DOI: 10.22531/muglajsci.1772250 ISSN: 2149-3596

Liquefaction Hazard Mapping: Statistical Approach for the Determination of the Most Effective Interpolation Method

Orkun Türe
Show PDF Cite
Liquefaction is one of the most devastating secondary effects of earthquakes and is responsible for widespread destruction during seismic events. Therefore, while detailed site-specific analysis is essential in liquefaction assessment, performing such analyses continuously over an entire region is impractical at the initial evaluation stage. Therefore, liquefaction susceptibility is typically assessed at selected locations and estimated for unsampled areas using spatial interpolation methods. However, despite the use of various interpolation methods in practice, each method produces different spatial patterns and predictions due to their methodological backgrounds, and there is still no standardization in their application. This has raised the research question of “which interpolation method yields the most accurate results for generating liquefaction susceptibility maps?”. To answer this research question, performances of four interpolation techniques (Inverse Distance Weighting, Ordinary Kriging, Natural Neighbor and Spline) were evaluated. Using 236 points (190 mapping, 46 validation), actual and interpolated LPI values were compared using R², MAE, and hierarchical cluster analysis. Overall, the Natural Neighbor method showed superior performance (MAE = 7.14, R² = 0.749), while Spline exhibited clear limitations, showing that the choice of interpolation method directly affects reliability of liquefaction hazard maps and has important influence on early-stage urban planning.

More from our Archive