[As published in Summer 2025/26 BayBuzz magazine.]
As we approach the third anniversary of Cyclone Gabriel, Keith Newman investigates whether updated warning and monitoring technology is sufficient to ensure we will never again be caught unaware, unprepared and under water.
Human and technological early warning systems proved horribly inadequate in February 2023 when a sustained deluge of wild water cut a devastating swathe through parts of Hawke’s Bay, taking thirteen lives, while key agencies were still debating whether to call an emergency.
Then the lights went out with mass power and communication blackouts. The Redclyffe substation was underwater with outages leaving 41,000 households without power, some more than a week.
Most cell sites in the region (80%) were out of action. No phone coverage, no internet or texting and no significant back up plan.
The scalding Gabrielle review (The Transformation Strategy) showed we were ill-prepared, didn’t cope well, Civil Defence and Emergency Management (CDEM) staff were overconfident, and a complete overhaul was needed – 70 plus recommendations. Good people were set up to fail, stronger connections, greater trust and better collaboration was needed, with local voices empowered.
Here are the technology ‘fixes’ underway so far to improve our safety and resilience.
Protecting the power
With power and communications systems blacked out and no adequate back up across multiple systems, communities were left to fend for themselves when the lights went out.
Today dozens of Community Emergency Hubs have become a mainstay for future preparation as councils engage and resource and train them, including providing radio equipment. Some have settled on private access to Starlink units.
Transpower and Unison moved quickly to address their inadequacies. The switchyard of Redclyffe substation was improved to protect equipment and the partially submerged control room elevated above flood levels.
A new high-capacity interconnecting transformer will nearly double input from the Transpower national grid into Hawke’s Bay and Gisborne as part of a ‘proposed’ $47 million substation upgrade. The boost from 120MW to 250MW is expected to be completed by December 2027.
Shane Briggs, director of HBCDEM says co-ordination, collaboration and information sharing has been strengthened with councils and utilities with two projects coordinated by the Lifelines Group set down for 2026:
• Priority Transport Routes will assess the vulnerability of primary road and rail access across the wider region, and
• Vulnerability Assessment will improve resilience across utilities including transport, fuel, electricity and communications.
Alerts strengthened
HBRC, in collaboration with local councils and emergency services, is the main responder for monitoring and reporting on weather and river conditions to assist Civil Defence with evacuation and early warning alerts.
The main telecommunications providers have rebuilt and improved the emergency audible mobile alert (EMA) system which sends evacuation details to mobile phones. Briggs is confident it’s up to the task as a national warning system.
Other disaster resilient technologies will keep communication open for emergency services. There’s now a secondary network between HBRC’s Napier and Hastings offices and a back-up ISP (Internet service provider) plus access to Starlink satellites in case the fibre path or other connections are compromised.
A mesh node off-grid decentralised network of small solar-powered devices will ensure essential staff can still send text messages, when all other networks are down.
HBRC also has access to high-resolution aerial and satellite imagery for rapid disaster assessment and recovery planning with detailed before and after views of affected areas.
Telemetry system crash
The telemetry network is the backbone of HBRC’s environment and climate monitoring system, with 160 sites broadcasting live data including rainfall, river flows and levels, groundwater, humidity and temperature, wind speed and direction and air quality.
During Gabrielle whole sections of the system crashed. The Kahuranaki repeater station for Civil Defence communication failed when the power went out, the back-up generator ‘broke’ and battery back-up didn’t kick in early on February 14.
That meant 54 telemetry sites were offline and live data about rainfall and river levels wasn’t transmitted to Esk Valley, Dartmoor and other locations who had little warning of the inundation coming their way.
The Kahuranaki equipment, sold by HBRC in 2011 to a communication provider that didn’t maintain it, has since been reclaimed.
HBRC chief executive Nic Peet says the hydrology team has repaired damaged telemetry sites and is strengthening, expanding and upgrading its reporting stations with back-up sensors, power supplies, additional rain gauges and access to radio, satellite, Internet of Things (IoT) and 5G communications.
Around $3 million is being invested on 30 plus sites at $16,000 each or $200,000 for more complex, instrument-heavy installations with work expected to be completed around June 2026.
The refreshed telemetry network will gather better, faster data on rainfall, rivers, and groundwater across Hawke’s Bay so “we can respond more quickly to changing conditions and make more informed decisions for our communities and environment,” says Peet.
The upgrades include 19 telemetry water level and rain sites, seven with trail cameras. It had planned 27 new river cameras recording hourly water but that’s been stalled after a spate of vandalism. They’re awaiting a more secure solution.
“We have added Starlink satellite backups to our on-call flood ute and to our backup flood server to ensure critical technology remains online in the event of a network outage,” says Peet.
Drones are also being used to measure current water flow and check sites and trigger thresholds have been adjusted and clearer communication established with Civil Defence.
Extreme weather resources are available through a dedicated hub including images from the river cameras.
Combined data from the cameras and rainfall forecasts will form part of “complex flood forecasting models” through an improved working relationship with MetService.
NIWA (National Institute of Water and Atmospheric Research) has also developed microsites for rapid impact assessment frameworks to model natural hazard exposure and improve forecasting of severe rainfall, storms, flooding and coastal inundation.
Cellular rethink
Most of us rely on cellphones to connect, but cellphone towers depend on wireless coverage and fibre optic backhaul into the network. Consequently there was dead air for thousands during Gabrielle.
Jon Brewer author of the Improving Telecommunications Resilience Report (July 2024, BayBuzz reported here and here) warned significant improvements including better battery back-up at cell sites were needed to avoid future multi-day outages.
