Geoinformatics-based assessment of land deformation and damage zonation for Gorkha earthquake, 2015, using SAR interferometry and ANN approach

Abstract

The Gorkha earthquake 2015 was one of the largest disastrous events that occurred in the main Himalayan thrust (MHT) region with epicenter at Gorkha region and magnitude of 7.8 which caused severe causality to life as well as property. The spatial statistics on vertical displacement and extent of damage zone is still too scarce to provide strong evidences of its hazard potential. In the present study, quantitative assessment on surface deformation has been carried out to compare land displacement of two different regions of Nepal (the Central Nepal and the Eastern Nepal), which are located at different distances from the epicenter, using InSAR technique on post- and pre-earthquake images from Sentinel 1A SLC product. The Central Nepal experiences an upliftment of 1.1 m and land subsidence of − 0.61 m, whereas for the Eastern Nepal the estimated upliftment and subsidence were 1.0 m and − 0.33 m, respectively. Further a regional earthquake-prone zone map was generated using the historical earthquake epicenter data and geographic information system (GIS) to understand the major vulnerability zones in the area. A total of 564 earthquake events were reported by USGS in Nepal region during 2000–2019, of which 476 (84.39%) were of magnitude greater than 4 on Richter scale and 376 events (66%) occurred at depth greater than 15 km. The damage assessment was done using machine learning (artificial neural network) back-propagation model in which the satellite imagery retrieved from the optical satellite Landsat 8 OLI sensor and digital elevation model was used to map slope, aspect, relief, drainage and lineament to be used as input layers to generate damage proxy map. The result obtained from ANN illustrated that despite being located comparatively at more distance from epicenter, the Eastern Nepal exhibited more damage-prone area (587 sq. km) in comparison with Central Nepal with 457sq. km damage prone in similar zone. Central Nepal evidences more damage-prone areas over compact build-up in contrast to Eastern Nepal, making greater risk potential in urban areas of Central Nepal during earthquake activity.