Feasibility of soil moisture monitoring with heated fiber optics

Abstract

Accurate methods are needed to measure changing soil water content from meter to kilometer scales. Laboratory results demonstrate the feasibility of the heat pulse method implemented with fiber optic temperature sensing to obtain accurate distributed measurements of soil water content. A fiber optic cable with an electrically conductive armoring was buried in variably saturated sand and heated via electrical resistance to create thermal pulses monitored by observing the distributed Raman backscatter. A new and simple interpretation of heat data that takes advantage of the characteristics of fiber optic temperature measurements is presented. The accuracy of the soil water content measurements varied approximately linearly with water content. At volumetric moisture content of 0.05 m3/m3 the standard deviation of the readings was 0.001 m3/m3, and at 0.41 m3/m3 volumetric moisture content the standard deviation was 0.046 m3/m3. This uncertainty could be further reduced by averaging several heat pulse interrogations and through use of a higher-performance fiber optic sensing system.

Full text

Click
Here
for
Full
A
rticle
Feasibility of soil moisture monitoring with heated fiber optics
Chadi Sayde,
1
Christopher Gregory,
1
Maria Gil‐Rodriguez,
2
Nick Tufillaro,
1
Scott Tyler,
3
Nick van de Giesen,
4
Marshall English,
1
Richard Cuenca,
5
and John S. Selker
1
Received 12 February 2009; revised 18 February 2010; accepted 29 March 2010; published 5 June 2010.
[1] Accurate methods are needed to measure changing soil water content from meter
to kilometer scales. Laboratory results demonstrate the feasibility of the heat pulse
method implemented with fiber optic temperature sensing to obtain accurate distributed
measurements of soil water content. A fiber optic cable with an electrically conductive
armoring was buried in variably saturated sand and heated via electrical resistance to
create thermal pulses monitored by observing the distributed Raman backscatter. A new
and simple interpretation of heat data that takes advantage of the characteristics of fiber
optic temperature measurements is presented. The accuracy of the soil water content
measurements varied approximately linearly with water content. At volumetric moisture
content of 0.05 m
3
/m
3
the standard deviation of the readings was 0.001 m
3
/m
3
, and at
0.41 m
3
/m
3
volumetric moisture content the standard deviation was 0.046 m
3
/m
3
. This
uncertainty could be further reduced by averaging several heat pulse interrogations
and through use of a higher‐performance fiber optic sensing system.
Citation: Sayde, C., C. Gregory, M. Gil‐Rodriguez, N. Tufillaro, S. Tyler, N. van de Giesen, M. English, R. Cuenca, and J. S. Selker
(2010), Feasibility of soil moisture monitoring with heated fiber optics, Water Resour. Res., 46, W06201, doi:10.1029/2009WR007846.
1. Introduction
[2] Soil water accumulation, storage, and depletion play a
central role in the hydrologic cycle and the global water
balance. Though many accurate methods are available for
point measurement of soil water content, there are currently
no precise in situ methods for measurement of soil water
content from meter to kilometer scales. The goal of this
article is to demonstrate the feasibility of the Active Heat
pulse method with Fiber Optic temperature sensing (AHFO)
to obtain precise, distributed measurements of soil water
content across these spatial scales and over a broad range of
soil water contents.
[
3] The ability of fiber optic Distributed Temperature
Sensing (DTS) systems to retrieve temperature readings
each meter along fiber optic cables in excess of 10,000 m in
length at high temporal frequency has afforded many
important opportunities in environmental monitoring [e.g.,
Selker et al., 2006a, 2006b; Tyler et al., 2008, 2009; Westhoff
et al., 2007; Freifeld et al., 2008]. Recently, Steele‐Dunne et
al. [2010] demonstrated the feasibility of using the thermal
response to the diurnal temperature cycle of buried fiber
optic cables for distributed measurements of soil thermal
properties and soil moisture content. Unlike the AHFO
method, the Steele‐Dunne et al. [2010] method does not
require an external source of energy. Nevertheless, its appli-
cation remains challenging under conditions where the
thermal response to the diurnal temperature cycle is not
large enough to allow accurate estimation of soil moisture
content (e.g., under dense vegetative canopy, at depths
beyond the top few centimeters of the soil column, cloudy
days, or other surface energy flux limited systems).
[
4] The principle of temperature measurement along a
fiber optic cable is based on the thermal sensitivity of the
relative intensities of backscattered Raman Stokes and anti‐
Stokes photons that arise from collisions with electrons in
the core of the glass fiber [see Tyler et al. , 2009]. A laser
pulse, generated by the DTS unit, traversing a fiber optic
cable will result in Raman backscatter at two frequencies,
referred to as Stokes and anti‐Stokes. The DTS quantifies
the intensity of these backscattered photons and elapsed
time between the pulse and the observed returned light.
