Four climate charts from the Southern Hemisphere

The ozone hole

The ozone hole is the area over Antarctica where the total ozone overhead falls below 220 Dobson units. NASA draws the line there because readings that low were not observed before 1979. A British Antarctic Survey team reported the losses in Nature in 1985, from ground measurements at Halley Bay.

The chlorine that does the damage comes mostly from chlorofluorocarbons (CFCs). CFCs are so unreactive in the lower atmosphere that they last long enough to drift up into the stratosphere, where ultraviolet light breaks them apart and frees their chlorine. Over Antarctica, the winter cold forms polar stratospheric clouds. Reactions on the cloud particles turn stable chlorine compounds into forms that sunlight splits apart when it returns in September. Each freed chlorine atom then destroys ozone in a catalytic cycle, breaking down thousands of ozone molecules before it is tied up again. The net reaction turns ozone into ordinary oxygen, 2 O₃ → 3 O₂. Bromine from halons adds to the loss.

The Montreal Protocol, signed in 1987, phased out CFC production worldwide by 2010. The hole stopped growing around 2000 but has not shrunk in a straight line since. The 2022 WMO/UNEP ozone assessment expects ozone over Antarctica to return to its 1980 level around 2066.

CFC-12 and the size of the hole

CFC-12 is the most abundant CFC in the atmosphere and lasts about a century there. In Southern Hemisphere air it climbed from 275 parts per trillion (ppt) in 1979 to a peak of 541 ppt in 2002, held almost flat through 2004, and has fallen slowly since, to 477 ppt in 2025. That is 12% below the peak.

The 2017 site called the link between CFC-12 and the size of the hole proportional. It isn't, for three reasons.

• The hole has an edge. It forms inside the polar vortex, a ring of strong winds that seals in cold air over Antarctica each winter. Through the 1980s the hole grew fast as CFC-12 rose. Its growth slowed in the 1990s because the hole cannot spread past the vortex, and in cold years chlorine was already destroying nearly all the ozone it could reach inside it.
• The stratosphere lags the surface. The chart below pairs each year's hole with CFC-12 measured near the ground that year. Air takes years to rise into the stratosphere and reach the pole, so the chlorine over Antarctica reflects surface air from several years earlier. Other chlorine and bromine gases matter too, so CFC-12 stands in for the total rather than measuring it.
• Weather moves single years. Cold winters form more polar stratospheric clouds, and warm, disturbed ones form fewer. In late September 2002 the Southern Hemisphere had its first recorded major sudden stratospheric warming, which split the vortex in two. That year's hole averaged 12.0 million km², the smallest since 1988, in the same year CFC-12 peaked. The next year it was back to 25.8. Similar conditions in 1988 and 2019 also gave small holes, and 2019's 9.3 million km² was the smallest since 1983. One year's hole says more about that winter's weather than about chlorine, so telling the two apart takes decades of data and models that account for temperature.

Winter in Chile

Pratyush's 2017 chart plotted Chile's June and July temperatures back to 1855, each month as its own dot, in absolute degrees. June and July are winter months in Chile. This version averages the two months for each year and shows how far that average sits above or below its 1951 to 1980 level (about 4.9 °C), with Berkeley Earth's 95% uncertainty band. Climate records are usually compared as differences from a baseline period, because those differences vary smoothly from place to place even when absolute temperatures do not.

Chile is a hard place to average. It spans 38 degrees of latitude, from the Atacama Desert to Tierra del Fuego, and the cold Humboldt Current runs along its coast. The national series still warms. June to July temperatures averaged 0.48 °C above the 1951 to 1980 level in 1991 to 2020, and 0.82 °C above it in 2011 to 2020. The straight-line trend from 1960 to 2020 is about 0.14 °C per decade. Before 1900 few stations reported, so the uncertainty band is wide. Berkeley Earth last generated its Chile file in January 2021, so the series ends in 2020.

Carbon dioxide from Chile

My 2017 chart used a World Bank series of carbon dioxide emissions per person, built on CDIAC data. The World Bank has since retired it. Its replacement, from the European Commission's EDGAR database, starts in 1970 and uses different methods: for 2013 the old series gives 4.73 metric tons per person and the new one 4.83. Joining them would invent a jump, so this chart uses only the new series and adds the world average for scale.

Chile's emissions per person fell from 3.15 metric tons in 1973 to 2.42 in 1977, across a 1975 recession in which its economy shrank 12.3%. They fell again in 1982, when it shrank 11.5%, and stayed below 2 tons until 1988. From then on they mostly rose, reached the world average in 2013, ran slightly above it from 2016 to 2020, and peaked at 4.96 metric tons in 2019. The series then drops 15% in 2023, in Chile's total emissions as well as per person. I haven't traced the cause, and the newest years in an inventory like this are the likeliest to be revised.

What changed since 2017

• Current data. Three of the 2017 sources are gone. CDIAC, which hosted the CFC-12 record, has closed. theozonehole.com no longer publishes data. The World Bank retired its CO₂ series. Each chart now uses one current source from start to end, retrieved on November 2, 2026.
• Units on every axis. The 2017 CFC axis read “CFC Mixing Ratio” with no unit. It is parts per trillion.
• Years as numbers. We fed whole-number years to a Bokeh datetime axis, which reads numbers as milliseconds since 1970, then covered it up with a plain number formatter. These charts use a numeric year axis.
• Data in files. The 2017 plots had their numbers typed into the Python code, and my CO₂ script listed years through 2014 for data that ended in 2013. Now a script downloads each source and writes the CSV files the charts read, and a check runs before every deploy.
• A steadier ozone measure. The 2017 numbers were each year's largest single-day hole area. These are NASA's September 7 to October 13 averages, which one unusual day can't swing.
• Measured CFC-12. The 2017 values came from a compiled history whose 2014 value was extrapolated from 2013. These are NOAA network measurements, on a slightly different calibration scale. Both records are flat within 1 ppt from 2001 to 2004. The old one tops out in 2003, the new one in 2002.
• Temperature as a difference from normal. June and July are averaged for each year and shown against 1951 to 1980, with uncertainty. The 2017 text said the record ran “since the 1700s”, but in that data Chile's record started in 1855.
• Corrected science. CFCs are chlorofluorocarbons, not “chlorinated fluorocarbons”. Chlorine turns ozone into ordinary oxygen (O₂), not a mix of O₂ and single oxygen atoms. And CFC-12 and the size of the hole are linked, but not in proportion.
• Works on more screens. The page fits phones and follows dark mode. Every chart has a data table and can be read with the keyboard: focus a chart, then use the arrow keys.

Credits

2017: Pratyush Kapur made the Chile temperature chart. Hannah Slocumb made both ozone charts. Siena Guerrero made the CO₂ chart and the references page and put the site together. It used the free “An Ocean of Sky” template by Bryant Smith.

2026 rebuild: Siena Guerrero.

Sources

• NASA Ozone Watch, Annual Records. Ozone hole area, average of September 7 to October 13.
• Dutton, G. S., et al. (2026). Combined Atmospheric Chlorofluorocarbon-12 Dry Air Mole Fractions from the NOAA GML Halocarbons Sampling Network, 1977-2026. doi:10.15138/PJ63-H440. Annual means of the Southern Hemisphere monthly series, for calendar years with all 12 months.
• Berkeley Earth, Chile regional land temperature (chile-TAVG-Trend.txt), analysis run January 2021. Licensed CC BY-NC 4.0.
• World Bank, World Development Indicators, EN.GHG.CO2.PC.CE.AR5, from the EDGAR Community GHG Database (JRC, European Commission, 2024). GDP growth from NY.GDP.MKTP.KD.ZG. Licensed CC BY 4.0.
• Farman, J. C., Gardiner, B. G., and Shanklin, J. D. (1985). Large losses of total ozone in Antarctica reveal seasonal ClOx/NOx interaction. Nature 315, 207-210. doi:10.1038/315207a0.
• WMO/UNEP, Scientific Assessment of Ozone Depletion: 2022.
• Yoo, J.-H., and Chun, H.-Y. (2025). Role of in situ-excited planetary waves in polar vortex splitting during the 2002 Southern Hemisphere sudden stratospheric warming event. Atmospheric Chemistry and Physics 25, 13651-13664.
• NOAA Climate.gov, Five questions about 2019's record-small ozone hole.