Provisional geodetic survey control coordinates (superseded)

Note: These provisional coordinates have been superseded by final post-earthquake coordinates available in the Geodetic Database. They remain on this page temporarily for archival reasons as they have been used in a large number of post-earthquake surveys.

Both summary and detailed datasets can be downloaded.

Simple Dataset

The simple dataset contains only data reflecting the position of a mark after all significant earthquakes which have affected it, including the most recent significant earthquake on 23 December 2011. It contains only ellipsoidal heights and is most useful to those who are only interested in current horizontal coordinates, or who are comfortable transforming between height systems.

This dataset is also available for download in a wide variety of formats from the LINZ Data Service.

Comprehensive Dataset

The comprehensive dataset contains data collected after each of the significant earthquakes and includes heights in Lyttelton Vertical Datum 1937 and (where applicable) Christchurch Drainage Datum. It is most useful to those who are interested in heights, or earthquake-related land movements.

This dataset is also available for download in a wide variety of formats from the LINZ Data Service.

Precise Levelling Height Dataset

The precise levelling height dataset contains normal-orthometric heights in terms of Lyttelton Vertical Datum 1937, updated for all significant earthquakes in Canterbury.

This dataset is also available for download in a wide variety of formats from the LINZ Data Service.

Geodetic Data

Updated coordinates are now available in Landonline and the Geodetic Database.

Specific queries should be directed to or 0800 ONLINE (0800 665463).

Why are these coordinates still provisional?
Authoritative coordinates have now been published in Landonline and the Geodetic Database. The provisional coordinates continue to be made available for historic purposes, as they were used as a source of coordinates for many surveys in the years immediately following the earthquakes. They should not be used for new survey work.
What origin of heights was used for the precise levelled height?
Four stations were held fixed in the precise levelling adjustment: AC9B, AFAQ, AG32 and B853. Both GNSS observations and earthquake models indicate that vertical movements at these marks are insignificant. These four marks are at the far extents of the network.
How accurate is the precise levelling data?
The misclose in millimetres between fore and back runs had to be better than 5*sqrt(k), where k is the distance of the run in kilometers. Based on adjustment results, the height error at each mark is less than 15mm at the 95% confidence level.
Why do the precise levelled heights differ from the heights calculated from GNSS data and the geoid model?
These differences may be up to 0.1m, even though both heights are in terms of the same datum. The differences are due to errors present in the GNSS data, precise levelling data, geoid model and vertical datum offsets used with the geoid model. For this reason it is important that precise levelling heights are not mixed with heights calculated using the geoid model.
What is the difference between the simple and comprehensive csv files?
The two files contain identical data.

However, for the convenience of users who are only interested in coordinates which reflect the current location of marks “canterbury_eq_linz_coords_simple_v0.6.csv” simply contains the first six columns from the comprehensive spreadsheet “canterbury_eq_linz_coords_comprehensive_v0.6.csv”.
Why do some marks not have an updated coordinate?
Some of the marks have not been resurveyed since the last major earthquake to significantly affect them. This may be because the mark was inaccessible, destroyed, or otherwise unsuitable for use as a post-earthquake control mark. In these cases, the mark has been replaced by new geodetic mark in the general vicinity.
Why do some marks have entries in the simple csv file, even though the comprehensive csv file shows that the mark has not been surveyed since the 23 December 2011 earthquake?
The file “canterbury_eq_linz_coords_simple_v0.6.csv” contains coordinates which reflect the current position of the mark (just after the 23 December 2011 earthquake). Therefore, if modelling indicates that a mark was not significantly affected by the 23 December 2011 earthquake, it may still have a coordinate in this file. The file “canterbury_eq_linz_coords_comprehensive_v0.6.csv” contains details of which earthquakes affected a particular mark.
Why are there three different heights in the detailed csv file?
There are three types of heights provided:
  • NZGD2000 ellipsoidal height
  • Lyttelton Vertical Datum 1937 orthometric height
  • Christchurch Drainage Datum orthometric height
Multiple height systems were in use before the Canterbury earthquakes. Ellipsoidal heights are used in the official geodetic datum, NZGD2000. Lyttelton Vertical Datum 1937 heights approximate mean sea level. They are the most widely used heights over the region as a whole and are suitable for engineering, topographic and cadastral purposes. Christchurch Drainage Datum heights are used by the Christchurch City Council.
Can I get coordinates in another coordinate system, such as a meridional circuit?
The best way to do this is using the coordinate converter on the LINZ website.
Why are the pre-earthquake coordinates in the csv and kmz files different from those in Landonline and the Geodetic Database?
Most of the LINZ geodetic surveys in Canterbury were carried out in the late 1990s, at least ten years before the September 2010 earthquake. Immediately after this earthquake, LINZ received a substantial amount of GNSS data from private surveyors and local authorities that had been collected during the previous two years meaning it was possible to recalculate more accurate pre-earthquake coordinates. During this process some errors in the existing data were found and corrected. Furthermore, a more accurate model of the regular deformation was used to reduce the impact of errors in the official deformation model.

