PAUL GRAD
A landslide monitoring system now under development in New Zealand is probably the most detailed in the world. GNS Science, a New Zealand government owned research organisation, is developing the system in the township of Taihape in central North Island. GNS Science operates a hazard monitoring system network called GeoNet, which is funded by the Earthquake Commission (EQC).
One of GeoNet's functions is to facilitate research into the damage caused by natural disasters. GeoNet also provides insurance under the Earthquake Commission Act.
The GNS Science team leader, Chris Massey, said landslides are monitored for two reasons: to track the movement patterns of the landslide for mitigation purposes, and to provide a warning system.
He said a wide variety of techniques have been employed to investigate landslide movement patterns, including survey marks, extensometers, inclinometers, terrestrial and aerial analogue and digital photogrammetry, and synthetic aperture radar interferometry (InSAR). Terrestrial and aerial LiDAR surveys are also used.
But these techniques present serious shortcomings, he said. For instance, poor spatial or temporal resolution can make it difficult to link specific periods of landslide movement to the factors that triggered the movement. There are also limitations in the ways the data is collected, transferred, processed and displayed.
Massey said solutions to such problems have been implemented at Taihape.
The Taihape system includes a comprehensive set of data, including field mapping, sub-surface investigation and near real time monitoring of rainfall. Ground-shaking intensity, groundwater levels and surface movement are also monitored. Instruments such as piezometers and rain gauges have been installed on the landslide.
Movement monitoring is achieved with reflectors – currently 35 prisms – located in the landslide to provide a high level of spatial resolution. A robotic total survey station has been placed inside a weather-proof hut. It seeks and measures the location of each reflector at hourly intervals, thus achieving high temporal resolution.
The positions of the reflectors have been chosen on the basis of geological mapping of the landslide features (at a scale better than 1:1000), and a review of the historical data. Information about rainfall is obtained by two tippingbucket rain gauges on the toe and near the back scarp of the landslide. Groundwater levels are recorded at five minute intervals with vibrating wire piezometers in four boreholes across the landslide.
The area is close to the active Taupo Volcanic Zone, and any movement caused by earthquakes is monitored by means of a strong motion accelerograph at the Taihape Rural Hospital.
Data loggers for the piezometers and rain gauges – together with batteries and radios – are inside custom-built cabinets on three metre timber poles.
All equipment obtains power from photovoltaic cells.
The data from the survey station, rain gauges, piezometers and strong motion sensors is transferred via radio to the Taihape Town Hall, and via the internet to GNS Science buildings located near Wellington, about 250 kilometres south of Taihape.
The data is automatically processed and made available in both human and machine readable formats.
The results are presented in an interactive web-based chart that is updated at 15 minute intervals. The chart can be viewed at www.geonet.org.nz on the GeoNet website.
The Taihape landslide is a large, deep-seated translational landslide that initially occurred between 2000 and 11,000 years ago. Latest data indicates that it is still moving and that the rate of movement has started to increase. This has caused damage to services, roads and residential properties. The affected area is about 45 hectares and includes 209 households, 388 residents and a primary school.
Massey said the tertiary rocks of the Taihape area are prone to large landslides, which are closely related to the area's geological features. Compression and rapid uplift has led to gentle folding of the rocks. As a result, the area's sandstone has a dip of about seven degrees towards the southsoutheast. The area is also traversed by a series of striking faults, including the Taihape Fault.
The fault is classified as 'active' – defined as a fault that has moved in the last 125,000 years and is likely to move again, causing a large earthquake.
Recent GeoNet monitoring has allowed GNS scientists to identify two main movement patterns, said Massey. Creep movement is characterised by slow displacements, with rates typically less than 0.1mm/day over many weeks or months. The second pattern, known as surge movement, is characterised by more rapid displacements. Rates are typically greater than 2mm/day over short periods of time – days rather than weeks or months.
Increased temporal resolution of the monitoring has allowed the surge movement patterns to be related to prolonged rainfall periods, where at least 60 mm of rain has fallen within 48 hours.
Massey explained that rainfall has caused groundwater levels across the landslide to rise, reducing the strength of the material along the landslide slip plane. This allowed linking of rapid displacements in Taihape to rainfall thresholds.
'The purpose of this project is to monitor the several landslides in similar spatial and temporal resolution. Specific periods of movement can then be linked with the factors that triggered the movement,' he said.
'By comparing data from other similar landslides it will be possible to assess the movement patterns. We can determine whether relationships exist between these patterns, triggering factors and triggering thresholds. We will be able to separate the effect of periods of prolonged rainfall or ground shaking due to earthquakes. We can also study the properties of the materials forming the landslide slip plane.
'If such relationships exist, this research can be used as the scientific basis for developing alert criteria which could be used for landslide warning systems. The systems would work in a similar way to the alert criteria currently used by GNS for volcano monitoring.'
GNS is currently using data from Taihape on a project to develop these alert criteria. It is anticipated that these criteria could be applied to similar landslides around New Zealand and overseas, once site-specific details are incorporated, Massey said.
'These tools would be of significant benefit to communities living on active deep-seated landslides.'
Paul Grad is an engineering writer living in Sydney.
Issue 34; April – May 2008
