|
 |
Geophysical surveys: principle,
method
This page provides a detailed description of the
principle and method for the geophysical surveys
seepage identification and measurement technique.
Pages in this section include:
Principle
Many techniques for seepage assessment identify seepage distribution
and rate by directly measuring a physical property at a single
location. For example, groundwater monitoring of water levels
in a bore allows a direct measure of the watertable, and infiltration
tests are direct measures of the soil properties at a particular
point.
In contrast, geophysical surveys and remote sensing, use high-density
sampling of subsurface and near-surface properties to provide
essentially continuous data along the channel.
Geophysical techniques applied to seepage measurement involve
measuring a contrast in terrain conductivity (or its inverse,
resistivity) in the subsurface profile around the channel. They
can be used in two ways:
- Direct measurement of conductivity of groundwater and
identifying the conductivity contrast of fresher channel
water as it seeps
into and dilutes saltier native groundwater. Decreasing groundwater
salinity causes a decrease in electrical conductivity (or an
increase in resistivity).
- Identification of contrasts in soil
properties and inference of the likelihood of greater seepage
through more permeable
materials in the zone above the watertable. Formations more
likely to allow
seepage, such as sands, are naturally lower in conductivity
(higher in resistivity) due to lower porosity and lower cation
exchange
capacity than tighter clay-dominated formations. In addition,
the higher permeability of such formations leads to better
drainage and lower salt content, further reducing conductivity.
The magnitude
of seepage is assumed to be related to unsaturated zone soil
properties beneath or adjacent to the channel.
Technically the second method of ‘detection’ is
not really detection, but the magnitude of seepage is assumed
to
be related to unsaturated zone soil properties beneath or adjacent
to the channel. In many cases this is a reasonable assumption,
however the unsaturated zone is not necessarily the controlling
influence on seepage. For example, Australian channels tend to
silt up over time and the resulting surface-clogging layer is
often more restrictive than the unsaturated zone. Therefore unsaturated
zone lithology may not be related to seepage rates, as seepage
is affected by the thickness and conductance of the clogging
layer. Nevertheless, the inferred method of identifying contrasts
in soil properties (i.e. where the watertable was deeper than
the penetration depth of the geophysical equipment) was successful
at most sites during the IAL trials (IAL, 2003).
There is less risk, however, in using the direct method of seepage
detection, because as the name implies it is not inferred, but
direct.
Figure 1 - Comparison of how
geophysical techniques can be used to identify channel seepage
(LHS – inferred
from soil property variations, RHS – direct measurement
of salinity impact on watertable)
Two possible limitations of the direct method of seepage detection
are:
- In a relatively non-saline groundwater environment the
fresh seepage water will not contrast with the native
groundwater. This is not a problem in most Australian conditions
- Groundwater
salinity may vary along the channel and this needs to be
allowed for in interpretation.
| Method |
 |
Technique selection
Different geophysical techniques measure the conductivity
(or resistivity) of different parts of the subsurface
profile. The methods most appropriate for channel
seepage detection in Australian water industry operations
are electromagnetics (specifically EM31 and EM34)
and resistivity. These methods achieve different
depths of penetration (the depth to which the instrument
set-up allows readings to be obtained) and have different
depth focuses (depth interval from which the bulk
of the response is derived).
Geonics EM techniques record average conductivity
across a certain depth interval. The depth of penetration
and the focus of the interval are dependent on
the type and dipole orientation of the instrument.
For
example, the depth of penetration for EM34 with
a 10m coil spacing and in vertical dipole is around
20m, so the conductivity measurement from available
tools is an average across this depth. In contrast,
existing resistivity tools (as used in the IAL
trials) (IAL, 2003) measure the resistivity of
the profile at a range of depths. This provides
information
on the change in resistivity through the profile.
The depth of penetration of resistivity techniques
can be varied by the addition or removal of dipole
arrays.
Depth to watertable influences the property being
measured by a given geophysical technique. When the
watertable is deep, unsaturated zone soil properties
may be measured, but when the watertable is shallow,
the same technique may be measuring changes in the
groundwater resulting from seepage.
