SITE CHARACTERIZATION, PORTFOLIO GRANULARITY AND ITS GEOSPATIAL DISTRIBUTION SIGNIFICANCE IN LOSS

Probabilistic loss integrates the loss exceedance probability conditioned on ground-motion intensity metric (IM) with the IM occurrence probabilities at the site. The first component requires a robust vulnerability model that accounts for the key physical attributes of the asset for which the loss is calculated. If the loss is to be computed for a portfolio, then the physical attributes of the portfolio, which are related to portfolio granularity, draw limitations on the rigorousness of portfolio vulnerability. The second component that yields the occurrence probabilities of ground motion intensities at the site(s) is related to proper source and ground motion models with reliable descriptions of the site conditions. In this paper, we study the significance of (a) portfolio accuracy in terms of building typology, (b) modeling uncertainty in vulnerability models, (c) geospatial portfolio distribution, and (d) uncertainty in soil conditions at portfolio sites for probabilistic portfolio loss assessment. We discuss the contributions of each of the above items as well as their combined effects on the TCIP (Turkish Catastrophe Insurance Pool) portfolio loss distributions in the Kahramanmaras province struck by the February 6^th, 2023, Kahramanmaras Earthquake sequence. We implement a practical loss calculation strategy to account for the consequential effects of two large earthquakes (moment magnitudes of M7.8 and M7.6) that occurred with nine hours difference on February 6^th. The performed studies have helped us to understand the importance of the site condition uncertainty, portfolio granularity, and geospatial portfolio distribution in the accuracy of loss calculations. We have also had the opportunity to understand the efficiency of our combined loss methodology with the reported Turkish Catastrophic Insurance Pool (TCIP) losses.