oak, and black gum. One tree of each species was
felled early on December 2, 1993 and the boles sectioned into
segments of about 30 to 40 cm in length. Water loss from
fresh-cut surfaces was minimized with a coating of paraffin
(Gulf Lite and Wizard Inc., Memphis, TN). Four segments
ranging in diameter from 16 to 22 cm were selected for each
species and two stainless-steel wave guides (15 cm long, 2.5
cm spacing) were inserted radially into pre-drilled holes. Wave
guides were also inserted into two additional segments of each
species and estimates of stem water contents for these samples
were used to verify the accuracy of the calibration curve.
Wave-guide length was considered 13 cm with 2 cm of the
stainless-steel rods exposed for sensor head attachment.
All calibration segments (16 total) and segments used to
verify the calibration curve (8 total) were weighed and then
allowed to dry on a greenhouse bench for 2 to 4 months. At
3-week intervals, each segment was weighed and the Ka measured with a time-domain reflectometer (Model 6050X1, Soil
Moisture Equipment Corp., Santa Barbara, CA). At completion, all segments were placed in an oven and dried at 60 °C to
constant weight. Segment volumes were determined by the
water displacement technique and then, knowing the wet
weight of each segment at any sampling date and the final dry
mass of each segment, wood densities (g cm −3) and volumetric
stem water contents (g cm −3) were calculated. A single calibration curve was developed from these data and those published
previously for sapwood blocks of Pinus radiata D. Don (P. insignis Dougl. ex Loud.) (Constantz and Murphy 1990). The
accuracy of this calibration equation was verified by comparing stem water contents estimated by TDR with those determined by gravimetric analyses on additional stem segments
MEASURING STEM WATER CONTENT BY TIME-DOMAIN REFLECTOMETRY
and on increment cores taken from randomly chosen trees at
the time of wave guide installation in the field.
Estimates of stem and soil water content
Seasonal and species-specific variations in stem water content
were examined for red maple, white oak, chestnut oak, and
black gum trees into which stainless-steel wave guides had
been installed at about breast height during mid-March 1994.
A total of 160 trees were examined, 35 red maples, 46 white
oaks, 49 chestnut oaks, and 30 black gum. These trees were
measured monthly during both 1994 and 1995 beginning in
early April and continuing until early December. Multiple sets
of wave guides were installed into a few trees (four of each
species) to investigate whether the time since wave guide
installation influenced estimates of stem water content.
Variation in soil water content (0 to 35 cm depth) was
determined for the study site with a TDR and 310 pairs of
stainless-steel wave guides. These were measured at least once
a month during 1994 and 1995. Soil water contents were
adjusted for percent coarse fragment (Drungil et al. 1987) and
soil matrix potentials (MPa) were calculated based on soilmoisture release curves generated for the A and B horizons of
these Typic paleudult soils (Peters et al. 1970).
Statistical analyses
Species-specific differences in stem water content throughout
the season were identified by a repeated measures analysis of
variance (Moser et al. 1990), with species as the between-subjects factor and sampling date as the within-subjects factor. A
one-way analysis of variance was used to test whether the
length of time since wave guide installation influenced estimates of stem water content, whereas a two-way analysis of
811
variance was used to test whether species differences in stem
water content were consistent across stem diameter classes.
Duncan’s multiple range test was used for mean separation
when differences between species were significant.
Results
Segments of stem used to construct calibration curves for the
four species had midwinter volumetric water contents that
ranged from 0.46 g cm −3 for white oak to 0.56 g cm −3 for red
maple (Figure 1). Over the 2- to 4-month period of drying,
these values decreased to about 0.12 g cm −3 for all species.
Apparent dielectric constants over this period ranged from
20.1 in red maple to 7.4 in black gum, and there was a
consistent, positive relationship between Ka and stem water
content for each of the four species. When these data were
combined with those of Constantz and Murphy (1990), a
curvilinear relationship was observed (Figure 2). A second-order quadratic was fitted to the pooled data, yielding the expression,
θ = −0.251 + 4.66 × 10 −2Ka − 4.93 × 10−4K2a ,
(1)
where θ is stem water content (g cm −3).
