A sound wave crossing 3,000 km of the East Indian Ocean was enough to pin down the average temperature of that whole section with a standard error of just 0.0060 kelvin (Wu et al., Science 2020). The sound came from small repeating earthquakes, not from a ship or a loudspeaker, and the result revived an idea first tested in 1991, when coded signals from Heard Island were monitored at 16 sites around the globe (Munk et al., JASA 1994). The stakes are large: ocean warming accounted for 91% of the heating in the climate system (IPCC AR6 WG1 Summary for Policymakers 2021), and the upper 2,000 m gained about 23 ± 8 ZJ in 2025 alone (Pan et al., Advances in Atmospheric Sciences 2026). We aggregated data from Science, the Journal of the Acoustical Society of America (JASA), Journal of Geophysical Research, the Proceedings of the Royal Society A, Caltech, the Argo Program, the IPCC, NOAA NCEI, the CTBTO and the University of Washington Heard Island archive. This roundup is about temperature; for shipping noise, sonar and marine mammals, see our separate underwater acoustics and ocean sonar statistics.
TL;DR
- The 1991 Heard Island Feasibility Test was monitored at 16 sites across six ocean regions (Munk et al., JASA 1994).
- Heard Island transmitted at source levels of 209-220 dB re 1 µPa from 175 m depth (Bowles et al., JASA 1994).
- Only about half of North Pacific seasonal and year-to-year sea level change came from thermal expansion (ATOC Consortium, Science 1998).
- A 1994 test at 3,250 km range estimated a six-day temperature change with an uncertainty of about 0.006°C (Worcester et al., JASA 1999).
- Seismic ocean thermometry resolved a 3,000 km section with a standard error of 0.0060 K over 2005-2016 (Wu et al., Science 2020).
- Warming in that region appeared to be as much as 69 percent greater than had been believed, depending on the comparison dataset (Caltech News 2020).
- The Kuroshio Extension warmed at 11.8 ± 5.0 mK per year (2σ) from 2008 to 2021 by seismic and in situ data (Peng and Callies, Proceedings of the Royal Society A 2025).
- Argo runs close to 4,000 active floats and collected its 3 millionth profile in July 2024 (Argo Program status page).
- Ocean warming accounted for 91% of the heating in the climate system (IPCC AR6 WG1 SPM 2021).
- The upper 2,000 m ocean gained about 23 ± 8 ZJ in 2025 relative to 2024 (Pan et al., Advances in Atmospheric Sciences 2026).
- The CTBTO hydroacoustic network has 11 stations: six hydrophone stations and five T-phase stations (CTBTO).
1. The Physics: Why Travel Time Works as a Thermometer
Acoustic thermometry rests on one stubborn fact: a 1°C change in water temperature changes sound speed by about 4 m/s (University of Washington Applied Physics Laboratory, Heard Island Feasibility Test archive). Warm water carries sound faster, so a path that warms delivers its signal a little earlier. The trick is that the clock does the averaging. A single thermometer reads one point, while a pulse crossing thousands of kilometers reports the mean of everything it passed through.
That averaging is the whole point, because the signal hunted for is tiny and the local noise is huge. The Heard Island archive puts the nominal climate signal at about 0.005°C per year at 1,000 m depth, while temperatures at a single point can swing by 1-2°C from one month to the next. On those figures, one month of local “ocean weather” (eddies and internal waves, the ocean’s equivalent of passing storms) is 200 to 400 times larger than a full year of the climate trend (1 / 0.005 = 200; 2 / 0.005 = 400). Point sensors have to average that noise away statistically; a long acoustic path averages it physically.
