When the Grid Fails: Building Resilient Comms for a Changing Climate

In an emergency, information is as vital as water. The official advice is clear: “leave early.” But how do you act on that advice when the power is out, the mobile network is congested to the point of failure, and the emergency broadcaster’s tower has been consumed by the very fire you’re trying to flee?

This isn’t a hypothetical. As Fiannuala Morgan chillingly documented in her article, “No power, no phone, no radio: why comms dropped out during the Central Victorian fires{target="_blank”}", this is the reality for communities across Australia. The wholesale replacement of resilient copper landlines with power-dependent NBN connections, coupled with the shutdown of the 3G network, has created a communications infrastructure that is dangerously brittle in the face of climate-fuelled disasters.

The pattern repeats across disaster types. During the devastating 2022 Lismore floods—Australia’s worst ever flood disaster{target="_blank"}—official warnings didn’t reach the Northern Rivers community. If it hadn’t been for neighbours undertaking rooftop rescues, the death toll would likely have been much higher. When flood response teams were later interviewed about the Hunter Valley floods{target="_blank"}, they identified communication as “the biggest problem to solve”—mixed messages, warnings that didn’t reach residents, and road isolation that made evacuation impossible even when people finally got the message.

When the grid goes down, our modern, centralised communication systems follow. We are left isolated at the very moment we need to be connected.

From Theory to Reality

For me, this isn’t hypothetical. It’s why I moved from exploring community networks in theory to deploying them in practice.

After years working in Open Source education and online learning—building digital platforms that connected students across continents—I thought I understood resilient communications. Then I moved to a remote Australian property and faced bushfire season. Suddenly, those abstract concepts became visceral: what happens when your only warning system depends on infrastructure that burns?

This drove my deep dive into LPWAN mesh technologies. I spent years experimenting—first with Meshtastic (quickly abandoned due to security flaws and terrible routing), then ClusterDuck Protocol (promising but slow-moving), then MeshCore (solid but complex). Each taught me what truly matters when everything else fails: cryptographic security, intelligent routing, and transport independence.

That journey led me to Reticulum—and it’s why I’m writing this article.

This is not a problem we can solve by asking individuals to buy their own satellite phone or backup generator. The solution must be as resilient and as community-focused as the challenge itself. This is where low-power, long-range mesh networks come in.

Technologies like LoRa (Long Range radio) enable the creation of decentralised, off-grid communication networks. These networks don’t rely on a central tower or a connection to the mains grid. Instead, small, battery-powered devices (nodes) talk directly to each other, relaying messages across vast distances. If one node is destroyed, the network automatically “heals” by finding a new path.

This creates a digital lifeline that can survive when everything else has failed.

The Technical Solution: Reticulum Over LoRa

In my comprehensive comparison of LPWAN mesh technologies (detailed in my upcoming technical series, publishing throughout February), I evaluated four protocols across five critical parameters: range, security, ease of use, scalability, and resilience. For the specific challenge of emergency communications, one stands out: Reticulum running over LoRa radio.

Here’s why Reticulum is different:

Security by Design, Not Afterthought

Unlike hobbyist protocols where encryption is optional, Reticulum builds cryptography into its foundation. Every packet is encrypted end-to-end using Signal-grade cryptography (Curve25519, Ed25519, AES-256). Your emergency messages can’t be intercepted, spoofed, or tampered with—even if attackers have physical access to relay nodes.

This matters in emergencies. When coordination is critical and misinformation is dangerous, you need to know your messages are authentic and private.

Transport Independence

Reticulum’s killer feature: it doesn’t care about the physical layer. The same network stack runs over:

• LoRa radio (my primary deployment—nodes 5-10km apart across bushland)
• Packet radio (AX.25/APRS)
• TCP/IP (when internet is available)
• I2P (for anonymous routing)
• Even sneakernet (USB drives physically carried between locations)

This means a community network can use LoRa for local communications, bridge to internet-connected nodes when available, and automatically fail over to radio-only when infrastructure collapses. It’s genuinely resilient.

Intelligent, Self-Healing Mesh

Reticulum is designed to be an intelligent, self-healing mesh. The protocol aims to learn the network topology, measure the quality of links, and automatically prefer the best paths for traffic. In theory, when nodes fail—whether they burn, run out of battery, or are otherwise destroyed—the mesh should automatically heal, finding new routes within seconds.

While my own testing of these features has been very promising, it’s important to be clear: this is an early-stage protocol that needs more widespread adoption and rigorous testing to prove its reliability in a true crisis. The more people who experiment with and help improve the network stack, the more resilient it will become.

Real-World Performance

Using LoRa parameters optimised for Australian conditions (915 MHz, SF8, BW125), my fixed nodes achieve:

• Range: 5-10 km through bushland, 15-20 km with elevation
• Battery life: 6+ months on solar with small panels
• Latency: Seconds for single-hop, under a minute for multi-hop messages
• Reliability: 99%+ message delivery in normal conditions, 95%+ in degraded conditions

This isn’t theoretical. These are production nodes running on my properties and those of neighbors who’ve joined the mesh.

What This Looks Like in Practice

Picture a rural region—say, a valley with a dozen properties spread across 50 square kilometres. Here’s the deployment:

Fixed Nodes (6-8 units): Solar-powered Raspberry Pi units with RNode LoRa interfaces, mounted on roof peaks or hilltops. These form the backbone, always-on relays providing 24/7 coverage. Total cost per node: ~$250-350 depending on solar setup.

Mobile Nodes (per household): Handheld devices running Sideband (the Reticulum Android app) or dedicated portable units. These connect to the mesh via LoRa radio. Cost: $50-150 per unit.

Gateway Node (1-2 units): One or two fixed nodes with internet backhaul (Starlink, NBN, or even cellular where available). These bridge the mesh to external networks but aren’t required for local communication.

Sensor Nodes (optional): Low-power environmental sensors (temperature, humidity, fire danger) that automatically broadcast to the mesh. These provide early warning data even when human communication is impossible.

It is crucial to note that the hardware ecosystem for Reticulum, while improving rapidly, is not yet at a consumer-ready level for “serious” emergency communications. The current generation of hardware is perfect for experimentation and development but requires a certain level of technical skill to deploy reliably. This is precisely where innovative designers, tinkerers, and local hacker/makerspaces can play a vital role in testing, refining, and developing the robust, user-friendly devices needed for a true crisis.

Total investment for a 10-household community: $3,000-5,000 for complete coverage. Compare this to the cost of a single Starlink dish ($600 + ongoing fees) multiplied across all households, and the economics become clear.

During normal times, this network handles routine communication, sensor data, and coordination. During emergencies—when power fails, internet goes down, and mobile towers are congested or destroyed—it becomes your lifeline. Messages still flow. Sensors still report. Coordination still happens.

Relying on centralised, corporate-owned infrastructure for our safety in an increasingly unstable world is a dangerous gamble. The technology to build our own resilient, community-owned alternatives already exists. It’s time we started using it.

• Resilience
• Emergency Communications
• Climate Change
• Lora
• Reticulum
• Community Networks