Therefore, it is important that the depth to watertable
at the site is known before selecting a geophysical
technique. This allows for the fact that geophysical
responses are related to direct measurements of seepage
impact in the groundwater or inferred seepage zones
based on interpreted soil properties.
Preferred geophysical techniques
The preferred technique for geophysical channel seepage
assessment is directly detecting the impact of seepage
on the groundwater. This means that the instrument
must focus on the zone immediately above and several
metres below the watertable.
It must be confirmed that seepage is controlled by
the unsaturated zone and not surface clogging processes.
Otherwise errors will potentially be introduced to
the assessment process.
A summary of geophysical techniques and how they
can be used in seepage measurement, including examples
from the IAL trials, is presented below. Work
is best undertaken by specialist geophysical contractors
who have the appropriate tools, including software
for analysis and documentation.
- For shallow watertables (surface to approximately
5m) EM31 is suitable for direct seepage detection.
- For
watertables deeper than 5m, EM34 (in vertical
dipole mode) or resistivity can be used. However,
particularly for deeper watertables, it is easier
to focus on a given depth with resistivity, and
this can be achieved independent of knowledge
of groundwater
depth. The significant advantage of resistivity
is that it provides a profile of the resistivity
beneath
the channel. The disadvantage is that resistivity
technology for channel seepage assessment is
relatively new and therefore more expensive.
- EM31 (vertical dipole mode) adjacent
to the channel can be used effectively
in areas with deeper
watertables, although it does not directly measure the seepage
impact on the watertable. It was effective
in the IAL trials (IAL, 2003) due to
fact that the
upper soil layers are the most influential on channel seepage
and the relatively shallow depth focus of
EM31 measures these upper soil layer properties.
The method infers
zones of likely channel seepage by identifying
materials in the unsaturated zone most susceptible to seepage.
A decision to use EM31 in deeper watertable
areas might be based on i) cost and required
accuracy
- If a potentially slightly lower level of accuracy
is considered acceptable then EM31 represents
a cheaper alternative than EM34 or resistivity; or, lack of
alternatives - EM34 or resistivity contractors
are not readily available.
Table – 1 Geophysical techniques
for seepage detection and measurement
| Watertable depth (m) |
Recommended technique [1] |
Detection method [2] |
Approximate depth of penetration
(m)[3] |
Depth focus (m)[4] |
| Surface to 1.5 |
EM31 (horizontal dipole)[5] |
Direct watertable impact |
3 |
0-1 |
| 1.5-5 |
EM31 (vertical dipole)[5] |
Direct watertable impact |
6 |
1-3.5 |
| 5-12 |
EM34 - 10|m coil spacing (vertical
dipole)[6]
OR
Resistivity[7, 9]
OR
EM31 (vertical dipole)[8] |
Direct watertable impact
Direct watertable impact
Soil property variations |
15
NA[10]
6 |
3-10
NA[11]
1-3.5 |
| 12-25 |
EM34- 20|m coil spacing (vertical
dipole)[6]
OR
Resistivity[7, 9]
OR
EM31 (vertical dipole)[8] |
Direct watertable impact
Direct watertable impact
Soil property
variations |
30
NA[10]
6 |
6-20
NA[11]
1+3.5 |
| > 25 |
Resistivity[9]
OR
EM31 (vertical dipole)[8] |
Direct watertable impact
Soil
property variations |
NA[10]
6 |
NA[11]
1-3.5 |
Note:
(Numbers in [ ] brackets)
-
It is recommended that EM techniques
are conducted adjacent to the channel (additional
survey runs can be conducted away from
the channel). Resistivity surveys should
be conducted
on-channel – (ie, on the channel
while the channel is running).
-
Direct detection
of seepage impacts on the watertable
is the recommended technique, but inferred
detection based on soil property variations
can provide
an adequate
simulation and may be more convenient for various reasons. Note that direct
detection relies on a salinity contrast between the channel water and the
groundwater.
It is recommended the groundwater should be at least 3 to 4 times more saline
than the channel water, a condition that is met in most Australian conditions.
-
Approximate
detection of penetration is referred
to in the Geonics manual (McNeil, 1980)
as the effective depth of exploration.