| Metric | Value | Source |
|---|---|---|
| Sound speed change for a 1°C change in water temperature | About 4 m/s | University of Washington APL, Heard Island archive |
| Nominal climate warming signal at 1,000 m depth | About 0.005°C per year | University of Washington APL, Heard Island archive |
| Depth of the sound speed minimum (SOFAR channel axis) in the deep mid-latitude ocean | About 800 to 1,000 m | DOSITS (University of Rhode Island), Sound Travel in the SOFAR Channel 2023 |
| Launch angle below which rays stay trapped in the deep sound channel | Less than about 12° from horizontal | DOSITS (University of Rhode Island) 2023 |
| Travel-time perturbation caused by mesoscale eddies | Roughly 100 ms | Howe et al., Frontiers in Marine Science 2019 |
| Travel-time change caused by a Gulf Stream or Kuroshio meander | About 1 second | Howe et al., Frontiers in Marine Science 2019 |
| Receiver timing and positioning accuracy needed for tomography | 1 ms and 1 m | Howe et al., Frontiers in Marine Science 2019 |
Context: the SOFAR channel (Sound Fixing and Ranging) is a depth band where sound speed is lowest, so sound bends back toward it from above and below instead of escaping. DOSITS places its axis at about 800 to 1,000 m in the deep mid-latitude ocean. Most recent available data for the timing and eddy figures: Howe et al., Frontiers in Marine Science 2019.
2. The Heard Island Feasibility Test (1991)
Heard Island was a proof of concept, and the failure statistics are as telling as the successes. The test made 35 transmissions, and its 10 acoustic sources failed with a half-life of 1.2 days (University of Washington APL, Heard Island archive, The Experiment). By the time 25 transmissions had been made, only three sources were still working, and the last transmission used the final two. The question it answered was not “how warm is the ocean” but “can a coded signal survive a trip around the planet and still be timed precisely”.
It could. The Munk et al. (1994) overview in JASA reports that the transmissions were monitored at 16 sites spread across the North Atlantic, South Atlantic, North Pacific, South Pacific, Indian Ocean and Southern Ocean. A separate marine mammal survey found no drop in the density of whale and dolphin groups near the source during transmissions, though sperm whales, heard during 23% of baseline listening time, went unheard during transmissions (Bowles et al., JASA 1994). ATOC later ran dedicated marine mammal research alongside its thermometry (see section 3).
| Metric | Value | Source |
|---|---|---|
| Start date, transmissions made, sources deployed and source failure half-life | January 26, 1991 (planned for 10 days); 35 transmissions; 10 sources; half-life of 1.2 days | University of Washington APL, Heard Island archive |
| Number of sites that monitored the transmissions | 16 | Munk et al., JASA 1994 |
| Source center frequency, maximum bandwidth and source depth | 57 Hz, 30 Hz, 175 m | Bowles et al., JASA 1994 |
| Source level of the transmissions | 209-220 dB re 1 µPa | Bowles et al., JASA 1994 |
| Cetacean group density before vs during transmissions | 0.0157 vs 0.0166 groups per km² | Bowles et al., JASA 1994 |
| Sperm whales heard: share of 1,181 baseline minutes vs 1,939 transmission minutes | 23% vs none heard | Bowles et al., JASA 1994 |
Outlier note: the idea was older than the test. A 1960 precursor detonated explosive charges off Perth and listened near Bermuda, about 19,820 km away (University of Washington APL, Heard Island archive). Transmissions ran one hour in every three (Bowles et al., JASA 1994). Most recent available data: 1991 experiment, published 1994.
3. ATOC and a Decade of North Pacific Thermometry (1994-2006)
ATOC turned the Heard Island demonstration into an observing system, and its key number is a precision figure. Acoustic travel times were measured with a precision of about 20-30 milliseconds at 3,000-5,000 km ranges, equivalent to about 0.006°C averaged over range and depth; at 5,000 km the signal takes nearly an hour to arrive (University of Washington APL, ATOC epilogue). That is timing to roughly one part in 100,000 of the travel time (0.03 s / 3,600 s is about 0.0008%).