The intensity of the Stokes backscatter is largely independent
of temperature, while anti‐Stokes backscatter is strongly
dependent on the temperature at the point where the scat-
tering process occurred. Temperature can be inferred from
the Stokes/anti‐Stokes ratio. The computed temperature is
attributed to the position along the cable from which the
light was reflected, computed from the time of travel for the
light [Grattan and Sun, 2000].
[
5] Heat pulse methods are well established for the
determination of soil thermal properties, soil water content
and water movement. These methods usually apply a line
source of energy to the soil with the resulting temperature
fluctuation monitored by one or more parallel probes
[Bristow et al., 1994]. The rate of radial transmission of heat
depends on the soil bulk density, mineralogy, particle shape,
and, principally, soil water content [e.g., Shiozawa and
Campbell, 1990]. Geometries where the thermal observa-
tions are colocated with the heated probe are referred to as
1
Department of Biological and Ecological Engineering, Oregon State
University, Corvallis, Oregon, USA.
2
Departmen t of Rural Engineering, Technical University of Madrid,
Madrid, Spain.
3
Department of Geological Sciences and Engineering, University of
Nevada, Reno, Reno, Nevada, USA.
4
Water Management Civil Engineering and Geosciences, Delft
University of Technology, Delft, Netherlands.
5
Hydrologic Sciences, Divi sion of Earth Sciences, National Science
Foundation, Arlington, Virginia, USA.
Copyright 2010 by the American Geophysical Union.
0043‐1397/10/2009WR007846
WATER RESOURCES RESEARCH, VOL. 46, W06201, doi:10.1029/2009WR007846, 2010
W06201 1 of 8

single probe methods [de Vries and Peck, 1958; Shiozawa
and Campbell, 1990; Bristow et al., 1994]. Heat pulse
methods have also been widely implemented in multiprobe
geometries, with one or more sensing probes in proximity of
the heat source [e.g., Lubimova et al., 1961; Jaeger, 1965;
Larson, 1988; Campbell et al., 1991; Bristow et al., 1993,
1994; Heitman et al., 2003; Ren et al., 2003, 2005].
[
6] Many analytical and numerical methods have been
developed for the interpretation of heat pulse experiments in
soils. Typically, the solutions assume an infinitely small
radius and infinitely long line source geometry. The thermal
properties of soil are calculated from the heat pulse response
via the solution of the radial heat conduction equation
[Carslaw and Jaeger, 1959]. During heating, a pulse of
duration t
0
(s) is applied to an infinite line heat source in a
homogeneous isotropic medium which is taken to be at
uniform initial temperature. The solution for the resulting
temperature change following the commencement of heating
that is given by [de Vries and Peck,1958;Shiozawa and
Campbell, 1990; Bristow et al., 1994]
!Tr; tðÞ¼$
q
0
4!"#c
Ei
$r
2
4"t
!"
f or 0 < t % t
0
ð1Þ
and during cooling
!Tr; tðÞ¼
q
0
4!"#c
Ei
$r
2
4" t $ t
0
ðÞ
!"
$ Ei
$r
2
4"t
!"#$
f or t > t
0
;
ð2Þ
where q′ is the energy input per unit length per unit time
(J m
−1
s
−1
), r is the density of the medium (kg m
−3
), c is the
specific heat of the medium (J kg
−1
°C
−1
), r is the distance
from the line source (m), " = l/rc is the thermal diffusivity
(m
2
s
−1
), l is the thermal conductivity (W m
−1
°C
−1
), and
Ei donates the exponential integral.
[
7] In this implementation of the line source transient
method, the radius of the heat source is assumed to be
infinitely small. A correction factor can be added to the long
time solution to account for the nonzero radius of the heat
source. The validity of such a correction decreases with an
increase of the probe radius [Blackwell, 1954]. To account
for the finite dimensions of the cable, the cylindrical tran-
sient method can be used as described by Jaeger [1965] for
a perfectly conducting cylinder with constant heat supply
per unit length per unit time (q′)
!T ¼
2q
0
$
2
!
3
%#c
Z
1
0
1 $ exp $&u
2
ðÞ
%&
du
u
3
! uðÞ
; ð3Þ
where
! uðÞ¼uJ
0
uðÞ$$ $ hu
2
'(
J
1
uðÞ
)*
2
þ uY
0
uðÞ$ $ $ hu
2
'(
Y
1
uðÞ
)*
2
;
ð4Þ
& ¼ "t=a
2
; ð5Þ
$ ¼ 2!a
2
#c=S; ð6Þ
h ¼ %=aH; ð7Þ
with a being the heat source radius (m), S the heat capacity
per unit length of the cylinder (J m
−1
°C
−1
), 1/H the thermal
contact resistance per unit area between the perfect con-
ductor and the surrounding material (m
2
°C W
−1
), and J
n
(u)
and Y
n
(u) the Bessel functions of u of order n of the first and
second kind (dimensionless).