This is the depth to
which approximately 75% of the response is attributed.
-
‘Depth focus’ describes
the depth (range) which is most influential
in terms of the relative contribution to the overall EM response (McNeil,1980).
-
These
can be conducted immediately adjacent
to the channel or on-channel. Both
are recommended if budget allows. If
on-channel
is used for a
watertable of 0-1.5m, the survey should preferably collect data
in vertical dipole
mode so the effects of channel water will be less influential.
For sites with a watertable
0-1.5m, EM31 on channel may be preferred if significant land salinisation
exists adjacent to the channel.
-
Horizontal and vertical dipole:
note that as applied to EM34, vertical
dipole does not refer to the coil
orientation with respect
to the ground,
and is in
fact opposite to the coil orientation. In vertical dipole mode
the coils should be horizontal to the ground, which is a slower
method
than horizontal
dipole
mode, in which they are held perpendicular to the ground.
-
Resistivity
is the preferred direct measurement
technique for this depth to watertable
but EM34 is a more accessible
alternative.
-
This should be conducted immediately
adjacent to the channel.
-
This should be
conducted on-channel.
-
The penetration
depth of resistivity depends of the
particular system
set-up (dipole spacing and
length).
-
Resistivity surveys measure
resistivity at a range
of depth intervals within the
profile
(i.e. there is
no fixed depth
focus).
|
EM systems
Ground conductivity can be obtained with frequency
domain electromagnetic (FEM) instruments such as
the Geonics EM38, EM31 and EM34 meters. EM38 depth
penetration is less than 1m and dominated by near-surface
conductivity changes. Trials have indicated it
is not an appropriate system for this work and
hence EM38 is not recommended for seepage assessment
(SKM, 1998).
Field operation
FEM systems can measure the electromagnetic properties
of the soil profile up to a depth of 100m, with
penetration depth dependent on frequency and coil
spacing. In the IAL trials, Geonics style FEM
units were used. For a given coil spacing, Geonics
EM systems can be used in horizontal dipole or
vertical dipole mode. The dipole mode affects the
relative contribution of the profile at different
depths to the overall response. Near-surface features
tend to dominate in the horizontal mode, while
the vertical mode is more influenced by the ‘mid’ part
of its depth range (McNeil, 1980).
Geonics EM34 systems can be used at various intercoil
spacings and orientations to vary the effective
depth of exploration, in contrast to Geonics EM31
systems which have a fixed coil spacing. EM34 is
slower than EM31, as the required coil spacing
means that the coils must be carried by hand. An
EM34 unit in operation (horizontal dipole) is shown
in Figure -2.
EM31 measurements were undertaken on a quad bike,
at some sites and also on-channel (Figure -3).
A digital outlet was linked to a data storage device
and the results linked to GPS location data. Sampling
could be set by time or by distance travelled,
and measurements could be taken as close as every
10cm along a transect line. In the IAL trials
a reading was taken approximately every 5m.
Various multi-coil, multi-frequency systems that
enable conductivity depth resolution are under
development.
Details of geophysical systems are described in
the Literature Review (IAL, 2000a) and the Technical
Report on the Project Trials (IAL, 2003).
Figure 2
EM34 in use (in horizontal dipole mode)
Figure 3
EM31 mounted on quad bike and on boat
Qualitative versus quantitative seepage
assessment using EM results
Geophysical tools can be applied in two ways:
- Mapping the distribution of relative seepage
zones. High- and low-seepage zones (relative
to each other) can be effectively mapped using geophysical
techniques alone. Greater confidence can be obtained
by confirmation with geological investigations.
- Quantification
of seepage rates. Quantification requires integration of
geophysical methods with
other techniques in order to calibrate results.
Geophysics can be used to provide an estimate
of seepage rate, provided a sufficiently strong relationship
can be developed between geophysical response
and pondage tests. The relative seepage rate can be
identified by the correlation of geophysics
with other data, particularly pondage tests.
The level of investigation needs to meet the project
objectives and budget and the level of confidence
required. These two applications of geophysical
surveys are described in further detail below,
as they relate to EM assessment.
