1816 · Tambora · Europe
In April 1815 Mount Tambora erupted. The next summer was exceptionally cold for western Europe: Geneva had not a single cloud-free day in June, and Prague rye rose to past four times its 1813 price by 1817. But the temperature reconstruction tells a different story about Bucharest: a near-normal summer. Live the summer week by week, then find out how it was in your town.
Begin the summer of 1816 Skip to your town
Map: NCEI data · 17 sources at the end
Each square is a half-degree cell of the reconstructed temperatures. Blue is cold against the 1901–2000 summer mean; red is warm. Press a marker to see what happened there.
A reconstruction, not a thermometer: a seasonal mean, with no per-cell error bar.
Not every week was as bad, and not everywhere. These are the moments with a date and a place — each quoted from its source.
5–11 April 1815 · Sumbawa, Indonesia (off the map)
After a week of rumbling, Mount Tambora reaches its climactic phase on 10–11 April: the summit collapses into a 6 by 7 kilometre caldera, releasing some 60 million tonnes of sulphur dioxide — a VEI-7 eruption.
Raible et al. 2016 · Whelley & Newhall 2015
December 1808 · Somewhere in the tropics (off the map)
Seven years before Tambora, another great volcano erupted somewhere in the tropics. Its location is unknown and nobody described the eruption; we know it from the sulphur left in the ice of both poles and from a veil seen in the sky over Bogotá and Lima in December 1808.
Cole-Dai et al. 2009 · Guevara-Murua et al. 2014
5–11 June 1816 · New England, America (off the map)
A cold wave with snowstorms strikes New England in full June.
Chenoweth 2009
June 1816 · Geneva
Geneva sees not a single cloud-free day in the whole of June. Afternoons run 3–4 degrees colder than those years' normal, and the rain comes not harder but more often: 80% more rain, from more rainy days.
Auchmann et al. 2012
17–18 June 1816 · Villa Diodati, near Geneva
Rain keeps Byron, the Shelleys and Polidori indoors. The ghost stories begin: on 18 June, Polidori notes in his diary that he began his story after tea, and by midnight they had «really begun to talk ghostly», over verses of Coleridge. Mary would begin Frankenstein that summer — but she did not finish it then.
M. Shelley 1831 · Polidori's diary
July 1816 · Villa Diodati
In July, Byron writes “Darkness” — a dream of a dead sun and a frozen world. It would appear in print on 5 December 1816. In Geneva, summer still does not come.
Byron, ed. Coleridge 1901
14 August 1816 · Geneva
The warmest afternoon of the summer in Geneva measures 26.9 °C: no day passed 30 degrees. And high-pressure days you could count on one hand: a single one, against 20 in an ordinary summer.
Auchmann et al. 2012
after August 1816 · Switzerland
After August, bread runs short in Switzerland: the price doubles, then rises again, and dinner guests are asked to bring their own bread. On 30 August, Geneva's officials discussed buying grain abroad.
Clubbe 1991
1816 · Transylvania
In Transylvania, 1816 brings catastrophic floods, as in the German-Austrian lands, followed by early frosts that destroy what was left of the harvests. It is the fourth year of the great famine of 1813–1817: wars, financial crisis and epizootics pile onto the broken weather.
Balog 2023 · Ciorba 2006
1817 · Prague
Wholesale prices, climbing since 1813, explode after Tambora and peak in 1817: in Prague, prices climbed to about 4.5 times their 1813 level for rye and barley, 3 times for wheat. The good harvest of 1817 stops the rise.
Brázdil et al. 2017
summer 1816 · Loire valley, France
In the reconstruction, the summer's coldest point sits over the Loire: −3.1 degrees against normal. Across Europe, 488 cells run over 2 degrees below normal — nearly one in ten.
this page's computation, from the Luterbacher grid
Pick a town. The page finds the grid cell beneath it and subtracts the 1901–2000 summer mean from the summer of 1816. The answer may surprise you.
In Bucharest, summer 1816: +0.3 °C near normal (summer mean 22.2 °C, against a normal mean of 21.9 °C). The value is a reconstruction for the grid cell — your street's real temperature may differ.
