Data from: Detecting altimetric changes in Arctic landscapes using historical aerial imagery-derived digital elevation models (hDEMs): Case study of the Black Mountain Alluvial Fan Complex, Canada
Data files
Jan 31, 2026 version files 257.43 MB
-
BMF_studyarea.zip
7.13 KB
-
BMF_transect_A.zip
4.66 KB
-
BMF_transect_B.zip
6.91 KB
-
DoD_randomerrors.zip
45.71 MB
-
DoD_totalerrors.zip
42.80 MB
-
hDEM_hillshade.zip
23.73 MB
-
hDEM_pseudoGCPs.zip
10.28 KB
-
hDEM.zip
145.15 MB
-
NAPL_A26752.zip
1.99 KB
-
README.md
8.56 KB
Abstract
In the rapidly changing Arctic, reconstructing landscapes over the last 50 years is essential to understanding impacts due to climate-induced geomorphic change. While region-wide warming became measurable in the 1980s, spatially extensive high-latitude elevation datasets extend temporally back to the 2000s. Historical aerial imagery archives provide datasets of high-resolution imagery from the mid- to late- 1900s with stereo-capability that can be harnessed to create historical digital elevation models, or hDEMs. Reconstructing a surface from the past is challenging due to a lack of ground control from that era to constrain it in space, especially at high latitudes. The main purpose of this study was to determine if an hDEM could be used to detect altimetric change in an area of poor ground control. We developed an hDEM from historical aerial imagery over the Black Mountain alluvial fan complex in NT, Canada, and used satellite imagery-derived ground control points to constrain the model in space. The resulting hDEM, when compared with the ArcticDEM, yields a vertical RMSE of 5.19 m. We were able to isolate approximately 30 m to 40 m of altimetric change from a landslide (c. 2013 - 2016) in the Black Mountain Fan catchment, supporting the supervised use of hDEMs for change detection studies.
Dataset DOI: 10.5061/dryad.mw6m90691
Description of the data and file structure
This data is referenced in "Detecting Altimetric Changes in Arctic Landscapes Using Historical Aerial Imagery-Derived Digital Elevation Models (hDEMs): Case Study of the Black Mountain Alluvial Fan Complex, Canada" in the journal Photogrammetric Engineering and Remote Sensing by Emma Menio, Hank Theiss, and Jackson Cothren.
Datasets
National Air Photo Library Aerial Film Frames
The digitized photographs used in this study are sourced from NAPL sequence 3351138 Survey Roll A26752 and have an acquisition date of 8 August 1985. The photographs were captured on 9-inch by 9-inch black and white film using a Wild RC10 camera and UAG 6007 lens with a 6-inch (152.935 mm) focal length.
Imagery download and metadata can be found on Natural Resource Canada's Earth Observation Data Management System (https://www.eodms-sgdot.nrcan-rncan.gc.ca/index-en.html). The specific frames and associated metadata used for this study can be found by searching:
- Select Data: All Collections / Aerial Photo / National Air Photo Library (NAPL)
- Select Data Options: Roll = A26752 (Optional: Frames 0003, 0004)
- Submit Search
ArcticDEM
The specific tile used in this study is 41_16_2_1_2m_reg_dem and can be found on the University of Minnesota's Polar Geospatial Center's Public HTTP Data Repository: https://data.pgc.umn.edu/elev/dem/setsm/ArcticDEM/mosaic/v3.0/2m/41_16/.
WorldView-3
The two WorldView-3 stereoimages used in this study were acquired through Maxar Technologies, Inc., capturing the field site on 7 May 2015.
- 15MAY07203919-P1BS-014287530020_01_P001.TIF
- 15MAY07204016-P1BS-014287530020_01_P001.TIF
Files and variables
Methods: Historical Digital Elevation Model (hDEM) Development
File: NAPL_A26752.zip/A26752_EO.txt
Description: Exterior orientation parameters for NAPL Roll A26752. Input for SOCET GXP. Fields include the NAPL roll and frame number [Image_File], the estimated camera center position coordinates in meters with respect to the NAD83 2011 UTM Zone 8N / NAD83 2011 Height Above Ellipsoid (HAE) [Easting, Northing, Height], and the camera orientation parameters in degrees [Omega, Phi, Kappa].
File: NAPL_A26752.zip/A26752_fiducials.txt
Description: Fiducial locations for frames 003 and 004. Input for SOCET GXP. Fields include the corner fiducial point names [ID], the projected focal length at the fiducial point in millimeters [h], the hypotenuse length between the the camera center and the fiducial point in millimeters [sqrt2/2], and the horizontal pixel coordinates in millimeters [x, y].
