Skip to content
Longwave
Donate

Seven hydrophones, moored in the dark, listening.

Longwave is a marine acoustics research and conservation nonprofit. We keep long-term listening stations in the Nordic seas, catalogue the whale calls they record, and publish every hour of audio as open data.

Recording since 2018. 61,000 hours archived. All of it public.

Pressure trace, Station LW-03, 20 February 2026, 06:12 UTC. The two bursts are fin whale pulses.

Why we listen

Seawater swallows light within a few hundred meters. Sound is what remains. A blue whale call near 25 hertz can cross an ocean basin in the deep sound channel, which means one quiet instrument, placed well, hears an enormous area.

Longwave builds and maintains moored listening stations, seven of them, from the Iceland Sea to the Norwegian Trench. Each records continuously for a year at a time. The result is a long, unbroken record of what the ocean sounds like, and of what is changing in it.

The record is the point. Whale populations shift on decadal scales; shipping noise climbs a fraction of a decibel a year. Neither shows up in a season of data. Both show up in ten years of it.

Three programs

Everything we run has to pass one test: it adds to the permanent record, and all of it can be published.

The call catalogue

4,112 catalogued calls · 6 species · updated monthly

Every detected baleen whale call is typed, timestamped, and filed with its source recording. Blue, fin, sei, humpback, minke, and bowhead so far. The catalogue is the reference set other groups use to train their detectors, which is exactly what it is for.

Not yet known: whether call types map to individual animals. We cannot count whales from calls alone, and we do not pretend to.

The noise budget

63 Hz third-octave band · 7 stations · 2019 baseline

We track ambient noise in the bands where ship traffic lives. At Station LW-03, median winter noise in the 63 hertz band has risen 1.8 decibels since 2019. Small numbers matter here: under the spreading conditions at our sites, three added decibels roughly halves the distance across which two whales can hear each other.

Not yet known: how much of the rise is traffic and how much is wind, ice, and climate. Attribution needs more winters.

Arrival dates

First and last calls · per species · per station

The first fin whale call of autumn is a date you can put on a chart. Across years, those dates trace how migration timing responds to warming water. We publish the series raw, with detection effort alongside, so the gaps stay visible instead of smoothed over.

Not yet known: whether the early arrivals are the same animals each year. A single station hears presence, not identity.

From the archive

One station, one call

Station LW-03 sits in the Iceland Sea at 66.402 N, 12.118 W, in 1,240 meters of water. The hydrophone hangs 22 meters above the seabed on a taut mooring, sampling at 2 kilohertz, around the clock, on battery, for a year at a time.

There is no cable and no live feed. The station stores sound locally and gives it back when a ship recovers the mooring, which is why our year runs ship to ship, not January to December.

Position
66.402 N, 12.118 W
Water depth
1,240 m
Hydrophone depth
1,218 m
Sample rate
2 kHz, continuous
Endurance
12 months on battery
Last recovery
April 2026, RV Skua

LW-D-114

Blue whale (Balaenoptera musculus)

D call, a social downsweep. Recorded 14 February 2026, 03:41 UTC. 92 to 36 Hz over 4.1 s.

LW-D-114fin whale chorus, 20 Hz04080120Hz0 s7 s14 s

7 seconds, at twice recording speed.

Drawn from the archive spectrogram. The faint band near 20 hertz is a fin whale chorus, present in most winter recordings at this site.

The button synthesizes the call from its catalogue parameters; at true speed the fundamental bottoms out at 36 hertz, beneath most laptop speakers. The archival audio is in the public archive as LW03-20260214-0341.flac.

We do not know which animal produced this call, how large it was, or how far away; past about 15 kilometers, a single station gives presence without a bearing. D calls are made by both sexes, so not even that much is certain.

The expedition log

Every station year ends with a ship. The log is complete for the past three years, including the parts that went wrong, because a mooring that fails silently teaches nothing.

DateVesselStationOutcome
Apr 2026RV SkuaLW-03Recovered and redeployed11.4 of 11.9 months of sound came back; a battery brownout cost 15 days in February.
Sep 2025RV SkuaLW-05Deployment abortedForce 9 on station for six days and the weather window closed. Deployed the following June.
Jun 2025MV Hafdis, charterLW-02Recovered, data lostThe housing flooded at depth around month eight. Twelve months of hardware, zero months of sound. The redesigned end cap is now on every station.
May 2025MV Hafdis, charterLW-01, LW-04Recovered and redeployedFull data return on both.
Apr 2024RV SkuaLW-04Release failedThe acoustic release did not answer. Recovered five weeks later by dragging, with the data intact and one bent frame.