Krakatoa 1883 · Hunga Tonga 2022
The air wave that circled the Earth: Krakatoa 1883 and Tonga 2022
Two volcanoes, 139 years apart, shoved the atmosphere hard enough that a pulse of raised air pressure, far too slow to hear, ran round the planet, met itself on the far side and came back, again and again. Barometers caught every pass. Here is the pulse as a ring on a globe, and the hour it crossed your town.
Scroll: the wave leaves the volcano↓
27 August 1883, just before 10 in the morning
Krakatoa, a volcanic island in the Sunda Strait between Java and Sumatra, blows up in a series of explosions that morning; four were the largest. The third, at 02:56 GMT, is “by far the most violent”, in the words of the Royal Society’s report.
It shoves the air outward. A ring of raised pressure leaves the island at roughly 310 to 320 metres a second (the report measured about 319 on the first pass), a little under the speed of sound at ground level.
How far was Krakatoa heard?
People up to a few thousand kilometres away heard the explosions: a ship in distress was suspected at Diego Garcia, and at Alice Springs two reports were “similar to the discharge of a rifle”. The farthest report is from Rodrigues Island, about 4,800 km away, where the Chief of Police wrote of “the distant roars of heavy guns”. The report does not tie them to the great explosion.
The gold dots are some of the places in the report. The ring did not stop there. It is a swell of pressure lasting an hour or two, far below the pitch the ear can hear.
Recorded in Calcutta, Vienna, Kew and Toronto
Barographs, instruments that record air pressure on a moving chart, caught it as a sudden rise, a deep dip, then a slow recovery over nearly two hours. Calcutta’s drew it at 06:47 GMT, Vienna’s at 12:34, Kew’s in London at 13:48, Toronto’s at 17:33.
The dots are the 44 stations with a position and passage times in the Royal Society’s tables. Each lights up as the ring reaches it.
The far side of the Earth is in Colombia
After about 18 hours the model’s ring has shrunk to a point on the opposite side of the planet, near Puerto Boyacá on Colombia’s Magdalena river, and spreads out again.
The 1888 report describes it the same way: the wave “advanced, gradually contracting again, to a node at the antipodes of Krakatoa; whence it was reflected”.
Back at the volcano after 35.9 hours, and round again
At 35.9 hours the ring is back over Krakatoa and starts a second lap. Over Kew, in London, the barograph drew it seven times in under five days, the last at 04:10 GMT on 1 September.
That is what “circled the Earth” means here: seven passes over one barometer, four from the volcano’s side and three from the far side.
How do we know it circled the Earth?
Because barometers counted it. The Royal Society’s 1888 report tabulates passage times from 45 stations, from Dunedin to Toronto; 44 of them have a position in its Table II, and those are the ones used here. A place d degrees from Krakatoa meets the first wave after d degrees of travel; the second wave reaches it the long way, through the far side, after 360° − d; the third after 360° + d; the fourth after 720° − d, and so on. The report calls the odd passes the wave travelling from Krakatoa and the even ones the wave travelling from its antipodes.
One speed does most of the work. 310 metres per second from 02:56 GMT places the first pass at a typical station within 32 minutes of its recorded time. The report’s times mark the deepest dip of the pulse, which follows the first rise by about 36 minutes in Strachey’s reckoning, so the recorded times run a little late.
The scatter grows with every lap. The report’s air-wave chapter, presented by Strachey, measured the gap: over Europe a lap against the Earth’s rotation took 36 h 24 min and one with it 34 h 46 min. It suggested westerly winds along the route as the probable cause. A single average speed cannot follow that, so later passes drift by hours.
| Pass | Arrives from | Recorded (GMT) | Model (GMT) |
|---|---|---|---|
| 1 | the volcano’s side | 27 Aug, 13:48 | 27 Aug, 13:22 |
| 2 | the far side | 28 Aug, 04:05 | 28 Aug, 04:25 |
| 3 | the volcano’s side | 29 Aug, 02:12 | 29 Aug, 01:17 |
| 4 | the far side | 29 Aug, 14:42 | 29 Aug, 16:20 |
| 5 | the volcano’s side | 30 Aug, 15:10 | 30 Aug, 13:12 |
| 6 | the far side | 31 Aug, 02:00 | 31 Aug, 04:15 |
| 7 | the volcano’s side | 1 Sep, 04:10 | 1 Sep, 01:07 |
The 1888 stations with a recorded first passage, nearest first. A question mark marks a time the report itself calls doubtful; several of the largest misses are among those.
