Building on Essex Clay: Shrinkage, Heave and Depth
The ground under Colchester is not one material, and which one you have decides how far down you go and how the building behaves for thirty years after.
Around Colchester the phrase “Essex clay” gets used as though it named one material with one answer attached to it. It does not. What sits under a garden here might be London Clay, or sand and gravel over London Clay, or glacial till, or river alluvium, and two houses in the same street can want different foundations. This guide is about the material rather than the foundation: why some ground moves and some stays still, how much and in which direction, and what that means for a building over the thirty years after it is finished rather than only on the morning the trench is dug.
There is no single thing called Essex clay
The bedrock is one question. What was left on top of it over the last half million years is what your trench will actually find.
The bedrock across this part of Essex is the London Clay Formation, a marine mud laid down roughly fifty million years ago. Where it is at or near the surface it is the material that governs foundation depth, and it is the reason the whole county has the reputation it has. Over a great deal of the district, though, it is not at the surface.
Three things put something else on top of it. The early Thames ran east across what is now north Essex and Suffolk before the ice diverted it south, and left a thick spread of sands and gravels behind it. The Anglian ice sheet then came down to roughly the line of the Blackwater and left chalky boulder clay, more sand and gravel, and fine laminated clays wherever meltwater ponded. Rivers cut through the lot afterwards and dropped alluvium along the valley floors.
Each of those materials behaves differently in a trench:
- Clean sand and gravel bears well and does not shrink or swell with the weather, so the seasonal problem largely disappears. It collapses into an open trench and runs when it is wet, which is a different problem on the day.
- Boulder clay is shrinkable, usually less so than London Clay but enough to matter, and it is variable within a few paces because of how it was deposited.
- Laminated lake clays, of the sort worked at the Marks Tey brickpit, are fine grained and among the more mobile materials in the county.
- Alluvium in a valley floor is soft, compressible and no use as a bearing material at shallow depth. You go through it.
The mapping is free and worth looking at. The British Geological Survey publishes its geology online, and its borehole record holds the logs from investigations already done nearby. Both tell you what is probable. Neither tells you what is under your lawn, because a gravel cap can thin from two metres to nothing across the width of a garden, and the boundary between the cap and the clay beneath frequently runs straight across a plot. The only thing that answers it for your address is a hole in the ground, which is why our groundworks page treats a trial hole as part of pricing rather than as an extra.
Why clay moves and gravel does not
It comes down to particle size, and to where the water is sitting.
In a sand or a gravel the grains are large enough to touch one another, and water simply drains through the spaces between them. Take the water away and nothing changes shape, because the grains were already carrying the load.
Clay particles are minute plates, small enough that surface chemistry matters more than weight. Water is held between the plates and holds them apart. Wet the clay and it takes on more water and expands. Dry it and the plates draw together and the whole mass contracts. That is not a marginal effect. Lawns on London Clay open into fissures you can get a hand into by late August, and the ground surface can rise and fall by tens of millimetres between a wet February and a dry September.
How much a particular clay does this is measured rather than guessed. The plasticity index is the range of moisture content over which a soil behaves as a plastic material rather than as a liquid or a solid, and it comes from a cheap laboratory test on a disturbed sample. Because a soil is rarely all clay, the figure that gets used is the modified plasticity index, which adjusts the result for the proportion of the sample that is genuinely fine, discounting the sand and gravel carried along with it. A soil is treated as shrinkable where it holds more than about 35 per cent fine particles and the modified plasticity index is 10 or above.
| Modified plasticity index | Volume change potential | What it means in practice |
|---|---|---|
| 60 and above | Very high | Substantial movement, large influencing distances around trees |
| 40 to 60 | High | Where London Clay usually lands |
| 20 to 40 | Medium | Much boulder clay and mixed drift |
| Below 20 | Low | Sandier material, still shrinkable at the top of the band |
You cannot arrive at that band by looking at a garden, and neither can anybody else. It comes off a test, and on a difficult site it is the cheapest piece of certainty available.
Where the 900mm figure comes from
It is not a rule of thumb somebody invented. It is the base depth for a medium volume change soil with nothing growing near it.
The warranty standards that engineers and building control bodies both work from set a minimum formation depth, meaning the underside of the concrete, against the volume change potential of the soil on a site with no significant vegetation influence. That is 0.75m for low, 0.9m for medium and 1.0m for high. Those figures are why 900mm is quoted so consistently across this area, and they are also why it is so often not the number that ends up on the drawing.
Read the condition attached to it. That depth applies to a clear site. Put a mature tree within influencing distance and you leave the base case entirely and enter a set of tables built around species water demand, mature height and distance, which run to about two and a half metres before the method has to change to something other than a trench. How that calculation is done, and by whom, is dealt with in trees, TPOs and foundation design, and the shape of the final answer in how deep extension foundations need to be.
Shrinkage and heave are two different movements
People use the words interchangeably. They are opposite directions, opposite timescales and opposite damage.
| Seasonal shrinkage | Heave | |
|---|---|---|
| Cause | Clay drying out, worst in a dry summer and worse again near a mature tree | Clay recovering water it has not held for years, usually after a tree comes down or a hedge is grubbed out |
| Direction | Down and inward, reversing each winter | Up and outward, one way |
| Timescale | An annual cycle, repeating for as long as the cause is there | Years, often more than a decade, slowing without ever quite stopping |
| What you see | A crack that opens in late summer and closes in winter, then opens wider the following year | Floors lifting, doors binding at the top, openings pushed out of square, nothing reversing |
| The answer | Found below the depth at which moisture content stops changing | Found below it and give the clay somewhere to expand into |
Heave is the more expensive of the two because a building is designed to carry load downwards. Every part of it, from the foundation to the roof, is arranged to pass weight to the ground. Nothing in a house is designed to resist being pushed up from underneath, and a slab lifted 30mm in the middle does far more damage than a wall that has settled by the same amount.