The science payoff was a correction to how sea level was being read. Comparing gyre-scale heat content from tomography with satellite altimetry and a circulation model, the ATOC Consortium (Science 1998) found that only about half of seasonal and year-to-year sea level change came from thermal expansion. Over 1996-2006 the program fed four independent estimates of the North Pacific temperature field (Dushaw et al., JGR Oceans 2009). It also came with dedicated marine life studies: juvenile elephant seals exposed to the source showed only subtle diving changes, judged likely to have minimal biological significance (Costa et al., JASA 2003). Chronic noise from shipping is a separate exposure question, covered in our ship propeller cavitation noise statistics.
| Metric | Value | Source |
|---|---|---|
| ATOC partners at launch | 11 institutions from 7 nations, 2 sources, 18 receivers | MIT News 1994 |
| ATOC source frequency | 75 Hz | MIT News 1994 |
| 1994 test range and receiver | 3,250 km, 700 m vertical array | Worcester et al., JASA 1999 |
| Uncertainty of a six-day temperature change at 3,250 km | About 0.006°C | Worcester et al., JASA 1999 |
| Uncertainty of range-averaged sound speed at 3,250 km | About 0.05 m/s | Worcester et al., JASA 1999 |
| Annual cycle of heat flux (peak-to-peak) in the gyre-scale North Pacific | 150 ± 25 W/m², a 0.2°C cycle in vertically averaged temperature | ATOC Consortium, Science 1998 |
| Elephant seal study: ATOC source depth and source level | 939 m, 195 dB re 1 µPa at 1 m | Costa et al., JASA 2003 |
| Seals whose descent rate shifted more than one standard deviation from control during exposure | 9 of 14 | Costa et al., JASA 2003 |
Context: received sound levels at the seals averaged 128 dB, ranging from 118 to 137 dB re 1 µPa in the 60-90 Hz band, against ambient levels of 87-107 dB (Costa et al., JASA 2003). Most recent available data: ATOC ended in 2006; see Dushaw et al., JGR Oceans 2009 for the full decade.
4. Seismic Ocean Thermometry: The 2020 Breakthrough
Dedicated sound sources need ships, moorings and environmental permits. The 2020 idea was to stop building sources and use the ones the planet already provides. Wu et al. used 2,047 pairs of repeating earthquakes off Sumatra to track temperature across a 3,000 km section of the equatorial East Indian Ocean from 2005 to 2016, with a standard error of 0.0060 kelvin (Wu et al., Science 2020; pair count as reported by Smithsonian Magazine 2020). Repeating earthquakes are quakes that rupture the same patch of fault again and again, so each pair gives two near-identical sound sources years apart.
The headline that travelled was a warming trend larger than earlier estimates: Caltech reported that ocean warming in the region appeared to be as much as 69 percent greater than had been believed, depending on which dataset is used. The same release carried the caveat that more data were needed before drawing conclusions. Note the T wave here: it is the part of an earthquake’s energy that converts to sound and travels through the water, not a wave in the solid rock.
| Metric | Value | Source |
|---|---|---|
| Length of ocean section averaged and period analyzed | 3,000 km (equatorial East Indian Ocean), 2005-2016 | Wu et al., Science 2020 |
| Standard error of the section-mean temperature anomaly | 0.0060 kelvin | Wu et al., Science 2020 |
| Time scales of fluctuations detected | 12 months, 6 months and about 10 days | Wu et al., Science 2020 |
| Repeating earthquake pairs used | 2,047 | Smithsonian Magazine 2020, reporting the Science paper |
| Sound travel time vs the part of it caused by water temperature | About half an hour vs about one-tenth of a second | Caltech News 2020 |
| Warming in the region vs earlier estimates | As much as 69 percent greater, depending on the comparison dataset | Caltech News 2020 |
| Depths sampled by the earthquake sound waves | Below 2,000 m, where conventional measurements are few | Caltech News 2020 |
Context: by October 2026 the paper had 78 citing works in OpenAlex. The 69 percent figure is a regional, dataset-dependent comparison, not a global revision of ocean warming (see Data watch in Methodology).
