|
Principle
- background theory
Time
domain reflectometry (TDR) methods are based on the differing
ability of different materials to conduct electromagnetic
radiation. Electromagnetic radiation consists of an electric
field and a magnetic field perpendicular to each other.
When an electric field is applied to polar materials, the
molecules of the material orientate uniformly in the field
and reduce its strength. This decreases the velocity of
propagation of the electromagnetic radiation in that material.
The tendency of molecules to orientate themselves in an
electric field is measured by the dielectric constant, e,
of that material. The greater the tendency of molecules
to orientate uniformly, the greater the dielectric constant
and thus the slower the velocity of propagation of electromagnetic
radiation in that material. The relationship between velocity
of propagation, v, and dielectric constant, where c is speed
of light through a vacuum, is (Charlesworth 2000):
Soil
is a mixture of soil solids, water and air. Soil solids
have a dielectric constant of between 4 and 8. Air, with
a very high percentage of non-polar constituents, has a
dielectric constant of 1, while for water, a bipolar molecule,
it is 81. The velocity of propagation of electromagnetic
radiation through soil depends on the relative volumes of
each of these three constituents. The moister a soil, the
larger its dielectric constant and the lower the velocity
of electromagnetic radiation through it. An empirical relationship
has been found between the dielectric constant of a soil
and its volumetric water content, q, and is valid for a
wide range of soil types (Topp cited in Charlesworth 2000)
:
The
predicted value of volumetric water content is only affected
to a small degree by soil type, bulk density, temperature
and electrical conductivity (Hillel 1998).
Top
of page
Principle
- TDR unit
A
TDR unit consists of a computerised measuring unit with
an attached probe known as a wave guide (Figure 1). The
wave guide is inserted into the soil and the unit generates
an
 
Figure
1 TDR unit (left) and probe ('wave guide')
(right)
electromagnetic
pulse which travels along the probes and through the soil
around them. At the end of the probes, the pulse reaches
the soil-probe interface, where there is a sharp difference
in conducting ability, and some of the pulse is reflected.
The TDR unit measures the time, t, taken by the pulse to
travel from when the pulse first encounters soil at the
base of the wave guide until the reflected signal is first
detected. As the distance travelled by the pulse is known
(viz. twice the probe length, L), its velocity and, hence,
the dielectric constant of the soil can be calculated (Charlesworth
2000):
The software
in the unit then calculates and displays volumetric water
content from the empirical equation above. The TDR technique
measures the dielectric property of soil in a cylinder of
diameter slightly larger than the spacing between the prongs
of the probe (Marshall et al. 1996).
Principle
– agricultural and environmental implications
The
growing importance of efficient water use has created a
need for practical and simple field methods of measuring
soil water content. The TDR technique measures volumetric
water content accurately. While the equipment is expensive
and only portable with some difficulty, it has been used
in irrigation scheduling partly because the measurement
of volumetric water content is valid for a range of soil
types.
TDR
can also measure soil salinity. The above prediction equation
gives erroneous results for volumetric water content if
the electrical conductivity (EC) of the soil solution is
greater than 8 dS/m (Marshall et al. 1996). The TDR is useful
in measuring the EC of the soil solution in broad survey
work as it samples larger volumes than laboratory sample
methods are capable of, although the latter methods are
more accurate (Marshall et al. 1996).
Top
of page
Method
– topsoil volumetric water content
The
TDR unit used was a TRASE brand unit, which is a purpose
built soil moisture meter. It includes an EM pulse generator,
pulse timer and inbuilt software which can detect and store
the wave form of the EM pulse. A wave guide with parallel
prongs is attached to the TDR and inserted into the soil
for measuring volumetric water content, which is displayed
almost immediately.
To measure
topsoil volumetric water content, the wave guide was inserted
vertically into the soil at each sampling point and the
measurement given by the TDR recorded. A new sampling point
was chosen by throwing a metre rule several metres from
the current sampling point and then sampling at the landing
point of the metre rule. The observers worked over the survey
area in this way in a reasonably systematic fashion and
attempted to avoid large spatial gaps in sampling points.
The easting and northing coordinates of each survey point
were recorded from a hand-held GPS unit which had an accuracy
of approximately ±3 metres.
Volumetric
water content was also measured at each of five peg points
in the cultivated and pasture areas. The five points in
each area were approximately equally spaced along a line
and the two peg lines were roughly parallel.
Top
of page
Analysis
of data and results - topsoil volumetric water content at
pairs of points in landuse classes
The
volumetric water content measured at the peg points was
analysed by a paired t-test assuming equal variances of
the two sets of data. This assumption was supported by an
F test of equality of the sample variances (P > 0.10).
From this test, we can conclude that the mean topsoil volumetric
water content of the cultivated ground is significantly
greater than that of the pasture area (P < 0.001). The
mean difference is 19.2% with a 95% confidence interval
of 15.9% to 22.6%.
The
mean topsoil volumetric water content for the cultivated
area sample is 35.8% and for the pasture area sample is
16.6%. As stated in the site description, the soil in the
survey area is a Black Vertosol. For a clay soil, volumetric
water contents of 35.8% and 16.6% indicate water potentials
in the vicinity of -10 kPa and -1500 kPa respectively. Water
evaporates slowly from the soil in the cultivated area but
the absence of vegetation in this area is the reason for
the higher water content of the soil there. The pasture
plants continually use water and so water is lost from the
pasture area soil through transpiration which occurs at
a higher rate than evaporation of water from the cultivated
ground.
