Imagine losing contact with your family during a major disaster. No phone calls, no text messages, no way to know if they’re safe. This nightmare became reality for thousands of New Zealanders during Cyclone Gabrielle [1]–[3].

On 13 February 2023, Cyclone Gabrielle knocked out communications across New Zealand’s East Coast. Within 12 hours, emergency teams couldn’t reach entire towns and districts. Cell phone towers died, internet cables flooded, and radio systems went silent [1], [2]. For three days, rescue coordinators operated without reliable information about survivor locations, road accessibility, or infrastructure damage. This communication breakdown delayed rescue operations and left isolated communities completely excluded from help [3], [4].

The Problem: When Help Can’t Find You

The disaster revealed a severe weakness in how New Zealand handles emergency communications. Currently, most of our system is tied to the ground: cell towers, internet cables, and radio stations. The problem is obvious: these are precisely the things that break when storms or earthquakes hit [3].

It’s like putting all your eggs in one basket. Once that basket breaks, there’s nothing left. Our communication network works similarly: strong when intact, but fragile when stressed.

During Cyclone Gabrielle, around 225,000 homes lost power, and some stayed disconnected for several weeks. Cell towers didn’t last either, as their backup batteries drained within half a day [1]. This wasn’t new. In the 2011 Christchurch earthquake, emergency teams spent hours trying to learn which buildings had fallen, where survivors might be trapped, and which roads were usable [5]–[7]. Once again, the problem wasn’t the lack of people willing to help; it was the lack of communication to guide them. The lesson is simple: the systems we count on most in disasters are the first ones we lose [6], [8].

The Solution: Communication That Flies Above the Chaos

What if our lifeline in a disaster didn’t sit on the ground at all? Imagine if the systems that connected us could float above the floods. That’s the idea behind aerospace networks using planes, drones, and satellites instead of towers and cables [9]–[12].

The strength of this approach is in its independence. A satellite isn’t pulled down by floodwater. A drone doesn’t rely on mains electricity. From the sky, they can keep sending and receiving signals no matter what’s happening on the ground [10].

It works like building several bridges across the same river. If one is swept away, people still have other ways to cross. By spreading communication across many independent flying machines, the system stays alive even if parts of it fail. Engineers call this a distributed network, a system with many small, connected parts instead of one central hub [13]. If one link is broken, messages can still find another path through the network. The result is simple: a disaster would no longer silence whole regions [11].

How Flying Communications Would Work

To visualise how this flying network operates, it helps to think of it in layers, where each type of machine has its own job.

• The High-Flying Backbone: Satellites and high-altitude aircraft would form the main communication highways, operating far above any weather [10], [14]. These can stay aloft for weeks or even months, covering areas larger than entire regions of New Zealand. Think of them as the main motorways of our aerial communication system.

• The Middle Layer: Medium-sized drones flying at heights between 1,000 and 10,000 metres would act as regional relays. They sit above the worst of the weather but are still close enough to give sharp, detailed coverage of disaster zones [11]. These drones would relay communications between the high-flying platforms and rescue teams on the ground.

• The Local Helpers: Small drone swarms operating below 500 metres would be like mobile cell towers, flying directly to isolated communities or rescue teams [11]. A single drone can circle over an isolated town and instantly restore service, or hover near a rescue crew to keep them in touch with headquarters.

The reach of even a modest drone is striking. At just one kilometre high, it can connect to radios on the ground for tens of kilometres [15], [16]. Push higher and the coverage grows, but with limitations due to terrain, power, and line-ofsight interference. Together, these layers create a mesh of overlapping safety nets that disasters cannot easily tear apart [10], [13].

Making It All Work Together

You might wonder how all these flying platforms talk to each other and manage the communications traffic. The reality is more complex than simply connecting devices.

The answer lies in what engineers call mesh networking, a system where each device can communicate with several others, creating multiple pathways for information to flow [13]. Picture a spider web where each strand can carry information. If you cut one, the information simply flows along others to reach its destination. Our aerial network would work the same way, with each flying platform able to communicate with several others [10], [11]. But this creates a fundamentally complex system where ground-based and airborne elements must work together seamlessly. The network would need to be smart enough to prioritise different types of communications automatically [10], [12]. Emergency voice calls between rescue teams would get top priority and instant connection. High-resolution photos of damage could wait a bit longer, but would eventually get through when bandwidth becomes available [11].