Stem water contents estimated with Equation 1 were in close
agreement with gravimetric analyses on independent stem
segments and on increment cores from the four species (Figure 3). There was, however, evidence early in the study (1994)
that overestimation of stem water content was possible with
TDR as a result of wounding following wave guide installation
(Table 2). In the case of chestnut oak, stem water contents
measured by TDR within 1 week of wave guide installation
Figure 1. Relationships between apparent dielectric constant and volumetric
stem water content for (A) red maple,
(B) white oak, (C) chestnut oak and
(D) black gum. The dashed line is the
calibration curve derived for soils by
Topp et al. (1980).
812
WULLSCHLEGER, HANSON AND TODD
Figure 2. Calibration curve of stem water content and apparent dielectric constant obtained from data for the four hardwood species used in
this study and from data of Constantz and Murphy (1990).
Little additional effect was observed on stem water content
when wave guides were left in place for up to 30 weeks.
Stem water contents estimated within 2 weeks of installing
the stainless-steel wave guides ranged from 0.28 g cm −3 for red
maple to 0.43 g cm −3 for black gum, with no evidence that
water content changed as a function of stem diameter (Table 3). There was an indication, however, that water content
differed (P < 0.01) among species. Stem water content averaged across stem diameters was 0.28 g cm −3 for red maple,
0.39 g cm −3 for white oak, whereas chestnut oak and black
gum values averaged 0.42 g cm −3 and above (Table 3).
Seasonal changes in stem water content were observed during 1994 and 1995, although the pattern of change was qualitatively different in the two years. Stem water contents
measured during 1994 (a wet year) increased from May to July,
reached maximum values in midsummer (0.41 to 0.50 g cm −3),
and then decreased gradually to a minimum in November
(Figure 4). Over the year, stem water contents averaged 0.36
g cm −3 for red maple, 0.47 g cm −3 for white oak, 0.49 g cm −3
for chestnut oak, and 0.46 g cm −3 for black gum. During 1995
(a dry year), stem water contents for red maple and black gum
(two diffuse-porous species) decreased from May to August,
reached a minimum value in September (0.29 to 0.37 g cm −3),
increased slightly in October and November, and then decreased in December (Figure 5). A somewhat different trend
was observed in white oak and chestnut oak (two ring-porous
species), with stem water contents being stable from May to
August, but then decreasing abruptly in September and again
in December. During 1995, water contents averaged 0.38
g cm −3 for red maple, 0.50 g cm −3 for white oak, 0.53 g cm −3
for chestnut oak, and 0.48 g cm −3 for black gum.
Discussion
Figure 3. Stem water contents estimated by time-domain reflectometry and by gravimetric analysis of increment cores. The 1/1 line
represents an essential agreement between the techniques for monitoring stem water content. Data were collected during the 1994 growing
season.
were higher than those determined gravimetrically on increment cores. For red maple, white oak, and black gum, differences due to the method of estimating stem water content were
apparent 12 weeks after wave guide installation (Table 2).
Techniques for measuring stem water content in trees range
from simple correlations with electrical resistance (Dixon et al.
1978, Davis et al. 1979) to rather complex associations between the absorption of ionizing radiation and the chemical
composition of the material being studied (Raschi et al. 1995).
In most cases, however, it appears that reliable spatial and
temporal estimates of water content are difficult to obtain on
woody tissues. Invasive sampling of stem water content, such
as the analysis of stem cores, is quick and repeatable, but has
been criticized because of the possible bias introduced as water
is forced out of the sample during extraction (Whitehead and
Jarvis 1981). More complicated methods, such as estimating
water content with the attenuation of gamma radiation (Ed-
Table 2. Stem water contents (g cm--3) (mean ± SD) of four deciduous hardwoods as affected by the time since wave guide installation. Values
within a row followed by the same letter are not significantly different.
