We had five summers in our house before we tackled the pond properly, including a trial dig in the third summer and autumn when we desilted a section of it. There was time for serious trialling and surveying, which are described in a separate post. It was important to estimate the amount of mud, so we could decide where it could be dumped.
We worked out that we could just about dispose of all the silt in the garden, most of it by shrinking the pond and using the silt to build out the banks, reducing the pond area of about 350 square metres by perhaps 100 square metres. The rest could be piled up to rot for a year and then spread over flower beds, the vegetable patch and parts of the garden. In shrinking the pond we decided that the new edge would go along a contour where the mud was about one spade, or 25 cm, deep. This line was marked out with sticks. The result of doing this was to narrow the pond in a way that made it look more like a moat.
Having estimated the scale of the job, we decided to do as much as possible of the work by hand, for several reasons: we wanted the exercise (seriously); the slow pace would make it much easier to determine the new outline of the pond and build up a wall to contain the mud as we went along, rather than have the silt dumped by an excavator in great slippery piles over a couple of days; we were concerned that the clay bottom might be fragile under the weight of a big machine, and that it would be difficult to repair; we wanted to examine the silt and the clay bottom as we went along, to check for interesting artefacts (sadly, there were hardly any); and if there was any other evidence of the pond’s history, we wanted to preserve it.
Our resolve was shaken when one of the contractors told us that to clear the pond with shovels and barrows would take 10 men 2 weeks, or 100 man days. This proved to be an exaggeration. When we got down to it, about two thirds of the mud volume and three quarters of the area was cleared by one person in about 80 hours of actual shovelling time. Pro rata, hand digging proved to be far faster than had been claimed. Our job was speeded up by the decision to dump most of the waste around the edge of the pond. Well over half could be thrown or walked a short distance and dumped, rather than loaded into wheelbarrows, which slows progress considerably.
We did eventually hire some mechanical help. This was after we had decided to drop an earlier plan to spread the digging over two years, which would have involved building a temporary sandbag barrier at the point where the first season’s dig ended, to prevent the remaining mud slipping back into the excavated areas (see planning, trialling and surveying). But by mid-October it appeared that hand digging the whole of the rest of the pond would not be finished until end November at the earliest, and winter rains were approaching – or so we thought at the time, though it turned out to be the dryest autumn anyone locally could remember.
We had bought a submersible electric pump with a long lead and hose from Screwfix to clear the water that accumulated where we had dug. (We emptied it onto a grassy area, which was by then so dry that it could absorb hundreds of gallons an hour without forming a pool). Our trial dig two years earlier had shown that once the rain started the mud quickly softened again and became difficult to handle, and especially hard to control when it was piled up, so a pump probably would not be enough to keep the mud firm enough to handle easily. There was also a lot of finishing work to do after the basic digging, and it seemed important to have the pond ready to fill again rather sooner than Christmas. We were also concerned at the effect of deep-freeze weather, like December 2010, because solidly frozen clay seems to become friable when it thaws, which might damage the water retaining properties.
So we called in a mechanical digger and expert driver from a neighbouring village for half a day.
The pond has good access, a feature of its central role in the farm. The contractor suggested he bring his biggest machine, a 13 ton monster, which has an arm long enough to dump silt in precise positions over a wide area without moving its tracks much. We asked that the digger’s tracks should stay on the hard area of the bottom that had been used by farm carts.
The result: the rest of the mud was shifted by an expert operator in 3 straight hours, with the rest of the hire time spent getting the digger in and out of the pond and cleaning it.
A lot of the mechanically-dug silt was put around the edge, but we did not wish to shrink the pond further. So we also accumulated a pile of silt about a metre deep in an area about 8 metres by 8 metres on a patch of rough grass in the garden. This is being left to rot down over the winter and it will then be shifted round the garden and distributed in thin layers. There will probably be a couple of hundred wheelbarrow loads to move.
The number of man hours we worked to finish the project totalled 165, which is 28 man days if you assume 6 hours hard graft during an 8 hour day (and the other two hours for meals, tea breaks and leaning on the shovel to contemplate the excavations or the moorhens, or to talk to the neighbours).
The overall cost of DIY in labour and materials, spread over three months, was £1,967. All but £400, for hire of the digger and driver, was for materials: timber, cement, gravel, hessian sandbags, geotextiles and tools (see the post on materials and costs). The finishing work, including the construction of containments to stop the mud slipping back into the pond, took about as long as the hand digging.
If we had a garden big enough to dump the silt out away from the pond instead of putting most of it along the perimeter, we could have relied on the existing banks and avoided purchasing timber, sand, cement and hessian bags. The total cost would then have been less than £800. Digging would have taken longer, because there would have been further to move the material.
Some might say we were mad to dig so much of it by hand (though there were good reasons for doing it, explained above.) It seems likely that the whole pond could have been cleared by the digger in 2 or 3 working days at £500 a day. This was a very large excavator, and with more restricted access a smaller machine would be needed, which would take longer. We estimate that if the whole pond had been dug mechanically with the big machine it would have taken the cost to £2,600 – £3,000. Again, if we had somewhere to put the silt in the garden away from the pond, we could have saved £1,200 on materials.