Landonline and the Geodetic Database have not been updated with these more accurate pre-earthquake coordinates as the date of the update would make it appear that the coordinates had been generated from post-earthquake surveys.
Why are the provisional coordinates in this version of the csv file slightly different from previous versions, even for the same coordinate set?
As new survey data is received, it is added to a combined least squares adjustment for the particular coordinate set. This new data may cause coordinates to change slightly, even for marks that are not directly included as new surveys. Usually this change is at the millimetre level. With each successive version of the csv file, the focus is on providing the best coordinates available at that time. However, we recognise that small coordinate changes are a nuisance for many users, which is one reason we have not been updating Landonline and the Geodetic Database after each survey.
When were the Landonline and the Geodetic Database updated?
Updates to Landonline and the Geodetic Database were completed on 14-15 December 2013.
What were the post-earthquake coordinates be used for?
Some examples of how these coordinates were used:
  • As control for survey observations made after one of the earthquakes, enabling further post-earthquake coordinates to be generated
  • As control for datasets such as LiDAR and InSAR which are used to assess land movements (noting though that most of these marks are below ground level)
  • To assess the nature of earthquake-related movements for cadastral survey purposes. For example, if a number of marks surrounding a cadastral survey had all moved by a similar amount, this could support an assessment of deep-seated movement.
How were these marks and provisional coordinates be used in cadastral surveys?
In a cadastral survey, the connections to control marks are more important than the coordinates that are on that control mark at the time of survey. Even setting aside the impacts of earthquakes, New Zealand is constantly moving, often in ways that are difficult to accurately model. A connection to a control mark means that the coordinates of the survey can be updated as required in the future.

The marks in the csv files were often connected to as part of cadastral surveys, for the benefit of other surveyors working in the area after Landonline had been updated.
The provisional coordinates were also often the best way of obtaining an NZGD2000 orientation when not using GNSS.
Are the movements at a mark reflective of movements in the surrounding land?
Not necessarily. Deformation caused by phenomena such as liquefaction and lateral spreading can be very highly localised. Land only a few metres away could have moved quite differently.

However, where movements of a large number of marks are consistent over an area of interest, this can indicate widespread, deep-seated movement.
Does the estimate of whether a mark has been significantly affected by a particular earthquake in the detailed csv file include consideration of localised deformation, such as liquefaction?
No, it only incorporates deep-seated movements. It is possible that a mark which has not experienced deep-seated movements has been subject to some kind of localised deformation.

It is also possible that the mark has been affected by deformation totally unrelated to the earthquake, such as localised subsidence due to groundwater extraction.
How should these coordinates be referenced?
There is no official referencing system for these coordinates. The important thing is that a future user is clear which coordinates were used for a particular piece of work and which coordinate set they come from.

Given the number of earthquakes, and therefore the number of separate coordinate sets, it is important that any coordinate is referenced to both its coordinate set, and the source (including the version) of that coordinate set. While this is rather cumbersome, it is the best way to avoid confusion in the future.

Example 1: If using the most up-to-date coordinates currently available, an appropriate reference would be:

“Coordinate origin from canterbury_eq_linz_coords_simple_v0.6.csv”

Example 2: If assessing the coordinate change at a mark after the 23 December 2011 earthquake, an appropriate reference would be:

“Pre-23 December 2011 coordinates sourced from postjun coordinate set in canterbury_eq_linz_coords_comprehensive_v0.6.csv. Post-23 December 2011 coordinates sourced from postdec coordinate set in canterbury_eq_linz_coords_comprehensive_v0.6.csv.”
My question is not covered here. How can I get some help?
The best way to get support is to send an email to us directly at