Assessing Seismic Impacts and Terrain Factors Relationships Using Statistical Models: The Case of Al-Haouz Earthquake, Morocco

Abstract

This study assesses the contribution of geological, geomorphological, and earthquake-related factors to terrain susceptibility during the 8 September 2023 Al-Haouz earthquake in Morocco. Terrain susceptibility, defined as the inherent tendency of a site to amplify seismic effects due to physical conditions, represents a key component of overall terrain vulnerability if buildings or settlements are present. Key variables include lithology, fracturing, proximity to stratigraphic boundaries, elevation, slope, aspect (slope direction), terrain curvature, proximity to ridges and streams, epicentral distance, and land surface temperature (LST) anomalies. Three statistical models were applied: Frequency Ratio, Event Susceptibility based on Frequency Ratio, and Event Susceptibility based on Weight of Evidence. These models were used to generate seismic susceptibility maps and evaluate spatial patterns of predicted damage. LST anomalies showed weak correlation with observed impacts and were excluded from final analyses. All models yielded consistent spatial patterns, with Event Susceptibility based on Weight-of-Evidence achieving the highest predictive performance with Area Under Curve=83.1%, followed by Event Susceptibility based on Frequency Ratio (81.2%) and Frequency Ratio (80.4%). The most severely affected zones represented 14.98% of the study area in the Event Susceptibility based on the Weight-of-Evidence model, 17.74% in Event Susceptibility based on Frequency Ratio, and 17.25% in Frequency Ratio. These findings highlight the importance of incorporating spatial building distribution into future analyses. The resulting maps support seismic risk reduction, recovery planning, and land-use management in the High Atlas and surrounding regions.