Building on a slope and difficult terrain – what I see on the plot before I say it can be done

In brief

A slope or difficult terrain does not rule out a residence – it moves the conversation from whether it can be done to what it will cost. Before I say it can be done, I want to know three things: where the water goes on this plot, whether the slope itself will hold once I load it with a house and cut into it, and how far down, all along the fall, you have to go to reach ground that will carry the house. The foundations, the retaining walls and the drainage come out of those answers, and on a plot like this they are what makes the difference in the budget; a retaining wall alone can be one of the largest items of the whole build, and on a visualisation you do not see it at all. Those answers should be on the table before the architect draws the first line, and if the design already exists, I check it against the same three questions. A slope handled well gives you something a flat plot cannot: the view, and a house that divides naturally into levels.

A slope is work to be costed

A slope has to be approached with respect. Does that mean every hillside is risky? No. It means a slope forgives less than a flat rectangle on the edge of town: the same mistake that would mean a repair on level ground can set the whole hillside moving here. Firms turn these plots down not because it cannot be done, but because the risk and the cost have to be worked out far more carefully than on flat ground, and not everyone wants to take that on. I do the sums on these plots – it is one of those subjects you simply have to sit down with properly.

A gradient changes the rules at the very first measurement. On a flat plot I set the level the house will stand at – its zero level, the finished ground-floor height from which every other height is measured – from the highest point of the ground along the whole perimeter of the building, and I add a margin to it so that water has no way in under the floor. If the plot clearly slopes and the road sits much higher than the spot where we will build the house, setting the zero level that way obviously makes no sense. There is no single level around the house then; there is an uphill side and a downhill side, sometimes a whole storey apart. A small detail, but it makes the point: on a slope nothing can be done by routine, and every item has to be worked out again for this particular fall.

That is why a plot like this is looked at more than once and in different weather, and on a slope you see the most after a downpour: where the water gathers, which way it runs off and where it washes the soil out. Then come the measurements. We have our own GPS receiver and plenty of our own surveying equipment, so we set out and check the levels ourselves. It is the surveyor’s level or the receiver that shows the height of the ground relative to the road, the height of the ground around the house and the height of the neighbouring plots (an architect or a site manager usually does not have that kind of equipment). Without those three figures, a conversation about the zero level on a sloping plot is a conversation about impressions.

Those visits and measurements give me three questions: about the water, about the stability of the slope and about the ground. Everything else – the design, the schedule and the price – follows from the answers.

First I ask where the water goes

On a slope, water has a direction. It runs downhill and looks for a place along the way where it can soak in or stop, and my job is to make sure that place is not the joint between the foundation and the wall.

I will start with what you cannot see. The place where water can get into a house is not the floor but the layer beneath it: the level of the slab or of the lean-concrete base, with the horizontal damp-proofing on top of it, where the ground floor begins. If from the top you have 2 cm of floor finish, 6 cm of screed and 15 cm of polystyrene, that level lies 23 cm below the floor you walk on, and with thicker layers lower still. Water that gets there has no way to evaporate, so after a while it turns into damp, and then into mould. On flat ground you defend against it with the ground falling away from the walls, a gravel strip around the house and a floor raised above the terrain. On a slope that is not enough, because the water does not wait in one place – it comes down from above and hits the house on the uphill side.

And here is the heart of the difference. On a flat plot, if the ground calls for it, a perimeter drain around the building takes the water away. On a slope that ring stays, but on its own it is not enough, because groundwater runs down through the layers from above and builds up on the side it comes from, like water in front of a badly placed dam. So higher up, across the slope, a second drain is laid – an intercepting one. Its only job is to catch the water and lead it off to the side before it ever reaches the foundations.

The drive is a subject of its own. Wherever you can, make sure water runs off the drive onto the road, not off the road onto your drive. On a slope with the road above the house, that cannot be turned around: the fall is what it is, and the rain from the whole surface will head for the building. What is left is to intercept it – a channel drain across the drive, above the house, and somewhere for that water to go.

