This page
Package-wide source registry
Shows where each provider is eligible to be asked, which provider is tried first, and what kind of starting evidence it can return. A rectangle is not proof of a valid pixel or nearby sample.
How seabed evidence becomes a propagation input
OceanData finds the evidence. OceanAcoustics turns it into one complete bottom definition
The world map below answers where OceanData can query and in what order. It does not show the seabed actually used by every CTBTO path. The second half of the chain converts one usable observation into engineering properties, binds them to an exact path and frequency, and only then computes transmission loss.
1Path sample
One exact marine position along a declared source-to-receiver path, at one frequency.
2OceanData
Queries eligible providers in rank order and retains the first usable class or grain result.
3OceanAcoustics
Selects one coherent engineering tuple: compressional sound speed, density and attenuation.
4RAMTools
Compiles the ordered path samples into a range-dependent bottom tied to the path, frequency and policy.
5RAM
The Range-dependent Acoustic Model combines water, bathymetry and bottom inputs to compute transmission loss.
Two maps, two questions
Global eligibility is not a CTBTO path result
The earlier OceanData explorer has been rebuilt inside this site so it remains usable without its detached worktree or internet-loaded libraries. Its 18 provider envelopes and ranks are unchanged in OceanData 0.1.27. The completion text uses the current 0.1.27 rule.
Project result explorer
Actual retained CTBTO path queries and their transmission-loss effect
Shows the sediment evidence and acoustic properties retained for the bounded H01W and H10 paths, alongside historical-fixed and geology-informed direct RAM results.
Open the exact-path result explorerInteractive source map
Eligibility narrows the queue. A usable observation decides the handoff
Move across the map to see the candidate provider order. Regional sources are preferred where their code-level envelopes apply. If a source returns no usable value, the resolver continues down the queue.
OceanData source registry
Loading the portable source map
Current completion rule
The solver receives one internally consistent bottom, not a patchwork of fields
OceanData 0.1.27 may refine a broad class with compatible same-point evidence. OceanAcoustics then selects all three engineering properties together from one class-specific preset or mixture.
Observed evidence
Class, fractions or occasional sound speed
The provider result remains identifiable. Its class or grain fractions are evidence about the seabed, not a measured complete RAM bottom.
Modelled completion
One coherent acoustic tuple
Compressional sound speed, density and attenuation come from one compatible class or mixture. Missing fields are not borrowed from unrelated sediment classes.
Explicit fallback
Only unresolved marine geology reaches the global prior
The fallback remains labelled as a prior. Known land is rejected, and an unknown domain requires independent evidence that the path is marine.
Current CTBTO use. H01 paths use Geoscience Australia mud, sand, gravel and carbonate mixtures. Bounded H08 and H10 paths use compatible mapped deep-sea class evidence and refinement where available. H03 remains explicitly unresolved or fallback-labelled; the map does not invent local geology there.
Current policy. The definitions and their retained results remain distinguishable, but only the geology-informed seabed satisfies desired RAM coverage. Historical fixed-seabed fields remain comparison evidence and do not create a parallel computation obligation.
Source note: native CTBTO adaptation of the OceanData sediment-coverage explorer created at revision 9d8cfcc for OceanData 0.1.26. Provider ranks and envelopes were checked against 0.1.27. The current completion rule follows the OceanData 0.1.27 sediment documentation and the CTBTO geology-first policy.
What the attenuation value means
The resolver still has to choose how quickly sound loses amplitude inside the seabed
Sound speed and density set the impedance contrast at the water–seabed boundary. Attenuation describes loss after sound enters the bottom. Published values differ widely because they concern different sediments, sites, frequency bands and kinds of evidence.
EntryOriginMeaning for 5–100 Hz work
Jensen clay, silt and sandEngineering handbook presets
Simple values of 0.2, 1.0 and 0.8 dB per wavelength. Jensen calls them indicative; they are not measurements made at 5–100 Hz.
Zhou sandSandy-bottom field regression
Fitted over 50–1000 Hz. At 63 Hz it gives about 0.067 dB per wavelength, but it is a sandy-site relation rather than a universal sand law.
Strict mud familySpecific soft-mud studies
About 0.040 dB/m/kHz, or 0.060 dB per wavelength at 1495 m/s. This is a plausible scale for some muds, not a global mud constant.
Hamilton low-frequency examplesSparse reported measurements
Only one printed value lies inside 5–100 Hz: sandy shale at 12.5 Hz and 0.001 dB/m. The other printed examples are at 150 and 600 Hz.
Stoll Figure 5Physical model fitted to mud data
The measurements are at 4–48 kHz. The line drawn through 5–100 Hz is a Biot–Stoll model continuation, not a low-frequency measurement.
Hamilton grain-size familyOceanAcoustics empirical relation
Supported mainly from 3.5–100 kHz. A value evaluated at 63 Hz is an extrapolation beyond the evidence band.
Current OceanAcoustics mud bridgePackage construction
A Jensen-normalized engineering scenario. It is almost Jensen-sand loss in native units, not measured low-frequency mud attenuation.
How the current mud bridge is built. A strict mud value of 0.0598 dB per wavelength is divided by Zhou sand at 63 Hz, 0.0669 dB per wavelength. OceanAcoustics applies that ratio to Jensen sand: 0.8 × (0.0598 / 0.0669) = 0.715097 dB per wavelength.
How to read it. The calculation preserves the mud-to-sand ratio while moving both onto Jensen’s larger engineering scale. In dB/m/kHz, the bridge is 0.478 and Jensen sand is 0.485: only about 1.3% apart. The bridge is roughly twelve times the strict 0.040 mud coefficient. This report therefore calls it a Jensen-normalized mud scenario, not a measurement or a universal sediment law.
This clarification changes the interpretation of a model input, not any retained transmission-loss field. The technical companion gives the direct Hamilton values, conversion assumptions, source boundaries and evidence hashes.