How many boreholes does a house block need?

Under AS 2870-2011, a house site needs at least one borehole or test pit, and at least three where the design depth of seasonal moisture change (Hs, how deep the soil wets and dries with the seasons) is 3 m or more, or where the soil profile is known to be highly variable. Each hole has to examine the soil to at least 0.75 times Hs below the surface, or below any cut known at the time, with 1.5 m as the shallowest allowed depth unless the hole reaches rock. Most firms drill two holes on a standard suburban lot.

Those are minimums. Above them, how many holes a block needs, where and how deep, is the engineer's call. The common beliefs about borehole numbers are mostly true somewhere and wrong somewhere else. Six of them follow the minimums.

How many holes, and how deep, city by city?

Hs comes from Table 2.4 of AS 2870-2011 for the listed cities, or from the climate zone elsewhere. The minimum depth is 0.75 Hs with a 1.5 m floor.

Location	Hs (AS 2870 Table 2.4)	Minimum depth examined	Minimum holes per site
Brisbane and Ipswich	1.5 to 2.3 m	1.5 to about 1.7 m	1
Toowoomba	1.8 to 2.3 m	1.5 to about 1.7 m	1
Sydney, Gosford, Newcastle	1.5 to 1.8 m	1.5 m	1
Melbourne	1.8 to 2.3 m	1.5 to about 1.7 m	1
Perth	1.8 m	1.5 m	1
Hobart	2.3 to 3.0 m	about 1.7 to 2.25 m	1, or 3 where Hs reaches 3.0 m
Albury-Wodonga	3.0 m	2.25 m	3
Adelaide	4.0 m	3.0 m	3

A site where the profile is known to be highly variable also needs three, and the Standard treats gilgais (the hummocky ground found on some cracking clay plains) as evidence of that. Across a subdivision the total can be reduced where soil profiling shows uniform conditions. A small extension or outbuilding, up to 9 m in each direction and articulated from the house, needs only one hole if the original classification has proved satisfactory.

A Class P site (fill, soft ground, landslip and other problem conditions) can need more than these minimums, and the Standard says so: an investigation that only meets them may not be enough to design footings on a Class P site.

Is one borehole enough for a house?

Sometimes. On a flat lot on one geological unit, with no fill and no slope, a single well-logged hole can represent the ground under a modest house, and AS 2870-2011 accepts it.

The trouble is variability. Li et al. (2016) built a shrink-swell database from 47 sites across Melbourne. Four samples from Point Cook, all from the same basalt-derived clay unit, ranged from about 1.6 to 6.2 percent strain per pF, close to a fourfold spread in reactivity within one suburb. Ground that varies that much between sites can vary across a block too, and on a cut-and-fill block one hole in the natural corner says nothing about the fill corner.

What to do with it: treat one hole as the floor (three where Hs is 3 m or more). The more the ground varies, or the less you know about it going in, the more holes you need.

Is two metres deep enough?

In coastal Queensland, Sydney, Melbourne or Perth, two metres clears the AS 2870 minimum of 1.5 to about 1.7 m. It does not in Adelaide, where the minimum is 3 m, or anywhere else Hs is 3 m or more. Where the profile shows deep, open shrinkage cracks, the Standard has Hs increased to at least the depth of cracking, which pushes the minimum down with it.

Cuts catch people out. The minimum depth is measured below any cut known at the time, not below today's surface. Hargreaves (2005), writing from twenty years of practice in South East Queensland, worked through a planned 1.2 m cut: where Hs is 1.5 m, the 1.5 m minimum below the finished pad means drilling to about 2.7 m from the original surface. He also noted that in parts of South East Queensland the clay gets more reactive with depth, so a cut can expose more reactive clay than the surface holes would suggest. Classification holes there once averaged 1.2 to 1.5 m deep; by 2005, 2 to 2.5 m was more typical.

What to do with it: let the climate set the minimum, the planned cut set the starting level, and the geology set the rest. Floodplains, swamp deposits and reclaimed land need holes that reach competent material. Residual clay over rock often does not.

Do more holes always give a better answer?