Species
Red maple
White oak
Chestnut oak
Black gum
Time since wave guide installation (weeks)
Gravimetric
30
12
1
0.44 ± 0.09 a
0.57 ± 0.01 a
0.51 ± 0.04 a
0.60 ± 0.03 a
0.45 ± 0.08 a
0.58 ± 0.06 a
0.51 ± 0.03 a
0.58 ± 0.04 a
0.27 ± 0.03 b
0.52 ± 0.01 b
0.51 ± 0.01 a
0.47 ± 0.08 b
0.29 ± 0.04 b
0.47 ± 0.01 b
0.41 ± 0.02 b
0.42 ± 0.05 b
MEASURING STEM WATER CONTENT BY TIME-DOMAIN REFLECTOMETRY
813
Table 3. Stem water contents (g cm--3) (mean ± SD) of four deciduous hardwoods as related to stem diameter. A diameter tape was used to measure
stem diameters during late March, whereas water contents were estimated by time-domain reflectometry in early April. Average values of stem
water content for the four species followed by the same letter are not significantly different.
Species
Red maple (n = 35)
White oak (n = 46)
Chestnut oak (n = 49)
Black gum (n = 30)
1
Diameter class (cm)
Average
20--30
30--40
40--50
> 50
0.26 ± 0.06
0.39 ± 0.03
0.42 ± 0.05
0.43 ± 0.06
0.30 ± 0.06
0.40 ± 0.05
0.43 ± 0.04
0.42 ± 0.07
0.31 ± 0.01
0.37 ± 0.05
0.41 ± 0.05
0.47 ± 0.04
0.32 ± 0.05
0.39 ± 0.02
0.45 ± 0.03
nd 1
0.28 ± 0.06 c
0.39 ± 0.04 b
0.42 ± 0.05 a
0.43 ± 0.06 a
No black gum trees of this diameter class were measured.
Figure 4. Seasonal estimates of stem water content during 1994 for (A)
red maple, (B) white oak, (C) chestnut oak and (D) black gum. Each
point represents the mean ± SD for 30 to 49 trees.
Figure 5. Seasonal estimates of stem water content during 1995 for (A)
red maple, (B) white oak, (C) chestnut oak and (D) black gum. Each
point represents the mean ± SD for 30 to 49 trees.
wards and Jarvis 1983) and nuclear magnetic resonance
(Byrne et al. 1986), may provide improved spatial and temporal estimates of stem water content, but these techniques suffer
from problems of expense, portability, and calibration.
Recent advancements in the use of TDR for measuring in
situ water content of soils have led some to consider whether
such methods can be applied to other porous media, such as the
woody tissue of forest trees. Time-domain reflectometry was
used by Constantz and Murphy (1990) to monitor daily and
seasonal changes in water content for a range of tree species
and it was concluded that TDR offered a rapid and convenient
technique for measuring sapwood water content. Short-term
(days) and long-term (months) studies indicated that stem
water content was responsive to flood irrigation in the genus
Juglans, and that water content in the genera Aesculus, Eucalyptus, Pinus, Quercus, and Sequoia exhibited annual variations that ranged from 0.61 to 0.52 g cm −3 (a 15% change) in
the ring-porous species Quercus argifolia Née (Q. oxyadenia
814
WULLSCHLEGER, HANSON AND TODD
Torr.) and 0.71 to 0.39 g cm −3 (a 45% change) in the diffuseporous species Aesculus californica (Spach) Nutt. (Constantz
and Murphy 1990). From the perspective of whole-plant water
balance, the magnitude of this annual variation reflects the
degree to which water is moved to and from storage compartments in the stem, or more specifically in the sapwood (Holbrook 1995).
We found little evidence that the magnitude of annual variation in stem water content differed among the four species
during 1994 (a wet year). Variation in water content ((maxmin)/max) ranged from 15 to 20% for all trees and the direction of change indicated a net movement of water into stem
storage. It was clear, however, that during 1995 (a dry year)
annual variation in water content was greater for stems of red
maple and black gum (two diffuse-porous species) than for
stems of white oak and chestnut oak (two ring-porous species).
Variation in stem water content was 39% in red maple, 35% in
black gum, 16% in white oak, and 19% in chestnut oak. Much
of this variation occurred between the early spring (May) when
soil water contents were high and late summer (September)
when soil water contents were low. In the ring-porous species,
there was evidence that stem water content decreased throughout the summer in parallel with soil water availability (data not
shown). The direction of change during 1995 suggested a net
movement of water out of storage, presumably to offset some
portion of the seasonal water requirements of a transpiring
plant canopy. Our data are not sufficiently resolved to address
changes in stem water content over shorter periods of time,
although we suspect that these cycles do occur (Constantz and
Murphy 1990).
