HazardNet is New Zealand’s first temporal hazard intelligence platform. Our quantum-entangled flux capacitor array observes earthquakes, tsunami, severe weather and volcanic activity up to 365 days before they happen, then presents them in a dashboard calm enough for a board meeting.
Live
Real geography, invented readings. Every marker sits on an actual place in Aotearoa; every reading arrives from a date that has not happened yet. Toggle hazard layers, click a site, and enjoy the serenity of knowing too much.
Seismic pre-echoes, wave intent, cloud sentiment and magma temperament, overlaid on the places they’ll eventually concern.
Each class has its own sensor family, its own vocabulary and its own flavour of plausible nonsense. All of them share a forecast skill score that conventional science describes as “not how percentages work”.
Seismic
We monitor sub-surface pre-memories: strain signatures a fault produces while remembering an earthquake it hasn’t had yet. Coverage includes the Alpine Fault, the Wellington Fault and anything in Canterbury that feels unresolved.
Marine
Shoreline intuition sensors along the Hikurangi margin read wave intent: the hydrodynamic equivalent of a sea that has started clearing its throat. Long or strong? Our array already knows which, and declines to say.
Atmospheric
Conventional models lose skill after about seven days. We bypass the problem by reading cloud sentiment directly from next season. Atmospheric rivers are detected at the moment they book accommodation.
Volcanic
Across the Taupō Volcanic Zone we model magma temperament and geothermal over-sharing. Our sensors can distinguish a volcano that is dormant from one that is merely being polite about it.
Everything about the QEFC-3 sounds just close enough to real physics to survive a product demo, before wandering gently off into the impossible. Here is the honest version, which is also the dishonest version.
Each HazardNet site hosts a pair of flux capacitors. One sits in the present. Its partner is temporally displaced by up to 365 days using a technique we call retrocausal phase stretching, which is a bit like pulling a rubber band through a calendar. Because the pair remain quantum-entangled, any hazard that perturbs the future capacitor perturbs the present one instantly.
A decoder then reads those perturbations and classifies them by hazard class. The decoder is very good. The physics are, strictly speaking, not available.
Two capacitors, one shared quantum state, 365 days apart. Measure one and the other remembers it, retroactively.
Three plasma conduits meet at a single node. 1.21 GW is required, for reasons we have stopped questioning.
A Novikov-compliant buffer discards any observation that would prevent itself. So far it has discarded three forecasts and one intern.
End-to-end latency: approximately −365 days.
| Parameter | QEFC-3 Kōtuku |
|---|---|
| Temporal reach | 1 – 365 days (configurable; longer horizons void warranty) |
| Entanglement fidelity | 99.97% (Bell-verified, self-reported) |
| Flux threshold | 1.21 GW nominal |
| Cryostat base temperature | 12 mK |
| Paradox handling | Novikov-compliant buffer, 3-branch quorum |
| Causality violations | 0 confirmed (4 under review) |
| Timekeeping | NZST/NZDT, Pacific/Auckland, bidirectional |
| Calibration medium | Espresso, double, no sugar |
| Form factor | 42U rack, cylindrical, mildly ominous |
Specifications are reproduced from a brochure the array printed for itself last Tuesday. Values have not been independently verified, because independent verifiers keep arriving before they are invited.
A known display characteristic. Engineering describes it as “aspirational” and has no plans to change it.
Every time someone presses “Collapse wavefunction”, the forecast changes. This is physics, not a bug, and also a bug.
Because the array is always right, verification is unnecessary. Because it is fictional, verification is also impossible.
Choose a region, a hazard class and a horizon up to one year out. HazardNet collapses the wavefunction and produces a polished advisory with enough confident jargon to unsettle a steering committee.
HazardNet has located a cluster of improbably informed hazard signals and converted them into a useful-looking national advisory.
The forecasting surface suggests multiple hazard classes are coordinating their schedules with unusual professionalism.
When you can see a year ahead, percentages stop being a limit.
How well this timeline agrees with the other two we checked.
Probability that publishing this advisory prevents it.
Less frantic sci-fi, more neatly packaged nonsense with just enough structure to get through a procurement process.
HazardNet establishes a low-latency relationship between the present day and several lightly supervised versions of the future. Latency is negative, which our network team finds upsetting.
The decoder listens for the early emotional states of faults, weather systems, volcanic reservoirs and coastal wave fields, then translates them into hazard classes, places and dates.
Results are rendered as a clean national dashboard with charts, metrics and a confidence score so high that it loops back around to being a warning sign.
Answered in advance, obviously.
Join the waiting list and we’ll let you know when HazardNet launches in your region. Given our latency, you may already have received the email.
