Aller au contenu principal
OceanData Consulting
Case study · Submarine cables

Channel cable atlas: from sediment risk to an A* corridor

A pre-survey screening study that turns 30 years of waves, currents and seabed data into a mobility atlas, scour levels and route alternatives.

All case studiesBy Matthieu Caillaud · Published on August 4, 2026
Measured results

The figures that summarize the study

500 mGrid resolution
30 years1993–2022 hindcast
51SHOM current points
908Dune crests compared

Decide where to investigate before commissioning an expensive survey

Bathymetry shows where a cable can physically pass, not whether the seabed will remain stable. Within a few hundred metres, sediment mobilisation frequency and exposure to scour can change markedly.

The objective was to produce a reproducible screening tool: classify risk cell by cell, expose uncertainty, then propose several corridors between two landfalls without presenting the result as a detailed geotechnical study.

Method

A verifiable computation chain

01

Align the sources

RESOURCECODE/HOMERE waves, Copernicus Marine currents, and EMODnet bathymetry and geology are projected onto a common 500 m grid.

02

Compute mobility

Combined wave-current shear stress follows Soulsby; the mobilisation threshold is derived from the Shields criterion and sediment d50.

03

Estimate extremes

A POT-GPD analysis produces return levels with confidence intervals instead of reducing risk to an average.

04

Route alternatives

A cost surface combines mobility, scour, traffic and exclusions. A* computes minimum-risk, balanced and minimum-length alternatives.

Validation

What the comparisons show

Current forcing is consistent with SHOM

Across 51 independent points, the median model/SHOM maximum-speed ratio is 0.98 at spring tide and 1.03 at neap tide. Five coastal points fall outside ±25% at spring tide and identify features the model grid does not resolve.

Mobility frequency carries the morphological signal

At dune crests, median mobile-time fraction reaches 0.73 versus 0.66 elsewhere on comparable sandy substrate, with p = 7.7 × 10⁻⁹. The class-based test does no better than the base rate; that negative result is retained.

Grain-size uncertainty becomes actionable

When d50 is shifted one class finer and one class coarser, 86% of nodes retain their mobility class. The remaining 14% become priority locations for sampling or grain-size analysis.

Validity domain

What this study does not claim

  • The atlas is an upstream screening tool; it does not replace coring, CPT, sonar or a detailed burial assessment.
  • Dune validation is one-way: a dune supports mobility, while the absence of a mapped dune does not prove a stable seabed.
  • Sub-grid headlands, banks and coastal features require finer models or measurements before detailed design.

Do you need to shortlist a corridor before survey?

We scope the area, available sources and validation level required for your decision.

Discuss a comparable project