Previous studies have shown that TDR can be used with
success to estimate stem water content in forest trees (Constantz and Murphy 1990) and an arborescent palm (Holbrook
and Sinclair 1992). Like these authors, we recognize that
various shortcomings of the technique must be addressed before fully understanding how data on stem water content are to
be interpreted. Perhaps most serious is that estimates of stem
water contents obtained with the TDR are averaged or integrated throughout the length of the wave guides and for an area
of wood roughly twice that of the wave guide separation (2.5
cm in our study). For diffuse-porous species this may not be a
problem given the preponderance of sapwood versus heartwood, but in ring-porous species such averaging of stem water
content will likely bias estimates because of the varying contributions of sapwood and heartwood. Constantz and Murphy
(1990) cautioned that when wave guides are installed radially
into a stem, as they were in both our study and theirs, the
resulting data must be interpreted with an appreciation for the
cross-sectional morphology of wood. Clearly, given the current wave-guide configuration, stem water contents estimated
for white oak and chestnut oak reflect changes not only in
sapwood water content, but changes (or lack thereof) in heartwood water content as well. If stem water storage is confined
to sapwood, then the annual variation in stem water content as
calculated in our study for the oaks in particular may show less
seasonal fluctuation than otherwise suspected. This partially
explains why annual variation in stem water content was lower
for the ring-porous species than for the two diffuse-porous
species.
Another concern related to the use of TDR to estimate
woody-tissue water content is that the relationship between
volumetric water content and Ka for trees is quite different from
that observed for soils. Samples of sapwood from Pinus radiata showed a much smaller change in Ka for a given change
in water content than is typical of most organic soils (Constantz and Murphy 1990). We also observed that the relationship between water content and Ka was apparently unique for
wood and agree that the calibration curve for soil is not appropriate for use in estimating woody-tissue water content. Constantz and Murphy (1990) suggested that species-specific
calibrations may provide more accurate estimates of stem
water content for a given species, but we note that their calibration curve was developed based solely on data from Pinus
radiata and the resulting equation applied to conifers and
hardwoods alike. Based on our analysis, however, which included data from deciduous hardwoods and the conifer data of
Constantz and Murphy (1990), it appears that a single ‘‘universal’’ calibration curve relating stem water content and Ka may
be warranted.
Acknowledgments
We thank R. Borchert, J. Constantz, and N.T. Edwards for their helpful
reviews of the draft manuscript. This research was sponsored by the
Program for Ecosystem Research, Environmental Sciences Division,
Office of Health and Environmental Research, U.S. Department of
Energy under contract No. DE-AC05-96OR22464 with Lockheed
Martin Energy Research Corp. Publication No. 4577, Environmental
Sciences Division, Oak Ridge National Laboratory.
References
Borchert, R. 1994. Electrical resistance as a measure of tree water
status during seasonal drought in a tropical dry forest in Costa Rica.
Tree Physiol. 14: 299--312.
Byrne, G.F., M.D. Fenn and M.I. Burgar. 1986. Nuclear magnetic
resonance studies of water in tree sections. Agric. For. Meteorol.
38:307--317.
Chaney, W.R. 1981. Sources of water. In Water Deficits and Plant
Growth, Vol. VI. Woody Plant Communities. Ed. T.T. Kozlowski.
Academic Press, New York, NY, pp 1--47.
Constantz, J. and F. Murphy. 1990. Monitoring moisture storage in
trees using time domain reflectometry. J. Hydrol. 119:31--42.
Davis, W., A. Shigo and R. Weyrick. 1979. Seasonal changes in
electrical resistance of inner bark in red oak, red maple, and eastern
white pine. For. Sci. 25:282--286.
Dixon, M.A., R.G. Thompson and D.S. Fensom. 1978. Electrical
resistance measurements of water potential in avocado and white
spruce. Can. J. For. Res. 8:73--80.
