Days, not decades: Auckland already has the data to settle its bridge argument
- Grant McLachlan
- May 13
- 7 min read
The Mayor and the Transport Minister are arguing about whether to put a bridge or a tunnel parallel to the existing harbour crossing. They could have an evidence-based answer in a week. The instrumentation is already deployed across the city. The data is being collected in real time. Nobody is choosing to ask it the question that matters.
Maps drawn by men who never lived there
In late 1915, with the Ottoman Empire collapsing, Mark Sykes and François Georges-Picot took a coloured pencil to a map of the Middle East and drew a line from Acre to Kirkuk. To the north, the French. To the south, the British. The men drawing the line had never lived in the territory it crossed and did not consult the people who did. The line they drew is still being fought over, in some form, eleven decades and four major wars later. Winston Churchill inherited the British side of the bargain at the Colonial Office in 1921 and redrew the borders again over forty days at the Cairo Conference, producing modern Iraq, Jordan and the British Mandate of Palestine. He, too, was working from maps.
Auckland is engaged in a faintly comic miniature version of the same exercise — except that, unlike Sykes and Picot, Auckland’s decision-makers have every piece of the relevant data already within reach.
This week, the Mayor of Auckland, Wayne Brown, told the country he understood bridges “better than professors of bridges” and called the Government’s plan a “bridge of fools.” The Transport Minister, Chris Bishop, told Newstalk ZB that a final decision between a bridge and a tunnel — both, by either party’s description, running close to the existing crossing — would come down by mid-year. Neither man has been able to answer a much simpler question: Where is the congestion, and what causes it?
Where the congestion actually is
Auckland’s worst traffic is not on the harbour bridge.
It is on State Highway 1 at Sylvia Park, where the southern motorway and the eastern catchment funnel into the same six lanes. It is at the Spaghetti Junction interchange, where SH1 meets SH16 and the freight corridor to the Port. It is on SH1 between the Harbour Bridge and Esmonde Road. Auckland Transport’s own modelling, conducted by EY and Arup, puts the economic cost at $2.6 billion a year by 2026, with Aucklanders collectively losing 29 million hours to delays. TomTom’s Traffic Index has Auckland’s drivers losing more time in rush hour than Sydney or Melbourne.
None of those choke points sit on the harbour bridge itself. They sit on the corridor the harbour bridge feeds into. Adding a parallel crossing to that corridor puts more vehicles into the same downstream bottleneck. A second crossing built where the Mayor and the Minister are arguing about it would, on the existing infrastructure, make Sylvia Park worse — not better.
The motorway that was designated in the 1950s
The route that would actually relieve those choke points has been designated for seventy years.
The Eastern Transport Corridor, drafted in the 1950s and protected on Auckland’s planning maps through every change of government since, runs roughly twenty-two to twenty-five kilometres from the Onehunga waterfront, along the southern edge of the Tamaki estuary, through Sylvia Park, up the rail corridor through Panmure, across Hobson Bay, and under the harbour to emerge near Bayswater. Much of the land is publicly owned. Some is rail corridor. Some is foreshore reserve. Its proper number is State Highway 17 — a designation NZTA quietly revoked in 2012 and has not reassigned in the thirteen years since.
A six-lane SH17, fully tunnelled where the geography requires, costs between $12 and $18 billion. The Waterview Tunnel precedent — which diverted seven percent of traffic off SH1 the day it opened in 2017 — suggests SH17 would remove up to 15,000 vehicles a day from the country’s worst pinch point. The Government’s current proposal is $35 to $45 billion to put a parallel bridge or tunnel next to a bridge that is already there. We are choosing to spend more for less.
The data is already being collected
Here is the part neither the Mayor nor the Minister will admit. Almost every input needed to model the question is already being collected in Auckland, in real time, by agencies they oversee.
Auckland Transport operates more than 4,870 cameras connected to the vGrid network, including 500 with automatic number-plate recognition installed in the last five years. Its parking-enforcement fleet of 25 ANPR-equipped vehicles scans roughly half a million number plates a day across the city’s streets. Since 2021, Auckland Transport has had a partnership with TomTom that provides consistent, granular, citywide real-time and historical traffic data — replacing what AT itself described as “fragmented” local datasets. AT has been publishing downloadable traffic-count records for every monitored site in the region since 2012; the current spreadsheet runs to 6.5 megabytes and covers July 2012 to September 2025.