[
8] Most of the existing heat pulse method literature
focuses first on calculating l and rc from the thermal
responses of the soil to a heat pulse. From these values, the
soil moisture content is then inferred, since both l and rc of
the soil monotonically increase with increasing water con-
tent. The well‐known advantage of using the dual‐probe
method for soil water determination is that both thermal
conductivity and volumetric heat capacity can be accurately
obtained from a single measurement, while the single probe
method is primarily sensitive to the thermal conductivity
[e.g., de Vries, 1952, 1963; Campbell, 1985; Kluitenberg et
al., 1993; Bristow et al., 1994]. The main advantage of
obtaining the volumetric heat capacity of the soil is that it
allows estimation of the change in soil water content without
information on soil‐specific thermal prope rties [Bristow et
al., 1993]. Some have tried to directly correlate soil mois-
ture content to the temperature rise during heating [e.g.,
Shaw and Baver, 1940; Youngs, 1956]. A disadvantage of
such methods is that a calibration curve that relates soil
moisture content to temperature change is needed for each
soil type, and for each probe design.
[
9] Systems using more than two probes provide addi-
tional information (e.g., direction of flux), and are an active
area of investigation [e.g., Bristow et al., 2001; Mori et al.,
2003, 2005; Hopmans et al., 2002; Ren et al., 2000; Green
et al., 2003; Kluitenberg et al., 2007]. Concerns regarding
the accuracy of the different heat pulse methods remain,
related to soil bulk density [Tarara and Ham, 1997], soil
mineralogy [Bristow, 1998], contact resistance between the
probe and the surrounding material [Blackwell, 1954], and
temperature sensitivity [Olmanson and Ochsner, 2006].
[
10] The use of actively heated fiber optic cable for obser-
vation of subsurface water movement has been demonstrated
[e.g., Perzlmaier et al., 2004, 2006; Aufleger et al., 2005],
though for determination of soil water content it was con-
cluded that (1) the method could only distinguish qualita-
tively between dry, wet and saturated soils [Perzlmaier et al.,
2004, 2006; Weiss, 2003] and (2) small changes in soil
water content could not be detected at levels above 6%
volumetric water content [Weiss, 2003]. Weiss [2003] con-
cluded that only with dramatic improvement of the signal‐
to‐noise ratio of the DTS instrumentation could sufficiently
accurate thermal conductivity be obtained by a DTS heat
pulse method to quantify soil water content above this level.
[
11] Although we agree that better DTS performance
improves accuracy, here we argue that the DTS method can
quantify moisture content more precisely than suggested
previously by using a different approach to data interpreta-
tion. Both Weiss [2003] and Perzlmaier et al. [2004] used
the long time approximation of either the line source or the
cylindrical source transient methods to calculate the thermal
conductivity of the soil, deriving the thermal conductivity
from the slope and intercept of a line fit to the temperature
response following an extended heat pulse. They then
computed the moisture content using a calibration equation.
Unfortunately, this fitting routine made use of data which
varied little between moisture contents (particularly the fit-
SAYDE ET AL.: RAPID COMMUNICATION W06201W06201
2 of 8

ted slope). Our approach was, in part, motivated by their
data, where it was evident that though the slope of heating
was rather insensitive to water content, the overall magni-
tude of the temperature change was quite sensitive to
moisture content. This is partially due to the impact of the
early time data that is not fully incorporated into the late
time analysis. In addition, there is an intrinsic improvement
in sensitivity found in integral methods compared to deriv-
ative (slope) approaches.
[
12] Recent work has shown that more robust estimates of
soil thermal properties are obtained using analyses that fit
the entire data set of temperature change with time to a
model [Mortensen et al., 2006]. In this article we do not
attempt to optimize the data interpretation, but rather dem-
onstrate the power of a simple interpretation methodology
that appears to make better use of information contained in
the heat pulse data obtained with a DTS system. Opportu-
nities for optimization of this method are manifold, and will
be the topic of further research.
2. Materials and Methods
[13] We seek a response variable that monotonically
varies with soil water content and is suited to the char-
acteristics of the DTS measurement method. To this end, we
propose quantifying the thermal response of the soil to the
heat pulse in the form of cumulative temperature increase
over a certain period of time
T
cum
¼
Z
t
0
0
!T dt; ð8Þ
where T
cum
is the cumulative temperature increase (°C.s)
during the total time of integration t
0
(s), and DT is the DTS
reported temperature change from the prepulse temperature
(°C). T
cum
is a function of the soil thermal properties. Higher
heat capacity and higher thermal conductivity, both of
which monotonically increase with soil water content ( '),
increase the rate at which heat is conducted away from the
probe and reduce the integral for sufficiently long heat
pulses. Thus, there exists a 1 to 1 function relating T
cum
to
' (under conditions where flow can be taken to be negli-
gible) for a given soil, heating rate, integration time, and
fiber optic cable characteristics.