Yet the machines aren’t just relays. Many drones today can do basic processing themselves. Instead of sending home a massive file of raw images, a drone might scan it on the spot and transmit only the crucial information [11]. That makes the whole system faster, lighter, and less likely to get clogged, but it also means we’re depending on artificial intelligence to make critical decisions about what information matters during life-anddeath situations [10], [17].

Keeping the System Powered and Running

One of the most challenging questions is how to keep all these machines running through days or even weeks of crisis. Fortunately, each type of platform has its own way of staying in the air.

Small drones usually run on batteries and can fly for just 30–90 minutes [11]. In practice, that means either swapping batteries constantly or setting up clever charging systems, similar to how delivery drone companies keep multiple units and charging pads ready to rotate in and out. Bigger platforms can draw power from the sun. The Airbus Zephyr, for example, managed to stay aloft for 26 days straight, powered only by solar panels [14]. During daylight, the panels run the aircraft and recharge batteries for the night. That makes solar especially attractive in places like New Zealand during long summer days.

Fuel-powered aircraft still play an essential role. They carry heavier equipment, more powerful radios, and advanced sensors, and can fly through rougher weather. The trade-off is obvious: they need refuelling, which can be difficult when roads are flooded or blocked [10]. The best solution may be hybrids that use more than one power source [10]. A drone could burn through battery power for quick takeoff and intense activity, then shift to solar mode for long, steady flights. This mix makes the system both flexible and reliable.

Smart Systems That Work Without Constant Human Control

In a disaster, emergency staff already face a flood of decisions. The last thing they need is to micromanage dozens of drones and satellites. That’s why the network must be smart enough to handle routine tasks independently, while still leaving humans in charge of key choices [11], [17].

Modern drone swarms can already fly in formation, avoid collisions, and spread out to cover large areas without someone at the controls of each unit [17]. If one drone drops out, the others shift their positions automatically. Even if a storm cuts off contact with human operators, the drones can keep following their pre-set missions, adjusting to conditions as they go [12], [17]. For the people in charge, the system should highlight only what matters. Instead of screens filled with dozens of drone status bars, managers would see a simple picture of overall coverage, with alerts only for urgent problems [12]. From there, they could set broad goals such as “restore communication for this community” or “scan this sector for survivors”. The machines handle the details while humans guide the mission [11], [17].

Real-World Challenges We Must Overcome

Of course, it’s not as simple as just launching drones into the sky. A system like this faces real hurdles. Radio space is crowded, and every drone or satellite needs a slice of it [15]. Bad weather, the very thing we’re preparing for, can ground smaller craft [11]. And then there are the rules: aviation approvals are slow and built for routine flights, not for the urgency of a cyclone response [18], [19]. Cybersecurity adds another layer. If communities depend on these flying networks, attackers might try to jam or hack them [17]. Strong protections are vital, but they can’t get in the way of fast-moving rescue work. Integrating existing emergency management systems requires careful attention to data formats, communication protocols, and operational procedures. Emergency managers need training on how to use these new capabilities effectively [9], [10].

The Limits of Engineering Solutions

However, we must be clear about what this system cannot do. Aerial networks represent an engineering approach to resilience, but they cannot guarantee safety or communication in the broader social and cultural sense that truly matters during disasters [9]. Communication infrastructure is merely the conduit. It cannot ensure that people will behave as we expect them to in crisis moments, have the skills to use the technology effectively, or make rational decisions under extreme stress [8]. We cannot simply engineer our way out of the impacts of natural events. The system’s complexity introduces new vulnerabilities: multiple technical systems could interfere with each other, competing drone operators could crowd the airspace, and the very decentralised structure that makes it resilient also makes it unpredictable [10], [17]. Moreover, satellites and high-altitude platforms raise thorny questions about data sovereignty and geopolitical dependence [8]. New Zealand’s emergency communications could become reliant on foreign-controlled satellites, creating new forms of vulnerability that transcend natural disasters [9], [14].

A 50-year contract with SpaceX or a major investment in Rocket Lab might solve immediate technical problems, but it also binds our national resilience to global power structures and commercial interests [8]. Every single component of this system, from individual drones to satellite constellations, represents a complex system in itself, each with its own failure modes, maintenance requirements, and unintended consequences [10], [12].

How It Would Have Helped During Cyclone Gabrielle

It’s worth asking what this would have meant if the system had been ready during Cyclone Gabrielle. Forecasts gave 2–3 days’ warning [1], [2]. In that time, high-altitude drones could have flown into position above the East Coast, building a backup communication spine that no flood could touch [10].