No. Extra holes on a small, uniform lot mostly confirm what the first ones showed. What matters is where they go and how well the work is done, and the Ballina Bypass on the NSW north coast shows it. Before mid-2007 alone, the ground under its 6 km of Richmond River floodplain was investigated with about 295 boreholes, 142 cone penetration tests and 169 test pits, plus two trial embankments. The road opened on budget and six months early. Even so, piles of the same length drove to refusal at very different levels on an irregular buried basalt flow, and several broke during driving, although a borehole had been drilled at every pier (Kelly et al. 2020). The authors' conclusion, with hindsight, was that quality matters as much as quantity, or more.

What to do with it: spend the effort on placement and depth, not raw count. Two well-sited holes on a variable block beat five clustered on the good half.

Do the boreholes down the street cover my block?

Not on their own. Nearby bore logs are the closest thing a desktop study has to ground truth, and a log three lots away showing rock at one metre is a useful hint. How far that hint holds depends on how uniform the ground is. River-laid soils are layered and change over short distances as old channels and swamps cross a floodplain. Residual soils, formed in place from weathered rock, grade unevenly into the rock below, so depth to refusal can swing by metres along one street. A neighbour's log tells you the unit and the likely range. It does not tell you where your lot sits within that range, and it says nothing about whether your lot was filled.

What to do with it: use nearby logs to scope the investigation, not to replace it.

Does the sampling method matter?

Yes, especially in soft clay, because how a sample is taken changes what the laboratory measures. On the Ballina Bypass, one early embankment had settled 3.5 m when the team judged it had done its job, and it kept going. The clay parameters had come from disturbed samples. A few years later, high-quality sampling at a national soft soil research site less than a kilometre away gave a compressibility figure about double the one first used (Kelly et al. 2020). Pineda et al. (2014) describe how that site was sampled: continuous holes, tube samplers from 50 mm open tubes to 100 mm pistons compared side by side, and block sampling planned to cut disturbance further.

For a reactive stiff clay under a house the concern is different but real: the sample has to be intact enough to run a valid shrink-swell index (Iss) test.

What to do with it: on soft or sensitive ground, the sampling method is part of the scope. Ask what will be recovered and how, not just how many metres will be drilled.

Can a desktop study tell me how many holes I need?

It can narrow the question. A desktop study shows the geological unit, the AS 2870 climate zone, whether the block sits on a floodplain or a filled gully, likely groundwater depth and what nearby logs show. From that it can suggest roughly how many holes, how deep, and whether to budget for shrink-swell testing or expect soft ground.

It stops short of the compliant number for your certifier. A desktop study never substitutes for the site classification. The final call on hole count and depth belongs to the engineer who signs the classification, informed by what the rig finds.

What to do with it: run the desktop first so the drilling brief is right the first time. Knowing where the soils came from also helps the logger read them, as the Gold Coast example in what a geotechnical desktop study covers shows.

Common questions

How many boreholes does AS 2870 require for a house?

At least one borehole or test pit per building site under AS 2870-2011, and at least three where the design depth of seasonal moisture change (Hs) is 3 m or more, as in Adelaide, or where the soil profile is known to be highly variable. Most firms drill two on a standard suburban lot.

How deep does a soil test borehole need to go?

To at least 0.75 times Hs below the surface or below any known cut, with 1.5 m as the shallowest allowed depth unless the hole reaches rock. That is 1.5 to about 1.7 m in Brisbane, Sydney, Melbourne and Perth, and 3 m in Adelaide. Soft ground, fill and Class P sites can need much deeper holes.

---

Before you brief a driller or price a soil test, generate a desktop report for the address to see the geology, nearby bore logs and indicative soil reactivity the investigation should be built around, 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, Victoria or South Australia.

Sources

• Jubert Pineda, Laxmi Prasad Suwal and Richard Kelly (2014). Sampling and laboratory testing of Ballina Clay. Australian Geomechanics, Volume 49, Number 4 (Dec 2014). geomechanics.org.au
• Li, J., Zou, J., Bayetto, P. and Barker, N. (2016). Shrink-swell index database for Melbourne. Australian Geomechanics, Vol 51 No 3. geomechanics.org.au
• Hargreaves, B. (2005). 20 years of AS 2870 in South East Queensland: a personal view. Australian Geomechanics, Vol 40 No 3. geomechanics.org.au
• Kelly, R. B., Drechsler, M. and Goldsmith, R. (2020). Connecting geotechnical investigations with project risk. Australian Geomechanics, Vol 55 No 1. 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