Because water you have caught still has to go somewhere, and on a slope that can be harder than collecting it. It is good when it can be led downhill by gravity into a storm drain or a ditch, with the consents that requires. And if there is no such outlet nearby? Then a soakaway is tempting – simply putting the water back into the ground on the plot. On flat, permeable ground it can be a good solution; on a slope it can be a trap, because the water you let in can saturate the hillside and set it moving by itself. So I do not decide on a soakaway alone – the geologist decides, on the basis of the ground investigation. What counts is the highest water level you can expect here in the course of a year, not a reading taken in late summer, when it is usually at its lowest.

There is one more thing that is easy to forget on land that used to be farmed: field drains. Many plots still have drainage pipes running through them from those days, and on a hillside those pipes carry water off the whole slope. If the excavator cuts through one, the cut pipe must not simply be buried. It has to be rebuilt around the building, otherwise the water that ran in it stays on the plot and looks for a new route. It usually finds one exactly where a fresh excavation has just been dug.

And the case where drainage will not help: water under pressure. A drain will not lower its level, so the part below ground is made watertight, able to resist the push. That is a different class of work, closer to what I described for a pool set in high groundwater. One thing is common to all of these cases: water you ignore does not go away – it comes back with the first heavy downpour or with the spring thaw.

Will the slope itself hold

The second question is one nobody asks on a flat plot at all: will the ground I am to build on stay where it is once I load it and disturb it?

A slope stands as long as the resistance of the ground is greater than whatever pulls the mass downhill. Water spoils both sides of that equation at once. Soaked ground is heavier, so it pulls down harder, and the water between its grains pushes them apart and reduces the friction. That is why a hillside that stood quietly for generations can move after one wet spring. Landslides rarely have a single cause – it is usually several things at once, and water is the one that tips the balance.

Building adds two more. The weight of a house standing higher up presses on ground that is already under strain. And an excavation at the foot of the slope, for a garage built into the hillside or for a lower storey, cuts away the ground that was propping up the mass above it. Undercut it without support and you can set off something that stood still for years: sometimes at once, sometimes after the first heavy downpour. Even the planting matters here, because tree roots work in a slope like reinforcement, so I am careful about felling trees on a hillside.

That is why on a slope I have the ground investigated at closer spacing and to greater depth than on flat land. A flat plot tends to be the same all over, while a slope can change its layers every dozen or so metres down the fall, so two boreholes in the corners settle nothing. How many boreholes, then? That follows from the footprint of the house and from how much the profile changes between the first ones. I would rather order a few more than learn something new about the ground from the excavation.

There are also areas formally marked as at risk of landslides. In Poland the maps of such areas are kept by the state geological survey; for many municipalities they are publicly available (in Polish), and you should look at them before you buy. Such a plot automatically means complex ground conditions, and the house usually falls into the highest geotechnical category; that brings separate, wider geological studies, which have to calculate the stability explicitly and design the drainage. Let me add at once, so that nobody relaxes too soon: a plot that is not on the map is not automatically safe, because that survey does not cover the whole country. I do not take the seller’s word for it, or the map’s. I trust what the boreholes show.

Where to check the rest of a plot’s parameters – the local zoning plan, utilities, access, boundaries – I described separately, in the text on vetting a plot for a residence. Here I deal only with what the fall itself does to a house.

The foundation steps down the slope and the wall holds it

The third question is this: how deep, all along the fall, is the ground that will carry the house? Because on a gradient the foundation does not lie at one level.

It is run in steps. Each length of strip footing has to go a set depth into load-bearing ground, below the frost line, measured from the terrain that happens to lie above it. Since the terrain drops, each next step has to go lower to keep that depth. The steps cannot be as steep as you like, either: too sharp a step concentrates stress and the footing can crack there, so they are kept gentle. The steeper the slope, the more steps, and on a really steep gradient piles or a slab come into play – and either way the slope itself then has to be secured separately.

Where the house goes into the slope, its uphill wall stops being an ordinary wall. The ground pushes against it, so it works as a retaining wall and is designed like a basement wall: for that push, not only for the weight from above. Add to that waterproofing and a drainage layer right against it, so that the pressure does not build up. The drain takes load off the waterproofing, but it does not replace it. They are two different things, and both have to be done.

