The builders were apparently not aware of the shrink-swell effect," says Tube Lines engineering manager for civils Karen Gates of the embankments made for the Tube network in the 1920s and 1930s.
"Materials were dumped, and not compacted. The composition is complex, with clods of clay, and a high moisture content, which is, by today's standards, a poor design. Clays work well in the short term, but long term the embankment are under-drained and the strength reduces.
The embankments are prone to 'shrink-swell', the opposite of thermal expansion. While most materials expand in summer, and contract in winter, clays will absorb moisture and expand in the wet of winter, but then dry and contract in the heat of summer.
"With up to 40 years of shrink-swell, this is a long-term effect," says Gates. Drier summers and wetter winters, as predicted by climate change models, will exacerbate this.
The wetter winters could also have an effect on older cuttings. "By 2020 we will have more high intensity rainfall events. Because cutting sides are much steeper than would be built now, with more high rainfall events you would expect more flow and more spills onto the track," says Gates. Tube Lines has now re-graded almost all of it's troublesome cuttings.
But there is an upside to climate change. "With drier and hotter summers, on embankments there could be 20% less moisture, and in clays this could give greater stability."
However, when embankments shrink, they may not do so uniformly, and cause the track to undulate, giving passengers a bumpy ride.
Tube Lines suspected that trees could exacerbate the summer shrink, and hired Mott MacDonald to investigate the effects of trees on embankments. The results confirmed that the suction from certain types of tree can be very significant in summer, but their roots can also give added stability.
Mott MacDonald geotechnical engineer James Scott led the research. "When trains were steam-powered, trees were routinely cutback because they were a fire hazard. Since electrification this has not been a problem, so many of these embankments have significant trees. Some oaks are 40 years old and more than 10m tall," he says.
"Water trickles through the ballast and sits on ash, laid on top of the clay. Ash stands up well in the wet, but starts unravelling in summer. In the 1990s, lime was added every 2m, which acts like concrete, designed to stabilise the clay. If we just had the embankment without trees, it would probably not be so bad. But, the suction of the trees exposes the lime, creating small crusts, which are immovable, making the ground under the rail corrugated."
In this case, the only solution was to remove all the trees in close proximity to the embankment. Modelling suggests that the ground will 'heave', and cover the corrugations again, smoothing-out the ride.
Mott MacDonald monitored the banks over three years, to take a snapshot of seasonal variations on banks with and without trees.
"Monitoring showed movement of up to 50mm across the seasons on banks with trees, and as little as 5mm on banks without," says Scott.
The position and type of tree is also significant – trees on the slope have more of an impact than those at the foot, and some species are 'thirstier' than others. "Some trees, like oak and poplar, are very thirsty in summer. Others are have much lower, or steadier, consumption, like birch or evergreens," says Gates.
Mott MacDonald's research will be used by other clients with identical problems, like Network Rail, Metronet and British Waterways.
But trees can also be an asset. "People like the trees – they hide the line. We have found that if trees are removed, people who live nearby report more noise from the track, but this seems to be just because they can see it," says Gates.
Councils may also takeout 'plantation orders' on some trees, to protect them. The ideal solution is a targeted approach. "The principle has to be sustainable management," says Scott.