Integrated seismic risk assessment using multi-criteria decision making, statistical, and machine learning approaches: a case study of Denizli, Türkiye

Abstract

Denizli is a province situated in T & uuml;rkiye's Western Anatolian Extensional Region, influenced by active fault lines and historically subjected to numerous destructive earthquakes. The area contains active seismic structures such as the Pamukkale, Buldan, and Honaz faults and, as both an industrial hub and a host to significant historical and touristic sites, seismic risk analyses here are of vital importance. This study presents a comprehensive seismic risk analysis for Denizli Province based on the integration of the Analytic Hierarchy Process (AHP), Random Forest, and Frequency Ratio methods. The objective is to achieve a more reliable and detailed risk assessment by combining the strengths of these different analytical approaches. Geological, topographic, and seismic data were used to generate the region's seismic susceptibility maps, and risk levels were classified into five categories (very low, low, moderate, high, very high). The integrated risk map shows that 29.3% of Denizli's total area falls into the very low category, 27.1% into low, 32.6% into moderate, 9.6% into high, and 1.5% into very high. In terms of population distribution, 48.21% of the province's residents live within the very high and high risk zones. Densely populated districts such as Pamukkale, Merkezefendi and districts prone to numerous earthquake occurences such as Ac & imath;payam, were specifically identified as being under elevated seismic threat. This work demonstrates that processing the parameters with suitible analytical methods, then integrating these results enhances the accuracy of seismic risk assessments and provides valuable insights for developing regional risk-mitigation strategies. The findings underscore the urgency of implementing effective earthquake risk management policies and urban planning strategies for Denizli Province.