Qualitative assessment - mapping
seepage zones
Geophysical surveys using EM31 and EM34 have shown
promising results in mapping the extent of seepage
zones along channels, based on both direct and
inferred seepage detection (IAL, 2003). Mapping
of relative seepage occurrences is valuable in
determining the extent of the seepage problem,
and the parts of the channel with the greatest
problems. Comparison of EM with direct measurement,
especially with pondage tests, is considered to
provide the best assessment of the relative significance
of the EM readings in mapping the distribution
of seepage zones along a channel.
Geophysical mapping is complemented by direct measurement
to evaluate the rate of seepage, although point
measurements can be also be used. In addition,
soil and geological strata information can provide a greater
level of confidence in this qualitative assessment.
The use of EM mapping is extremely valuable for indicating
parts of a channel that are most likely to be seeping
at the highest rate. Mapping should be the first
task conducted in an investigation. This can provide
the basic framework on which to locate other work
such as pondage tests, groundwater and soil bores.
Quantitative analysis of seepage
Good correlations were obtained between pondage
test seepage rates and average EM conductivity
of the subsurface profile along the pond, in the
trials. EM conductivity values were correlated
against the pond seepage rates with correlation
coefficients as high as 0.9 for sections of channel
measured. Care must be made in interpreting trends
from a local scale because there may only be a
narrow range of seepage rates and conductivity
values. Too much emphasis on trends under these
circumstances can lead to invalid interpretations.
The relationships between seepage rate and conductivity
measurement can be used to attempt to quantify
the seepage rate. Statistical analysis is described
in detail in the Project Trials Report (IAL,
2003). The relationships obtained are included
in this website as a possible tool to provide preliminary
estimates of seepage rates. The relationships are
not definitive and there are limitations on the
extent to which conclusions may be drawn.
It must be stressed that these trends and confidence
limits are general and are based on data sets derived
from locations where there are different soil and
groundwater properties. While the trends are valid,
the width of the error bands indicates that at
best they can provide a range of possible magnitudes
of seepage.
Detailed quantification requires more detailed
site-specific testing to identify local trends
that can be extrapolated beyond the immediate test
sites. In some locations there is a very limited
range of readings and relying solely on developing
a correlation coefficient from these trends may
be misleading.
Conductivity has been plotted against pondage test
seepage for most IAL test sites, representing
a total of 57 ponds. Results for EM31 are shown
in the figure below (General plot of EM conductivity
against measured seepage ) There appear to be two
categories based on depth to watertable. These
concur with the understanding of the EM31 inferring
seepage based on soil type for the deeper watertable
site and directly detecting seepage for the shallow
watertable site. Details of the data and interpretation
are provided in IAL (2003).
Using the two different categories, the limited
information suggests that within each category,
there may be a trend. Confidence limits based on
each trend may allow estimation of a likely range
of seepage from a pond with a particular average
EM31 response.
Figure 4 General
plot of EM conductivity against measured seepage
Figure 5 Trend
line for watertable depths of 5-10m
The trend line seepage based on EM31 response when
the watertable depth is 5-10m is shown in figure
above (Trend line for watertable depths of 5-10m)
with surrounding 80% and 90% confidence limits.
If an average EM31 measurement represents a certain
stretch of channel, we would be 80% confident that
the seepage rate in that section would lie within
the confidence bands at that average EM31 value.
The 80% intervals are narrower than the 90% intervals,
as there is less certainty in pinning the value
down to its ‘true’ seepage rate. This
figure shows that the prediction equation is accompanied
by quite broad prediction intervals.
For example, an EM31 survey with an average response
of 30|mS/m, would suggest a seepage value of 10|mm/d.
Using the prediction bands, we would be 80% confident
that the seepage in that section would be between
5-13mm/d and 90% certain that it was between 4-14mm/d.
Although these bands are wide, the prediction equation
is a useful tool for broadly classifying
seepage rates (e.g. into low, medium and high categories).
Figure 6 presents prediction bands (80%
and 90%) for estimating seepage based on EM31 response
when the watertable is less than 2m. While the
prediction intervals are broader (approximately
20mS/m and 15mS/m for the 90% and 80% intervals
respectively) than for those for the deeper watertable
scenario, they are actually narrower relative to
the seepage range covered by the equation. This
pattern is based on limited data.
