Impact of the July 2009 Earthquake on New Zealand’s Geodetic System
The July earthquake centred in south-west Fiordland was the biggest since the Hawkes Bay earthquake of 1931. It is also the first earthquake since the introduction of the New Zealand Geodetic Datum 2000 (NZGD2000) to have a significant impact on the geodetic system in populated areas.
Role of the PositioNZ network
The PositioNZ stations in the Otago/Southland area were critical in enabling rapid estimates of land movements to be made by GNS Science. Of particular importance for this earthquake was the PositioNZ site near Puysegur Point, installed only two years previously.
Data from five PositioNZ sites (Puysegur, Bluff, Dunedin, Mavora Lakes and Alexandra) was used in the preliminary modelling. This data was used by LINZ firstly to confirm that the datum had been impacted, and secondly to determine what is required to restore the accuracy of the datum in affected areas.
Station prediction models
The PositioNZ network is used in a number of applications, some of which require highly accurate (sub-centimetre) coordinates. The movements of up to 30cm seen at some sites mean this requirement cannot be met. The proposed solution to this problem is to use a model to predict the location of a PositioNZ station at any desired time. These models will incorporate information about the Fiordland earthquake to account for the observed movements.
Localised Deformation Model (LDM)
At a more general level, the LDM can be incorporated into the National Deformation Model currently implemented as part of NZGD2000. The LDM accounts for movement due to the earthquake, meaning that pre-earthquake and post-earthquake observations can be combined to produce consistent coordinates.
In August, LINZ and GNS Science surveyors carried out a survey of 27 marks in Fiordland National Park and Western Southland. Each existing geodetic mark was occupied for 48 hours with GPS to measure precisely how much movement had occurred. This data will form the basis for the LDM.
Impact on surveyors
The solutions outlined above are not expected to be implemented for several months. In the meantime, surveyors working on jobs covering large areas may notice that their post-earthquake observations do not agree with the pre-earthquake coordinates in the geodetic database.
The diagram above shows that for surveys over a small area (represented by the 5km observation), movements caused by the earthquake are almost identical over the area of the survey. Therefore, any post-earthquake observations should fit well with the pre-earthquake coordinates (or observations). However, for surveys covering a large area (represented by the 100km observation), movements caused by the earthquake are not uniform across the survey. In the diagram, the pre-earthquake observation is clearly different from the post-earthquake observation. In such situations, the best approach is to connect to sufficient ‘local’ geodetic marks to ensure that coordinates generated are consistent with those in the ‘local’ area.