If you want to do it all mechanically, don’t go to an expensive specialist pond contractor but find a skilled local builder who can excavate under your supervision at a fraction of the cost. One caveat is that it may not be the right solution if your silt is very liquid or if you can’t fully pump out the pond.
There are a number of differentways to contain silt once it has been dug out, explained in the planning, trialling and surveying post. When under the water in the spring, our pond mud was almost liquid, so a bamboo pole dropped vertically (as an experiment) went well over a metre down at the deep end under its own weight alone, which was why we were so scared of falling in there. Once dug out, mud of this consistency needs to be contained in extensive ditches or held back by geotextile barriers to drain. This is time consuming and potentially expensive.
But the key to our job was that in 2011 the pond dried early, helped by some pumping in the early stages. From August, the ooze consolidated rapidly, and within two or three weeks much of it began to behave like damp garden soil, and could be stacked more than a metre deep without running quickly away.
The stacked mud still needed a timber containment to stop it creeping along slowly over the days. First, we used scrap timber we had saved from building work on the house to build a rough and ready revetement, or timber retaining wall, pinned back by softwood stakes driven into the bottom, to hold each section as the mud was put behind it while the dig progressed. The fresh mud, while moving slowly, exerted tremendous force. This first defence collapsed in one 5 metre long section of the 80 metre perimeter, and had to be remade, and the mud shovelled back again. But a month later that mud pile had dried to the point where it was stiff enough to walk on.
The biggest materials cost was new treated timber to build the main revetement in front of the scrap timber barrier, using treated softwood planks laid three, four and five high on top of each other (depending how deep the pond was at each point). This was the timber which would be visible around the edge when the pond was finished. The timbers were to be supported by round softwood posts driven into the pond bottom.
We knew the maximum depth of the water above the mud (assumed to be flat) in the spring, which was about 40 cm. So the timber needed to be 40+25 cm high, ie 65 cm. We settled for 60cm of timber, which is four 15 cm wide planks on edge. The shape of the pond meant that in one short stretch five planks were needed, and in several other short stretches only three were needed. We could then estimate the length of planking to order from the length of the sides of the pond.
We included two concrete sections, one 4.5 metres long and the other 6 metres. They were built from hessian sandbags filled with a dry mix of 5 sand, 1 gravel, 1 cement. The bags will biodegrade, and waterside plants will be used to camouflage the concrete. The walls were in two places where the bottom went down sharply. This structure followed the contour better. We placed the bags in position as soon as they had been filled, because they hardened overnight. We did not bother with foundations, but did level the hard clay bottom before placing the bags. We used a single line of sandbags laid lengthwise.
One of the concrete sandbag walls was built to hold back the mud and roots of a reed bed which we wanted to preserve at one end of the pond. We also used gravel and permeable geotextiles behind the timbers along the short section of the edge where groundwater enters the pond, to help that flow.
In the shallowest part of the pond where the new work would be most visible, we used felled timber and rough sawn pine sleepers as edging. The felled timber we had available was willow, which is far from ideal, and tends to sprout, but it was that or ash, which was more useful as firewood.
Finally, we used 3 tons of gravel to firm up the bottom of the pond by making paths, which will be used for maintenance, wearing waders of course. This was essentially what the farmers had done, by dumping tons of gravel and stone in the middle to prevent their water carts sinking into the bottom clay.
In mid-November, three months after we started, the pond had been completely cleared and the timber and sandbag work finished, together with a gravel path all the way round the edge and across the middle. The path is to be underwater when the pond fills, to make maintenance easier.
There was much finishing work to do.
In March, the posts were tapped with a sledgehammer to check that ice, rising water and a softening bottom had not loosened them. The tops were trimmed off at the same time.
Some of the top planks loosened as the water rose, and two or three floated loose. They were put back and nailed. But nails proved unnecessary, because it was easier to secure loose planks simply by ramming down the mud immediately behind them and raking mud from higher up into any voids. The friction of the rammed mud was enough to keep the planks down. To be sure, we rammed the mud down behind all the planks, not just the loose ones.
We also moved many hundreds of barrow loads of mud from the large pile of left-over spoil which had had to be stacked in the garden (see picture above). This pile proved very useful for adjusting the profile of the pond banks, because in some places the new mud settled further than in others, mainly – we think – because of varying amounts of clay in the mud itself. Mud into which a lot of clay had fallen from the pond sides consolidated very quickly, while soft mud from the deep water continued to shrink for longer, a process that is likely to go on for several years, judging by the results of the 2009 trial.
Even after this fine tuning there were still tens of tons of spoil left. This was taken to the far end of the garden and deposited to form a low terrace, which will be ready for planting next year.
These landscaping jobs were spread out over six months, so we could judge how the spoil was settling. So the final barrowload of pond mud was not shifted from the pile until early June 2012.