浅層地下水流動モデルの不確定要因とその時・空間伝播特性

Abstract

There are many studies to estimate parameters in numerical models of groundwater flow.However, before parameter estimation, numerical analysis of uncertainty propagation ofparameters is required in order to know which parameter to estimate, and how to estimateit.In this paper, an approximate solution to the distribution of dependent random variablesis applied to numerical models of groundwater. First-order second-moment method ofmultidimensional Tayor-series expansion, that is, is presented in matrix form for such analysisso as to calculate the mean and the covariance of predicted groundwater heads due to manyuncertainty parameters.This algorithm is applied to an assumed groundwater basin and its results are presentedin graphic formula, and are compared with those calculated by Monte Carlo simulation.