And then the retaining walls, the item most often underestimated, because from a distance a wall like that looks innocent. A retaining wall on a slope can cost so much that it becomes one of the largest items of the build – it all depends on what you want to build on such a plot. Where does that come from? Earth pressure is smallest at the top of the wall and greatest at its base, so the total force does not grow in proportion to the height: a wall twice as tall carries roughly four times the resultant earth pressure, and the moment trying to overturn it grows eightfold. That is why a wall that looks like a plinth on the drawing can be the single most expensive element on the plot.

Then there is what you cannot see. The wall has its own foundation, hidden below the frost line and bearing on load-bearing ground, and on weak ground it also needs that ground strengthened, or piles. It has drainage at the back, too, without which water builds up behind it and adds pressure to pressure, and sooner or later the wall bulges or slides. The repair then costs many times more than doing it properly in the first place.

There are several forms, and they are matched to the situation. A reinforced concrete wall holds with a slender stem standing on a wide base, with soil sitting on that base and helping to hold it down, but the base needs room. Gabions – wire cages filled with stone – let water straight through, so water pressure does not build up behind them the way it does behind a solid wall; in exchange they take up more space, and they also need the water drained from the ground beneath them. Sometimes there is no room for a base at all, because the wall stands on the boundary, and then the structural engineer turns to solutions that do not need a wide base: ground anchors, micropiles or a pile wall. Which one to choose is decided for the particular slope.

And one more thing, a formal one. A retaining wall holding up a slope is not a fence. It is a structure, with an engineer’s design and formalities, often with a building permit, and on top of that it must not make the run-off of water onto the neighbour’s land any worse.

A building site on a gradient

One thing tends to be thought about last, and on a slope it can account for half the trouble: whether the build can physically be serviced at all.

Heavy equipment does not like a gradient. A concrete pump and a crane set up on outriggers, which have to stand on level, load-bearing ground. On a slope there is usually no such platform, so it has to be made first: levelled, hardened, sometimes laid with road plates. The pump operator has the right to refuse to work if he sees that the machine will stand crooked or in mud, and he is right to.

Then there is the soil. An excavation cut into a slope produces far more of it than one on flat ground: some of it goes to level the terrain on the other side of the house, and the surplus has to be taken away. A spoil heap on a gradient will not hold itself up – it can slide, block the access or press on the edge of the excavation – so the place for the spoil is planned in advance. The excavation itself needs its sides secured, and left open it will not stand for long, because the first downpour washes it out. That is why the work is sequenced so that what is done higher up does not disturb what has already been done below.

And there is the drive, which stays with you for years. A steep drive can be impassable in winter. Either it is laid on a longer, gentler route, and then it takes up more of the plot, or its surface is heated – and that is a real electrical load to allow for when the grid connection is sized, as I set out in the text on a residence’s grid connection capacity. The routes of the utility connections run over this slope as well, and they also have to be on the plan before the excavator arrives.

Each of these things can be done, as long as it is foreseen before the price is set rather than discovered during the build.

When the slope becomes an asset

I have written a lot about the trouble, so now let me say why anyone takes it on. The same fall that forces a stepped foundation lets you set the house into the slope. Down below, set into the hillside, go the things that are meant to be closed and cool anyway: the garage, the plant rooms, a wine cellar, rooms without windows. Above them, already at the level of the upper garden, sits the living space, opening straight onto the view – the very thing a plot like this is chosen for. One house, and two ground floors on two levels of the terrain, each with its own way out. On a flat plot you would have to stack all that upwards; here it lays itself out along the slope, and the roof of the lower storey often becomes the terrace of the upper one.

There is a catch in this, and it ties this whole text together. The excavation that gives you the lower storey and the view is exactly the same excavation that undercuts the slope. Whether the slope turns out to be an asset or a problem is not decided by luck. It is decided by a retaining wall designed for the right pressure, by drainage laid where it is needed, and by a sequence of works that does not disturb the ground before there is something to hold it. With a house like this, what I care about most is that everything under the ground is boring and predictable.

