Nothing transmits, and nothing here is invented.
CableWatch reads public instruments that already watch the ocean, the seabed and the Sun, then calculates what each reading means for every cable. Every number carries one of three labels.
- measured came from an instrument: a seismometer, a magnetometer, a tide gauge, a spacecraft, a ship's AIS transponder. Most are minutes old; the Feeds tab shows the age of each source.
- model came from an operational forecast model (waves, currents, wind).
- computed is physics applied to the two above: arrival times, induced voltage, how far a cable can hear.
How readings are cross-checked
- Earthquakes: the US and European catalogues locate each event independently. CableWatch matches them and shows the difference.
- Solar storms: a spacecraft 1.5 million km upstream, the planetary Kp index and ground magnetometers must agree before a storm is called.
- Tsunami: a large shallow quake, an official alert, a tide gauge off its prediction and a source fitted to the gauges alone are four separate lines of evidence.
- Each line adds decibels of evidence to a prior set by how rare the event is. 10 dB is ten times more likely.
Methods that locate and measure, not just display
- Source location on the Oregon cable: signals from the Axial Seamount seismometers are stacked along predicted travel times for every map point and three wave speeds. The speed that focuses best says whether the sound came through water or rock.
- Seabed monitoring: background ocean noise at two stations is correlated to recover the wave between them; its slow stretching tracks changes in the seabed. The public archive runs one to three days behind, so this is recent history.
- Sea-level source locator: when gauges leave their predictions, one origin and time is fitted to all of them at tsunami speed, with no input from seismology.
- Storm field: electric-field strength is fitted against geomagnetic latitude across twelve magnetometers and evaluated along every cable.
What is not connected, and why
- Strain, polarization and power-feed voltage from commercial telecom cables. No operator publishes them. For every cable CableWatch shows what it should be experiencing, computed from live inputs; the Network tab adds whether traffic is actually getting across.
- Worldwide ships need an AIS key held by the CableWatch server. Without it, ships are shown for the northern Baltic, which Finland publishes openly.
- Cosmic-ray monitors and deep-ocean tsunami buoys are not connected yet.
- Cable routes come from TeleGeography's public chart. They are drawn routes, accurate to a few kilometres, not surveyed positions; depths come from a 1° grid.
Physics used
- Wave arrivals: P at 8 km/s, Rayleigh at 3.5 km/s, tsunami at √(g × depth).
- Strain from a distant quake: surface-wave magnitude relation, converted to strain rate and compared with the published noise floor of ocean-wide fibre sensing.
- Electric field under a storm: magnetometer dB/dt convolved with the response of a uniform conducting Earth, reduced by seawater shielding with depth.
- Voltage from ocean flow: current × vertical magnetic field × cable length (motional induction).
- Hearing range: passive sonar equation with Thorp absorption, wind noise and fibre self-noise.
- Shaking at a cable: Bakun–Wentworth intensity relation, MMI = 3.67 + 1.17 M − 3.19 log₁₀(distance), at the closest point of the route. Derived for California; elsewhere it is an estimate.
- Network paths: light in fibre crosses 1 km in 4.9 µs (refractive index 1.468); a measured round trip is compared with twice the cable's length at that speed.
- Repair weather: hourly significant wave height from the Open-Meteo marine model at five points per route, against a 2.5 m working limit.