Keeping the water in
We look at how the Canal and River Trust’s engineers try to ensure that 200-year-old embankments, built without today’s engineering theory, carry on doing their job – and why, just occasionally, they don’t…
They knew how to build things in those days”, exclaims George Ballinger, head of engineering at the Canal & River Trust, as we stand next to the massive bulk of the embankment carrying the Macclesfield Canal high above the town of Bollington.
He’s being ironic. When it came to understanding the science and engineering of how to build embankments that would stay up, it appears that they knew pretty damn close to zero. Even as late as the 1820s, when this latecomer of the canal era was under construction, building earthworks seems at times to still have been very much a case of trial and error. If it stayed up, fine. If it didn’t, they rebuilt it. And at times they didn’t even seem to learn from their errors. Take this extract from a letter from engineer William Crosley to the canal company…
“A slip has taken place in this embankment, which has not only retarded the work, but has injured the masonry of the Culvert…”
Like any decent engineer, he’s keen to find out the reason …
“The cause appears to be this:- Finding that the embankment would be chiefly composed of stone, I concluded it would stand upon a much narrower base than was intended by the original Contract, and that … a saving of about £1,700 might be effected”
…which sounds like cutting corners. But…
“…in the event of its not answering, I considered that the original plan might be resorted to and the work completed at a sum not exceeding the original Contract.”
…so at least he had a fallback plan. And when, in due course, it collapsed…
“…a stop was put to it for about six weeks, in order that the effect produced by time and the heavy rains then falling, might be seen…”
…which once again seems prudent. But what happened in the end?
“The work was resumed several weeks ago; and, as no further slip has taken place, I am of opinion that the embankment will stand without any extension of the base.”
That’s right – they rebuilt it to the same narrower, more steep-sided profile that had just failed, in the hope that it would be ‘second time lucky’. And it was: that’s the embankment we’re standing next to today.
The result, all over the network, was earthworks built to an angle of slope that would (and does) make today’s civil engineers wince. Look at a typical motorway cutting or embankment: the sides slope at maybe 1 to 3 (the ratio of the vertical distance to the horizontal). On canals and early railways 1 to 2 was more usual, and 1 to 1 (45 degrees, as here at Bollington) wasn’t unknown. Sometimes they’ve lasted well, despite breaking all the rules of the soil science that has been developed since then; in other places there’s been a history of collapses and breaches – including here on the Macclesfield Canal, where there was a major breach close to Bollington in 1912.
George Ballinger isn’t being ironic, though, when he describes Bollington Embankment as “a beautiful big structure; a real work of art”. And that’s the problem. Unlike today’s highway engineers, whose aim is to build for minimum maintenance for the design life of a structure, CRT’s team has been entrusted (“duty-bound and privileged” are George’s words) with looking after our waterways heritage – even if it was built to standards which would be unacceptable today. So (even if it weren’t for the fact that some of the land around it is built on) rebuilding Bollington to a more reliable slope angle just isn’t an option.
Having said that, Bollington does have something going for it. It’s founded on rock and built from stone, albeit uncompacted rough stone, piled up to the angle that it will rest at – which is about 45 degrees. We can see that it’s settled a certain amount over the years: the slopes are no longer flush with the walls of the aqueduct which takes a road through it. But at least this ‘self compaction’ has finished. Other materials will continue to settle for much longer: clay is slower; peat even worse. Without even starting to think about the effects of any wear-and-tear, problems can develop. Voids can open up; subsidence can occur.
Now add to that the effects of 200 years of boat movements; landowners not always respecting the boundaries and sometimes digging away at the ‘toe’ of the embankment; neglect of drainage ditches at times of poor maintenance; moored craft (banging pins into an embankment isn’t always the best for it!) and so on. Even if it was originally stable, problems can develop. And it all gets a lot trickier if the canal channel at the top of it starts leaking, because water changes all the properties of the material. So keeping the channel walls intact is crucial.
So you can maintain the channel, and you can do things like putting in mooring rings to stop boaters puncturing the towpath. But you can’t do anything about the slope. How do you attempt to stop disastrous breaches from happening?
A good start is to look at how most of them occur, and (subject to the proviso that it’s not easy to figure out the cause when a major burst has literally washed away the evidence), CRT’s engineers reckon that the biggest single proportion, 40 percent of all breaches, relate to failed culverts under the canal. And sure enough, when we look on the east side of Bollington there’s an overspill channel leading to a small culvert through the base which merges with a larger culvert carrying the River Bollin. Make sure they’re well maintained and you’ve reduced the risk.
That leaves the other 60 percent, which split into several different categories: base erosion by rivers and overtopping by floods (both of which were implicated in the recent Croxton burst) are actually relatively unusual causes (3 percent each); leakage, sluice or weir failure, and third party works are much more common. But how do you spot the problem areas for these and then prioritise and plan any work to stop them collapsing, when you have 650 major embankments (defined as either more than 6m high, or more than 3m over a length of 200m) and another 2,000 minor ones?
