Setting Foundation Depth Near Trees on Shrinkable Clay

Three Numbers, and This Page Will Not Supply the Fourth

The first is the modified plasticity index of the clay, and it is the one nobody has. Take a disturbed sample from the trial hole and have the Atterberg limits run — BS 1377-2 or BS EN ISO 17892-12 in the UK, ASTM D4318 in North America — which gives a plasticity index. That raw index is then multiplied by the fraction of the soil passing the 425 micron sieve, because the sand and gravel coarser than that take no part in the swelling. A clay with a plasticity index of 45 that is a third sand does not behave like a clay with a plasticity index of 45 that is not, and the modified figure is what NHBC Standards Chapter 4.2 bands into high, medium and low volume change potential. A soil test on a domestic extension costs less than a day of the digger it is about to book.

The second is the water demand of the species, which is a property of the tree rather than of the site. Oak sits in the high water demand group in the Chapter 4.2 classification, alongside the willows, poplars and elms that turn up in every subsidence case study. Read a high-demand species as moderate and nothing changes on the drawing; everything changes at the building three dry summers later.

The third is the ratio D over H: the distance from the stem to the nearest part of the new foundation, divided by the tree's mature height. Mature height is the published figure for that species — the height it is expected to reach — not a measurement of the tree standing there today, and that distinction is the most common error in the whole exercise. A young high-demand tree measured at eight metres and entered as eight metres produces a generous D/H, a modest trench and a foundation designed for a tree that will not exist in twenty years. The tape supplies D. The table supplies H.

BRE Digest 298, Low-rise building foundations: the influence of trees in clay soils, sets out the mechanism with the reasoning attached, and BRE Digests 240 to 242 cover shrinkable clay behaviour more generally. Between them and Chapter 4.2 there is a table converting volume change potential, water demand and D/H into a minimum depth to the underside of the foundation. It is not reproduced here, deliberately: it gets revised, it needs all three inputs entered correctly, and a depth copied from an article is the single figure on this job that nobody downstream can check. On this extension the engineer came back with 2.6 m — deep enough to change the method and the price, shallow enough to still be a trench.

What a 2.6 Metre Trench Fill Is Actually Made Of

Most people picture concrete in a slot. What gets built is a mass concrete fill with both vertical faces treated, a stratum at the bottom somebody has to look at and record, and a ground floor above it that must not be sitting on the clay at all — six distinct things, four of them invisible once the ground is closed up, and three of them regularly missing from a quotation. Trench fill rather than a strip footing is not a preference at this depth, it is the only method available: laying 2.6 m of blockwork off a strip in a 600 mm trench means a person standing in a supported excavation two and a half metres deep for several days, and there is no version of that which is either safe or affordable. Trench fill converts all of that labour into concrete, which is why the concrete bill here looks alarming next to a conventional strip and is still the cheap answer. Past roughly two and a half to three metres the arithmetic tips again and bored piles with a suspended ground beam start to win — a conversation to have with the engineer before the depth is fixed, not after.

The quantity is straightforward geometry and one trap. The perimeter of new foundation is two returns of 4.0 m plus the 5.2 m rear wall, so 13.2 m of trench at 600 mm wide. Excavation goes to 2.6 m; the concrete stops about 150 mm below finished ground level so there is something to build the below-ground brickwork off, which puts 2.45 m of concrete in the trench. That is 19.4 cubic metres before a single allowance for over-dig.

Twenty-One Cubic Metres Out of a Slot Nobody Can Stand In

The dig is 13.2 m by 0.6 m by 2.6 m, which is 20.6 cubic metres of clay in the bank. That is the figure on the drawing. What comes out is more, because a 600 mm bucket does not leave a 600 mm trench in fissured clay — the sides break back along the fissures, the corners get squared out twice, and every trench fill job over-digs somewhere between five and fifteen per cent depending on the ground and the driver. It is worth measuring the over-dig at the first corner rather than discovering it when the concrete runs out, because the concrete goes wherever the hole went.

The other thing that separates this from a service trench is that nobody may enter it. A 2.6 m excavation with vertical sides in clay is a supported excavation or a battered one, and clay's habit of standing up beautifully for two days and then failing without warning is exactly what makes it dangerous. In the UK the governing document for the earthworks is BS 6031 and the duties sit under CDM 2015; in the US it is OSHA 29 CFR 1926 Subpart P, which puts a competent person on the inspection and treats a foundation trench as an excavation like any other. Trench fill helps here, because the sequence can be dig-and-pour without anybody going down.

