This layer contains the DEM for LiDAR data for Kaikōura captured in 2016.
- The index tiles are available as layer [Canterbury - Kaikōura LiDAR Index Tiles (2016)](https://data.linz.govt.nz/layer/110607)
- The LAS point cloud and vendor project reports are available from [OpenTopography](https://portal.opentopography.org/datasets?loc=New%20Zealand)
LiDAR was captured after the 14 November 2016 Christchurch earthquake. The dataset was generated by AAM and their subcontractors. Data management and distribution is by Toitū Te Whenua Land Information New Zealand.
- DEM: tif or asc tiles in NZTM2000 projection, tiled into a 1:1,000 tile layout
- Point cloud: las tiles in NZTM2000 projection, tiled into a 1:1,000 tile layout
Pulse density specification is at a minimum of 3.5 pulses/square metre.
Vertical Accuracy Specification is +/- 0.1m (95%) Horizontal Accuracy Specification is +/- 0.5m (95%)
Vertical datum is NZVD2016.
LiDAR was commissioned by Toitū Te Whenua Land Information New Zealand on behalf of GNS science, Environment Canterbury Regional Council, Marlborough District Council, Kaikoura District Council, Hurunui District Council, Waka Kotahi NZ Transport Agency and the Earthquake Commission for earthquake recovery purposes for Kaikōura.
The acquisition design was planned on recording a minimum 3.5 pulses per square metre (NPS) for each flightline swath.
GPS base station support was sourced from GeoNET CORS. The CORS sites KAIK, CMBL, LRR1 and SEDD were used for sensor position and orientation system processing. Ground check points were acquired by Opus International Consultants Limited. Opus surveyed ground check points at 29 locations distributed across the wider project area. CORS stations LLR1, SEDD, KAIK, MRBL and HAMN were utilised. The ITRF2008 daily coordinates made available from GNS Science and the current LINZ deformation model (20160701) were utilised for positioning.
Reduction of the ALS data proceeded without any significant problems. Laser strikes were classified to ICSM Level2 standard. The focus of the classification was to ensure that only ground points were assigned to the ground class. Following automated classification, additional manual classification editing involved removal and adding of points to the ground class. Target classification accuracy for ground points is 98%.
The classified point cloud product contains the following classes: Default, Ground, Water, Bridges DEM and Contour datasets were derived from the classified point cloud data. The DEM were created with 1m grid spacing and interpolated from TIN created from the ground classified points.
GNS Science provided AAM and Opus with daily updates of ITRF08 coordinates for the earthquake affected CORS that were used for sensor position and orientation system processing and ground check point surveying.
The point cloud data was transformed into NZTM utilising the LINZ NZGD2000 deformation model 20160701 and date 12 December 2016. Changes to the CORS coordinates for the period of survey were observed to be insignificant in terms of the LiDAR point cloud error budget.
The horizontal accuracy of the LiDAR point cloud was check by visual inspection of the test point data overlaid on the point cloud data displayed with height and intensity attributes. The point cloud data was observed to fit all against the test point data. The vertical accuracy of the LiDAR point cloud was checked by computing height difference statistics between the test point data surveyed as spot heights and a TIN created from the ground classified points. For this test the field survey points are assumed to be error-free.
Notes on Expected Accuracy:
- Accuracy estimates for terrain modeling refer to the terrain definition on clear ground. Ground definition in vegetated terrain may contain localized areas with systematic errors or outliers which fall outside this accuracy estimate. - Laser strikes have been classified into “ground” and “non-ground”, based upon algorithms tailored for major terrain/vegetation combinations existing in the project area. The definition of the ground may be less accurate in isolated pockets of dissimilar terrain/vegetation combinations.
- Test point sites: 31 - No. Points: 983 - Mean Difference: 0.00 - Std Deviation (m): 0.04 - RMS (m): 0.04
The point data classification has been manually reviewed and checked for completeness.
Features in the orthophotos were also checked against field survey data. The datasets were found to align well with one another, although measurement vectors have not yet been observed.
All product deliverables supplied in terms of NZTM map projection and NZVD2016 vertical datum.
Classification of the point cloud followed the classifications scheme below:
1- Unclassified 2 - Ground 9 - Water 10 - Bridges
Disclaimers with this dataset:
- This datasets was captured and delivered quickly post 2016 Kaikoura earthquake.
- Incorrect format of GPS time for 80 point cloud files.
- The DEM files do not exactly match LINZ 1:1k tile index layer. Some discrepancies between tiles exists throughout the dataset and include gaps, overlaps and misaligned pixels.
The deliverables to LINZ were:
1m gridded bare earth digital elevation model (DEM) Classified point cloud