Digital elevation model and orthophotographs of Greenland based on aerial photographs from 1978-1987

Abstract

Digital Elevation Models (DEMs) play a prominent role in glaciological studies for the mass balance of glaciers and ice sheets. By providing a time snapshot of glacier geometry, DEMs are crucial for most glacier evolution modelling studies, but are also important for cryospheric modelling in general. We present a historical medium-resolution DEM and orthophotographs that consistently cover the entire surroundings and margins of the Greenland Ice Sheet 1978-1987. About 3,500 aerial photographs of Greenland are combined with field surveyed geodetic ground control to produce a 25 m gridded DEM and a 2 m black-and-white digital orthophotograph. Supporting data consist of a reliability mask and a photo footprint coverage with recording dates. Through one internal and two external validation tests, this DEM shows an accuracy better than 10 m horizontally and 6 m vertically while the precision is better than 4 m. This dataset proved successful for topographical mapping and geodetic mass balance. Other uses include control and calibration of remotely sensed data such as imagery or InSAR velocity maps.

Figure 1. Footprints of aero-triangulated photographs by year of recording.

Figure 2. Ground control used in the aero-triangulation of the photographs.

Triangulation points indicate ground control with coordinates and heights assigned by GPS or Doppler-based adjustment of the terrestrial triangulation network. Ground control outside of these networks is not connected to triangulation networks.

Figure 3. Examples of the data products from the head of Nuup Kangerlua (Godthåbsfjorden).

(a) Orthophotograph, (b) hillshade DEM, and (c) reliability mask where white=measured heights, and black=interpolated or outside of boundary.

Figure 4. A posteriori mean errors from the aero-triangulation.

Mean errors for both ground control points (GCP) and tie points are shown. (a) horizontal mean errors, (b) mean errors on height. Plot of the result files. Modified after Engsager et al..

Figure 5. Histograms of the horizontal (a) and vertical (b) co-registration displacements for each 50 km×50 km grid cell show that the aero-photogrammetric DEM compilation is generally accurate to within 10 m horizontally and 6 m vertically with a precision greater than 4 m (1σ confidence level) (c).

The red bars show the fraction of displacements determined from 200 elevation difference samples or less.

Figure 6. Map of the horizontal (a) and vertical (b) components of the co-registration vectors between 50 km by 50 km sections of the aerophotogrammetric DEM compilation and ICESat laser altimetry.

(c) The RMSE of stable terrain differences after adjusting for the 3D mis-registration.

Figure 7. Magnitude and direction of the co-registration.

There is some spatial consistency of the vertical adjustments between the aerophotogrammetric DEM and ICESat, which is likely to be related to the density of the original input ground control that is used to constrain the aerotriangulation during the adjustment of the photogrammetric model.

Figure 8. Decadal elevation change calculated using the G150 DEM and IPY-SPIRIT SPOT5-HRS DEM products.

(a) Kangerlussuaq Glacier 1981–2008 and (b) Dauggaard-Jensen Glacier 1987–2014. Note the elevation difference legend has been saturated at −30 m, and elevation difference transects are plotted in the insert show actual values. Analysis of elevation change on Kangerlussuaq Glacier can be found in Khan et al. and Kjeldsen et al..