5. Seismic Thermometry 2023-2025: Hydrophones, Fiber Cables and Real Trends
The follow-up papers turned a single demonstration into a toolkit, and the most consequential number is a trend with an honest error bar. Seismic and in situ data together show steady subsurface warming of 11.8 ± 5.0 mK per year (2σ uncertainty) in the Kuroshio Extension from 2008 to 2021, with no substantial trend between 1997 and 2008 (Peng and Callies, Proceedings of the Royal Society A 2025). Converted to the units climate reports use, 11.8 mK per year is 0.118 K per decade (11.8 × 10 / 1,000).
The same paper shows why seismic data is a complement rather than a rival to floats. Relative to the prior, eddy-driven error variance in area-mean temperature fell by 30% with in situ data, 39% with seismic data and 50% with both; for path-mean estimates the combined reduction was 83% versus 45% from in situ data alone. Meanwhile, sensors keep getting more sensitive: CTBTO hydrophones allow smaller earthquakes (magnitude below 4.0) to be used (Wu et al., JGR Solid Earth 2023), and a fiber-optic cable on the seafloor detected 92 oceanic earthquakes in four days, most missing from existing catalogs (Shen and Wu, JGR Solid Earth 2024).
| Metric | Value | Source |
|---|---|---|
| Kuroshio Extension subsurface warming rate, 2008-2021 | 11.8 ± 5.0 mK per year (2σ) | Peng and Callies, Proceedings of the Royal Society A 2025 |
| Eddy error-variance reduction in area-mean temperature: in situ only vs seismic only vs both | 30% vs 39% vs 50% | Peng and Callies, Proceedings of the Royal Society A 2025 |
| Path-mean error-variance reduction: seismic plus in situ vs in situ alone | 83% vs 45% | Peng and Callies, Proceedings of the Royal Society A 2025 |
| Start of the Wake Island station’s digital record used for the Kuroshio work | 1997 | Peng et al., JGR Oceans 2024 |
| Earthquake magnitude usable with CTBTO hydrophones | Below 4.0 | Wu et al., JGR Solid Earth 2023 |
| Oceanic earthquakes detected by an ocean-bottom fiber-optic cable in a 4-day experiment off Oregon | 92 | Shen and Wu, JGR Solid Earth 2024 |
| Smallest repeating earthquakes usable with the fiber-optic cable | M 3.5 | Shen and Wu, JGR Solid Earth 2024 |
| Frequency bands used in a published step-by-step workflow for passive deep-ocean thermometry | 3-6 Hz and 6-12 Hz | Smets et al., JGR Oceans 2022 |
Context: distributed acoustic sensing (DAS) turns an ordinary seafloor fiber-optic cable into thousands of vibration sensors along its length. The method is also spilling into ocean dynamics: seismic tomography in the East Indian Ocean picked up biweekly temperature signals carried by Yanai waves (a type of wave trapped along the equator), and the variance in that band differed by up to a factor of 2 between ocean models (Peng et al., Journal of Physical Oceanography 2025).
6. Argo and Ocean Heat Content: The Yardstick Acoustic Methods Are Checked Against
Every acoustic result is ultimately compared with Argo, so its coverage defines what acoustics can add. Argo runs close to 4,000 active floats and collected its 3 millionth profile in July 2024 (Argo Program status), at about 13,000 profiles a month from 30 countries (Argo Program, About). But a standard float descends to 2 km and no further. The deep ocean below that is exactly where earthquake sound samples (see section 4), and where the Deep Argo mission is working toward a standing array of 1,228 floats (Argo Program, Deep Argo Mission).
