Studies on an updating procedure for efficiently estimating spatial distribution of earthquake ground motion indices fro early damage assessment

最近，已经建立了几个地震监测网，以便利用地震信息对城市系统进行早期预警或破坏评估。然而，可以安装的地震仪数量是有限制的。在这种情况下，在考虑地震波形和地面条件的空间分布及其不确定性的同时，通过内插和外推估计地震运动，以及评估估计的精度是很重要的。本文提出了一种更新方法，当一个或多个观测数据连续可用时，有效地估计地震动指数的空间分布。推导了简单克立格法和普适克立格法的更新方程。地震动特征值服从对数正态分布。当地震运动的随机特征(均值和协方差)按位置指定时，简单克里格法是一个内插问题，而当均值的空间变化未知时，泛克里格法是这样一个问题。数值结果表明，随着观测数据数目的增加，递推得到更好的估计，同时误差方差减小。结果发现，与朴素的克里格法实现相比，该过程提供了显著的计算节省。

Recently, several earthquake monitoring networks have been established in order to use earthquake information for early warning or damage assessment of urban systems. However, there is a limit to the number of seismographs that can be installed. In such a case, it is important to estimate earthquake motion by interpolation and extrapolation while taking into account the spatial distribution of seismic wave-forms and ground conditions and its uncertainty, as well as to evaluate the accuracy of the estimation.in this paper, an updating procedure was efficiently estimating the spatial distribution of earthquake ground motion indices when a one or more observed data are sequentially available. The update equations of simple and universal kriging were derived. The characteristic values of earthquake motion follow a lognormal distribution. A simple kriging is an interpolation problem when stochastic characteristics (mean values and covariance) of earthquake motion are specified in terms of location, and a universal kriging is such a problem when the spatial change in mean values in unknown. Numerical results show that as the number of observed data increases, better estimates are recursively obtained, and at the same time the error variances reduce. It is found that this procedure offers a significant computational saving when compared to a naive kriging implementation.