It is also slow enough to be misdiagnosed. Movement that starts three years after handover and is still going eight years later does not look like a foundation problem to most people, and by then the builder is a memory and the tree is firewood. The measures that prevent it are cheap while the trench is open and impossible afterwards: a compressible layer against the vertical faces of the foundation so swelling clay can move sideways instead of gripping the concrete, and a floor suspended over a void rather than bearing on ground that is about to rise. Those are set out on the foundations guide.
One cause of heave gets missed almost universally, and it is not a tree at all. Ground that has been drying and wetting seasonally for a century stops doing either the day you cover it with a building. If it was on the dry side of its natural moisture content when the slab went down, it will slowly recover underneath the new floor whether or not anything was ever felled. A summer build on desiccated clay is the classic set-up for it.
Reading a crack
Width, shape and season together say more than any one of them alone.
Not every crack is the ground. Fresh plaster shrinks through its first heating season and gives fine hairlines, particularly at the board joints, and a hairline over a lintel is usually the same thing. Stepped cracking through the mortar joints with rust staining in it points at wall ties rather than at anything below.
Clay movement has a signature. It is usually diagonal, it tapers rather than running at even width, and it is worst where two structures founded at different depths meet, which on an extension means the junction with the old house. A crack wider at the top suggests a corner dropping away. Wider at the bottom suggests something lifting. And it is seasonal: a shrinkage crack opens through August and closes through the winter, and each year opens a little wider than the last.
The categories a surveyor will use. Damage from foundation movement is classified by crack width: negligible below about 0.1mm, very slight up to about 1mm, slight up to about 5mm, moderate at 5 to 15mm, severe at 15 to 25mm, and very severe above that. Anything in the first three bands is a redecoration question. From moderate upwards it is worth paying for a structural opinion rather than a builder’s opinion, including ours.
If something appears, resist the urge to fill it. Photograph it with the date and a rule in the frame, pencil a mark across each end, and measure the same point in February and again in September. A crack that has been measured twice tells an engineer something. A crack that has been filled and has reopened tells them almost nothing except that it moved.
What clay changes besides the depth
The material affects the method, the drainage and the programme, not only the number on the drawing.
A freshly cut clay trench stands up remarkably well, which is exactly why it gets left open. Rain then softens the base and the sides, and a trench that has sat through a wet week frequently has to be dug out again to reach sound material before anything can be poured. The gap between excavation and concrete is a quality item, not a scheduling convenience.
Drainage is the other consequence, and it runs both ways. Clay drains slowly or not at all, so a soakaway here has to be sized against a percolation test on the actual plot rather than assumed, and on heavy clay it often fails the test outright and the surface water has to be dealt with another way. Meanwhile the ground movement itself breaks rigid drainage, and a leaking drain wets the clay around it, which allows more movement, which opens the joint further. It is a loop that gets worse. Flexible jointed pipe on a proper bed is the cure, and the wider drainage question is covered in build over agreements and extension drainage.
The decisions that come after the builder leaves
The foundation was designed against the conditions around the building. Change those conditions and you have changed the design.
Planting is the one that catches people. A foundation designed on a clear plot assumed a clear plot, so putting a willow, a poplar or an oak in near a new extension a few years later gradually creates the exact problem the depth was chosen to avoid. The tree is fine. It is drying ground that nothing allowed for.
Taking things out deserves the same pause. A mature hedge along a boundary has been drawing water for decades, and grubbing it out to put a fence in starts a slow recovery in the clay behind it.
Water going where it should not is the third. A downpipe discharging at the foot of a wall, a blocked gully, or a new soakaway close to the foundation all concentrate wetting on one spot of shrinkable ground, and paving laid tight to the wall does the reverse and seals it. Watch the first two summers, particularly if one of them is dry. That is when a foundation taken deep enough quietly proves it, and when one that was not starts to say so.
Subsidence, and the word that follows a house around
Subsidence is an insured peril on an ordinary buildings policy, and it is what responds when an existing house starts moving on clay. New work that moves because it was founded too shallow is a construction defect, which is a different conversation with different people in it. Knowing which of the two you have matters, because a recorded subsidence claim attaches to the property and turns up at every renewal and every sale afterwards.
What protects you on a new extension is documentary rather than dramatic: a foundation designed by a structural engineer against a soil that was actually tested, and a building control record of the excavation inspection showing what the trench exposed and at what depth the concrete went in. Those answer the question years later, when nobody remembers the summer it was dug.
We look at the ground before we price the job, dig a trial hole where the site is uncertain, and have the foundation designed off what is there rather than off the number everybody quotes. If you want a realistic view of what your plot is likely to demand, the survey and the quotation cost nothing.
Soil behaviour varies from plot to plot and foundation design is a matter for the structural engineer and building control on your own project. Treat this as an explanation of how the ground behaves rather than as a specification for your extension.
Last reviewed August 2026. Planning and Building Regulations change, and your property may differ.
Thinking about the job itself?
Most people who call us already have drawings and permission and want a builder to price the work and build it. Either way the survey and the written quotation cost nothing.