The heat being measured is enormous and accelerating. Ocean warming accounted for 91% of the heating in the climate system, and the observed heating rate rose from 0.50 [0.32 to 0.69] W m⁻² over 1971-2006 to 0.79 [0.52 to 1.06] W m⁻² over 2006-2018 (IPCC AR6 WG1 Summary for Policymakers 2021). In 2025 the upper 2,000 m gained about 23 ± 8 ZJ relative to 2024 by the IAP/CAS estimate, with about 33% of the ocean area in its top three warmest conditions on record (Pan et al., Advances in Atmospheric Sciences 2026).
| Metric | Value | Source |
|---|---|---|
| Active Argo floats | Close to 4,000 | Argo Program status page |
| Argo profile milestones | 1 millionth in November 2012, 2 millionth in November 2018, 3 millionth in July 2024 | Argo Program status page |
| Deep Argo target array size and maximum float depth | 1,228 floats; some models capable of 6,000 m | Argo Program, Deep Argo Mission |
| Share of climate-system heating that went into the ocean | 91% | IPCC AR6 WG1 SPM 2021 |
| Global ocean heat content increase, 1971-2018 | 0.396 [0.329 to 0.463] yottajoules (likely range) | IPCC AR6 WG1 Chapter 9, 2021 |
| Upper 2,000 m ocean heat gain in 2025 vs 2024 (IAP/CAS estimate) | About 23 ± 8 ZJ | Pan et al., Advances in Atmospheric Sciences 2026 |
| Share of global ocean area in its top three (and top five) warmest conditions on record in 2025 | About 33% (about 57%) | Pan et al., Advances in Atmospheric Sciences 2026 |
Context: NOAA NCEI publishes its 0-700 m and 0-2,000 m heat content series from 1955 to the present, updated quarterly (NOAA NCEI Global Ocean Heat Content). Global sea surface temperature in 2025 was 0.49°C above the 1981-2010 baseline (Pan et al. 2026).
7. Arctic Thermometry and the Global Listening Networks
The Arctic is where acoustic thermometry has delivered some of its clearest climate signals, in an ocean where sea ice gets in the way of most other instruments. In 1994, 19.6 Hz signals sent about 2,630 km across the Arctic showed the Atlantic Layer to be about 0.4°C warmer than expected, and a 1999 follow-up found about 0.5°C of further warming (Howe et al., Frontiers in Marine Science 2019). The Atlantic Layer is the warm, salty water from the Atlantic that flows into the Arctic beneath the colder surface layer.
The newest data come from CAATEX, a US-Norwegian campaign whose results were published in 2026. Six moorings transmitted and recorded every three days during 2019-2020, and signals carried over 2,600 km (AIP Scilights 2026). The paper concludes that long-range transmissions can give precise, year-round under-ice measurements, although large mode-2 travel-time uncertainty prevented a definitive comparison with the 1994 data (Dzieciuch et al., JASA 2026). Lower latitudes depend on infrastructure built for other reasons, chiefly the CTBTO hydroacoustic network, the same kind of always-on listening now used to score coral reef soundscape health.
| Metric | Value | Source |
|---|---|---|
| 1994 Transarctic (TAP) signal frequency and path length | 19.6 Hz over about 2,630 km | Howe et al., Frontiers in Marine Science 2019 |
| Atlantic Layer temperature in 1994 vs expectation, and further warming by April 1999 | About 0.4°C warmer; about 0.5°C further warming | Howe et al., Frontiers in Marine Science 2019 |
| CAATEX moorings, center frequency and bandwidth (2019-2020) | 6 moorings; 35 Hz; about 4 Hz full bandwidth | Dzieciuch et al., JASA 2026 |
| CAATEX transmission interval and distance covered | Every three days; over 2,600 km | AIP Scilights 2026 |
| CTBTO hydroacoustic stations | 11 (6 hydrophone stations, 5 T-phase stations) | CTBTO, Hydroacoustic Monitoring |
| Depth of CTBTO hydrophones and design lifetime | 600 to 1,200 m; 20 to 25 years | CTBTO, Hydroacoustic Monitoring |
| North Atlantic tomography array | 2 sources and 7 receivers | Dushaw, Journal of Atmospheric and Oceanic Technology 2018 |
| North Atlantic temperature uncertainty: one data type vs acoustic plus other data | About 50% of prior vs about 25% of prior (94% of variance resolved) | Dushaw, Journal of Atmospheric and Oceanic Technology 2018 |
Context: a T-phase station is a seismometer on a steep oceanic island that picks up sound waves converting into ground motion at the shore. The CTBTO record has value for history as well as for monitoring: researchers point to a nearly 20 year record of treaty recordings as a basis for historical ocean thermometry (Dall’Osto and Heyburn, JASA 2019). For the North Atlantic array, see Dushaw, Journal of Atmospheric and Oceanic Technology 2018.