For the quantity itself, the excavation calculator below is being used for one line of its output. Its headline answer is loose backfill for a pipe trench and this trench is not being backfilled — it is being filled with concrete. Read the first line of the breakdown, Excavation volume (the depth entered), which is simply length times width times depth: neither the pipe diameter box nor the compaction box touches it, so set them to any valid value and ignore them. The headline figure does become the right answer later on the same job, for the drain run that follows the extension out to the existing manhole.

The Hole Measures 21 and the Spoil Measures 27

Clay dug from 2.6 m comes out in lumps with voids between them and does not go back to its in-situ density on the back of a lorry. A stiff fissured clay lifted from a deep trench commonly bulks by a quarter to a third, so 20.6 cubic metres in the ground becomes something near 26.8 loose at thirty per cent, and everything downstream of the bucket — the muck-away, the grab loads, the space it occupies in a side return while it waits — is counted in that larger state. None of it is coming back, either. A pipe trench gets backfilled with what came out of it; a trench fill is filled with concrete, so the entire bank volume leaves site and roughly the same volume of concrete arrives to replace it. The only backfill on this job is the narrow strip outside the wall above the concrete, and cohesive clay is a poor choice for it: it compacts badly against new brickwork, it holds water where you least want it, and it goes on moving seasonally right where the drainage and the damp proof course are.

Take the swell figure from the ground investigation or from whoever hauls muck in your area, not from a round number. Wet-dug clay and dry-dug clay from the same field behave differently in a barrow. One unit note on the tool below: its bank volume box follows the metric and imperial switch at the top of the page, reading cubic metres or cubic yards, so the hole goes in the unit the box names and the answer comes back in the same system.

The Skip Fills by Weight Long Before It Fills by Volume

Thirty-five cubic yards of loose clay sounds like three large skips. It is not, because clay is heavy enough that the vehicle runs out of payload before the container runs out of space, and haulage firms will not put heavy inert muck in their largest skips for exactly that reason. Convert the volume to a weight before booking anything, and expect to be talking about grab lorries or small skips with a lot of visits rather than a builder's skip on the drive.

There is a trap in the conversion worth naming, because it moves the answer by a third. Bulk density figures for clay are measured on material in a fairly natural state, so multiplying the loose, bulked volume by a bank-ish density double-counts the air you just added. Mass does not bulk. Put the bank volume through the conversion to get the weight that has to be lifted, and use the loose volume only for the space it occupies. On this job the bank 20.6 cubic metres is 26.9 cubic yards, which at the planning density below is about 40 short tons, near enough 37 tonnes — three grab loads on an eight-wheeler, give or take what the vehicle is plated for. Wet clay is heavier again, and clay dug in February off a site with nowhere to drain sits at the top of any published range. The figure that governs in the end is the weighbridge ticket, which is why the tonnage matters commercially: tipping is charged by weight, and clay is usually classified and priced differently from clean fill.

Deep Is Not the Same Thing as Checked

None of the above proves the foundation bears. It is easy to assume a trench fill this deep is so far inside its capacity that the check is a formality, and on the building load alone it would be — 30 kN per metre run spread over a 600 mm wide footing is 50 kPa, which almost any stiff clay carries without noticing.

The self-weight is what makes it interesting. Each metre run of this foundation contains 0.6 by 2.45 metres of concrete, which is 1.47 cubic metres weighing about 35 kN at 24 kN per cubic metre. The foundation is heavier than the building standing on it. Total load at formation is therefore around 65 kN per metre run, and over 0.6 square metres of bearing area that is a little over 108 kPa — better than double the figure the superstructure alone suggested, and close enough to a typical assumed allowable of 150 kPa that the margin is worth knowing rather than assuming. Anyone whose experience is 900 mm strips will not have met this, because at 900 mm the concrete is a minor term.

That 108 kPa is a gross pressure: it takes no credit for the 2.6 m of clay that was removed to make room for the concrete, which was already pressing on that horizon at something like 49 kPa before anyone arrived. The net increase the clay actually feels is much smaller. Whether the check is run gross or net, and against which allowable value, is the geotechnical designer's decision and it needs to be stated on the drawing — the two conventions differ here by a factor that matters.

Driving the checker below takes one deliberate move: work per unit run. Put 1 in the length box, the trench width in the width box, and the load for that one length of wall in the load box. Every box follows the unit switch — the lengths, the load in kN or kips and the allowable pressure in kPa or psf — so on the imperial page the 1 is a foot of wall and the load is the load on that foot.

Where an Allowable Bearing Pressure Comes From

A number like 150 kPa gets quoted for stiff clay so routinely that it is easy to forget it is derived rather than given. For a foundation loaded quickly relative to how fast clay drains, the governing case is undrained: the friction angle is taken as zero, the bearing capacity factor Nc becomes 5.14, the width term drops out entirely, and the ultimate capacity reduces to 5.14 times the undrained shear strength plus the overburden pressure at founding level. Which means the width of a trench fill contributes nothing to its bearing capacity in this case — the depth does all the work, through the surcharge term.

