Seismic hazard sensitivity assessment in the Ethiopian Rift, using an integrated approach of AHP and DInSAR methods

Seismic related natural hazard would be a major disaster as a result of the recent development and planning of new reservoirs, road networks, and the establishment of new urban settlements in previously uninhabited areas. Producing seismic hazard sensitivity maps could provide very fundamental knowledge to understand the probable occurrence of disasters in a seismic prone area. The aim of this paper was to produce seismic hazard sensitivity maps by integrating the Analytical Hierarchy Process (AHP) and the differential interferometric Synthetic Aperture Radar (DInSAR) methods. For the AHP method, six parameters have been generated and evaluated, such as distance from the fault, Lineament density, Lithology, Soil texture, Proximity of historical earthquake epicenters, and slope. The result of the sensitivity mapping showed that 8% (561.51 km2) of the study area was low-risk, 53% (3801.99 km2) medium, 35 per cent (2526.04 km2) high and 2% (142.82 km2) very high. On the other hand, SLC products from Sentinel-1A data sets were analyzed using a six-month interval master and slave image co-registration using the DInSAR method. A mean annual vertical displacement map showing max subsidence of 9.8 mm/year and uplift of 1.7 mm/year was obtained after processing. The acceptability and accuracy of the resulting seismic hazard sensitivity maps was cross-validated using the Receiver Operating Characteristics (ROC) curve. Validation results showed a strong predictive value with an AUC of 0.848. Therefore, output maps will assist policy and decision makers in identifying natural disaster-prone areas. (c) 2021 National Authority for Remote Sensing and Space Sciences. Production and hosting by Elsevier B. V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/).

Automated summary

The study aimed to produce seismic hazard sensitivity maps using AHP and DInSAR methods, revealing different levels of risk across the study area. Additionally, SLC products from Sentinel-1A data sets were analyzed to determine annual vertical displacements, showing subsidence and uplift patterns. The resulting seismic hazard sensitivity maps were validated using ROC curve, demonstrating strong predictive value with an AUC of 0.848.