Computational & Technology Resources
an online resource for computational,
engineering & technology publications
PROCEEDINGS OF THE TWELFTH INTERNATIONAL CONFERENCE ON CIVIL, STRUCTURAL AND ENVIRONMENTAL ENGINEERING COMPUTING
Edited by: B.H.V. Topping, L.F. Costa Neves and R.C. Barros
Characterising Soil Moisture as an Indicator of Transport Corridor Slope Instability using Remotely Sensed Data
A.J. Hardy, P. Miller, J. Mills and S. Barr
Civil Engineering and Geosciences, Newcastle University, United Kingdom
A.J. Hardy, P. Miller, J. Mills, S. Barr, "Characterising Soil Moisture as an Indicator of Transport Corridor Slope Instability using Remotely Sensed Data", in B.H.V. Topping, L.F. Costa Neves, R.C. Barros, (Editors), "Proceedings of the Twelfth International Conference on Civil, Structural and Environmental Engineering Computing", Civil-Comp Press, Stirlingshire, UK, Paper 242, 2009. doi:10.4203/ccp.91.242
Keywords: remote sensing, transport earthworks, slope stability, soil moisture, terrain analysis, hyperspectral analysis.
Transport earthwork structures are susceptible to instability which is caused by a build up of pore water pressure. As a result, earthwork instability has been linked to higher antecedent soil moisture conditions . Mapping soil moisture conditions would therefore be of significant value to help prioritise maintenance work. Despite this, there are currently no operational techniques for characterising soil moisture in vegetated earthworks over wide areas, such as a transport corridor. This paper explores the potential for using high spectral resolution imagery and high spatial resolution digital elevation models to characterise soil moisture on a test embankment.
Field spectral reflectance data was gathered over three grass covered plots on a test embankment; one covered, one left as a control, and one wetted periodically over a thirty-one day period. Spectral analysis techniques that have shown to depict vegetation stress including derivative stress ratios , continuum removal and Lagrangian red-edge position  were applied and compared against coincidental soil moisture measurements. These analysis techniques were then applied to spectra simulated to have a similar response to that of the CASI airborne sensor. The results indicated a reasonable correlation for both the original and CASI simulated spectra, achieving R2 values of approximately 0.6.
Terrestrial laser scanning data was collected for the test embankment and used to generate a 46cm digital elevation model (DEM). A topographic wetness index and potential solar radiation were then applied to the DEM, as well as DEMs degraded to 1m and 2m resolution. These techniques have been shown to delineate soil moisture distribution  and were compared to soil moisture measurements using regression analysis and geographically weighted regression. The results indicated the calculations could be replicated at coarser resolutions. However, poor correlations were found with soil moisture, possibly due to the homogenous nature of the test embankment's terrain. Improvements were made using the geographically weighted regression approach indicating that the relationship between the calculations and soil moisture varies over space.
This study identifies the potential for using vegetation spectral reflectance for characterising soil moisture which would be valuable for slope stability analysis in transport corridors. Predictions made using DEM terrain analysis were less successful but show the potential for using geographically weighted regression.
purchase the full-text of this paper (price £20)