File: NAPL_A26752.zip/A26752.cam
Description: Camera calibration file from the calibration report. Input for SOCET GXP. Fields include focal length in millimeters, film size in millimeters, principal point of autocollimation [ppa] coordinates in millimeters, principal point of symmetry [pps] coordinates in millimeters, fiducial locations in millimeters, lens distortion pairs in degrees, and flight orientation direction.
File: hDEM_pseudoGCPs.zip
Description: Spatial distribution of pseudo-GCPs (.SHP). This shapefile contains the geographic locations of the pseudo-ground control points used to constrain the historical imagery for the generation of the historical digital elevation model. The geographic coordinate system is NAD 1983, and projected coordinate system is NAD 1983 UTM Zone 8N / NAD83 2011 HAE. All units are in meters.
Methods: DEM of Difference (DoD)
File: BMF_studyarea.zip
Description: Polygon of the Black Mountain Fan (BMF) study area used for clipping geometry (.SHP). We manually delineated the BMF area of interest as a rectangle encompassing the full alluvial fan system, from the catchment to the termination at the delta. The geographic coordinate system is NAD 1983, and projected coordinate system is NAD 1983 UTM Zone 8N / NAD83 2011 HAE. All units are in meters.
File: BMF_transect_A.zip
Description: Line feature transecting the landslide of interest (.SHP). Transect A–A' shows an inverse surface level change, where the upslope surface decreased in elevation over time and the downslope surface increased in elevation over time. The elevation residuals of the DEM of difference between the hDEM and the ArcticDEM (DoD_randomerrors.zip) show a positive trend along-track in which the elevation surface decreased and shows a negative trend along-track where the elevation surface increases, consistent with the topographic signature of a landslide. The geographic coordinate system is NAD 1983, and projected coordinate system is NAD 1983 UTM Zone 8N / NAD83 2011 HAE. All units are in meters.
File: BMF_transect_B.zip
Description: Line feature transecting an area of processing anomaly (.SHP). Transect B–B' shows the hDEM displaying a planar slope decrease, with a more constant offset from the ArcticDEM. The elevation residuals of the DEM of difference between the hDEM and the ArcticDEM (DoD_randomerrors.zip) show a positive trend along-track where the elevation surface decreased over time. The cross-section does not show any mirrored surface change or along-track inverse trend of the residuals. This indicates that this area of gross change was not from a landslide, but a processing error. The geographic coordinate system is NAD 1983, and projected coordinate system is NAD 1983 UTM Zone 8N / NAD83 2011 HAE. All units are in meters.
Results
File: hDEM.zip
Description: Raster output (.TIFF) from SOCET GXP processing and main product of this study. The Terrain/Surface Automatic Terrain Generation tool in SOCET GXP was used to generate the hDEM from the triangulated aerial imagery and pseudo-GCPs. The algorithm used to generate terrain was the NGATE (Next Generation Automatic Terrain Extraction). The digital terrain model (DTM) setup included a tiled grid (1023 x 1024 posts), 32-bit float for terrain values, with a precision output format of 32-bit float with a cutoff of 2 and autoscaled. The coordinate system for the DTM was UTM with the horizontal datum in reference to NAD83(2011) and vertical datum in reference to the ellipsoid (HAE). Both vertical and horizontal units were in meters. The saving format of the DTM was a GEOTIFF with 2m spacing in X and Y and the boundary corners and post adjustment were adjusted to the nearest outside whole grid coordinate.
File: hDEM_hillshade.zip
Description: Hillshade vizualization of the hDEM (.TIFF). The cell size is 2m x 2m. The geographic coordinate system is NAD 1983, and projected coordinate system is NAD 1983 UTM Zone 8N / NAD83 2011 HAE. All units are in meters.
File: DoD_totalerrors.zip
Description: Raster product (.TIFF) from the subtraction of the ArcticDEM from the hDEM over the study area. This raster includes the full distribution of errors resulting from the DEM of difference, including both random and systematic errors. The cell size is 2m x 2m. The geographic coordinate system is NAD 1983, and projected coordinate system is NAD 1983 UTM Zone 8N / NAD83 2011 HAE. All units are in meters.
File: DoD_randomerrors.zip
Description: Raster product (.TIFF) from the subtraction of the ArcticDEM from the hDEM over the study area with the systematic error removed. A change detection threshold of 2 sigma was used to filter out random errors (<2 sigma) from gross errors (>2 sigma), resulting in areas we define as "change" (>2 sigma). For this study, the 2-sigma threshold was +/- 10.38 meters, meaning any vertical change greater or less than 10.38m would be considered as "detecting change". The cell size is 2m x 2m. The geographic coordinate system is NAD 1983, and projected coordinate system is NAD 1983 UTM Zone 8N / NAD83 2011 HAE. All units are in meters.
Software
- The photogrammetric processing for the hDEM generation was conducted using BAE Systems SOCET GXP v.4.5.0.
- The change detection calculations were performed using ESRI’s ArcGIS Pro v.3.1.3.