In line with other types of structure, the approach by British Waterways (continued by CRT) has been to set up and then refine an inspection regime – first checking every embankment, then moving to individual 100 metre lengths, grading both the risk (the likelihood of failure) and the consequences (the possible damage to property and life if it did) – and prioritising work on those that score highest on a combination of the two.
But what are they looking for when it comes to the regular inspections to assess the likelihood of failure?
There are a whole range of ‘usual suspects’: bank erosion; signs of settlement of the crest or slope (such as irregularities or breaks); indicators of instability (for example leaning trees, ground cracking); water seepage; damage by badgers, foxes or other animals; problems with retaining walls (including bulging masonry); excavation by users of adjacent land; vegetation damage.
The last named of these leads me to ask George whether roots actually help to bind aqueducts together and make them stronger, or create holes in them that weaken them – I’ve heard both claimed. The answer is that it depends what type and where they are. Tree growth at the bottom of the embankment can add weight and suck out moisture, aiding stability; further up it can have the opposite effect – especially if exposed to high winds – but grass and small vegetation can help to hold the ground surface together.
That’s a simplified version: different tree species have different root structures; the material that the embankment is built from has an effect. But by and large it’s a good idea to allow some vegetation up the banks, but strip it back before it gets too big (every three years at Bollington).
Other than that, tackling the various problems doesn’t seem hard to understand, or to prioritise from limited funding. Raise eroded banks; track down and seal leaks; excavate, replace and compact damaged ground; eradicate burrowing animals; repair walls; deal firmly with adjacent landowners. So has a methodical approach to this reduced the number of burst banks?
George believes it has. Back in the 1970s and early 1980s he feels BW was ‘fire-fighting’, dealing with a number of major collapses (Mon & Brec, Peak Forest, a long-running series on the Llangollen), and with no data to go on that might give any idea of where it would happen next. But although they may be less frequent today, breaches still happen, on some of the same canals: the Llangollen in 2004; at Gilwern on the Mon & Brec in 2007, the Stourbridge around the same time, and the Leek Arm two years later.
Why aren’t they being spotted?
One issue is that leakage and culvert failure can be surprisingly hard to spot. Many culverts under canals are no longer in use, with streams dried up or diverted, the portals buried, and no mention in such old canal company records as have survived. Leaks are easy to spot if water’s pouring out of an embankment base or there’s always a damp patch in the adjacent field; less so if it emerges in a river or some distance away on somebody else’s land. Short of digging out (or perhaps ultrasound testing) every embankment to look for problems that almost certainly don’t exist in most cases, is there anything that can be done?
The answer, at least in some cases, is that yes there is – and surprisingly it involves checking the ground of the canal banks not for dampness but for temperature.
Anyone who’s been in a cave (or for that matter a canal tunnel) will know that once you get a certain distance underground the temperature hardly varies at all with the weather. That’s because the ground doesn’t conduct heat very well.
If you drill a deep narrow hole in a canal bank, lower a thermometer down it, and check it every day at various depths, you will find that at (say) 1 metre below ground level it will follow the temperature of the canal water moderately closely, but lagging behind it a little and not reaching the same heights or depths. At 2m below ground, it follows it rather less closely; the deeper you go, the less variation there is, until by 5 or 6m there’s barely any change from day to day, and it will take a sustained warm or cold spell to make a significant difference.
But what if water’s leaking down into the ground? See the graph above. The line indicating the temperature 1 metre down follows the canal water temperature but lags behind; the 3m, 4m and 5m plots follow it less closely until the 6m one barely changes by half a degree in three weeks. But what’s happening with the 2m line? It’s following the water temperature much more closely than either the 1m or 3m line. That was enough to alert CRT to possible trouble; sure enough, there was a leak. Water was seeping through the canal bed, passing under the bank about 2m below ground, and disappearing off to emerge some distance away. Thanks to the temperature sensors a possible potential breach was averted that otherwise might have been very hard to spot.
It’s not a method that the engineers can afford to use everywhere, but certainly on a vulnerable waterway like the M&B with 16 miles of hillside channel it’s justifiable. And it’s just one of a number of methods being developed.
So armed with these techniques, can CRT be confident of no more major breaches? They may not have “known how to build things” in those days, but do we “know how to maintain them” today? Well, even in the case of the Mon & Brec, George is cautious: while he believes that those places with very serious consequences (such as housing or schools downhill from the canal) have been secured, he warns that “we’re not saying it won’t breach again and flood (say) a field”.
Quite rightly, too: between our visit to Bollington and my writing up this piece, that’s exactly what happened at Dutton. Read our Questions & Answers for an idea of how, despite all these precautions and advances, a major breach still happened.