Summary: Ocean Acoustic Tomography and Thermometry by the Numbers
| Metric | Value | Source |
|---|---|---|
| Sound speed change per 1°C of water temperature | About 4 m/s | University of Washington APL, Heard Island archive |
| Heard Island 1991: sites monitoring the transmissions | 16 | Munk et al., JASA 1994 |
| Heard Island source level | 209-220 dB re 1 µPa | Bowles et al., JASA 1994 |
| Six-day temperature change uncertainty at 3,250 km | About 0.006°C | Worcester et al., JASA 1999 |
| Share of North Pacific seasonal and year-to-year sea level change due to thermal expansion | Only about half | ATOC Consortium, Science 1998 |
| Seismic thermometry standard error over a 3,000 km section | 0.0060 kelvin | Wu et al., Science 2020 |
| Repeating earthquake pairs in the first study | 2,047 | Smithsonian Magazine 2020, reporting the Science paper |
| Regional warming vs earlier estimates | As much as 69 percent greater | Caltech News 2020 |
| Kuroshio Extension warming, 2008-2021 | 11.8 ± 5.0 mK per year (2σ) | Peng and Callies, Proceedings of the Royal Society A 2025 |
| Path-mean error-variance reduction, seismic plus in situ vs in situ alone | 83% vs 45% | Peng and Callies, Proceedings of the Royal Society A 2025 |
| Oceanic earthquakes found by a seafloor fiber-optic cable in 4 days | 92 | Shen and Wu, JGR Solid Earth 2024 |
| Active Argo floats | Close to 4,000 | Argo Program status page |
| Share of climate-system heating absorbed by the ocean | 91% | IPCC AR6 WG1 SPM 2021 |
| Upper 2,000 m ocean heat gain, 2025 vs 2024 | About 23 ± 8 ZJ (IAP/CAS) | Pan et al., Advances in Atmospheric Sciences 2026 |
| Arctic Atlantic Layer, 1994 acoustic measurement vs expectation | About 0.4°C warmer | Howe et al., Frontiers in Marine Science 2019 |
| CTBTO hydroacoustic stations | 11 | CTBTO, Hydroacoustic Monitoring |
Methodology and Sources
We collected about 75 raw data points from peer-reviewed papers, agency pages and institutional press releases, kept the 60-plus that could be traced to a primary document, and dropped figures we could only see in search snippets (including a widely repeated Heard Island reception distance we could not confirm in a primary page). Abstract-level figures for papers behind paywalls were read from their indexed abstracts.