[
14] One may ask about the advantage of the integrated
parameter compared to the maximum temperature increase
approach described by Shaw and Baver [1940] and Youngs
[1956]. The variance of the computed parameter is mini-
mized by taking advantage of the fact that the DTS read-
ings are fundamentally based upon cumulative photon
counting. The standard deviation of DTS temperature mea-
surements reduces with the square root of reading time
[Selker et al., 2006a]. This method allows use of relatively
long reading times (photon integration) and low sampling
rates. In fact, the value of T
cum
is largely unaffected by
sampling rate since the DTS will internally compute this
integral as it reports lower time resolution data requiring, for
example, a less expensive DTS recording instrument. It
will be shown later that T
cum
allows for more accurate
estimation of soil water content than DT in our experi-
mental setup.
[
15] The high‐speed DTS unit used in this experiment
(Sensortran DTS 5100 M4) allows high frequency data
collection for comparison of more traditional interpretations
of the integral method. This DTS unit recorded temperature
every 0.5 m along the fiber optic cable, with a spatial res-
olution of 1 m for each measurement. The average reading
frequency was 0.2 Hz.
[
16] A 0.61 m diameter sand column was supported by
a 1.46 m tall smooth interior, corrugated exterior HDPE
pipe (Figure 1). The bottom of the pipe was sealed with a
rubber membrane, and an outlet was installed 0.05 m above
the membrane seal. A 0.012 m diameter perforated hose
was fitted to the inside of the drainage port and wound in a
spiral laying flat on the bottom of the rubber seal to provide
an easily controlled lower boundary condition. The drain-
age was actively controlled using a peristaltic pump.
[
17] Within the column, 31.5 m of BruSteel (Brugg
Cable, Brugg, Switzerland) fiber optic cable was distributed
Figure 1. Images showing (a) the sand column and (b) the
fiber optic section (in helical coils) before inserting into the
sand column.
SAYDE ET AL.: RAPID COMMUNICATION W06201W06201
3 of 8

in a helicoidal geometry supported by five vertical 0.006 m
diameter fiberglass rods (Figure 1). The 3.8 × 10
−3
m outer
diameter cable made twenty‐one 0.48 m diameter helical
coils, spaced 0.06 m vertically, starting 0.05 m from the
bottom and ending at the surface of the sand (1.30 m from
the bottom). The fiber optic cable employed was composed
of two optical fibers encased in a central stainless steel
capillary tube (OD 1.3 × 10
−3
m / ID 1.07 × 10
−3
m) sur-
rounded by stainless steel strands (12 4.2 × 10
−4
m OD
stainless steel wires), all of which were enclosed in a 2 ×
10
−4
m thick nylon jacket. The metal components were used
as an electrical resistance heater (0.365 W/m).
[
18] Air‐dried medium sand (d
50
= 0.297 mm) was added
in 0.30 m deep lifts with vibration of the entire column using
a rubber mallet to settle the sand between lifts. No further
settling was observed during the remainder of the experi-
ment. The total depth of sand in the column was 1.30 m with
0.12 m of the HDPE pipe extending beyond the top of the
sand.
[
19] Computation of T
cum
requires a precise value of the
temperature before the start of the heat pulse. A 5 min DTS
reading preceding each heat pulse was used as the baseline
temperature. Thereafter, a 44.5 m section of the cable
(including the section in the sand column) was heated by
connecting the stainless steel windings at both ends of the
heated section to a variable voltage AC current source
(Staco® Variable Autotransformer Type 3PN1010). The
drop in voltage along the 12 AWG copper connecting wires
was ∼0.1% of the total, and thus was assumed to be negli-
gible. A digital timer with a precision of ±0.01%
(THOMAS® TRACEABLE® Countdown Controller
97373E70) controlled the duration of the heat pulse. A wide
range of combinations of power and time were tested,
though in this article we discuss only the results of 2 min
heat pulses at 20 W/m (120.2 VAC) which appeared to
provide an appropriate balance of temperature response and
duration relative to the DTS resolution. The measurements
were repeated three times. T
cum
was calculated using the
data obtained over the entire heating period of 120 s. The
temperature increase observed at the end of the heating
period (DT
120s
) will also be reported to compare its per-
formance in predicting soil water content with that of T
cum
.
We chose to employ DT
120s
because among all values of
DT for heating and cooling it had the highest signal‐to‐
noise ratio. A reference temperature reading was obtained
from a 33 m coil of fiber optic cable kept in an ice‐filled
water bath (0°C) (Figure 2).
[
20] DTS readings were taken in dry, saturated and
drained conditions. The drained condition was obtained one
month after establishing the water table at 0.4 m above the
bottom. Following the final DTS measurements in the
drained column, triplicate volumetric samples were obtained
from eight depths between the sand surface and the water
table (spanning 0.9 m) for calibration.