A seasonal mean doesn't tell how the weeks went. The twist: on this map, where you live wasn't cold.
| Bucharest | +0.3 °C near normal | 22.2 °C |
| Cluj-Napoca | +0.0 °C near normal | 18.4 °C |
| Iași | +0.4 °C near normal | 20.5 °C |
| Timișoara | −0.5 °C below normal | 20.6 °C |
| Brașov | +0.3 °C near normal | 17.5 °C |
| Constanța | +0.5 °C near normal | 21.9 °C |
| Craiova | +0.1 °C near normal | 21.9 °C |
| Sibiu | +0.1 °C near normal | 17.8 °C |
| Oradea | −0.4 °C near normal | 19.9 °C |
| Galați | +0.4 °C near normal | 21.7 °C |
| Chișinău | +0.4 °C near normal | 20.6 °C |
| Chernivtsi | +0.4 °C near normal | 18.6 °C |
| Geneva | −2.6 °C below normal | 14.4 °C |
| London | −2.2 °C below normal | 13.6 °C |
| Paris | −2.9 °C below normal | 15.7 °C |
| Vienna | −1.3 °C below normal | 17.6 °C |
| Berlin | −1.2 °C below normal | 16.2 °C |
| Rome | −1.4 °C below normal | 21.5 °C |
| Madrid | −1.7 °C below normal | 21.8 °C |
| Warsaw | +0.0 °C near normal | 17.5 °C |
| Budapest | −1.1 °C below normal | 17.7 °C |
| Belgrade | −0.6 °C below normal | 21.3 °C |
| Sofia | +0.3 °C near normal | 18.9 °C |
| Athens | +0.2 °C near normal | 25.1 °C |
| Lisbon | −0.8 °C below normal | 20.4 °C |
| Dublin | −1.5 °C below normal | 12.5 °C |
No single culprit. Three causes lined up in the same decade — and the models say how much each pulled.
−5 W/m²
The volcano's peak forcing: about 60 million tonnes of emitted sulphur dioxide. In models, with Tambora's forcing, the Central European cold becomes up to 100 times more likely — and the rain 1.5–3 times.
Raible et al. 2016 · Schurer et al. 2019
the second volcano
The tropical eruption nobody saw had already cooled the decade. In many detailed Northern Hemisphere and tropical records, 1810–1819 is probably the coldest decade in at least 500 years — with two volcanoes to blame, not one.
Cole-Dai et al. 2009
the Dalton Minimum
The sun sat in the Dalton Minimum (1780–1840 in Anet et al.'s window), and the atmospheric circulation kept the clouds over the west. Models say circulation alone explains only part of the cold — but much of the rain comes from it.
Anet et al. 2014 · Schurer et al. 2019
The models' verdict: Tambora plays the leading role in the Central European cold — without volcanic forcing, an equally cold summer was highly unlikely.
Look at the map again. The cold sits in the west: the western-Europe box (43–52°N, 5°W–15°E, 627 cells) averages 2.2 degrees below normal. The Romania box (43.5–48.5°N, 20–30°E, 196 cells) — 0.1 degrees above normal. Bucharest +0.3. Iași +0.4. Cluj +0.0.
But a season's mean is not the weather people lived. In Transylvania, the summer of 1816 came with catastrophic floods and early frosts — on top of a famine already underway since 1813, from wars, financial crisis and epizootics. A near-normal mean can hide exactly the weeks that destroyed the harvest.
The narrative chronicles we searched are silent: the Wallachian chronicle stops in early 1815, the Moldavian verse chronicle is about 1821, and systematic registers begin only in 1831–1832. But their silence doesn't mean nothing happened: in Transylvania, the famine historiography tells another story — our first search had missed it.
The map shows an average; people lived through weeks.
Four questions from the page. The answers are above, sources included.
1 In which month of 1816 did Geneva have no cloud-free day?
2 When did grain prices peak in Prague?
3 Which town had the colder summer of 1816, on the reconstruction?
4 What did the summer of 1816 bring in Transylvania?
The map and verdicts use the seasonal series from the NCEI archive of the European reconstruction: 0.5° cells, summers 1500–2002. The winter and summer values were recomputed against Luterbacher et al. 2004 (fitted to Mitchell and Jones 2005) and differ slightly from the paper's. Before 1901 the values are reconstructed from instrumental, documentary and natural proxy series; 1901–2002 are observational estimates.
The “Romania” and “western-Europe” boxes are grid rectangles (43.5–48.5°N, 20–30°E, n=196; and 43–52°N, 5°W–15°E, n=627), not border-clipped means.
The anomaly is the summer of 1816 minus the 1901–2000 summer mean, following the publishing team's convention. Each town gets its nearest land cell. The below/above-normal labels apply from ±0.5 °C on the unrounded value.
The grid ships no per-cell error margin — which is why the page says “reconstruction”, not “measurement”, and shows the seasonal mean, not weeks.
The New England snowstorm is cited from the title of Chenoweth's 2009 study, via Auchmann's bibliography — that is all that could be verified.
The page about skies painted after eruptions is a different chapter: see cerul-pictat.
The neighbouring chapter: cerul-pictat.
All links were checked at publication.