Using historical aerial photographs to measure earthquake deformation: Testing the effects of scan resolution

Historical aerial photographs are a valuable source of data for investigating past earthquakes, volcanic activities, ice mass balance and land changes in the pre-modern era. Scan resolution for digitising the historical archives is an important factor in photogrammetric processes, but the effects have not been quantified yet. In this paper, we use the 1978 Ms 7.4 Tabas-e-Golshan earthquake in eastern Iran as a case study to test how scan resolution affects digitisation and hence the measurements of deformation. We scanned 6 aerial photographs that were acquired in 1956 with 150, 300, 600 and 1200 dots per inch (dpi), and generated four sets of pre-earthquake digital elevation model (DEM) covering the Tabas fold. By differencing the pre-earthquake DEMs and the post earthquake DEM derived from SPOT-6 data acquired in 2013, we made four measurements of slip on the Tabas fold. We compared our measurements to the published model which was derived from matching the pre earthquake Hexagon Keyhole-9 US image and the post-earthquake SPOT-2 and SPOT-6 images, and found that the 600 dpi scans yield the most consistent result among the four different resolutions, with a misfit of 0.7 m. A different resolution might well be better for a different set of aerial photographs, depending on the quality and photographic scale of the images. We scanned 20 aerial photographs that cover the Kurit fold to the south of the Tabas fold with 600 dpi and used them to investigate the slip on that segment. Similarly, DEMs were extracted first to produce the DEM of difference (DoD). Five swath profiles taken from the DoD show clear discontinuities across two fault strands, i.e. the main Kurit fault (the western strand) and a back thrust (the eastern strand). Based on the profile measurements, we implemented Monte Carlo modelling of elastic dislocation and obtained a slip of 1.4 m and 2.2 m on the western and eastern strands respectively. By combining all the information about the earthquake and postseismic slip, we inferred a coseismic slip of 3.5 m on the detachment fault by which most of the seismic moment was released. This slip enhanced the applied load on part of the Kurit fault, which is partly responsible for the long-lasting postseismic deformation in the region.

Automated summary

Historical aerial photographs are valuable for studying past earthquakes, with scan resolution affecting digitisation and deformation measurements. In a case study of the 1978 Tabas-e-Golshan earthquake in Iran, scans at 600 dpi yielded the most consistent results, with a misfit of 0.7 m, showcasing the importance of resolution for accurate analysis.