3. Results and Discussions
[21] Volumetric soil moisture content of samples taken
from the drained column varied from 4% to 41% (saturated),
with a sharp transition 0.3 m above the water table, typical
of sands (Figure 3). Repeatable, distinct values of T
cum
were obtained up to saturation (Figure 3). The slope in the
' − T
cum
and ' − DT
120s
relationships decreased with water
content (Figures 4 and 5), suggesting lower sensitivity at
higher water contents, as found in previous studies [e.g.,
Weiss, 2003].
[
22] To estimate the error in soil water content (' )
obtained from T
cum
, a function f (') was fitted to the T
cum
versus ' data using least squares regression (Figure 4). For
each value of ', the estimated error ( s
'
) was calculated as
(
'
¼
(
T
cum
df 'ðÞ
d'
+
+
+
+
+
+
+
+
;
ð9Þ
Figure 2. Temperature readings along the fiber optic cable before (solid line) and at the end (dashed
line) of a 2 min , 20 W/m heat pulse for the drained soil column condition. The “before” temperature
was obtained by averaging all readings during the 5 min directly proceeding the heat pulse start.
SAYDE ET AL.: RAPID COMMUNICATION W06201W06201
4 of 8

where s
T
cum
is the standard deviation of T
cum
,
df ð'Þ
d '
is the local
slope of the T
cum
response evaluated at '. In general, the
standard deviation of DTS‐measured temperature depends
on the distance from the DTS recording unit, increasing with
light loss as it potentially travels kilometers of distance from
the unit [e.g., Tyler et al., 2009]. However, over shorter
cable distances, such as the 50 m span employed here, this
effect is negligible. Therefore, the standard deviation of T
cum
,
s
T
cum
, should only depend on the performance of the DTS
system. In this experiment, s
T
cum
was computed as the aver-
Figure 3. Measured soil water content (circles) and cumulative t emperatu re increase (triangles) as
function of depth for a 2 min, 20 W/m heat pulse.
Figure 4. Average cumulative temperature increase (T
cum
)integratedover120sasfunctionofsoil
water content (') for three 2 min, 20 W/m heat pulses and fi tted function. For each soil water content
value, the error bars are obtained from the standard deviation of three repetitions. The R
2
of the fitted
function is 0.994.
SAYDE ET AL.: RAPID COMMUNICATION W06201W06201
5 of 8

age of all standard deviations of T
cum
observed along the 30
m cable section in the sand column. The same method was
employed to estimate the error in soil water content ob-
tained from DT
120s
. The error analysis shows that s
'
ob-
tained from either T
cum
or DT
120s
increased approximately
linearly with soil water content (Figure 6). As expected, the
error in soil water content obtained from T
cum
was much
smaller than that obtained from DT
120s
(Figure 6). This
error could be further reduced by increasing the signal‐to‐
noise ratio, which could be accomplished by averaging
several heat pulse results, using a more precise DTS unit,
increasing the heating intensity, or increasing the duration
of heating.
[
23] A large heat pulse could cause water to evaporate
and/or diffuse away from the cable [Farouki, 1986]. To
avoid this, and to minimize the energy required to complete
Figure 5. Average temperature increase at 120 s (DT
120s
) as function of soil water content (') for three
2min,20W/mheatpulsesandfittedfunction.Foreachsoilwatercontentvalue,theerrorbarsare
obtained from the standard deviation of three repetitions. The R
2
of the fitted function is 0.987.
Figure 6. Calculated err or (s
'
) in soil water content derived from T
cum
(solid line) and from DT
120s
(dashed line) as function of soil water content (').
SAYDE ET AL.: RAPID COMMUNICATION W06201W06201
6 of 8

a measurement, it is desirable to reduce both the magnitude
and duration of temperature increase. An important advan-
tage of the integral method is that a relatively good estimate
of soil water content can be obtained with a brief heat pulse.
In this experiment, the maximum cable temperature never
exceeded 17°C over the ambient soil temperature (Figure 5).
The injected energy was less than 2.4 kJ/m, compared to the
11.7 kJ/m for Weiss [2003] and greater than the 72 kJ/m
employed by Perzlmaier et al. [2004]. The much shorter
heating interval employed here (120 s), compared to 626 s
used by Weiss [2003] and 7200 s by Per zlmaie r et al.
[2004], greatly reduces the potential for such disturbance.
That said, Weiss [2003] indicated that his approach did not
give rise to water displacement, and our experiment
showed no change in T
cum
with replication, suggesting
there were no significant distortions due to the heat pulse
measurements. Sequential measurements did not show
persistent cumulative heating in our experiments, but this
would ultimately provide a practical limit to the feasible
sampling frequency using this method. Fortunately, this
cumulative heating can easily be measured with DTS.
[
24] Currently marketed DTS systems have both a tenfold
higher speed of reading performance and four times better
spatial resolution than that employed here. The magnitude
of the heat pulse required to obtain a particular level of
precision is scaled linearly with reading speed, thus we have
by no means explored the instrumentation limitations on
accuracy or energetic requirements of the DTS approach.