For this trench, taking an undrained shear strength of 90 kPa for the stiff clay at 2.6 m and an overburden of about 49 kPa from 2.6 m of soil at 19 kN per cubic metre, the ultimate figure comes out near 512 kPa. Divided by a factor of safety of three that is around 170 kPa, which lands in the same territory as the 150 kPa assumed in the previous section without agreeing with it exactly — and the honest response to that gap is to use the site's own number, not to average them. The undrained strength has to come from the ground investigation for this plot. Undrained shear strength varies by a factor of several across ordinary firm-to-stiff clays and it increases with depth, so a value taken from 1.5 m is not the value at 2.6 m.

Two assumptions are being stretched and should be said out loud. Terzaghi's expression was derived for a foundation whose depth does not much exceed its width, with the soil above founding level replaced by a simple surcharge; here the depth is over four times the width, and the shear strength of the clay along those 2.6 m faces is being thrown away. Both simplifications are conservative, so the answer is safe to compare against — but it is not a design. The formula is also strictly a strip footing formula, which is the one thing on this page a continuous trench fill genuinely is. Eurocode 7 in BS EN 1997-1, or BS 8004, sets what an actual design has to demonstrate, and neither is a single line of arithmetic.

Ordering Concrete Into a Trench You Cannot Look Into

Specify before you quantify, because on clay the mix is not a default. Ground containing pyrite — London Clay and the Lias among others — oxidises to produce sulfates, and concrete cast directly against it may need specifying for aggressive ground. BRE Special Digest 1, Concrete in aggressive ground, classifies the site and produces a design chemical class; BS 8500-1 then converts that into a designated FND concrete rather than the general-purpose mix a merchant will assume you want. Where the International codes apply, ready-mixed concrete comes under ASTM C94 and the sulfate exposure classes sit in ACI 318. A trench fill is unformed, permanent and unvibratable in any meaningful sense, so it is a poor place to save money on a mix.

Quantity is 13.2 m by 0.6 m by 2.45 m of concrete, which is 19.4 cubic metres nominal. Waste on a trench fill is not the five per cent a formed footing takes: the trench is the formwork, the trench is not a rectangle, and the over-dig discussed earlier all has to be filled. Fifteen per cent is a defensible allowance in fissured clay and it brings the order to about 22.3 cubic metres — three full loads on a standard six-metre truck and a part load behind them. Running short mid-pour on a trench fill is worse than on a slab, because the cold joint lands somewhere the design may not permit one and there is no practical way to reach it.

Two notes on driving the calculator. Its thickness box stops at 1 m, so the 2.45 m depth will not go in that box — but the three dimensions simply multiply, and the other two boxes accept up to 50 m. Put the 13.2 m run in the length box, the 2.45 m depth in the width box, and the 0.6 m trench width in the thickness box, which reads centimetres, so 60. Second, ignore the bag count in the breakdown entirely: 22 cubic metres is a ready-mix order in every conceivable circumstance, and the number of 80 lb bags it would take is a figure with no use on this job.

The Tree and the Neighbour Both Have Standing

The oak is not simply an input to a table. BS 5837, Trees in relation to design, demolition and construction, defines a root protection area as a circle whose radius is twelve times the stem diameter measured at 1.5 m above ground, capped at 707 square metres. A 620 mm stem therefore gives a radius of 7.44 m and an area of about 174 square metres. The trench at 9.1 m sits outside it, which is the answer everyone wants and is worth confirming with a tape rather than assuming — a stem 150 mm thicker would have brought the protected area to the corner of the dig, and excavation inside a root protection area is an arboricultural method statement question before it is a groundworks one.

It may also carry a Tree Preservation Order or stand in a conservation area, both of which sit under the tree provisions in Part VIII of the Town and Country Planning Act 1990 and both of which make work on it a consent matter with real penalties. Here the point is close to academic, since the tree belongs to the neighbour and is not going anywhere by agreement — worth saying plainly to a client who has started pricing a foundation and is beginning to wonder how much cheaper the job would be without an oak in it.

The neighbour has a second interest as well. Excavating to 2.6 m within three metres of an adjoining owner's structure, to a level below their foundations, triggers a notice under section 6 of the Party Wall etc. Act 1996 in England and Wales, and the notice period is not something to discover in the week the digger is booked. One last detail has nothing to do with the tree and everything to do with living with the result: this extension sits on a foundation 2.6 m deep, against a house almost certainly on a strip at less than a metre. The two will move differently every year for the life of the building, and where the new work meets the old that difference wants a designed joint rather than a bonded-in crack that opens next spring and gets blamed on the plasterer.