Testing the Applicability of Correlations between Topographic Slope and VS30 for Europe
Abstract
In the past few years a series of articles have been published concerning the use of topographic slope from digital elevation models (DEMs) constructed through remote sensing (satellite imaging) to give first-order estimates of National Earthquake Hazards Reduction Program (NEHRP) site classes based on the average shear-wave velocity in the top 30 m, VS30 (Wald and Allen, 2007). We evaluate the potential applicability of these methods taking advantage of a large (706 sites) new database of measured and estimated VS30 values and their topographic slopes for locations in Europe and the Middle East. Novel statistical tests are performed to evaluate the predictive power of the procedure in this region. We evaluate the percentage of sites correctly classified/misclassified for each site class for active and stable regimes. We also analyze the marginal distributions of the input VS30 and slope values and their impact on the VS30-slope correlations and we evaluate whether the method performs better than does chance. We also consider the surface geology of sites and investigate whether differences in geology can help explain why certain sites are poorly classified by the method. Finally, we use the city of Thessaloniki, Greece, as a test case for comparison between the results of a recent microzonation and the site classes predicted by VS30-slope correlations. Our results show that the method does a better job than blind chance for all site classes in active regions, but only for class B (rock) and to a lesser extent class C (stiff soil) sites located in stable areas, although the conclusions for stable areas are based on limited data. We recommend that site classifications based on the VS30-slope correlations proposed by Wald and Allen (2007) be used only for regional or national (and not local or site-specific) first-order studies in active parts of Europe and only in the absence of other more detailed information, excluding sites inside small basins or those with special geological conditions that may affect results (e.g., flat-lying volcanic plateaus, carbonate rocks, continental glaciated terrain, or a coastal location if slope is not calculated using bathymetric data).