Probabilistic methods for the estimation of seismic Fa and Fv maps—application to Montreal

Abstract

In most national building codes, site classification is based on the Vs30 parameter, the average shear wave velocity for the first 30 m of soil below the surface, and defines amplification factors (e.g. Fa and Fv) to be applied to ground motions for a reference condition. Seismic microzonation mapping is usually achieved by combining information from various sources, each with varying degrees of uncertainties. A preliminary microzonation can be derived from surface geology or surface elevation maps, while a more detailed and accurate map is usually based on extensive seismic surveys. A procedure is proposed that progressively allows the integration of information from various sources and to estimate the degree of uncertainty on the microzonation. This allows planners to determine where microzonation maps require further investigations given current or future urban development plans. The proposed procedure uses conditional second moment estimation and provides the best linear unbiased estimates of Vs30 and its uncertainty. Next, these estimates are used to derive soil classification probability maps and to compute the expected values and variance of soil amplification factors Fa ad Fv to be used in probabilistic seismic risk analyses. The proposed procedure is demonstrated for the seismic microzonation of the island of Montreal.