Figure 6 Trend
line for watertable less than 2m
Quantification of seepage rates can be
made by assessing the relationships between pondage
tests and average EM response over a pond length.
A preliminary approximation can be obtained using
the relationships illustrated and described above,
recognising that the accuracy of rates inferred
have limitations.
Confirmation with site-specific testing is strongly
recommended. This includes targeting pond locations
in areas inferred to be high-seepage based on geophysical
results and comparing the results. Results of distributions
based on EM results can be complemented by soil
and groundwater investigation.
Implications of
EM results for remediation
There is a general correlation between EM results
and seepage measured by pondage tests. It is therefore
generally assumed that the lowest conductivity
is an indicator of zones of highest seepage and
is a guide to channel managers of locations in
need of remediation. The relationship does not
apply in all areas investigated and before extensive
channel works programs are conducted it is highly
recommended that the correlation be locally verified
by direct measurement. Being able to assign approximate
seepage rates to conductivity response for the
region will assist in refining this process.
The figures presented above can assist in improving
the decision-making process as to the EM response
which represents a cut-off criterion for determining
which areas of channel are the highest seepage
sections in a particular area.
Drilling at sites where channel refurbishment is
proposed is often undertaken to provide greater
definition of the lateral and vertical extent of
materials for excavation. It also assists in definition
of high-seepage areas by obtaining more strata
information for comparison with EM results, and
additional information on the relationship between
conductivity and material type. In addition, detection
of the watertable during this drilling may assist
in determining locations of greatest seepage. (See
also groundwater
assessment)
Resistivity surveys
Resistivity surveys provide a depiction of resistivity
(and indirectly its inverse, conductivity) variation
through the profile. They are therefore ideally
suited to the direct method of seepage detection,
where seepage-induced changes in the watertable
are targeted. An advantage is that it is not necessary
to know the exact depth to watertable before conducting
the survey.
Resistivity can be measured using grounded (or
immersed) electrodes to impress an applied voltage
across a section of the ground. Differences in
voltage distribution can be used to calculate apparent
ground resistivity. The method depends upon good
electrode connection and can be slow if extensive
electrode preparation is necessary. Rates of acquisition
are usually only around 5|km per day depending
upon conditions. The exception to this is when
the electrodes can be immersed in water, overcoming
the need for electrode preparation. Systems can
be linked with a recording device and GPS positioning
for rapid survey procedure. In such circumstances
continuous recording can be achieved at rates of
greater than 5km/h or 40km per day. Hotchkiss
et. al. (2001) employed such a device for measuring
seepage from irrigation channels in Nebraska, USA.
Similar devices are commonly used down bore holes
to measure formation resistivity.
The advantage of resistivity systems is that a
single transmitting dipole can be used with a number
of receiving dipoles. Dipoles positioned at increasing
distances from the transmitting electrodes can
be used to calculate the depth and conductivity
relationships of the subsurface. This allows a
conductivity profile to be established, whereas
conventional frequency domain EM systems provide
only a single average conductivity for the profile.
In the IAL trials a multi-electrode array was
built in a dipole-dipole configuration. A pair
of current electrodes separated by a distance x
are followed by a series of receiver electrodes,
all separated by the same distance. The closest
receiver electrodes sample the resistivity in the
near surface (around one third to half x) and the
more distant electrodes ‘see’ deeper
into the ground. Using an array of receiver dipoles
allows for a resistivity section to be created.
In the trials the array was immersed (floating)
on the channel. Contact of electrodes in the water
and thus to the underlying ground was good and
the array could be towed at speeds of around 5-8km/h
while data was collected. This allowed data collection
of 2km sections in around 20 minutes. The set-up
used in the trials is shown in Figure 7.
Figure 7
Resistivity configuration used in the trials
Resistivity profiles were prepared
which could be used to identify seepage plumes.
An example from the trials is shown in Figure 8.
Figure 8 Resistivity
section (blue = high; red = low) under pond on
Toolondo channel
from (a) south-north and (b) north-south traverse.
Arrows show probable seepage zone under pond.