[
25] While the laboratory results are encouraging, field
measurements of soil water content using the DTS‐based
heat pulse method are expected to bring additional sources
of uncertainty. Expected primary sources of error include
poor contact between the probe and soil, and the spatial
variability of soil thermal properties.
[
26] Finally, in addition to varying with moisture content,
T
cum
is expected to be a function of the convective flow of
water around the heated cable. An increase in convective
flow will further increase the rate at which heat is dissipated
away from the probe and thereby reduce T
cum
. Thus, this
method has the potential to not only detect soil water content
but also to monitor water fluxes in saturated soils, as dem-
onstrated by Perzlmaier et al. [2004], with long heated
durations. The ability to use shorter pulses based on the
method proposed here allows greater separation between
measurements of moisture content and flux.
4. Conclusion
[27] We have shown that the heat pulse method using
coaxial heating and a DTS system is feasible for determi-
nation of soil water content across a much broader range
of values than previously reported. This result was found
by using a response metric that has not been previously
employed: the time integral of temperature deviation. This
strategy is especially appropriate to the DTS method wherein
precision of temperature reporting is a direct function of the
interval of photon integration. Though we have used high
temporal resolution in the DTS measurements, this method
can provide the same level of precision with less expensive,
slower DTS instruments since the data can be integrated in
time for analysis. Further, using more sensitive DTS sys-
tems, the technique could be more accurate and use shorter,
lower energy heat pulses which may be of importance in
remote application of the method.
[
28] While this study demonstrates feasibility, additional
work is required to develop optimal heating and interpre-
tation strategies for DTS‐based heat pulse methods, building
upon the rich literature related to needle heat pulse systems.
The key finding of this work is to confirm the potential to
employ DTS systems to monitor soil water content at tem-
poral resolutions well under one hour and at high spatial
resolution (≤1 m). In principle, this DTS method could
monitor soil moisture along cables exceeding 10,000 m in
extent. This would allow for concurrent observation of
thousands of adjacent locations, which will likely provide
new insights into the spatial structure of infiltration and
evaporation. Such measurements could be transformative in
our understanding of soil hydrology in natural and managed
systems at field and watershed scales. Many challenges
remain (e.g., installation in the presence of stones and
roots), calling for significant further effort in developing
this methodology. For example, we presented only results
from a single‐probe DTS approach, though multiple probe
approaches using DTS are expected to be of utility just as
they have been in other soil heat pulse applications.
[
29] Acknowledgments. The authors gratefully thank Atuc Tuli, Jan
Hopmans, Jim Wagner, Mark Hausner, and Christine Hatch for their very
helpful conversations about this method. We also wish to express our
appreciation to Water Resources Research reviewers, Editor, and Associate
Editor for their valuable suggestions and comments concerning this manu-
script. We acknowledge the NationalScienceFoundation(grantsNSF‐
EAR‐0930061, NSF‐EAR‐07115494) and the Oregon Experiment Station
for their critical financial support.
References
Aufleger, M., M. Conrad, S. Perzlmaier, and P. Porras (2005), Improving a
fiber optics tool for monitoring leakage, Hydro Rev. Worldwide, 13(4),
18–23.
Blackwell, J. H. (1954), A transient‐flow method for determination of
thermal constants of insulating materials in bulk, J. Appl. Phys., 25,
137–144, doi:10.1063/1.1721592.
Bristow, K. L. (1998), Measurement of thermal properties and water con-
tent of unsaturated sandy soil using dual‐probe heat‐pulse probes, Agric.
For. Meteorol., 89, 75–84, doi:10.1016/S0168-1923(97)00065-8.
Bristow, K. L., G. S. Campbell, and C. Calissendorff (1993), Test of a heat‐
pulse probe for measuring changes in soil water content, Soil Sci. Soc.
Am. J., 57, 930–934.
Bristow, K. L., R. D. White, and G. J. Kluitenberg (1994), Comparison of
single and dual probes for measuring soil thermal properties with tran-
sient heating, Aust. J. Soil Res., 32, 447–464, doi:10.1071/SR9940447.
Bristow, K. L., G. J. Kluitenberg, C. J. Goding, and T. S. Fitzgerald (2001),
A small multi ‐sensor probe for measuring soil thermal properties, water
content and electrical conductivity, Comput. Electron. Agric., 31, 265–
280, doi:10.1016/S0168-1699(00)00186-1.
Campbell, G. S. (1985), Soil Physics with BASIC, Elsevier, New York.
Campbell, G. S., K. Calissendorff, and J. H. Williams (1991), Probe for
measuring soil specific heat using a heat‐pulse method, Soil Sci. Soc.
Am. J., 55, 291–293.