| Station | Distance, km | Recorded, GMT | Model, GMT | Recorded − model, min | Passes |
|---|---|---|---|---|---|
| Kolkata (Calcutta) | 3,688 | 06:47 | 06:15 | +32 | 3 |
| Zikawei (Shanghai) | 4,483 | 07:27 | 06:57 | +30 | 3 |
| Mumbai (Bombay) | 4,527 | 07:26 | 07:00 | +26 | 4 |
| Melbourne | 5,323 | 08:14 | 07:43 | +31 | 4 |
| Mauritius | 5,391 | 08:17 | 07:46 | +31 | 7 |
| Sydney | 5,628 | 08:24 | 07:59 | +25 | 5 |
| Tokyo | 5,869 | 08:51 | 08:12 | +39 | 3 |
| Dunedin | 7,615 | 10:15 | 09:46 | +29 | 4 |
| Wellington | 7,841 | 10:20 | 09:58 | +22 | 4 |
| Pavlovsk | 9,761 | 12:18 | 11:41 | +37 | 5 |
| Saint Petersburg | 9,779 | 12:14 | 11:42 | +32 | 5 |
| Luanda | 10,156 | 12:18 | 12:03 | +15 | 7 |
| Budapest | 10,237 | 12:30 | 12:07 | +23 | 5 |
| Vienna | 10,443 | 12:34 | 12:18 | +16 | 3 |
| Palermo | 10,607 | 13:37? | 12:27 | +70 | 2 |
| Berlin | 10,688 | 12:52 | 12:31 | +21 | 6 |
| Rome | 10,703 | 13:00 | 12:32 | +28 | 3 |
| Leipzig | 10,747 | 14:00? | 12:34 | +86 | 2 |
| Munich | 10,798 | 12:35 | 12:37 | -2 | 5 |
| Magdeburg | 10,801 | 12:59 | 12:37 | +22 | 6 |
| Modena | 10,840 | 12:51 | 12:40 | +11 | 7 |
| Brussels | 11,310 | 13:28 | 13:05 | +23 | 6 |
| Paris (Parc Saint-Maur) | 11,462 | 13:41 | 13:13 | +28 | 5 |
| Paris (Montsouris) | 11,475 | 13:34 | 13:14 | +20 | 3 |
| Geldeston | 11,484 | 13:39 | 13:14 | +25 | 7 |
| Greenwich | 11,603 | 13:45 | 13:21 | +24 | 7 |
| Kew | 11,625 | 13:48 | 13:22 | +26 | 7 |
| Aberdeen | 11,628 | 13:52 | 13:22 | +30 | 7 |
| Oxford | 11,685 | 13:53 | 13:25 | +28 | 4 |
| Stonyhurst | 11,723 | 13:55 | 13:27 | +28 | 7 |
| Liverpool | 11,772 | 13:58 | 13:30 | +28 | 7 |
| Glasgow | 11,791 | 14:02 | 13:31 | +31 | 7 |
| Armagh | 11,977 | 14:15 | 13:41 | +34 | 7 |
| Falmouth | 11,980 | 14:04 | 13:41 | +23 | 5 |
| Valentia (Ireland) | 12,285 | 14:25 | 13:57 | +28 | 6 |
| San Fernando (Cádiz) | 12,341 | 13:22? | 14:00 | -38 | 4 |
| Serra da Estrela | 12,387 | 13:50? | 14:03 | -13 | 4 |
| South Georgia | 12,390 | 15:24 | 14:03 | +81 | 4 |
| Coimbra | 12,462 | 14:20 | 14:07 | +13 | 4 |
| Lisbon | 12,553 | 14:17 | 14:12 | +5 | 4 |
| Toronto | 15,814 | 17:33 | 17:07 | +26 | 7 |
| New York | 16,166 | 17:53 | 17:26 | +27 | 4 |
| Washington | 16,360 | 18:00 | 17:37 | +23 | 7 |
| Havana | 17,942 | 19:00? | 19:02 | -2 | 4 |
Download the 1883 passage times as a CSV file (our transcription of Tables II to IV, public domain)
15 January 2022, 04:15 UTC: Hunga Tonga–Hunga Haʻapai
A volcano that is mostly under the sea in the South Pacific erupts. The pulse leaves at about 315 metres a second. In Alaska, 9,400 to 9,800 km away, people hear booms; instruments there show the sound arriving after the pressure wave.
Krakatoa’s bang was reported at about 4,800 km. Tonga’s was reported about twice as far.
Six barometers, one arrival time
Barometers across Europe and in New York drew the same ring. Lined up on the model’s arrival time, each record peaks 10 to 45 minutes after it.
The dots are the six open records used below.