Important
geophysical survey variables
IAL trial results suggest that local conditions
such as depth to watertable, soil type, salinity
of the groundwater, channel operations and rainfall
can all potentially have some affect on the geophysical
response. Very local conditions such as high-water-use
vegetation adjacent to the channel also affect
the results. All of these need to be considered
when planning or undertaking an EM investigation,
and accounted for in the analysis of results.
The effectiveness of EM as a seepage mapping technique
can be influenced by site conditions, which can
vary with time at the same site. Repeat tests under
similar conditions showed very similar results
(IAL, 2003). However, under slightly different
conditions, while the same overall trends were
obtained, slight variations occurred.
These variables can however be complementary and
used to improve the interpretation of the EM results.
Some of the key variables are as follows:
- Survey timing - The timing of the geophysical
survey will depend on the method of seepage detection
being used. If seepage is being inferred from
soil properties then the timing of the survey
is not
critical and can be conducted whether the channel
is running or empty. However, if direct measurement
of seepage is used, the survey must be conducted
while the channel is running, and preferably
after it has been running for a least one month
(depending
on depth to watertable and vertical hydraulic
conductivity), to ensure seepage has impacted
the groundwater.
- On-channel versus on-land -
During the trials, on-channel (i.e. in a boat)
EM31 surveys:
- Did not work at one site where
the watertable was beyond the range of the
EM31 and returned similar
(reasonable) results to the on-land survey
at another site (Waranga).
- Did work at sites
with a shallow watertable.
- Were partially
successful when the watertable was located
at the edge of the depth penetration
capacity
of the EM31.
Further work is required in this area, but
the evidence collected in this investigation
suggests
on-channel geophysical surveys should only
be conducted where the geophysical technique
can
penetrate into
the watertable, and ideally target the top
of the watertable. However, there is some conflicting
evidence, as demonstrated by the trial results
summarised above. The most consistent results
were
returned on-land and this is considered the
safest
option. Results from on-land surveying at sites
that contain significant land salinisation
adjacent to the channel may be limited. Therefore,
if
the budget allows, it is recommended that both
on-land
and on-channel surveys be conducted. Resistivity
surveys can be conducted on-channel because
of their depth penetration capacity.
- Off-set
distance and location for on-land surveys -
The best off-set location for on-land
surveys
is generally adjacent to the outside toe of
the down-slope side of the channel. This applies
to both direct measurement of seepage (as immediately
adjacent the channel is the zone of greatest
seepage
impact on the watertable) and to seepage inferred
from variations in soil properties (as the
soil
type adjacent to the channel is most likely
to be representative of the soil type beneath
the
channel). For direct measurement based on seepage
impacts on the groundwater, with increased
distance from the channel, the impact on the
groundwater
is diluted. However, at most sites without
a steep gradient or high transmissivity, an
average
of
survey traverses up to 50|m on each side of
the channel was found to improve the correlation
between seepage and the geophysical survey.
Traverses
beyond
this distance are not recommended, except to
determine background conditions.
Traverses on both sides of the channel are
recommended. However, if budget is a constraint,
a traverse
on the down-slope side of the channel should
be preferred.
- Groundwater salinity - For direct
measurement of seepage impacts on groundwater
salinity,
significant changes in background groundwater
salinity along
the length of the channel need to be allowed
for in the interpretation process.
- Other variables
- Other variables relevant to interpretation
of geophysical survey results
include:
- Trees - In two surveys tree plantations
adjacent the channel appeared to interfere
with the
survey results. The postulated mechanism is that the
trees consumed seeped water and lessened the impact observed
in the geophysical survey on native groundwater.
- Rain
- Rainfall did not interfere with the
surveys conducted in these trials. Surveys
conducted after
heavy rainfall on light-to-moderate soil
types (allowing significant infiltration) should
possibly be avoided due to effects of rainfall on conductivity
and geophysical response. Surveys inferring
seepage based on shallow soil properties or direct measurement
in shallow watertable environments are
most likely to be affected.
| Related
pages |
|
Geophysical
surveys:
summary
Geophysical surveys: applicability, practical implementation,
experience from the trials, indicative costs |
|
