Hydrogeophysical characterisation for improved early warning of landslides

Boyd, Jimmy and Chambers, Jonathan and Binley, Andrew and Wilkinson, Paul (2024) Hydrogeophysical characterisation for improved early warning of landslides. PhD thesis, Lancaster University.

[thumbnail of 2024BoydPhD]
Text (2024BoydPhD)
2024boydphd.pdf - Published Version
Available under License Creative Commons Attribution-NonCommercial.

Download (21MB)

Abstract

Landslides are gravity driven movements of earth material that can have major economic and societal consequences. Most of these movements are driven by changes in subsurface moisture, usually resulting from rainfall and consequently are likely to become more frequent in regions where more extreme wetting events occur due to climate change. This work focuses on moisture-driven landslides in clay rich unstable slopes. Conventional methods of characterising landslides include intrusive sampling methods, such as boreholes and point samples; however, these are only sensitive to discrete portions of the subsurface. Remote sensing can be used to map the external geometry of landslides but offers little information on internal structure and condition. Geophysical methods can enhance the internal characterisation of landslides as they are spatially sensitive to subsurface properties like electrical resistivity and seismic wave speed. Furthermore, repeated geophysical surveys can reveal how these properties change with time. Here the use of direct current (DC) electrical resistivity measurements for monitoring landslides is explored, as near-surface changes in moisture tend to drive changes in electrical resistivity. The research is applied to the Hollin Hill Landslide Observatory, composed of Jurassic rocks in the north of England, as this site is representative of many vulnerable slopes in the United Kingdom. By their nature, landslides move downslope. This continuous movement poses a challenge for the long-term processing of data from permanently installed sensors entrained within these slopes. Here the goal is to process long term DC resistivity data using time-lapse Electrical Resistivity Imaging (ERI) to aid understanding seasonal moisture content dynamics and internal geometry of the Hollin Hill landslide. The topography of the slope and locations of the electrodes at the surface move with landslide movements, which introduces artefacts in conventional ERI processing results. Global positioning systems (GPS) were used to track landslide movements via permanently deployed markers (pegs) on the slope. A thin plate spline algorithm was used to interpolate changes to the slope topography and electrode locations through time, allowing for geophysical modelling to account for changes of the slope surface. These efforts culminate in a time series of geophysical models with a dynamic surface that captures both geomorphological and electrical resistivity changes at Hollin Hill, useful for illuminating landslide moisture content dynamics. Time series ERI models show low resistivities, linked to sustained high moisture contents, are present in an area of the landslide actively undergoing movement during this study. Clay rich rocks are particularly susceptible to landslides due to their low resistance to shearing at high moisture contents. Petrophysical relationships between electrical resistivity and moisture content have been established for decades, hence ERI can facilitate volumetric imaging of moisture content in the field. Such a conversion is useful as it presents geophysical properties in an engineering context and makes geophysical models more accessible for decision makers (or engineers). Electrical resistivity is sensitive to formation lithology, porosity, moisture content, pore fluid conductivity, and temperature. Clay formations have some unusual properties from a petrophysical perspective as they conduct electricity, and their porosity can increase as the clay grains swell with increasing moisture content. It is found that accounting for the swelling of clay is necessary for reliably fitting established petrophysical models. Relationships between matric potential, or negative pore pressure, and electrical resistivity are also explored as the former can be directly related to the unsaturated shear strength of a geological formation. Although, volumetric models of matric potential derived from ERI processing are not always realistic, tending towards negative or negligible matric potentials. Hydrological models of landslides can be used to understand fluid dynamics within slopes and predict crucial hydro-mechanical parameters controlling slope stability. Given the relationship between electrical resistivity and moisture content, electrical resistivity measurements can be used to calibrate hydrological parameters in these models. The soil retention parameters controlling unsaturated fluid flow are calibrated via coupled geophysical and hydrological (hydrogeophysical) modelling, in two formations at the Hollin Hill landslide. Parameter sampling is achieved using a Markov chain Monte-Carlo approach to find most likely soil retention parameters. The workflow is firstly tested against a synthetic case study with known parameters and then applied to Hollin Hill. The results are promising and show agreement with other (conventional) methods of determining these parameters, demonstrating that hydrogeophysical modelling can be used successfully for calibrating landslide models. However, there are limitations with this approach as assumptions with petrophysical relationships and modelling domain must be made. Overall ERI is a valuable tool for enhancing the understanding of landslide structures and moisture content conditions. Time-lapse processing can illuminate moisture content dynamics, and with appropriate petrophysical calibration ERI volumes can be mapped into moisture content and matric potential. Coupled hydrogeophysical approaches can be further used to constrain unsaturated fluid flow parameters in landslide models. As such, geoelectrical monitoring of landslides is a viable tool, alongside pre-existing conventional methods, for continued assessment of unstable slopes and model construction.

Item Type:
Thesis (PhD)
ID Code:
214692
Deposited By:
Deposited On:
21 Feb 2024 13:45
Refereed?:
No
Published?:
Published
Last Modified:
06 Dec 2024 00:56