Your browser is no longer supported

For the best possible experience using our website we recommend you upgrade to a newer version or another browser.

Your browser appears to have cookies disabled. For the best experience of this website, please enable cookies in your browser

We'll assume we have your consent to use cookies, for example so you won't need to log in each time you visit our site.
Learn more

THE EFFECTS OF CLAY SHRINKAGE ON EMBANKMENT MOVEMENT: Implications for earthworks asset management

Report on the BGA workshop 'Maintenance of ageing earthworks' held at the Institution of Civil Engineers, London, on 16 January. Summary article by Gary Robinson, Atkins Geotechnics.

Cyclic shrinking and swelling of clay earthworks causes a major maintenance headache for UK asset managers. Millions of pounds are spent each year to realign and tamp sections of track twisted by embankment movement, a problem compounded by vegetation that increases clay shrinkage in the summer. Vegetation removal could ease the problem, however, it could also increase the risk of slope instability.

The British Geotechnical Association (BGA) brought together a number of those researching this area, to find a better balance between the level of track maintenance and vegetation management.

This article gives a flavour of the issues discussed during the afternoon's workshop and evening lecture. Much of this research is ongoing and only apparent current trends are discussed. Readers are referred to the BGA's website ( for further details.

Current research
Research presented and discussed at the workshop was:
 Pound Green and Magnolia
Road monitoring sites (Arup/University of Southampton/GeoObservations/Network Rail)
 Northolt vegetation management
trial (Atkins/GeoObservations/Network Rail)
 BIONICS (The Biological and
Engineering Impacts of Climate Change on Slopes) trial embankment project led by Newcastle University (see for collaborators)
 CLIFFS (Climate Impact
Forecasting for Slopes), part of a research and development network led by Loughborough University (see for collaborators)
 Seasonal Preparedness Project
– numerical modelling of progressive failure and hydrological modelling of pore-water pressure changes (Network Rail/Mott MacDonald)
 CRANIUM (Climate change
Risk Assessment: New Impact and Uncertainty Methods) statistical modelling research project led by Newcastle University and University of East Anglia ( for collaborators)
 Centrifuge model testing of root
stabilisation of slopes, University of Dundee
 Soil Moisture Deficit as an asset
management tool (Network Rail)
 The numerical modelling of vegetation impacts on potential for progressive failure (Mott MacDonald/London Underground.

Many of the studies followed a broadly similar approach. Piezometers (with tensiometer tips) and moisture meters were installed in vegetated embankments to monitor changes in water content and pore-water pressure, with extensometers and inclinometers recording contemporaneous vertical and horizontal movements (Figure 1).

The studies showed variations in pore-water pressure, moisture content and embankment movement are cyclic and seasonal, although the wet summer of 2007 produced less variation in some instances.

In general, high moisture content, positive pore-water pressure at shallow depth and embankment heave were recorded during the winter. Low moisture content, negative pore-water pressure (suctions) and settlement were recorded during the summer. But last year's wet summer produced some variation. Vegetation was shown to increase the amount of soil suction during the summer when plant activity is greatest.

Significantly, where vegetation was present and summer suctions increased, a remnant suction was preserved at depth during the winter (Figure 2). A number of factors were discussed as important to the preservation of this remnant suction.

Permeability was discussed as fundamental to this problem and a largely forgotten parameter in site investigation. Most railway embankments were built by the Victorians using end tipping construction methods and, as a consequence, comprise clods of stiffer clay within a matrix of softer material and potentially voids.

The permeability of this matrix is greater than that of the clods, which increases the overall permeability of the embankment. It was suggested that the permeability of the embankment could be at least one order of magnitude greater than the permeability of the clods alone.

The type of soil and vegetation were also important in controlling the water content within embankments. High plasticity soils such as London Clay have a greater swelling capacity and take up more water during wet periods.

The size, maturity and species of vegetation was shown to be important. Slopes covered by larger mature high water demand trees generated greater soil suction than grassed slopes.

Consequently, pore-water pressures in grassed slopes tended to return to a hydrostatic condition during wet winters, while slopes vegetated by mature trees retained a residual suction.

