Every line of physics, every byte of bundled data, traced to its public-domain or permissively-licensed source. No GPL, no proprietary code, no licence ambiguity — just published equations and open datasets, re-implemented in Swift.
Ægir Prediction does not link Bellhop, Kraken, RAM, or any other GPL-encumbered acoustics library. The ray solver, transmission-loss model, ambient-noise spectrum, and absorption calculation are all re-implemented from scratch in Swift against the canonical published equations cited below. The bundled data is either explicitly public-domain (US-government datasets) or permissively licensed (CC BY 4.0). So you can put Ægir to work anywhere — offline, in the field, in commercial projects — with no licence-clearance friction.
The General Bathymetric Chart of the Oceans is the canonical global ocean-depth dataset, jointly maintained by the IHO and IOC of UNESCO. The "sub-ice" variant uses BedMachine and Bedmap2 ice-shelf draft estimates so the Ross and Filchner-Ronne shelves resolve to bedrock rather than ice surface — material when planning Southern-Ocean acoustics. Ægir uses GEBCO for the depth lookup at every pin and as the bottom boundary for the ray solver.
View at gebco.netNOAA NCEI's quarter-degree climatology of temperature, salinity, dissolved oxygen, nitrate, phosphate, and silicate — averaged over decades of in-situ casts and re-gridded at 1° (Compact) and 0.25° (HQ). Ægir uses the temperature and salinity columns as the input to Mackenzie's nine-term sound-speed equation and as the basis for Francois-Garrison absorption.
View at ncei.noaa.govMackenzie's nine-term polynomial relating temperature, salinity, and depth to sound speed is the workhorse formula for ocean-acoustic codes. Accurate to about 0.07 m/s over the full oceanographic envelope — finer than WOA23's underlying observations warrant, so the bottleneck is data, not formula. Ægir applies it column-by-column at every pin.
View DOI 10.1121/1.386920The Dutkiewicz dataset is a digital map of seafloor lithology — calcareous ooze, siliceous ooze, pelagic clay, etc. — derived from 14,000 surface samples and gridded for global coverage. Ægir uses it to look up sediment class under every pin, which then keys into the APL-UW reflection-loss model.
View at earthbyte.orgThe Applied Physics Laboratory at the University of Washington's "High-Frequency Ocean Environmental Acoustic Models Handbook" is the standard reference for converting sediment class to acoustic reflection loss as a function of grazing angle and frequency. Ægir uses APL-UW's tabulated coefficients per sediment class to compute bottom-bounce loss in the ray solver — a documented, public-release model with no Bellhop / Kraken-style GPL entanglement.
View at apps.dtic.milThe two-paper Francois-Garrison set is the canonical model for sound absorption in seawater as a function of temperature, salinity, depth, pH, and frequency. It captures the boric-acid relaxation around 1 kHz and the magnesium-sulphate relaxation around 100 kHz, both critical for sub-100-kHz marine-acoustic work. Ægir applies it at every range bin in the transmission-loss integral.
View Part I (DOI)Wenz's 1962 paper laid out the spectral structure of ocean ambient noise — the wind-driven component above ~500 Hz, the shipping component in the 50-500 Hz band, biological and seismic at lower frequencies — and his curves remain the operational reference 60+ years later. Ægir composes the wind term from live wind speed and the shipping term from a modelled shipping-density climatology to produce a frequency-dependent NL at the receiver.
View DOI 10.1121/1.1909155The standard graduate-level textbook on ocean-acoustic numerical methods. Ægir's ray solver is built from the eikonal-tracing scheme described in Chapter 3, with surface-bottom-CZ classification and dynamic-ray-equation amplitude tracking per the same chapter. No code from the textbook's accompanying Bellhop/Kraken/RAMGeo distribution is used — only the equations.
View at SpringerNMFS's 2024 Update (Version 3.0) sets the marine-mammal hearing-group auditory-weighting functions and the received-level thresholds for the onset of temporary and permanent threshold shift (TTS / auditory injury). It is the operative US regulatory criterion behind incidental-take analysis and most marine-EIA underwater-noise assessment. Ægir applies the v3.0 weighting functions and onset thresholds to turn a transmission-loss field into hearing-group-weighted impact ranges.
View at fisheries.noaa.govThe Southall et al. 2019 review underpins the hearing-group structure (low-, high- and very-high-frequency cetaceans, sirenians, and pinnipeds in water and in air) and the auditory-weighting and exposure recommendations that the NMFS guidance builds on. Ægir uses its hearing-group definitions when weighting the field for non-US assessments and cross-checks.
View DOI 10.1578/AM.45.2.2019.125The ASA-accredited Popper et al. 2014 guidelines provide the fish and sea-turtle exposure criteria — by hearing category and injury / mortality endpoint — that complement the marine-mammal criteria in a full EIA. Ægir applies them to produce fish and sea-turtle impact ranges alongside the marine-mammal ranges.
View DOI 10.1007/978-3-319-06659-2The Hybrid Coordinate Ocean Model run by the US Naval Research Laboratory provides daily nowcast and forecast fields of temperature, salinity, and currents at 1/12° resolution. Ægir's HYCOM proxy (running on the same NAS as the marketing site) fetches fresh T/S columns to refine the WOA23 climatology when a connection is available; falls back to the bundled climatology when offline.
View at hycom.orgSurface wind and waves from NOAA's GFS and GFS-Wave models are the input to the wind-driven term of the Wenz noise spectrum. When online, Ægir reads the 10-metre wind speed at the pin location, derives a sea-state, and feeds it into the noise calculation; offline, the user selects sea-state manually. The fields are served through Ægir's own weather proxy (weather.aegirprediction.app), not fetched from any third-party weather API.
View at nomads.ncep.noaa.govA modelled global shipping-density climatology underpins the shipping component of the Wenz ambient-noise spectrum. Higher-density basins (e.g. the Strait of Malacca, North Sea) get a 5-15 dB correction in the 50-500 Hz band over near-empty water.
The Biot-Stoll theory treats sediment as a saturated porous medium with both fluid and solid phases, capturing frequency-dependent attenuation that the simpler APL-UW class-based model misses — particularly relevant in the low-frequency band (sub-500 Hz). Deferred to v1.1+ because Biot-Stoll requires class-specific calibration parameters (porosity, grain density, frame moduli) that aren't yet in the Dutkiewicz pipeline; the simpler APL-UW model is closer to operationally deployable for v0.1.
Roadmap item — not in v0.1