Providers often shared common routes and even the same fibre cable infrastructure, increasing exposure to flood, landslides, earthquake and liquefaction.
Part of the resilience plan is to move the physical network including fibre optic cables away from bridges and vulnerable areas to tall poles upstream, or under riverbeds where feasible.
Brewer remains cautious about “commodity infrastructure” like cell towers and fibre backhaul for cameras, river level monitors and weather stations. In some cases these are now operated on independent infrastructure.
No cellular hub sites had permanent generators and only some cell towers had 3-8 hours battery back-up and relied on generators which may not be available in an emergency.
The Resilience Report commissioned by Hawke’s Bay Regional Economic Development Agency (REDA) and the Hawke’s Bay Regional Recovery Agency (HBRRA) called for trunk microwave towers, gateway sites and locations used by multiple carriers to have emergency power for one or two weeks.
Battery bickering
The Report urged the big three mobile operators, OneNZ (former Vodafone), Spark and 2degrees and the joint venture Rural Connectivity Group (RCG) to install up to 48 hours backup for 167 towers providing 700 MHz coverage across Hawke’s Bay.
Brewer insisted this was a fair standard for other local and international systems, however the idea of investing around $20 million didn’t go down well with the industry’s Telecommunications Forum (TCF).
TCF contended the recommendation was well-meaning but not efficient or cost-effective at $90,000-$180,000 per site and cited restrictive regulatory conditions for battery units (up to the size of a shipping container) and preferred generators.
Brewer claimed the initial knee-jerk reaction from TCF was from non-technical managers who hadn’t kept up to date with technology. “The math doesn’t lie – when it’s practical and cost-effective to drive a few hundred km in an electric car, it’s practical and cost-effective to kit out the cell towers communities rely on with 48 hours of standby power.”
He wondered whether it wasn’t time for the government to require carriers to become more resilient and for councils to develop policy guidelines and set new resilience requirements for preferred telecoms providers.
TCF also balked at his suggestion all towers should also have satellite backup, “it’s such an obvious and inexpensive thing to do it wasn’t even worth mentioning in the report”.
TCF’s CEO Paul Brislen, claims batteries are a bit of a red herring, good for “a few hours at most”. Generators were the answer, although switching off data and prioritising voice, text and 111 calls can potentially extend battery life for 12-24 hours.
He says carriers try to avoid batteries in streets or public spaces as they need ‘noisy’ cooling, present an accident risk and the space allocated is limited by law although the TCF is working with MBIE and MfE to update this.
Brislen says a new game in town will further address resilience. “Low Earth Orbit (LEO) satellites are starting to offer direct to cell (D2C) services, meaning you’ll soon be able to send TXT messages and hopefully soon make voice calls via satellite … even if the towers are without power.”
He’s hopeful the government will address the need for 111 call centres to accept emergency txts in the near future.
Human factor
Former Civil Defence group controller Ian Macdonald told the Coronial Inquiry a common cross-agency operating platform (COP) with a live overview co-ordinating and managing resources in emergencies could have been “the most important tool in the system” during Gabrielle.
Bringing information together from all agencies including Fire and Emergency with the ability to see where 111 calls were coming from, would have provided pivotal intelligence for answering the ‘what if’ questions.
Macdonald said authorities had known for 20 years that the lack of a common operating platform posed a big safety risk. It was proposed after the 2011 Christchurch earthquake, reiterated in a 2018 government report which put plans for a COP technology on fast track but dumped in 2020 amidst Covid budgetary debates. There’s still no centralised emergency platform.
Arguably current technical responses – next generation technology aided by artificial intelligence (AI), and the Internet of Things (IoT) – mean we are better connected and protected from inadequate weather forecasts, back-up sensor and telemetry failure, and electrical and comms breakdowns.
The big challenge however is to ensure this improved technology infrastruture to assist critical decision making isn’t let down again by a lack of standard operating procedures and humans trying to outguess each other in an emergency.


This is an excellent informative article by Keith Newman – particularly on the early warning systems that proved quote: “horribly inadequate in February 2023”. I followed river levels at various sites where rivers reached extreme levels and caused bridges to collapse and gauges to stop recording. I was familiar with most of these bridges and knew extreme rainfall and normal water ‘runoff’ does not and should not destroy bridges.
Unfortunately, the Regional Council overlooked key information on the extreme flooding incident at the Redcliffe Substation that caused 41,000 households to be without power. It was not just the major flood event that caused ‘the lights to go out’ because the site of the Power Station is 16m above MSL and 5m above the adjacent riverbed of the Tutaekuri River (at 11m above MSL). The elevation of the site (which is just 10kms from the river mouth) should never be exposed to flooding due to the lack of drainage.
Without ‘forestry slash’ blocking and collapsing the Puketapu River that released a wall of water around 6am and without the same ‘slash’ reaching and blocking the Redcliffe Bridge until it collapsed at 11am, the normally ‘high and dry’ Substation would not have been exposed to such devastation. With better river management – the raising the equipment at the Power Station could prove to be an expensive pointless exercise.
After many years working heavy equipment and a fleet of trucks on the three local rivers for the HB Catchment Board back in the 1970’s, many involved back then would agree river management has been lacking for decades and also agree this disaster was preventable.
The Regional Council should have protected our environment and had rules in place for modern harvesting methods that produce the huge volumes of ‘ready for export’ pine logs. This would ensure all off-cuts, stripped branches and reject logs are turned into chips instead of left on hillsides where they can enter streams, then the rivers and eventually end up on the seafloor.