read out directly in the field and that could be
surement by solving both the water and the metha-
interfaced directly to automatic data logging
nol equations simultaneously.
equipment. The performance limits of these cir-
6. It is possible to use insulated probes to make
cuits when used with the specific probes that were
accurate soil moisture measurements, but a care-
designed for these test programs have never been
ful calibration program is required. The test data
determined for the specific application of the mea-
(although somewhat small in sample size) indi-
surement of soil moisture.
cate that better results are likely possible with the
It should also be noted that the accuracy ques-
calibration technique of assuming a series capaci-
tance to correct the readings to those that would
tion may be circumvented entirely by the expedi-
be expected from uninsulated probes.
ent method of calibrating each measurement
7. Data from the insulated probes used for the
system to each individual soil and applying an
tests in this report are valid only up to a molarity
accuracy limit that is equivalent to the repeatabil-
of 0.01 N unless calibrated directly for a specific
ity of the measurement circuit. If such a proce-
saline soil.
dure is used, it would be easier to calibrate directly
8. The dielectric constants found by using the
to a capacitance measurement since a dielectric
capacitance measurement procedures together
constant measurement is not required. For indi-
vidual calibrations such as this, it is probable that
with the circuits and conversion techniques de-
acceptable soil moisture measurements could be
scribed in this report produce too high a soil mois-
made with only a single output voltage; for ex-
ture measurement when used in the equations of
ample, soil moisture by percent dry weight vs.
Topp et al. or Roth et al. The probable reason for
Vph. If percent dry weight were required, it would,
the high readings is the difference in the frequen-
of course, necessitate tests to find the site's dry
cies used in this technique and the TDR tech-
soil density, since the capacitance measurement
nique. Nevertheless, these two measurement
approach is a volume calibration method.
techniques produce such consistently different
values that there is almost certainly a constant
multiplier that would allow the use of the Topp et
al. or Roth et al. equation and still yield an accu-
CONCLUSIONS AND
racy equal to that of the gravimetric calibration
RECOMMENDATIONS
equations.
9. At a dielectric constant of K = 32 (the metha-
Based on the tests performed to date on this
nol checkpoint), an error of 2.0 in the calculation
soil moisture measurement system, the following
conclusions appear to be supported by the data:
for this dielectric constant results in an error of
3.3% in the volume moisture content. The same
1. The circuits used for these tests provide poor
2.0 error in the dielectric constant measurement
accuracy but are highly repeatable. As such they
at K = 10 causes an error of 2.14% in the soil
are perfectly acceptable for capacitance or dielec-
moisture calculation. Either error seems to be rea-
tric constant measurements, provided that an ad-
sonably acceptable, but this does point out the
equate calibration procedure is developed that
necessity of making dielectric constant measure-
produces acceptable accuracy.
ments as accurately as possible.
2. This capacitance method for the determina-
tion of soil moisture is capable of producing results
Based on these conclusions, the following rec-
that are equally as good as the TDR techniques.
ommendations seem appropriate:
3. Water may be used as a reference material to
1. This system works quite well to measure soil
convert capacitance measurements to dielectric
moisture. CRREL should continue to develop it
constant measurements. Worst-case accuracies of
and report its use. Ultimately it will prove to be
5% may be reasonably expected using this tech-
better than TDRs, tensiometers, soil moisture
nique.
blocks, or similar devices in meeting the ideal
4. The circuits used for the tests in this report
performance criteria, listed in the beginning of
are more accurate and repeatable when their cable
this report and repeated here:
capacitances are nulled.
5. Methanol is a good check on the procedure
Accuracy
for dielectric constant measurements. It could also
Good long-term stability
be used as a reference material, the same as water.
The calibration procedure would then be to find
Easily read using untrained observers
the dielectric constant from the capacitance mea-
Low cost
21
Previous Page