Drungil, C.E.C., T.J. Gish and K. Abt. 1987. Soil moisture determinations in gravelly soils with time domain reflectometry. Am. Soc.
Agric. Engineers Paper No. 87--2568.
Edwards, W.R.N. and P.G. Jarvis. 1983. A method for measuring
radial differences in water content of intact tree stems by attenuation of gamma radiation. Plant Cell Environ. 6:255--260.
MEASURING STEM WATER CONTENT BY TIME-DOMAIN REFLECTOMETRY
Ewers, F.W. and P. Cruiziat. 1991. Measuring water transport and
storage. In Techniques and Approaches in Forest Tree Ecophysiology. Eds. J.P. Lassoie and T.M. Hinckley. CRC Press, Boca Raton,
FL, pp 91--115.
Fellner-Feldegg, J. 1969. The measurement of dielectrics in the time
domain. J. Phys. Chem. 73:616--623.
Hinckley, T.M., J.P. Lassoie and S.W. Running. 1978. Temporal and
spatial variations in the water status of forest trees. For. Sci.
Monogr. 20.
Holbrook, N.M. and T.R. Sinclair. 1992. Water balance in the arborescent palm, Sabal palmetto. II. Transpiration and stem water storage.
Plant Cell Environ. 15:401--409.
Holbrook, N.M., M.J. Burns and T.R. Sinclair. 1992. Frequency and
time-domain dielectric measurements of stem water content in the
arborescent palm, Sabal palmetto. J. Exp. Bot. 43:111--119.
Holbrook, H.M. 1995. Stem water storage. In Plant Stems: Physiology
and Functional Morphology. Ed. B.L. Gartner. Academic Press,
New York, pp 151--174.
Johnson, D.W. 1989. Site description. In Analysis of Biogeochemical
Cycling Processes in Walker Branch Watershed. Eds. D.W. Johnson
and R.I. Van Hook. Springer-Verlag, New York, pp 6--20.
Moser, E.B., A.M. Saxton and S.R. Pezeshki. 1990. Repeated measures analysis of variance: application to tree research. Can. J. For.
Res. 20:524--535.
Peters, L.N., D.F. Grigal, J.W. Curlin and W.J. Selvidge. 1970. Walker
Branch Watershed Project: chemical, physical and morphological
properties of the soils on Walker Branch Watershed. Oak Ridge
National Laboratory, Technical Manual, ORNL-TM-2968, 96 p.
815
Raschi, A., R. Tognetti, H.-W. Ridder and C. Béres. 1995. Water in the
stems of sessile oak (Quercus petraea) assessed by computer tomography with concurrent measurements of sap velocity and ultrasound emission. Plant Cell Environ. 18:545--554.
Reynolds, E.R.C. 1965. Transpiration as related to internal water
content. Nature 208:1001--1002.
Roberts, J. 1976. An examination of the quantity of water stored in
mature Pinus sylvestris trees. J. Exp. Bot. 27:473--479.
Topp G.C., J.L. Davis and A.P. Anan. 1980. Electromagnetic determination of soil water content: measurements in coaxial transmission
lines. Water Resour. Res. 16:574--582.
Turner, N.C. and P.E. Waggoner. 1968. Effects of changing stomatal
width in a red pine forest on soil water content, leaf water potential,
bole diameter and growth. Plant Physiol. 43:973--978.
Tyree M.T. and S. Yang. 1990. Water-storage capacity of Thuja, Tsuga
and Acer stems measured by dehydration isotherms. Planta
182:420--426.
Waring, R.H. and S.W. Running. 1978. Sapwood water storage: its
contribution to transpiration and effect upon water conductance
through the stems of old-growth Douglas-fir. Plant Cell Environ.
1:131--140.
Waring, R.H., D. Whitehead and P.G. Jarvis. 1979. The contribution of
stored water to transpiration in Scots pine. Plant Cell Environ.
2:309--317.
Whitehead, D. and P.G. Jarvis. 1981. Coniferous forests and plantations. In Water Deficits and Plant Growth. Vol. VI. Woody Plant
Communities. Ed. T.T. Kozlowski. Academic Press, New York, NY,
pp 49--152.