Cars law, H. S., and J. C. Jaeger (1959), Conduction of Heat in Solids,
2nd ed., 262 pp., Clarendon Press, Oxford, U. K.
de Vries, D. A. (1952), A nonstationary method for determining thermal
conductivity of soil in situ, Soil Sci., 73,83– 90, doi:10.1097/
00010694-195202000-00001.
de Vries, D. A. (1963), Thermal properties of soils , in Physics of Plant
Environment, edited by W. R. Van Wijk, pp. 210–235, North‐Holland,
Amsterdam.
de Vries, D. A., and A. J. Peck (1958), On the cylindrical probe method of
measuring thermal conductivity with special reference to soils. Part 1.
Extension of theor y and discussion of prob e characteristics, Aust. J.
Phys., 11, 255–271.
SAYDE ET AL.: RAPID COMMUNICATION W06201W06201
7 of 8

Farouki, O. T. (1986), Thermal Properties of Soils, Ser. Rock Soil Mech. ,
vol. 11. Trans Tech, Clausthal‐Zellerfeld, Germany.
Freifeld, B. M., S. Finsterle, T. C. Onstott, P. Toole, and M. Pratt (2008),
Ground surface temperature reconstructions: Using in situ estimates for
thermal conductivity acquired with a fiber‐optic distributed thermal per-
turbation sensor, Geophys. Res. Lett., 35,L14309,doi:10.1029/
2008GL034762.
Grattan, K. T. L., and T. Sun (2000), Fiber optic sensor technology: An
overview, Sens. Act uators A, 82, 40–61, doi:10.1016/S0924-4247(99)
00368-4.
Green, S., B. Clothier, and B. Jardine (2003), Theory and practical applica-
tion of heat pulse to measure sap flow, Agron. J., 95(6), 1371–1379.
Heitman, J. L., J. M. Basinger, G. J. Kluitenberg, J. M. Ham, J. M. Frank,
and P. L. Barnes (2003), Field evaluation of the dual‐probe heat‐pulse
method for measuring soil water content, Vadose Zone J., 2, 552–560,
doi:10.2113/2.4.552.
Hopmans, J. W., J. Simunek, and K. L. Bristow (2002), Indirect estimation
of soil thermal properties and water flux using heat pulse probe measure-
ments: Geometry and dispersion effects, Water Resour. Res., 38(1),
1006, doi:10.1029/2000WR000071.
Jaeger, J. C. (1965), Application of the theory of heat conduction to geo-
thermal measurements, in Terrestrial Heat Flow , Geophys. Monogr.
Ser., vol. 8, edited by W. H. K. Lee, pp. 7–23, AGU, Washington, D. C.
Kluitenberg, G. J., J. M. Ham, and K. L. Bristow (1993), Error analysis of
the heat pulse method for measuring soil volumetric heat capacity, Soil
Sci. Soc. Am. J., 57, 1444–1451.
Kluitenberg, G. J., T. E. Ochsner, and R. Horton (2007), Improved analysis
of heat pulse signals for soil water flux determination, Soil Sci. Soc. Am.
J., 71(1), 53–55.
Larson, T. H. (1988), Thermal measu rement of soils using a multineedle
probe with a pulsed point‐ source, Geophysics, 53,266– 270,
doi:10.1190/1.1442461.
Lubimova, H. A., L. M. Lusova, F. V. Firsov, G. N. Starikova, and A. P.
Shushpanov (1961), D etermination of surface heat flow in Mazesta
(USSR), Ann. Geophys., XIV, 157
–167.
Mori, Y., J. W. Hopmans, A. P. Mortensen, and G. J. Kluitenberg (2003),
Multi‐functional heat pulse probe for the simultaneous measurement of
soil water content, solute concentration, and heat transport parameters,
Vadose Zone J., 2(4), 561–571, doi:10.2113/2.4.561.
Mori, Y., J. W. Hopmans, A. P. Mortensen, and G. J. Kluitenberg (2005),
Estimation of Vadose zone water flux from multi‐functional heat pulse
probe measurements, Soil Sci. Soc. Am. J., 69(3), 599– 606,
doi:10.2136/sssaj2004.0174.
Mortensen, A. P., J. W. Hopmans, Y. Mori, and J. Simunek (2006), Multi‐
functional heat pulse probe measurements of coupled vadose zone flow
and transport, Adv. Water Resour., 29 , 250–267, doi:10.1016/j.advwatres.
2005.03.017.
Olmanson, O. K., and T. E. Ochsner (2006), Comparing ambient temper-
ature effects on heat pulse and time domain reflectom etry soil water
content measurements, Vadose Zone J., 5, 751–756, doi:10.2136/vzj2005.
0114.
Perzlmaier, S., M. Aufleger, and M. Conrad (2004), Distributed fiber optic
temperature measureme nts in hydraulic engineering—Prospec ts of the
heat‐up method, paper presented at 72nd ICOLD Annual Meeting Work-
shop on Dam Safety Pro blems and Sol utions‐Sharing Experience,
Korean Natl. Comm. on Large Dams, Seoul, 16–22 May.