That is just Auckland Transport. NZTA operates a public Auckland Traffic API that returns real-time congestion levels, travel-time estimates between intersections, variable-message-sign status and a live feed from over 100 traffic cameras. NZTA is expanding its national camera network from 150 to 800 and has previously used point-to-point ANPR for speed-corridor surveys. On the mobility side, Spark NZ provides anonymised, aggregated mobile location data — held for up to seven years — to its wholly-owned analytics arm Qrious, which already sells journey-pattern intelligence as a commercial product. Two New Zealand-built ANPR platforms, Auror and SaferCities, host data from between five and ten thousand cameras nationally and are queried around 600,000 times a year by police alone.
Auckland Transport even runs its own public developer portal with 35,000 free API calls a week available to anyone who registers.
This is not a city short of data. This is a city sitting on a comprehensive observational dataset — and not asking it the question that matters.
An experiment that would take days, not decades
Here is what a competent week of work would look like.
Combine six months of AT’s existing TomTom feed with the AT count data and the vGrid ANPR layer. Add a sample of Qrious’s anonymised cellphone mobility records for the morning and evening peaks. Strip identifiers at source. Build an origin-destination matrix for every harbour-crossing trip and every SH1 trip through Sylvia Park, Spaghetti Junction and Esmonde Road on observed movement — not on the inferred trip-generation rates that every previous business case has relied on.
Now bring in the modelling. Feed those observed flows into a calibrated transport simulation — AIMSUN, VISUM, or a modern machine-learning equivalent are all off-the-shelf, used routinely overseas. Test the counterfactuals one by one. What happens to Sylvia Park if SH17 exists? What happens to Esmonde Road? What happens at Spaghetti Junction? What happens if the existing Harbour Bridge is tolled at $5 a crossing while SH17 is being built? What happens if it is tolled at $10? Every one of those questions can be answered against the actual movements of actual vehicles, not against a planner’s assumption about what a hypothetical commuter ought to be doing.
A competent contractor working on that brief — given the data is already there, the simulation tools are already there, and the route corridor is already designated — could deliver a defensible, peer-reviewable answer in weeks. Six months at the outside. The cost is well under what the NZTA board’s next ten meetings will spend on venue hire and Cabinet-paper drafting.
Compare that to where Auckland actually sits. Thirty years of business cases. The eleven-option study of 1997. Labour’s five-option suite of 2023. The 2025 announcement that committed $35 to $45 billion to a parallel crossing without knowing where the bridge’s 155,000 daily users actually start their day.
That is not a planning failure. That is a refusal to find out.
What competent looks like elsewhere
In the same fifty years that Auckland has spent producing reports, Sydney has spent building harbour crossings.
The Sydney Harbour Tunnel was proposed by Transfield and Japanese contractor Kumagai Gumi in 1984, signed off by Cabinet in 1987, broken ground in 1988 and opened in 1992 — four and a half years from contract to opening, A$750 million, privately financed, on budget. Sydney is now simultaneously building a third road crossing and a second rail crossing. By 2032 it will have five harbour crossings. Auckland, on current trajectory, will still be deciding whether to start.
It is not as if Auckland cannot do this when forced. The Nippon clip-ons that doubled the harbour bridge’s capacity in 1969 were prefabricated in Japan and came in under budget. The second Mangere Bridge was delivered seven months ahead of schedule in 2010 under the MHX alliance model. The institutional knowledge exists. The recent muscle memory exists. What is missing is the political will to apply it to the right project.
What is actually being decided
The Mayor and the Minister are not negotiating about engineering. They are negotiating about who gets to claim the announcement. The Auckland City Deal signed in April requires the two to coordinate. That is the constraint Wayne Brown is invoking when he says decisions cannot be made by June. It is a procedural objection to a political schedule, not an engineering objection to the underlying plan.
Both sides will be worn down. A decision will be made too late, in panic, after another generation of further congestion. It will probably be the wrong decision, because the question on the table is the wrong question — and the means of answering the right question is being ignored by the same agencies that already operate it.
Sykes and Picot at least had the excuse of being five thousand kilometres from the territory they were drawing. Auckland’s decision-makers have not even that excuse. The cameras are mounted. The APIs are live. The data sits in databases at TomTom, Qrious, AT and NZTA, refreshed by the second. The artificial intelligence required to model the counterfactuals is the same kind that already routes Uber drivers around traffic in real time.
All of which means the experiment proposes itself. Run it. Publish the result. Re-designate SH17 by Gazette notice — a Minister’s signature, ten minutes’ work — if the model confirms what the geography already suggests. Stop pretending the question is unanswerable.
Auckland has the instrumentation of a smart city and the planning instincts of a colonial cartographer. The first could fix the second within a week. Nobody in Wellington, or in the Mayor’s office, has chosen to ask it.
Here are the results of a model based on the available data:
