Climate Change: Simple, Serious, Solvable › Models and Data
Hothouses and Icehouses: 485 Million Years of CO₂ and Temperature
For most of the last 485 million years, the Earth was much warmer than it is now, and there was much more CO₂ in the air. Zoom in and pan through deep time, one geologic stage at a time, and see how the two have moved together.
Loading the data…
Temperature and CO₂, stage by stage
- Temperature, middle estimate
- CO₂, middle estimate
- Two chances in three
- Nine chances in ten
- The Holocene
- CO₂ today
Each step is one geologic stage, as wide as the stage lasted, from 11,700 years (the Holocene) to more than 20 million. Its height is an average over the whole stage, so short events are smoothed out. Click or tap a stage to read its numbers, or use the arrow keys on the graph.
Higher CO₂, warmer world
- Paleozoic Era
- Mesozoic Era
- Cenozoic Era
- Outside the stretch of time shown
- CO₂ today
Each dot is one stage. The lines through the colored dots cover two chances in three, for temperature (up and down) and CO₂ (across). Zoom and pan above to choose which stages are colored. Click or tap a dot to read its numbers.
About this page
What the data show
The Phanerozoic Eon, the age of visible life, began about 539 million years ago. This page covers all but its first period, the Cambrian: the last 486.85 million years, in 85 steps. Each step is one geologic stage (or two put together), and gives the global mean surface temperature and the CO₂ in the air, averaged over the stage.
The world has been much warmer than it is now for most of that time. The middle estimates run from 11.4 °C, in the Upper Pleistocene stage, the time of the last ice age, to 35.6 °C, in the Turonian stage of the Cretaceous, 93.9 to 89.4 million years ago. The Holocene, the last 11,700 years, averaged 14.0 °C. Counted by how long each stage lasted, 94% of the time was warmer than the Holocene, and only six stages were colder.
CO₂ ran from 239 ppm to 3,330 ppm, and when there was more CO₂, it was warmer. The authors of the reconstruction find “a strong correlation” between the two, which identifies CO₂ “as the dominant control on variations in Phanerozoic global climate” (Judd and others, 2024). Other things mattered too, such as the slowly brightening Sun and the drifting continents, and CO₂ itself rose and fell as volcanoes and the weathering of rock got ahead of each other. Chapter 5 of the book tells that story.
The dashed purple line on the CO₂ graph is the CO₂ in the air today: 426 ppm, the world's average for 2025. It's already more than the average of any stage in the last 11.6 million years of this record. Averages over long stages hide shorter swings, though.
How to use this page
The bar above the graphs chooses the stretch of time to show, for both views. "Jump to" goes straight to an era, a period or an event. The zoom slider sets how much time to show, from all 485 million years down to half a million. The pan slider moves through time, with older times to the left, like the graphs. "Older" and "Younger" move half a window at a time.
"Over time" draws temperature at the top and CO₂ below. Each stage is a flat step, as wide as the stage lasted. The bold line is the middle estimate: the 50th percentile, with even odds of the true value being above or below it. The darker band covers two chances in three (the 16th to the 84th percentile), and the paler band nine chances in ten (the 5th to the 95th). The bars along the bottom name the eras, the periods and, when there's room, the stages. The vertical scale fits the stages in view, so zooming in also stretches the graph up and down.
The CO₂ graph has a scale on which each doubling of CO₂ takes the same height, because each doubling adds about the same push to the climate (Chapter 4). "Temperature against CO₂" draws each stage as one dot, with the same kind of scale for CO₂ across the bottom. The stages in the stretch of time you chose are colored by era; the rest are gray.
Click or tap a stage, on either view, to read its numbers below the graph. With the graph selected, the left and right arrow keys move to the next older or younger stage. Under every graph, a few lines say in numbers what it shows.
The address in your browser changes as you move the controls, so you can bookmark a view or send it to someone. Nothing is sent anywhere as you work.
This page is part of the website for Climate Change: Simple, Serious, Solvable, an open textbook by Scott Denning. It goes with Chapter 5, "Past Climate Change".
How the data were made
No thermometer reaches back 485 million years. The temperatures come from chemical clues in fossils and sediments laid down in the ocean, which record the temperature of the seawater they formed in. Judd and others gathered more than 150,000 of these measurements, from five kinds of clue, into a database called PhanSST (Judd and others, 2022).
The clues are scattered: a few places at a time, often far apart. So the team combined them with a climate model, HadCM3L, run for the geography and conditions of each stage (Valdes and others, 2021). The method, called data assimilation, starts with many model runs and nudges them toward the clues, trusting each clue according to how certain it is. The result is a whole-Earth temperature, and a spread of possible values, for each stage. The reconstruction is called PhanDA.
The CO₂ comes from other clues, in fossil leaves, soils and shells, which the authors compiled for the paper. For the Paleozoic and Mesozoic, they drew heavily on the compilation of Foster and others (2017).
Keep three things in mind. The temperatures depend on the climate model and on how each clue is turned into a temperature. Stage averages smooth out short events. And the uncertainty is real: the bands are wide, especially for CO₂ in the Paleozoic.
How the numbers are worked out
The page shows the numbers in the authors' file, and works out two comparisons with the Holocene.
1. How much warmer or colder than the Holocene?
T is a stage's middle estimate of temperature, in °C, and TH is the Holocene's, 13.99 °C. ΔT is the difference in °C, and ΔTF the same difference in °F: a change of 1 °C is a change of 1.8 °F.
2. How much more CO₂ than the Holocene?
C is a stage's middle estimate of CO₂, in parts per million (ppm), and CH is the Holocene's, 275.91 ppm. R says how many times as much CO₂ there was. n is the number of doublings: 2 to the power n is R, so n is the logarithm of R to the base 2. One doubling is twice the CO₂, two doublings four times, three doublings eight times. Less CO₂ than the Holocene gives a negative n. On a calculator, take the natural logarithm (ln) of R and divide by ln(2), as above.
A worked example
- The Turonian stage, 93.9 to 89.4 million years ago, has a middle estimate of T = 35.56 °C.
- ΔT = 35.56 − 13.99 = 21.57 °C warmer than the Holocene, and ΔTF = 1.8 × 21.57 = 38.83 °F.
- Its CO₂ is C = 753.99 ppm. R = 753.99 ÷ 275.91 = 2.733 times the Holocene's.
- n = ln(2.733) ÷ ln(2) = 1.450 doublings.
- The page says the Turonian was 21.6 °C (38.8 °F) warmer than the Holocene, with 2.7 times the CO₂, or 1.5 doublings.
The percentiles
The authors give five numbers for each stage, for temperature and for CO₂: the 5th, 16th, 50th, 84th and 95th percentiles of their estimates. The 50th percentile is the middle estimate. Between the 16th and the 84th percentiles lie two chances in three, and between the 5th and the 95th, nine chances in ten. The page shows them as they are in the file.
Things to try
- Start with all 485 million years. Find the warmest and coldest stages, and the times the world was colder than the Holocene. How long did those cold times last, compared with the warm ones?
- Jump to the PETM, 56 million years ago: a burst of warming that lasted a few hundred thousand years. It was 33.6 °C, 8.0 °C warmer than the Thanetian stage just before it, and CO₂ rose from 958 to 1,270 ppm.
- Jump to the end-Permian extinction, 252 million years ago, the worst in the history of life. The temperature jumps by 10.5 °C from the Changhsingian to the Induan, but the middle estimates of CO₂ don't rise. Look at how wide the CO₂ bands are there. What could a stage average miss?
- Jump to the late Paleozoic ice age. Which stages were colder than the Holocene, and how much CO₂ did they have?
- Choose the Cenozoic Era, the last 66 million years, and open "Temperature against CO₂". Follow the orange dots from the hot Eocene down to the ice ages. Do temperature and CO₂ fall together?
Where the data come from
The data are the authors' file PhanDA_GMSTandCO2_percentiles.csv, from their GitHub repository, EJJudd/PhanDA, at commit 9f06c0d (downloaded 12 November 2026), shared under the Creative Commons Attribution 4.0 licence (CC BY 4.0). It has the 85 stages, their age bounds and midpoints, and the five percentiles of temperature and CO₂. The ages are the authors'. The PETM, 56.0 to 55.7 million years ago, is a step of its own, so the Ypresian stage after it starts at 55.7 million years ago. The colors of the eras and periods are those of the International Chronostratigraphic Chart, from Macrostrat's list of geologic time.
CO₂ today is NOAA's global average for 2025, from the Global Monitoring Laboratory.
Foster, G. L., Royer, D. L. and Lunt, D. J. (2017). Future climate forcing potentially without precedent in the last 420 million years. Nature Communications, 8, 14845. doi:10.1038/ncomms14845
Judd, E. J., Tierney, J. E., Huber, B. T., Wing, S. L., Lunt, D. J. and others (2022). The PhanSST global database of Phanerozoic sea surface temperature proxy data. Scientific Data, 9, 753. doi:10.1038/s41597-022-01826-0
Judd, E. J., Tierney, J. E., Lunt, D. J., Montañez, I. P., Huber, B. T., Wing, S. L. and Valdes, P. J. (2024). A 485-million-year history of Earth's surface temperature. Science, 385, eadk3705. doi:10.1126/science.adk3705
Valdes, P. J., Scotese, C. R. and Lunt, D. J. (2021). Deep ocean temperatures through time. Climate of the Past, 17, 1483–1506. doi:10.5194/cp-17-1483-2021
The code
Two files run this page, and they use no outside software. The comments in numbers.js use the same step numbers and letters as above.
numbers.js is the arithmetic: the comparisons with the Holocene, the window of time and the stages in it
app.js is the page: the controls, the graphs and the lines of numbers
To keep a copy, save numbers.js or app.js. The page's data are in data/stages.json.
Credits
By Scott Denning, for Climate Change: Simple, Serious, Solvable.
The reconstruction is the work of Emily Judd and her colleagues (Judd and others, 2024). Claude, from Anthropic, built the page from an entry in Scott Denning's list of open climate data.