When a slope is not worth it

Are there slopes where it cannot be done? Very few. Engineering copes with water, weak ground and a steep fall, so the question is hardly ever whether it can be done; it is what the solution costs and whether it makes sense for this particular house. The real limit tends to be ground that is still moving – an active landslide.

There are, however, slopes where building does not pay, and here the conversation depends on the stage you are at. If you are only buying the plot, checking it before the purchase costs a fraction of what the slope can add to the build, and that is the moment when it is simplest to walk away. If you already own the plot and what you want to build on it would bring astronomical extra costs, it is sometimes more sensible to sell it – or to take those costs on knowingly. Either way the decision is yours, made knowing how much this slope can add to the build. Sometimes this particular spot is worth it, because there simply is no second view like it. And sometimes, once the retaining walls are added up, the sentiment fades by itself.

One caveat, so that we understand each other: investigation and a sound design take away most of the risk, not all of it. A slope can surprise you even after a thorough investigation – with a local seep of water or with a layer that was not in the borehole next to it. That is why I have the ground investigated before I give a price, and the risk that investigation cannot remove is described separately in the contract, before it is signed. The price itself takes shape in conversation: at the start there is a range, and the more finished the concept, the narrower it gets – how that works with me I described in the text on the lump sum.

What I ask before I say it can be done

The three questions from the start of this text have consequences for the design, and they are what sets the cost:

Those three answers turn into a list I go through on every sloping plot:

1. Water – where it comes from, how high it rises in the course of a year and where I will send it.
2. Slope stability – whether the slope will stay stable under the weight of the house and with an excavation at its foot, and whether the plot lies in an area at risk of landslides.
3. Ground along the fall – how many boreholes and how deep, given that the profile changes down the slope.
4. Retaining walls – where, how tall, for what pressure and with what drainage behind them; whether there is room for a base or anchors are needed.
5. Foundations – a stepped strip footing, a slab or piles; walls pressed against the slope designed and waterproofed like basement walls.
6. The building site – access for heavy equipment, room for the outriggers of the pump and the crane, how much excavated soil can be reused and how much has to be taken away, the order of works on the gradient, and finally the drive itself in winter.

The results then go into the feasibility study – the work I do before I agree to take a design on. For how we work together from there, the simplest thing is to ask me directly; I have collected those questions in the Questions section.

Frequently asked questions

Can a house be built on a slope at all?

In nearly every case, yes. Water, weak ground and a gradient are jobs for engineering, not dead ends, so the question is what solving them costs. The exception tends to be an active landslide – ground that is still moving. Beyond that it comes down to the sums: whether the result on this particular plot is worth what has to go into the foundations and the walls.

What raises the cost of building on a slope the most?

Usually whatever holds the house up and keeps it dry: retaining walls, special foundations and drainage. The building itself can be a smaller part of the bill than the slope beneath it. The wall is the treacherous one, because its cost rises much faster than its height – every extra metre means disproportionately more structure. Then there are the earthworks, of which there is far less on flat ground: platforms for the equipment, taking away the surplus soil, securing the excavation.

How do I check whether a plot is in an area at risk of landslides?

Start with the local plan and with the public landslide maps kept by the state geological survey; for many municipalities you can look them up without leaving home. If the plot lies in such an area, the design needs wider investigation and an explicit stability calculation. If it does not, that settles nothing yet, because those maps do not cover the whole country – only boreholes on this particular plot settle it.

Is a house built into a slope safe in the long run?

It is, if what lies under it was designed and built for this particular slope. The walls pressed against the ground are designed for the pressure and waterproofed as in a basement, the slope is secured with a wall designed for the right force, and the water is led away from the foundations. A house is not safe because it happened to turn out that way – it is safe because someone worked out what acts on it before the first excavator arrived.

If you are looking at a plot on a slope, or already own one, and want to know what really stands behind the words “it can be done” – I will go over that ground and work out whatever can be worked out before a concept is drawn. Let’s talk about your project.