Faced with broken highways in Waikato and Wellington regions, the Transport Agency has staked $1 billion on a roadbuilding method that is difficult to get right, hard to repair and that lacks a proven track record.
The method is being rolled out in Waikato at the same time as the liability on the taxpayer when roads go wrong has increased.
"A disaster is looming," one industry insider who RNZ agreed not to name, said.
The agency is using its new method called Hi-lab for the first time, on three sections of the Waikato Expressway worth $1.1 billion.
This is being done even though a 2020 review for NZTA makes clear the jury is out on Hi-lab.
"Hi-lab basecourse pavements have supply and construction complexities which may adversely affect reliability integrity, risk and economic viability," said the review by two independent engineers, which NZTA is now using to try to improve its road building.
"They do not yet have an established long-term performance history."
Hi-lab is a road foundation that has bigger stones and less sand in it than traditional road foundations.
Hi-lab cracked in an early test in 2015, around the same time that the agency decided to roll it out en masse.
It insisted on the new design when awarding three contracts for the Huntly, Longswamp and Hamilton sections of highway, covering 170km on some of the busiest parts of State Highway One.
It had not finished trialling Hi-lab at the time, but the agency told RNZ that trials had gone very well and it was "confident in its performance".
Contractors are nervous of the complexities of the 12-step Hi-lab process and have pushed back against taking any liability for it.
On small sections at Huntly and Longswamp, they have already had to pull it up after laying it wrong, despite NZTA's supervision.
It is so hard, it has to be ground off with a milling machine, at high cost, and cannot just be dug up.
"It is designed as a one-hit solution," said another industry source RNZ agreed not to name.
"If it works well, it's fine, but if it fails you have a big problem."
Stemming highway failures
That hardness is also Hi-lab's big attraction.
It was "comparable to concrete, however is not as expensive", NZTA said.
On trials, it "has performed well, even under the same conditions that caused other pavements to fail".
It was developed, and championed strongly since 2010 by the agency's chief pavement engineer, Gerhard van Blerk.
"The theory is, we're trying to build perpetual pavements," said van Blerk.
He is a native of South Africa, where roadbuilding conditions are much drier, as they are too in Australia, where the rock for building is generally much stronger than in New Zealand.
A good highway here lasts 25 years without major repairs.
NZTA has a good reputation and has built many good roads, the industry says - but it has also built bad ones that are failing now, within five years.
- Repeated ongoing failures of road surfaces on Te Rapa and Ngaruawahia sections of the Waikato Expressway
- Repeated failure of road surfaces on the near-new Kāpiti highway, costing at least $25m.
- Having to rip up parts of billion-dollar Transmission Gully highway near Wellington.
Those huge repair bills are partly covered by contractors.
To try to find a remedy to the failures, NZTA in 2017 took over all highway pavement design, in a little-reported move that increases the liability for itself, and taxpayers, if a road goes wrong.
But they were not going wrong, not any longer, the agency said.
"The pavement issues we are currently experiencing are on projects where contracts were awarded prior to July 2017.
"On contracts awarded since, improved quality testing and monitoring is picking up risks before they become issues," the agency told RNZ.
But this assertion does not sit well with the 2020 independent review.
It recommended that the agency "review and upgrade risk profiles" of types of pavement - and specifically mentioned Hi-lab.
It detailed many weaknesses, including that quality controls had not extended into the design phase.
The agency accepted all nine recommendations in the review.
Roadbuilding weakness laid bare
Having accepted it needs to upgrade its roadbuilding abilities, it is working on that urgently with the Civil Contractors industry group, which lauded the "massive" changes that led to the 2020 review.
Yet its Hi-lab work over the last several years, laid at a cost of $25m or more per kilometre, is absolutely okay, the agency asserted in various ways to RNZ.
"Pavement specifications, including the Hi-lab specification, are subject to review and change as we learn more and techniques improve.
"The construction methodology has changed substantially, with the implementation of a method specification and a vastly improved quality control process during construction and production."
The period of Hi-lab's rollout beginning in 2016, till now, is bookended by the afore-mentioned 2020 review - and a 2015 report that similarly laid bare weaknesses in building new roads and maintaining old ones.
The 2015 report to NZTA said the combination of poor performance goals and price pressures "consistently produced thin, high-risk, low-cost pavements".
NZTA responded with the 2017 quality controls and design shake-up.
But the January 2020 review found, once again, that too often penny-pinching was winning out.
"Procurement tensions and Value for Money challenges have tended to minimise pavements without appreciation of the elevations in risk."
Also, it found the agency had weak risk assessment and did not investigate hard enough when things went wrong.
It identified weaknesses that could raise questions about how the Hi-lab rollout has been handled. It found:
* Technical challenge and risk evaluation of pavements was variable
* The risk of failure may not be fully appreciated and mitigated in later stages
* Environmental and climatic conditions barely influence what design to use "and this is a significant shortcoming"
The NZTA pavement guide provides "very little explanation of risk level acceptability or scrutiny".
The stakes were always high, and have risen again: More and better roads are an expectation that has been stoked by both major political parties in the run-up to the election and in their Covid-19 response plans.
Hi-lab is not NZTA's only hope. It has several design remedies for failing roads.
An economic analysis is being done to compare Hi-lab with two other methods: Foam bitumen and very expensive structural asphalt.
It is unclear why this was not done years ago, at least before Hi-lab's rollout in Waikato.
These remedies are especially vital in the North Island where there is so much weak rock that can compromise road bases.
The 2015 roadbuilding report to NZTA rated existing pavement quality at just 4.9 out of 10.
Introducing new designs was critical to raise performance, but "is generally uncharted water", it said.
NZTA charted that water for Hi-lab by testing it, first on a lane 60km long in the King Country.
Next, in tests since 2011, on four sections of the Expressway in small lengths up to 2km long - at Taupiri, Te Rapa, Ngaruawahia and Cambridge.
Hi-lab cracked at Taupiri in 2015.
But the agency said this was expected - that the land would settle, and the cracking was only superficial and easily fixed.
Core samples have recently been taken on the Cambridge highway section, but NZTA said that had nothing to do with Hi-lab.
The tests were still going on when NZTA decided to roll the design out, en masse, on the 15km-long, $384m Huntly section.
At that point, said Gerhard van Blerk, "Treasury asked us to do further testing", which he said they began.
"So it's not like the agency ... went into this thing blindly with taxpayers' money without doing any homework," van Blerk said.
"So if anything, this design has been more thoroughly tested, than probably any other new innovation."
However, when RNZ asked Treasury about this, it said that did not sound like something it would do, and "the person ... could have been mistaken in identifying Treasury as the agency involved".
RNZ has asked for Hi-lab test results.
The design also passed smallscale lab tests.
However, the 2020 review said there was a "real issue", in general, with over-reliance on controlled lab tests that did not translate to the realities of construction in all sorts of conditions on site.
Close eye being kept
NZTA is trying to balance its greater design responsibility with better construction and material quality control.
It is now fully committed to a Hi-lab design that it admits few contractors have experience of laying, but which it is keeping a very close eye over.
Even so, the design's advocate, van Blerk said: "The construction managers, they say that it's easier than traditional pavement."
Quality control was intense, he said.
These controls failed to stop two lots of Hi-lab being laid wrong, but did identify the failings later.
Construction monitoring, though, has been a weak point of New Zealand building. And the 2020 review described how it was still slow and patchy.
Hi-lab is at the extreme end of needing close control: It even needs its own quarry in Waikato to make sure the rocks are just right.
The 2020 review asked how roadbuilding risk can be managed. It said in part: Only by "adopting lower-risk pavement designs".