The far side in 2022 is southern Algeria
About 18 hours after the eruption the ring closed on a point in the Sahara, about 270 km from Tamanrasset. Five days later, instruments in Britain were still picking up its passes.
Did the same thing happen in 2022?
Yes, and modern barometers recorded it in more detail. Matoza and colleagues count “four (plus three antipodal) passages around Earth over 6 days” for the Hunga Tonga eruption, the same seven that the 1888 report counted for Krakatoa. They give the speed as about 315 metres per second, 1,134 km/h. The pulse over Britain measured a little over 2 hectopascals, peak to trough, against 1.7 to 2.7 for Krakatoa.
The plot lines up six open barometer records on the model’s arrival time, taking the eruption as 04:15 UTC and the speed as 315 m/s. The dots mark each trace’s highest reading, which comes 10 to 45 minutes after the leading edge. Zürich and Hamburg show the biggest swings.
New York’s one-minute record shows the return pass as well: on 16 January the wave came back from the opposite direction, after 27,603 km, and peaked at 04:59 UTC.
We found no open data finer than hourly from Romania’s weather service, and one reading an hour cannot resolve a pulse that lasts half an hour. The model puts the first pass over Bucharest at 18:59 UTC, 20:59 local time. Nyíregyháza in Hungary, about 55 km from the border, recorded its peak at 19:30 UTC.
One barometer, eight passes
At the Reading University observatory in England the 2022 wave was picked out eight times in 127 hours, with the swing falling from 2.38 hectopascals to under 0.2. The model, with nothing fitted to these records, is within two hours of every pass.
| Pass | Observed, UTC | Model, UTC | Observed − model, min | Swing, hPa |
|---|---|---|---|---|
| 1 | 15 Jan, 18:52 | 15 Jan, 18:50 | +2 | 2.38 |
| 2 | 16 Jan, 01:05 | 16 Jan, 00:57 | +7 | 1.72 |
| 3 | 17 Jan, 07:00 | 17 Jan, 06:08 | +52 | 1.28 |
| 4 | 17 Jan, 13:00 | 17 Jan, 12:15 | +44 | 0.87 |
| 5 | 18 Jan, 18:35 | 18 Jan, 17:26 | +69 | 0.42 |
| 6 | 19 Jan, 01:15 | 18 Jan, 23:33 | +101 | 0.40 |
| 7 | 20 Jan, 05:30 | 20 Jan, 04:44 | +46 | 0.26 |
| 8 | 20 Jan, 11:30 | 20 Jan, 10:51 | +38 | < 0.2 |
When did the wave pass your town?
Search for a town or tap the globe. Nothing you type leaves your device.
Questions people ask
How far away was Krakatoa heard?
The farthest report in the Royal Society’s 1888 report is from Rodrigues Island in the Indian Ocean, “very nearly 3,000 English miles” (about 4,800 km) from Krakatoa. It calls this “the only instance on record of sounds having been heard at anything like so great a distance”. By the report’s own rough estimate the sounds were heard over “rather less than one-thirteenth” of the Earth’s surface.
One caution from the report and one from reading it closely. The Rodrigues account describes guns heard at intervals of three to four hours until 3 p.m. and does not tie them to the great explosion. And the report notes that “there is no direct evidence that the great final explosion … was accompanied by sounds heard at any considerable distance”. A letter from Diego Garcia, 3,650 km away, records dull but violent detonations between 10 and 11 in the morning, local time, which fits the great explosion once the sound’s travel time is allowed for; the report itself does not make that link.
Was Krakatoa heard in London, America or Australia?
In Australia, yes: from Cossack and Geraldton on the west coast to Perth, Daly Waters in the Northern Territory and Alice Springs, about 2,100 to 3,600 km away. From London or the Americas the report has no sound report: its table of places that heard the explosions lists none in Europe or the Americas, and it left out a report from the Cayman Islands as too vague. London’s barograph did record the pressure wave, at 13:48 GMT.
How many times did the pressure wave circle the Earth?
Barometers counted at least seven passes at many stations, four from the volcano’s side and three from the far side. By the last pass at Kew the front from the volcano’s side had gone round the Earth three times and a third of the way round a fourth; the front from the far side had gone round nearly three times. Popular accounts say “four times” or “three and a half times”. The Royal Society’s own count is seven passages, which is the safe number. For Tonga, the 2022 papers count the same, four plus three over six days, and at Reading in England one barometer picked out eight passes in 127 hours.
How loud was Krakatoa in decibels?
Nobody measured it. No instrument measured the sound. The only pressure reading near the volcano is a gas-works gauge in Batavia, 150 km away, whose pen ran off the scale. Figures such as 172 dB or 180 dB that circulate online are later estimates, and the 1888 report contains none of them. What the report does record is how far the sounds were reported, and that distance is the main basis of Krakatoa’s reputation as the loudest sound in history.
