|
 |
Remote sensing: principle,
method
Pages in this section include:
This page provides a detailed description of the principle and method for
the remote sensing seepage identification and measurement technique.
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.
Remote sensing techniques are an efficient method
for detecting and locating seepage without interfering
with channel operations. Remote sensing refers to
any kind of data recording by a sensor which measures
energy emitted or reflected by objects located at
some distance from the sensor and includes aerial
photography and satellite imagery.
Remote sensing techniques can provide a cost-effective
means of assessing long sections of channel seepage
by evaluating soil moisture, vegetation vigour and
soil profile properties, especially in dry periods
when channels are operating.
In particular, airborne night thermal infrared imagery
can provide an indication of shallow soil saturation
resulting from lateral seepage occurring along channels,
which may be a precursor to waterlogging and soil
degradation from soil salinisation.
There are currently no documented studies of remote
sensing being used to quantify channel seepage; although
it has been used for detecting and predicting channel
seepage in New South Wales (McGowen, et al. 2001)
and in the United States (Nellis, 1982).
| Method |
 |
Remote sensing techniques for channel seepage detection
assume that channel seepage has a surface expression adjacent to
the channel. This may be detected as increased soil moisture, vegetation
vigour and water status. Techniques are limited to detecting seepage
that migrates laterally through channel banks and surfaces near
the channel toe.
Data must be of sufficient resolution to allow definition of seepage
zones. Typical seepage zones may be 10-20m wide, adjacent to
a 10-20m channel. Therefore ground resolutions of less than 10
x
10m are required. The regions most useful for channel seepage
detection include visible, reflected (near) infrared and thermal
infrared.
Remote sensing data can be evaluated in conjunction with other
spatial data such as that from EM surveys and soil survey assessments.
This can be done using GIS.
At this stage there is little evidence that spatial data analysis
is useful for quantifying seepage. However, it may be worthwhile
considering combining the results from seepage measurement such
as pondage test results, and using GIS to compare actual seepage
with areas mapped by remote sensing.
Attempts to incorporate remote sensing data into channel seepage
studies have not been successful so far. This needs careful planning
of the details of the data to be obtained, in what format and in
what range of the spectrum. An experienced specialist is needed
to provide guidance and define the program and the expected outputs
before any data gathering work.
As with all channel seepage investigations, the general approach
depends on project requirements. Remotely sensed data analysis
techniques could be used to identify potential seepage sites. This
should be followed by on-ground verification of these sites (using
say drilling and groundwater monitoring) and an evaluation of the
accuracy of the techniques implemented.
Remote sensing investigation
of channel seepage
Data source review and image acquisition
A review of available remotely sensed data, including published
literature, needs be undertaken, culminating in a comparison of
spatial and spectral resolutions and costs of acquisition and analysis.
Channel seepage analysis requires source data with high spatial
resolution (10m or less) and that it is multispectral (i.e. has
data collected from more than one distinct region of the electromagnetic
spectrum). In most cases infra-red data is the most beneficial,
as this area of the spectrum is strongly absorbed by water and
will be able to most distinctly distinguish between areas of
varying soil moisture and plant water and growth status.
Increased surface moisture and vegetation growth due to channel
seepage are likely to be evident during late summer and early autumn
when surrounding areas (apart from irrigation) would be distinctly
drier. This is likely to be the optimum data collection time. Imagery
from more than one date would be useful to remove the seasonal
variations such as irrigation or rainfall.
Remotely sensed image data sources may include:
- Digital infra-red aerial photography
- Airborne high-resolution
sensor data
- Satellite imagery
Automated image analysis and GIS techniques can be used to
identify and map potential channel seepage sites. Requirements
include image
analysis software such as ER MAPPER and ERDAS IMAGINE and
GIS software such as ArcInfo and Arcview.
Accuracy assessment and evaluation
Accuracy of remote sensing techniques should be assessed
by comparison to field surveys. Assessment measures include:
- Percentage of sites identified as potential seepage locations
that are actually seepage locations.
- Percentage of known
seepage locations identified using the techniques.
- Characteristics
of inaccurately identified sites (to make recommendations
that may reduce errors).
Spatial data can be derived from a number of sources,
including:
- Electromagnetic data (EM data)
- Soil survey assessments
- Airborne radiometric data
- Channel flow and width
- Pondage test data
Data should be combined and analysed using GIS. The accuracy
with which various types of input data can quantify
seepage at known
locations can be assessed by comparison to pondage
test data.
| Related
pages |
|
Remote sensing: summary
Remote sensing: applicability, practical implementation,
experience from the trials, indicative costs |
|
