Earthquake-Resistant Submarine Cables: Engineering for the Ring of Fire

The 2006 Hengchun earthquake severed nine cables south of Taiwan in a single afternoon, knocking out internet service across Southeast Asia for weeks. That event, more than any other, forced the submarine cable industry to confront a question it had largely ignored: what does it mean to build cable infrastructure that survives the seafloor moving?

This article examines how submarine cables are engineered for seismic environments — the failure modes that actually break them, the materials that resist abrasion when the seabed slides, the route choices that avoid known fault zones, and the case studies of cables that broke (or didn't) when the earth shook. GeoCables continuously monitors more than thirty cables in the Pacific Ring of Fire through RIPE Atlas measurements; the patterns we observe inform much of what follows.

What Actually Breaks a Cable During an Earthquake

Direct seismic shaking is rarely the cause of cable damage. The fiber and copper inside a deep-sea section can withstand decades of vibration without measurable signal degradation. The real killer is the submarine landslide: a sudden mass of sediment, sometimes hundreds of cubic kilometres, sliding down the continental slope in response to seismic shock. A passing landslide of even modest size shears any cable in its path with overwhelming mechanical force.

The 2006 Hengchun event is the canonical case. The earthquake itself did little to the cables. The submarine slope failures it triggered, however, cut cables along their entire run — not at one point, but in seven separate locations across multiple systems. Repair operations took 49 days because the cable ships had to wait for the seafloor to stabilise before they could even locate the breaks.

The Noto entry is the most informative. A magnitude 7.6 quake on the Sea of Japan side, in a region with several international cables, produced zero international cable damage. Two factors explain it: the fault was on the landward side of the cable corridors, and the Sea of Japan continental shelf is shallow enough that landslide-generating depths are largely absent. Geography sets the limit on what engineering can achieve.

Materials: What Stops a 100-km Underwater Landslide

The honest answer is: nothing. No realistic cable design survives a direct hit from a major submarine landslide. What engineering buys you is resilience to smaller events and predictable failure modes for the large ones.

Modern submarine cables are built from concentric layers, working outward from the optical core:

• Optical fibres — typically 8–24 fibres in the core, single-mode glass.
• Steel strength member — a stranded steel wire that absorbs longitudinal tension during laying and recovery.
• Copper conductor — carries 3–15 kV DC to power amplifiers along the cable.
• Polyethylene insulator — high-density PE rated for 8000 m water depth.
• Aluminium water barrier — prevents hydrogen ingress, which degrades fibre over 25-year life.
• Steel armour — single, double, or rock-armoured wire layers, applied selectively.
• Polypropylene yarn jacket — abrasion protection on armoured sections.

The armour layer is where seismic engineering happens. In deep-water sections (below ~1500 m), the cable is unarmoured — a thin, flexible 17 mm tube. The deep ocean is mechanically benign and armoured cable would be unnecessary weight. In shelf and slope sections (0–1500 m), single armour is standard. In zones with documented landslide risk, double armour or rock-armoured cable is specified — typically increasing diameter to 50 mm and crush strength to over 60 kN.

The JGA North and Apricot systems both use double-armoured sections through their Sagami Trough crossings. The trough is one of the world's most seismically active subduction zones. The cables run through it because there is no alternative route between Tokyo and the Pacific basin.

Design Principles: Routing Around Faults

Where you can't engineer through a hazard, you route around it. Modern subsea cable routing involves multibeam bathymetric surveys at sub-metre resolution, side-scan sonar mapping of the seabed, and historical earthquake catalog analysis for every kilometre of the proposed track. Three principles dominate:

1. Avoid the 500–2000 m depth band on subduction-zone slopes. This is the depth range where slope failures most often originate. Where transit through this band is unavoidable, the cable is buried where the seabed permits.

2. Cross faults at high angles. A cable running parallel to a fault line is exposed for kilometres. A cable crossing it perpendicularly is exposed for metres. Even on active faults, perpendicular crossings have shown low historical fault rates.

3. Build redundancy at the system level. Japan's 70 landing stations are not duplication for its own sake — they are a routing topology that lets traffic route around any single broken cable within 50 ms of failure detection. The continent's connectivity does not depend on any one cable.

Testing: How You Validate a Cable for Seismic Service

Cable systems are validated against IEEE 1631 (mechanical) and IEC 60794-3 (optical environmental) standards before deployment. The relevant tests are:

• Tensile load to break — the cable is pulled until the steel strength member fails. Modern deep-sea cables are rated to 30–60 kN.
• Crush resistance — a roller applies up to 5 kN/cm; the cable must continue carrying signal during and after.
• Bend cycle fatigue — the cable is bent through its minimum bend radius 1,000+ times.
• Hydrostatic pressure — testing to 1.25× rated water depth for 24 hours.
• Repeater housing pressure — titanium housings tested individually to pressures equivalent to 10,000 m depth.

None of these tests reproduce a landslide. The industry assumption is: tests prove the cable is sound at deployment; landslides are an unforced loss event handled by repair logistics, not by design.

Case Studies

(...to be continued — sections on Hunga Tonga 2022 isolated single-cable nation, Tōhoku 2011 redundancy success, Noto 2024 routing-by-geography, with cable links to Tonga Cable, JIH-1, etc.)

What Operators Are Building Now

(...to be continued — Apricot routing strategy, BtoBE redundancy, dual-trough crossing requirements in Japan post-2011, Tonga's 2024 backup cable contract.)

Source Data

This article draws on RIPE Atlas measurements collected by GeoCables across more than 30 Pacific cables, public failure reports from operators, and the IEC/IEEE submarine cable testing literature. Browse all cables to see specific landing stations and measured latency for any system mentioned above.