3D Probabilistic Modelling and Uncertainty Analysis of Glacial and Post-Glacial Deposits of the City of Saguenay, Canada

Mohammad Salsabili,Alain Rouleau,Ali Saeidi,Miroslav Nastev

doi:10.3390/geosciences11050204

Mohammad Salsabili, Alain Rouleau + Show 2 more

Open Access

https://doi.org/10.3390/geosciences11050204

Copy DOI

Abstract

Knowledge of the stratigraphic architecture and geotechnical properties of surficial soil sediments is essential for geotechnical risk assessment. In the Saguenay study area, the Quaternary deposits consist of a basal till layer and heterogeneous post-glacial deposits. Considering the stratigraphic setting and soil type heterogeneity, a multistep stochastic methodology is developed for 3D geological modelling and quantification of the associated uncertainties. This methodology is adopted for regional studies and involves geostatistical interpolation and simulation methods. Empirical Bayesian kriging (EBK) is applied to generate the bedrock topography map and determine the thickness of the till sediments and their uncertainties. The locally varying mean and variance of the EBK method enable accounting for data complexity and moderate nonstationarity. Sequential indicator simulation is then performed to determine the occurrence probability of the discontinuous post-glacial sediments (clay, sand and gravel) on top of the basal till layer. The individual thickness maps of the discontinuous soil layers and uncertainties are generated in a probabilistic manner. The proposed stochastic framework is suitable for heterogeneous soil deposits characterised with complex surface and subsurface datasets.

Highlights

The soil stratigraphy and geotechnical characteristics are important factors in geotechnical risk evaluations over a region
This study aims to develop a methodology for probabilistic regional 3D modelling of soil deposits by considering soil type heterogeneity as the main source of uncertainty
The spatial interpolation of the total soil thickness that represents the depth to bedrock map is performed by using the Empirical Bayesian kriging (EBK) method in addition to triangulated irregular network (TIN)

Summary

Introduction

The soil stratigraphy and geotechnical characteristics are important factors in geotechnical risk evaluations over a region. Soil heterogeneity is attributed to two main sources: one is rooted in the lithology and the other is the inherent spatial soil variability [1]. The so-called lithological (soil type) heterogeneity is related to the substantial differences in the mineralogy, grain size and others, within a relatively uniform soil mass. This heterogeneity is qualified using descriptive terms (i.e., soil types), such as sand, clay and stiff/soft soil layers. The second source of heterogeneity is rooted in the inherent spatial soil variability, which modifies the spatial variation of soil properties due to different deposition conditions and different loading histories [1]

Objectives

Methods

Results

Conclusion