130
9 Hazard Identification
Intermittent Water Analysis
Intermittent Water Analysis (IW) layers characterize water extent on the Earth’s
landscape over time. IW products are derived from a stack of over thirty years of
LANDSAT imagery, often totalling upwards of 500 historical images per location.
A sophisticated process for very accurately extracting water at the pixel level is used
to map water extent on each image individually. The individual water masks are
then stacked and analysed to count the number of water observations and frequency
at each pixel. Frequency is summarized on the full dataset, decadal subsets of the
images, and monthly subsets of the images. Finally, the frequency derived from the
full set of imagery is used to generate a vector geodatabase of the naturally occurring
lakes and ponds in the footprint.
There are three general categories of deliverable layers for each 30 m pixel:
• layers related to the frequency that standing water is detected;
• layers related to the frequency that snow or ice is detected;
• relative soil moisture.
Indeed, the space observation techniques described above seamlessly integrate
with modern risk approaches, supporting tailings dams management for the twentyfirst century for medium to large dam portfolios, where the costs of traditional monitoring techniques would be prohibitive.
Each of the categories above has additional statistical layers related to observations
clustered by month or decade, as well as average or extreme values per pixel.
• Historical InSAR Deformation Analysis: analysis covering historical observation
period (depends on extant coverage) using archived RADARSAT-1 satellite images
designed to establish historical deformation trends within the area of interest.
• Forward InSAR Deformation Monitoring Program: this monitoring can be
deployed for a selected observation period (for example, one year) using
RADARSAT-2 satellite images.
• Persistent Change Monitoring (PCM
® ) Analysis.
• Intermittent Water Analysis (IW) layers characterize water extent on the Earth’s
landscape over time.
The techniques lead to probabilistic analysis of dams and dykes with an automated
or semi-automated probability updating system, based on the dynamic link between
space observation and QRA mentioned earlier. Of course, like all systems of this
kind, caution will have to be exerted during all phases of the deployment. This is not
a universal panacea, but, to use an automotive metaphor, it provides a good set of
lights to drive through the night and a couple more instruments on the dashboard to
alert the driver in case of an emerging problem, which is an impartial, fact-driven,
emotionless aid to driving. Thus, if we cannot ensure that sudden, unforeseeable failures will never occur, we can certainly say that the results brought by this approach
will enhance planning and mitigation capabilities. It is indeed possible to develop
probabilistic updating of various types of data which may include, among others,
deformation velocity (for example, cm/year), and number of events of a certain magnitude (for example, number of events exceeding a certain magnitude per year), etc.
9 Hazard Identification
Intermittent Water Analysis
Intermittent Water Analysis (IW) layers characterize water extent on the Earth’s
landscape over time. IW products are derived from a stack of over thirty years of
LANDSAT imagery, often totalling upwards of 500 historical images per location.
A sophisticated process for very accurately extracting water at the pixel level is used
to map water extent on each image individually. The individual water masks are
then stacked and analysed to count the number of water observations and frequency
at each pixel. Frequency is summarized on the full dataset, decadal subsets of the
images, and monthly subsets of the images. Finally, the frequency derived from the
full set of imagery is used to generate a vector geodatabase of the naturally occurring
lakes and ponds in the footprint.
There are three general categories of deliverable layers for each 30 m pixel:
• layers related to the frequency that standing water is detected;
• layers related to the frequency that snow or ice is detected;
• relative soil moisture.
Indeed, the space observation techniques described above seamlessly integrate
with modern risk approaches, supporting tailings dams management for the twentyfirst century for medium to large dam portfolios, where the costs of traditional monitoring techniques would be prohibitive.
Each of the categories above has additional statistical layers related to observations
clustered by month or decade, as well as average or extreme values per pixel.
• Historical InSAR Deformation Analysis: analysis covering historical observation
period (depends on extant coverage) using archived RADARSAT-1 satellite images
designed to establish historical deformation trends within the area of interest.
• Forward InSAR Deformation Monitoring Program: this monitoring can be
deployed for a selected observation period (for example, one year) using
RADARSAT-2 satellite images.
• Persistent Change Monitoring (PCM
® ) Analysis.
• Intermittent Water Analysis (IW) layers characterize water extent on the Earth’s
landscape over time.
The techniques lead to probabilistic analysis of dams and dykes with an automated
or semi-automated probability updating system, based on the dynamic link between
space observation and QRA mentioned earlier. Of course, like all systems of this
kind, caution will have to be exerted during all phases of the deployment. This is not
a universal panacea, but, to use an automotive metaphor, it provides a good set of
lights to drive through the night and a couple more instruments on the dashboard to
alert the driver in case of an emerging problem, which is an impartial, fact-driven,
emotionless aid to driving. Thus, if we cannot ensure that sudden, unforeseeable failures will never occur, we can certainly say that the results brought by this approach
will enhance planning and mitigation capabilities. It is indeed possible to develop
probabilistic updating of various types of data which may include, among others,
deformation velocity (for example, cm/year), and number of events of a certain magnitude (for example, number of events exceeding a certain magnitude per year), etc.