- University of Washington Applied Physics Laboratory (B. Dushaw), Heard Island Feasibility Test archive: thermometry, experiment and ATOC epilogue pages
- Munk, Spindel, Baggeroer and Birdsall, The Heard Island Feasibility Test, JASA 1994
- Bowles et al., Relative abundance and behavior of marine mammals exposed to transmissions from the Heard Island Feasibility Test, JASA 1994
- MIT News, sound and climate experiments article, May 18, 1994
- Worcester et al., A test of basin-scale acoustic thermometry at 3250-km range, JASA 1999
- ATOC Consortium, Ocean Climate Change: Comparison of Acoustic Tomography, Satellite Altimetry, and Modeling, Science 1998
- Dushaw et al., A decade of acoustic thermometry in the North Pacific Ocean, JGR Oceans 2009
- Costa et al., ATOC sound source and juvenile northern elephant seal diving, JASA 2003
- Howe et al., Observing the Oceans Acoustically, Frontiers in Marine Science 2019
- DOSITS (University of Rhode Island), Sound Travel in the SOFAR Channel
- Wu, Zhan, Peng, Ni and Callies, Seismic ocean thermometry, Science 2020
- Caltech News, Undersea Earthquakes Shake Up Climate Science, September 18, 2020
- Smithsonian Magazine, Undersea Earthquakes May Help Us Take the Ocean’s Temperature, 2020 (used only for the 2,047 pair count from the Science paper)
- Wu et al., Seismic Ocean Thermometry Using CTBTO Hydrophones, JGR Solid Earth 2023
- Shen and Wu, Ocean Bottom Distributed Acoustic Sensing for Oceanic Seismicity Detection and Seismic Ocean Thermometry, JGR Solid Earth 2024
- Peng, Callies, Wu and Zhan, Seismic Ocean Thermometry of the Kuroshio Extension Region, JGR Oceans 2024
- Peng and Callies, Estimating temperature variability and trends from a combination of seismic and in situ data, Proceedings of the Royal Society A 2025
- Peng, Callies, Nagura and McPhaden, Yanai Waves in the Deep East Indian Ocean Observed with Seismic Ocean Tomography, Journal of Physical Oceanography 2025
- Smets, Weemstra and Evers, Hydroacoustic Travel Time Variations as a Proxy for Passive Deep-Ocean Thermometry, JGR Oceans 2022
- Dall’Osto and Heyburn, sound sources of opportunity on the CTBT hydroacoustic network for historical ocean thermometry, JASA 2019
- Argo Program: status page, About Argo and Deep Argo Mission
- IPCC AR6 WG1 Summary for Policymakers 2021 and IPCC AR6 WG1 Chapter 9: Ocean, Cryosphere and Sea Level Change
- Pan, Cheng, Abraham et al., Ocean Heat Content Sets Another Record in 2025, Advances in Atmospheric Sciences 2026
- NOAA NCEI, Global Ocean Heat and Salt Content
- Dzieciuch et al., Transarctic acoustic transmissions during the coordinated Arctic acoustic thermometry experiment in 2019-2020, JASA 2026 and AIP Scilights summary 2026
- Dushaw, Ocean Acoustic Tomography in the North Atlantic, Journal of Atmospheric and Oceanic Technology 2018
- CTBTO, Hydroacoustic Monitoring
- Data watch: (1) The share of excess heat held by the ocean varies by source: 91% (IPCC AR6 SPM), more than 90% (Wu et al., Science 2020 abstract) and as much as 95 percent (Caltech News 2020); we use the IPCC figure. (2) The 2025 upper-ocean heat gain differs sharply by product within the same paper: about 23 ± 8 ZJ (IAP/CAS), 13 ± 5 ZJ (NOAA NCEI, data only through June 2025) and 70 ± 39 ZJ (Copernicus Marine, near-real-time data); we headline the IAP/CAS value the paper itself uses. (3) The 69 percent figure is regional (East Indian Ocean, 2005-2016) and depends on the comparison dataset; the authors cautioned against immediate conclusions. (4) ATOC dates differ slightly between sources: the UW archive gives Pioneer Seamount transmissions from late 1995 to the end of 1998 and Kauai from late 1997 to October 1999, while Dushaw et al. 2009 describes a 1996-2006 decade with California transmissions in 1996-1999 and Kauai transmissions in 1997-1999 and 2002-2006. (5) The 1994 TAP path is given as about 2,630 km at 19.6 Hz by Howe et al. 2019 and as 2,610 km at 20 Hz by MIT News 1994. (6) Heard Island, ATOC and TAP figures are historical by nature (most recent available data: 1991-2009); CAATEX temperature estimates along the acoustic paths are still pending, and Argo float counts change monthly.
Last updated: October 25, 2026. We update this roundup quarterly, and the next refresh is expected when the IAP/CAS-led annual ocean heat content update for 2026 is published in Advances in Atmospheric Sciences in January 2027.