The indication from these studies was that the influence on soil moisture content and pore-water pressure extended to a greater depth for tree-covered slopes and to a shallower depth for grassed slopes composed of London Clay. For the studies presented at the workshop, an approximate guide suggests a depth of 4m to 5m for tree-covered slopes and 1.5m for grassed slopes. However, it must be noted that other sites could vary significantly from these values.

Soil Moisture Deficit

Soil Moisture Deficit (SMD) was a commonly used measurement for soil water content. SMD, available from the Met Office, is the height of water in millimetres required to return a partially saturated soil to its field capacity water content. It has been demonstrated that periods of slope instability can be correlated to times of low SMD, and periods of high shrinkage correlated with times of high SMD (Figure 3).

It was discussed that SMD data could be a useful instrument for predicting the overall incidence of future instability. However, this data is based on the MORECs grid with spacing 40km by 40km. Results from the BIONICS site indicated that SMD values varied significantly on opposing sides of an embankment; this highlighted the need for data to be as local to the site as possible.

Failure mechanisms

The shrinking and swelling movement is thought to result in progressive embankment failure. Strain softening at the toe results in deformation retrogressing back into the embankment, resulting in a deep-seated failure. Finite element modelling indicates a critical factor is an oscillation between positive and negative pore-water pressures.

The presence of vegetation increases soil suction generated in the summer and can potentially increase the range of pore-water pressure during a cycle. But where this suction is preserved during the winter as a remnant suction, studies have suggested the number of seasonal cycles to slope failure is greater. If pore-water pressure changes from positive to negative during seasonal cycles, these studies show the number of cycles to slope failure reduces, (Figure 4). The presence of vegetation could therefore be important in increasing the operational life of an embankment.

A second mechanism was also discussed and described as a "ratcheting" movement produced by the opening and closing of desiccation cracks in the down-slope direction. This is thought to be an important mechanism on slopes without trees.

It is thought likely that vegetation is a benefit in terms of slope stability, but results in increased shrink and swell issues and increased track maintenance. London Underground has previously employed a solution to the problem where trees are removed and the slope angle reduced to give a factor of safety > 1.3. However, there are significant costs with this approach and a less costly solution could be to remove sufficient vegetation to reduce the shrink and swell issue but not cause significant slope instability – a balance between the two.

Other research

Much of the research focused on variations in moisture content and pore-water pressure and how this was influenced by vegetation and increased soil strength. However, vegetation increases soil strength through mechanical binding of the soil by the root network. This strength can be expressed as an enhanced cohesion (c'R).

One study showed a technique using centrifuge modelling to allow a scaled-down model to be subject to many years of climate oscillations in a matter of days. Using this technique, different root structures were modelled to observe their effect in a failing soil. But it was shown that this effect was complicated and had many variables that were not yet fully understood.

Another study attempted to look at the shrink and swell issue using a statistical approach. Nevertheless, there were many variables that are not adequately understood, meaning this type of approach carries uncertainty.

The way forward

Derek Butcher (Network Rail) identified the balance between the level of track maintenance and the level of vegetation management as crucial and for which no definitive guidance exists. The consensus at the workshop was, given current climate change predictions, that the problem would get worse.

Maintainers urgently required guidance for both tree removal and new planting. Safety is always the priority and more work is needed to fully understand the effects of vegetation removal on stability, in terms of soil type, permeability and type of vegetation. At the same time, more work is needed to understand the effects of tree removal on track performance.

Future work needed to consider these in tandem, with an emphasis on managing vegetation to reduce the cost of track maintenance. Consultants and researchers were encouraged to approach maintainers for the funding of such projects as this is considered an important problem that needs answers.

Evening lecture
Following the workshop, William Prowie (University of Southampton) and Tony O'Brien (Mott MacDonald) delivered an interesting lecture on the subject of vegetation and slopes. Prowie summed up the current situation, stating there was still much that we do not understand, but that research is encouraging and broadening our experience.

Future work should take a theoretical approach as well as an experience-based approach. In understanding the two, we will be in a much better position to understand the problem as a whole. Once this has been done, more accurate and informed recommendations can be made to asset maintainers.

Have your say

You must sign in to make a comment

Please remember that the submission of any material is governed by our Terms and Conditions and by submitting material you confirm your agreement to these Terms and Conditions. Please note comments made online may also be published in the print edition of New Civil Engineer. Links may be included in your comments but HTML is not permitted.