Perzlmaier, S., K.‐H. Straß er, T. Strobl, and M. Aufleger (2006), Integral
seepage monitoring on open channel embankment dams by the DFOT
heat pulse method, paper presented at 74th Annual Meeting, Int. Comm.
on Large Dams, Barcelona, Spain.
Ren, T., G. J. Kluitenberg, and R. Horton (2000), Determining soil water
flux and pore water velocity by a heat pulse t echnique, Soil Sci. Soc.
Am. J., 64, 552–560.
Ren, T., T. Y. Ochsner, and R. Horton (2003), Development of Thermo‐
time domain reflectometry for Vadose zone measurements, Vadose Zone
J., 2, 544–551, doi:10.2113/2.4.544.
Ren, T., Z. Ju, Y. Gong, and R. Horton (2005), Comparing heat‐pulse and
time domain reflectometry soil water contents from thermo‐ time domain
reflectometry probes, Vadose Zone J., 4,1080–1086, doi:10.2136/
vzj2004.0139.
Selker, J., N. van de Giesen, M. Westhoff, W . Luxemburg, and M. B.
Parlange ( 2006a), Fiber optics o pens window on stre am dynamics,
Geophys. Res. Lett., 33, L24401, doi:10.1029/2006GL027979.
Selker, J., L. Thévenaz, H. Huwald, A. Mallet, W. Luxemburg, N. van de
Giesen, M. Ste jskal, J. Zeman, M. We sthoff, and M. B. Parlange
(2006b), Distributed fiber optic temperature sensing for hydrologic sys-
tems, Water Resour. Res., 42, W12202, doi:10.1029/2006WR005326.
Shaw, B., and L. D. Baver (1940), An electrothermal method for following
moisture changes on the soil in situ, Soil Sci. Soc. Am. Proc., 4, 78–83.
Shiozawa, S., and G. S. Campbell (1990), Soil thermal conductivity,
Remote Sens. Rev., 5, 301–310.
Steele‐Dunne, S. C., M. M. Rutten, D. M. Krzeminska, M. Hausner, S. W.
Tyler, J. S. Selker, T. A. Bogaard, and N. C. van de Giesen (2010), Fea-
sibility of soil moisture estimation using passive distributed temperature
sensing, Water Resour. Res., 46, W03534, doi:10.1029/2009WR008272.
Tarara, J. M., and J. M. Ham (1997), Measuring soil water content in the
laboratory and field with dual‐probe heat‐capacity sensors, Agron. J.,
89, 535–542.
Tyler, S., S. Burak, J. McNamara, A. Lamontagne, J. Dozier, and J. Selker
(2008), Spatially distributed temperatures at the b ase of two mountain
snowpacks measured with fiber optic sensors, J. Glaciol., 54(187),
673–679, doi:10.3189/002214308786570827.
Tyler, S. W., J. S. Selker, M. B. Hausner, C. E. Hatch, T. Torgersen, C. E.
Thodal, and G. Schladow (2009), Environme ntal temperature sensing
using Raman spectra DTS fiber‐optic methods, Water Resour. Res.,
45, W00D23, doi:10.1029/2008WR007052.
Weiss, J. D. (2003), Using fiber optics to detect moisture intrusion into a
landfill cap consisting of a vegetative soil barrier, J. Air Waste Manage.
Assoc., 53, 1130–1148.
Westhoff, M. C., H. H. G. Savenije, W. M. J. Luxemburg, G. S. Stelling,
N. C. van de Giesen, J. S. Selker, L. Pfister, and S. Uhlenbrook (2007),
A distributed stream temperature model using high resolution tempera-
ture observations, Hydrol. Earth Syst. Sci., 11, 1469–1480.
Youngs, E. G. (1956), A laboratory method of following moisture content
changes, Trans. Int. Congr. Soil Sci., 6, 89–93.
R. Cuenca, Hydr ologic Sciences, Division o f Earth Sciences, Natio nal
Science Foundation, 4201 Wilson Blvd., Arlington, VA 22230, USA.
M. English, C. Gregory, C. Sayde, J. S. Selker, and N. Tufillaro,
Department of Biological and Ecological Engineering, Oregon State
University, 116 G ilmore Hall , Corvallis, OR 97331, USA. (say dec@
onid.orst.edu)
M. Gil‐ Rodrig uez, Department of Rural Engineering, Technical
University of Madrid, ETSI A grónomos, Ciudad Universitaria s/n,
E‐28040 Madrid, Spain.
S. Tyler, Department of Geological Sciences and Engineering, University
of Nevada, Reno, MS 172, Reno, NV 89577‐
00172, USA.
N. van de Giesen, Water Management Civil Engineering and
Geosciences, Delft University of Technology, Stevinweg 1, NL‐2628 CN
Delft, Netherlands.
SAYDE ET AL.: RAPID COMMUNICATION W06201W06201
8 of 8