Site costs on a new home build: the ground behind rock, fill and slab upgrades

The site works allowance in a fixed-price building contract is a prediction about ground nobody has looked inside yet. It is written from a site inspection, a set-out plan and experience of the suburb, and it has to cover the slab, the excavation, the drainage and whatever the excavator finds at 600 mm. When it moves, it is usually not the builder shifting the goalposts. It is the ground correcting the assumption.

Most of the correction comes from four places: the reactivity of the clay, fill left by whoever shaped the lot, rock, and water. Each of them can be narrowed before the soil test comes back, and each of them has a specific thing worth checking at tender stage.

The class letter is a movement number, and band edges are close together

AS 2870 sorts residential sites by characteristic surface movement (ys), which is how far the ground surface is expected to lift and drop through one dry-to-wet cycle. Lochaden and Haberfield (2018) set the bands out plainly: up to 20 mm is Class S, slightly reactive; above 20 and up to 40 mm is Class M; above 40 to 60 mm is H1; above 60 to 75 mm is H2; and anything past 75 mm is Class E, extremely reactive.

The gap between a Class M price and a Class H2 price is real money in beam depth, reinforcement and concrete volume. The gap in ground behaviour between 39 mm and 41 mm of predicted movement is not. Two lots in one street, or two corners of one lot, can land on either side of a band edge, which is why a builder who has priced twenty slabs on the same estate can still get caught on the twenty-first.

What narrows it before drilling is the mapped geology and the climate zone, because those two set the plausible range. A block on deeply weathered basalt and a block on coastal sand are not the same bet, and you can know which one you are pricing on the day you quote.

Fill: the 400 mm line that decides whether there is a standard design at all

Fill is the item that turns a slab price into an engineered footing price, because a site with uncontrolled fill under the footings generally becomes Class P, and Class P has no standard design in the Standard. An engineer designs the footing from first principles for that lot.

There is a threshold worth carrying in your head. Delaney (2005) notes that the Standard tolerates a limited depth of uncontrolled fill, around 400 mm, before Class P is warranted, and that the allowance is tied to the assumption that ordinary slab footing excavations reach roughly that depth anyway. The same paper points out how easily estate earthworks reach that thickness: topsoil, mulch and material that was never suitable for founding get spread across a lot as part of finishing it.

That matters at pricing because the fill you are told about is the fill in the earthworks certificate. The fill that hurts is usually older: a filled gully, a dam that got pushed in, a batter that was trimmed. Historical aerial imagery, mapped contours and nearby borehole logs are the practical way to see it before you sign, and it is the same evidence trail used for spotting uncontrolled fill on a block.

Rock, and the difference between "rock" and refusal

Rock does not arrive as a flat surface at a known depth. In weathered profiles the material grades from clay through progressively less weathered rock, and the depth at which a bucket stops depends on which part of that sequence sits under the trench. Two piers ten metres apart can hit refusal metres apart in depth.

The tender question is therefore not "is there rock" but "how deep is competent material likely to be, and how variable". Mapped geology answers the first half. Nearby bore logs answer the second, and if the logs within a few hundred metres disagree with each other, that variability is the finding.

Trees, drains and the landscape that arrives after handover

Reactive-clay damage is rarely caused by the design alone. Cameron (2018) is explicit that footings built to the Standard can still show some movement and distress, usually cosmetic, and that a client who wants better performance than the Standard targets can ask for an upgraded footing design and pay for it. The Standard also carries a method for allowing for the extra drying caused by nearby trees, so a large tree that is staying is a design input, not a landscaping detail.

Delaney (2005) reported that after the 1989 Newcastle earthquake, a review of one major insurer's files found that reactive soils were the primary cause behind 51 per cent of claims, and separately observed that damage is rare where the correct site class was adopted and severe moisture conditions, such as large trees hard against the building, were avoided. The failure mode is nearly always the pairing of an optimistic classification with a site that then gets drier or wetter than the design assumed.

What "compliant" is designed to deliver

Worth knowing before a defect conversation starts: the Foundation and Footing Society Victoria (2022) practice note states that footing systems built to AS 2870 are an acknowledged compromise between cost and performance, that the Standard defines expected performance using its own damage categories, and that Category 0 to 1 is the normal limit with Category 2 accepted under adverse conditions. A hairline crack in a cornice is not, by itself, evidence that the slab was wrong. Being able to say that early, with the site class and the design assumptions in front of you, is worth more than saying it after the first summer.

What to settle before you price the site works

• The geological unit under the lot, not the suburb. It sets the plausible reactivity range and whether rock is a realistic risk.
• The fill history. Historical aerials and contours show gullies, dams and pads. Anything that shaped the lot after subdivision is a fill question.
• Nearby borehole logs. Depth to refusal, depth to clay, how much the logs disagree.
• The fall across the building platform. Cut and fill drives retaining, drainage and the depth of the deepest beam.
• Groundwater. A shallow water table changes excavation, drainage and the practicality of deep beams or piers.
• Trees staying on or beside the lot, with rough distances, because they are a design input.
• Acid sulfate soils on low coastal ground, since they change what disturbed soil can be moved and where.

None of that replaces the soil test. It sets the number you write in the contract, the qualifications you attach to it, and the brief you give the driller so the investigation answers the questions your price depends on. Our geotechnical report cost breakdown covers what each tier of investigation buys.

Common questions

Why did my site costs go up after the soil test?

Usually because the classification came back higher than the allowance assumed, because uncontrolled fill or shallow rock was found in the footing zone, or because the fall across the pad turned out to need retaining and drainage. All three are ground findings, not pricing decisions.

Can a builder know the site classification before drilling?

No. A classification requires a site investigation with sampling and testing by a geotechnical engineer. What can be known beforehand is the mapped geology, the climate zone, the fill and terrain history and what nearby bores found, which together give a defensible expected range to price against.

Is a Class P site always expensive?

Not always, but it is always engineered. P means the site sits outside the standard designs, so a footing is designed specifically for it. The cost depends on why the site is P: shallow uncontrolled fill can sometimes be removed, while deep soft ground or landslip risk is a different order of problem.

Before you commit a site works figure, generate a desktop report for the address so the geology, terrain, fill history and nearby bore logs are on the table while the price is still yours to set, or see a sample report to check what it covers.

LayeredGeo compiles geology, soils, groundwater and site data into an automated geotechnical desktop report for any address in Queensland, New South Wales or Victoria.

Sources

• Andrew L.E. Lochaden and Chris M. Haberfield (2018). Assessment of the AS2870 standard designs for residential rafts on reactive clay. Australian Geomechanics, Volume 53, Number 1 (Mar 2018). geomechanics.org.au
• M. G. Delaney (2005). An overview of engineering geology and geotechnical challenges in the Newcastle Region. Australian Geomechanics, Volume 40, Number 1 (Mar 2005). geomechanics.org.au
• Donald A. Cameron (2018). Dealing with reactive clay soils through a national standard. Australian Geomechanics, Volume 53, Number 1 (Mar 2018). geomechanics.org.au
• Foundation and Footing Society Victoria Inc. (2022). Practice Note No. 8. Australian Geomechanics, Volume 57, Number 2 (Jun 2022). geomechanics.org.au

About this article. Published by LayeredGeo and written from the published research cited in the Sources section above. It is general information about how property and ground conditions are assessed in Australia, not engineering, planning, legal or financial advice, and it is not specific to any property. Check anything that matters against the source dataset or a suitably qualified professional before you rely on it. If you spot something wrong, tell us at hello@layeredgeo.com.au and we will fix it.

← All articles