Analysis
of data - TDR topsoil volumetric water content over survey
area
Measurements
were made at points chosen by the method described above.
As the TDR unit failed for some time, volumetric water content
was only measured at 35 of the 101 sampling points in the
survey area, 18 points in the cultivated area and 17 points
in the pasture area.
The
data were analysed in the same way as the electrical conductivity
data: firstly, by one-way ANOVA to examine the effect of
landuse, then kriging the residuals and finally adding the
residuals, weighted for landuse, to the ANOVA means for
the two landuse classes. The details of this analysis are
as follows.
These
data were analysed by one-way analysis of variance against
landuse (viz. pasture or cultivated ground). This analysis
yielded a mean volumetric water content for each landuse
type. The residuals at each sampling point were calculated
by subtracting the appropriate landuse mean value from the
measured value at each point. This set of residuals were
then analysed by kriging using the Vesper (ACPA
2003) software package. A grid was created for the survey
area and an exponential variogram was fitted. The kriging
produced kriged residual predictions at one?metre intervals
over the grid area.
The
kriged residuals were then added to the ANOVA means, which
were weighted for landuse type at each grid point in the
following way. Sampling points were designated with an indicator
variable of 0 for cultivated ground or 1 for pasture. The
indicator variables were then kriged using the same settings
as for the volumetric water content residuals, producing
probability predictions for landuse at each grid point.
The overall mean-kriging prediction of volumetric water
content was then obtained from the following formula:
Volumetric
water content prediction = Kriged residual + (Landuse probability
x Pasture mean) + [(1 - Landuse probability) x Cultivated
mean)]
These
predictions were then displayed on a contour plot using
the Minitab statistical package.
Top
of page
Results
- TDR topsoil volumetric water content over survey area
A
contour plot of predicted volumetric water content over
the survey area is shown in Figure 2. The volumetric water
content in the cultivated area is generally greater than
25% and in the pasture area less than 25%. The noticeable
gradation in predictions over the boundary between the two
areas is the result of kriging the landuse factor. The actual
boundary between the two areas in the field was sharper.
Figure
2
The
volumetric water content measured over the survey area was
analysed by one-way ANOVA assuming equal variances of the
two sets of data. This assumption was supported by an F
test of equality of the sample variances (P > 0.10).
The ANOVA revealed a significant difference between volumetric
water content between the two landuse classes (P < 0.001).
The mean for cultivated land is 29.9% and for pasture is
20.7%. The soil on the survey area is a Black Vertosol and
for a clay soil these volumetric water contents correspond
to soil water potentials of approximately -35 kPa and -250
kPa. Soil water in the pasture area is being withdrawn by
the vegetation cover and is lost through evapotranspiration.
In the cultivated area, the lack of vegetation means soil
water is lost only by evaporation, which occurs more slowly.
While
more soil moisture is conserved in soil with no vegetation
cover than in soil with vegetation cover, the risk of erosion
by wind or water is less where there is vegetation cover.
A cover crop (e.g. legume) can also benefit the soil by
increasing soil nutrients even though it may use soil moisture.
Leaving stubble from a previous crop is an effective way
of reducing both the risk of erosion and soil water loss
by evaporation, as well as maintaining the soil organic
carbon level.
Top
of page
Problems
with method - peg lines
The
TDR unit can be used to measure volumetric water content
in the range from 0.05 cu. cm/cu. cm to 0.55 cu. cm/cu.
cm (Charlesworth 2000). The values obtained in this experiment
were well within this range and so should be accurate. Bulk
soil electrical conductivity (EC) can also affect the accuracy
of the TDR measurement when pore water EC reaches 800 mS/m
(Charlesworth 2000). This is a high level of solute and
as there was no evidence of such a high soil water solute
level over the site, this distortion would not have occurred.
The
peg points within each area were possibly too close together
to capture the variability evident across the whole site.
This is shown by the differences in the two sets of means
from the peg points and the survey site as a whole.
Another
issue related to the experiment is replication. The two
sets of sampling points were taken from only one field each
and so there is no replication. The five points sampled
in the cultivated area, for example, are subsamples from
one field, as are the five points in the pasture area. There
is no way in this experiment to estimate the variability
of volumetric water content in each landuse class. The observed
variation may be due to the random variation in the fields
sampled or to difference in landuse; however, a firm conclusion
in this regard cannot be drawn from the experiment. The
mapping of volumetric water content from the experimental
results is valid with the proviso that no statistically
significant difference is read into the map. The same constraints
apply to the survey area wide results and conclusions.
Top
of page
References
Charlesworth
P (2000) ‘Soil water monitoring.’ Irrigation
Insights No. 1, Land and Water Australia, Canberra.
Hillel
D (1998) ‘Environmental soil physics.’ (Academic
Press: New York)
Marshall
TJ, Holmes JW, Rose CW (1996) ‘Soil physics.’
(Cambridge University Press: Cambridge)
|