As the cyclone closed in, medium-altitude drones would have launched from safer bases outside the danger zone. Carrying radios and sensors, these craft would have linked the backbone above with the emergency crews below [11]. While in reality drone endurance in extreme storm conditions is uncertain, research suggests that higher-altitude platforms could remain aloft while smaller drones sheltered and resumed operations once winds dropped [14]. The hours immediately after landfall are often the most critical. This is when drone swarms could have flown over cut-off valleys and flooded plains, mapping damage and searching for survivors [11]. Instead of waiting days for road crews to clear access, emergency teams could have had a live overview within hours [10]. Whole towns, suddenly cut off, could have been reconnected by drones hovering overhead with cell and internet relays [11]. The network would then direct its resources smartly, sending more support to places where survivors were found or damage was most severe [12]. In short, communication wouldn’t have vanished just when people needed it most.

The Economics and Path Forward

None of this comes free, and the true costs extend far beyond initial investment. Building and maintaining an aerial communication fleet is a significant investment, but the hidden expenses lie in system lifecycle management, how these technologies are deployed, how they evolve while in service, and how they’re eventually decommissioned [8]. However, the comparison matters: hardening land-based infrastructure is also hugely expensive and fails when nature pushes hard enough [4]. Flying systems, by contrast, can be moved wherever they’re needed, but they also create new dependencies on global supply chains and foreign expertise [8].

One way forward is partnership, but this approach brings its own complexities. Many tools, such as drones, satellites, and even high-altitude aircraft, are already used by private companies for mapping, delivery, or connectivity [10], [14]. With agreements in place, these resources could be switched into emergency mode when disaster strikes [8]. However, this means New Zealand’s emergency resilience becomes intertwined with commercial interests that may not align with our national priorities during crises. The technology itself is only getting stronger, but it’s also becoming more complex [10]–[11]. Smarter AI means drones that can make better decisions independently, but it also means less human oversight and more unpredictable system behaviours [17]. Advances in batteries keep them in the air longer, but they also create new environmental and resource dependencies [11]. Manufacturing techniques are speeding up, so specialised craft could be built for particular emergencies, but this rapid evolution makes long-term planning and investment decisions increasingly difficult [12].

These systems don’t exist in isolation. They’re part of local and global infrastructure networks that operate across different timescales and serve multiple functions beyond disaster response [8]. The same satellite constellation that provides emergency communications might also support commercial internet, military surveillance, or agricultural monitoring [14], [20]. What seems like a focused emergency tool becomes part of a much larger, more complex web of technological and political relationships that New Zealand cannot fully control.

The Choice Before Us

Cyclone Gabrielle made one thing painfully clear: New Zealand’s communication lifelines are far too fragile [1]–[3]. Aerospace networks offer a way to change that, building resilience in the sky instead of depending only on the ground [10], [11]. But this choice is not simply technical. It’s about what kind of dependence we’re willing to accept and what new vulnerabilities we’re willing to create [8].

Yes, the engineering challenges are serious, but they’re not impossible. The pieces fit together with the right mix of autonomy, reliable power, weatherproofing, and strong partnerships. However, what’s needed most is not just commitment from the government, researchers, private operators, and emergency managers, but also honest recognition that this approach has limits [9]. We need behavioural change from communities, realistic expectations about what technology can deliver, and careful consideration of how increased system complexity might create new failure modes we haven’t yet imagined [12], [17].

New Zealand is a perfect testing ground precisely because our size makes full coverage realistic, and our isolation means fewer conflicts over crowded skies [8]. But that same isolation makes us more dependent on external partners for satellite coverage, technical expertise, and system maintenance [9]. If done right, we could lead the world with this technology, but we could also become more vulnerable to global supply chain disruptions, geopolitical tensions, and the commercial decisions of foreign corporations [8], [14].

The next disaster will come; it’s only a matter of when. The real question is not just whether we’ll act now to prepare, but whether we can build systems that enhance our self-reliance rather than simply shifting our dependencies from ground-based to sky-based infrastructure [10], [11]. The tools are ready, but they come with their own risks [17]. The need is undeniable, but so are the limitations of purely engineering solutions [9]. What remains is not just the choice of whether to build this system, but how to build it in ways that serve our long-term interests as a nation while acknowledging that no system, no matter how sophisticated, can eliminate the fundamental uncertainties that come with living in a dynamic, unpredictable world.