Was the sound of Krakatoa ever recorded?
We found none. The 1888 report has no recording, and the videos titled “real sound of Krakatoa” cite no source for their audio. What was recorded is the pressure wave, as pen traces on barographs.
Krakatoa or Hunga Tonga: which air wave was bigger?
Similar. Over Britain both gave a pressure swing of about 2 hectopascals, 1.7 to 2.7 for Krakatoa and a little over 2 for Tonga. The Tonga papers call the number of passes “approximately the same” as Krakatoa’s and the Krakatoa pulse about 30% longer. Tonga’s bang was reported farther away, nearly 10,000 km in Alaska against 4,800 km, though many more people were listening and reporting in 2022.
Where the explosions were heard
| Place | Distance, km | What the report says |
|---|---|---|
| Singapore | 840 | “Two steamers were sent on the 27th to look for the vessel which was supposed to be firing guns as signals of distress” (p. 82) |
| Perth | 3,072 | “sounds as of the firing of guns inland” (p. 84) |
| Ceylon (Mannar) | 3,288 | “Loud sounds, resembling the report of distant cannon” (p. 86) |
| Alice Springs | 3,616 | “Two distinct reports, similar to the discharge of a rifle” (p. 84) |
| Diego Garcia | 3,647 | “supposed to be those of guns fired by a vessel in distress” (p. 80) |
| Rodrigues | 4,780 | “like the distant roars of heavy guns” (p. 87) |
Distances are recomputed as great circles from today’s coordinates; the report’s own figures, in miles, agree within 1%.
How the model works, and what it does not know
The wave is one ring that leaves the volcano at its start time, grows at a fixed speed, closes on the far side and returns. For Krakatoa the speed, 310 m/s, is the least-squares fit to 220 recorded passages at 44 stations, with the start fixed at the explosion, 02:56 GMT. For Tonga the speed, 315 m/s, and the start, 04:15 UTC, come from the papers, and the model is then checked against six open barometer records.
The model gives when the leading edge of the ring reaches a place. A barometer’s deepest dip or highest reading comes later, by about 36 minutes in 1883 and 10 to 45 minutes in 2022. The model has one speed for all directions, and the real wave was faster with the Earth’s rotation than against it, so the model is closest on the first passes. The glow on the globe is a drawing of the pulse. Its length, the distance from rise to dip, follows the records, and it fades with each lap by the swings measured at Reading in 2022, 2.38 hectopascals falling to under 0.2; for 1883 that fade is an assumption, because the 1888 report prints no amplitudes for the passes.
Times in 1883 are local mean time: Greenwich time moved by four minutes for every degree of longitude, because time zones were not yet in general use. Times in 2022 are in the town’s time zone today.
Sources
- Symons, G. J. (ed.), 1888. The Eruption of Krakatoa, and Subsequent Phenomena. Report of the Krakatoa Committee of the Royal Society. Part II, “On the air waves and sounds caused by the eruption of Krakatoa in August 1883”, prepared in the Meteorological Office and presented by Lt-Gen. R. Strachey, pp. 57–88. Public domain. archive.org/details/eruptionofkrakat00roya
- Matoza, R. S. et al., 2022. Atmospheric waves and global seismoacoustic observations of the January 2022 Hunga eruption, Tonga. Science 377, 95–100. doi.org/10.1126/science.abo7063
- Wright, C. J. et al., 2022. Surface-to-space atmospheric waves from Hunga Tonga–Hunga Haʻapai eruption. Nature 609, 741–746. doi.org/10.1038/s41586-022-05012-5
- Burt, S., 2022. Multiple airwaves crossing Britain and Ireland following the eruption of Hunga Tonga–Hunga Ha’apai on 15 January 2022. Weather 77(3), 76–81. doi.org/10.1002/wea.4182
- Pressure records, 15–16 January 2022: Deutscher Wetterdienst (Potsdam, Hamburg), MeteoSwiss (Zürich), Finnish Meteorological Institute (Helsinki), HungaroMet (Nyíregyháza), NOAA ASOS one-minute data via the Iowa Environmental Mesonet (New York LaGuardia). Graphically reproduced data. opendata.dwd.de/climate_environment/CDC
- Volcano positions: Smithsonian Global Volcanism Program (Krakatau 262000; Hunga Tonga–Hunga Haʻapai 243040). volcano.si.edu
- Coastlines: Natural Earth, public domain. Towns: GeoNames (cities of 15,000 or more), CC BY 4.0. www.geonames.org