The history of Earth’s climate is punctuated by a succession of named intervals associated with prolonged shifts to warmer, colder, wetter or drier conditions. During the Common Era (the past 2,000 years), the two best-known such climate epochs are the Little Ice Age1 and the Medieval Climate Anomaly2 (also called the Medieval Warm Period; Fig. 1). The former was a cool period that extended from the sixteenth to the late nineteenth centuries; the latter was a warm, dry period between ad 950 and 1250. Many assume that these intervals had a global impact. But in a paper in Nature3 and in a companion paper in Nature Geoscience4, Neukom et al. demonstrate that these and earlier climate epochs in the Common Era were much smaller in scope than the near-global reach of current human-induced warming.
Because thermometer measurements of air near Earth’s surface before ad 1850 are not widely available, we rely on archives of proxy data to extend our perspective on climate further back in time. Trees in cold Arctic or alpine forests have annual rings with widths and wood densities that reflect year-to-year variations in summer temperature5. And because the chemical make-up of seawater depends on its temperature, massive corals build endoskeletons that contain a permanent geochemical record of past warming and cooling6. Other geological and biological archives that encode temperature information into their physical structure, substance or geochemical composition include lake sediments, glacier ice and bivalve molluscs (such as clams, oysters and mussels). Such archives likewise serve as ‘palaeothermometers’ that record temperatures stretching hundreds or thousands of years into the past.
Neukom et al. weave all of this evidence into a detailed global portrait of surface temperatures that spans the past two millennia. The foundation for their work is provided by the PAGES 2k proxy temperature database7. This community-sourced compilation includes nearly 700 records from trees, ice, sediment, corals, cave deposits, documentary evidence and other archives. Partly because the database incorporates so much information, the authors can chart the geographical extent of unusually warm or cold conditions across the entire planet by year.
The team reports in Nature that, although the Little Ice Age was the coldest epoch of the past millennium, the timing of the lowest temperatures varied from place to place. Two-fifths of the planet were subjected to the coldest weather during the mid-nineteenth century, but the deepest chill occurred several centuries earlier in other regions. And even at the height of the Medieval Climate Anomaly, only 40% of Earth’s surface reached peak temperatures at the same time. Using the same metrics, global warming today is unparalleled: for 98% of the planet’s surface, the warmest period of the Common Era occurred in the late twentieth century — the authors’ analysis does not encompass the continued warming in the early twenty-first century, because many of their proxy records were collected more than two decades ago.
In 2005, palaeoclimatologists John Matthews and Keith Briffa1 cautioned against deeming the Little Ice Age an “uninterrupted, globally synchronous, cold period”. These new results certainly bolster their point of view. And we can be confident in that conclusion because Neukom et al. carried out an exhaustive set of experiments to confirm that their findings were unaffected by their choice of statistical tools to relate the proxy network to thermometer measurements.
Unfortunately, limitations inherent in the proxies themselves probably still hamper our ability to compare warm or cool intervals with each other throughout the entire Common Era. Tree-ring records, the most frequently used proxy archive in the PAGES 2k database, are sometimes unreliable in registering slow climate changes over several centuries or longer8. Moreover, some other proxies — particularly records from marine and lake sediments — exaggerate variations at multidecadal or centennial timescales9,10. It is still an open question how well we can compare global temperatures across this entire 2,000-year span.
We can be more certain of how and why Earth warms or cools over decadal and multidecadal timescales. In their companion paper in Nature Geoscience, Neukom et al. show that, in the pre-industrial period (ad 1300–1800), major volcanic eruptions (or the lack of such eruptions) were the main factor behind cold (or warm) swings that persisted for several decades. Shifts in greenhouse-gas concentrations had a smaller, but still detectable, imprint. The team found no indication that variations in the Sun’s radiation output affected mean global temperature over the same time frames.
In general, physics-based climate models accurately reproduce proxy estimates of our climate’s history over the past millennium. However, these models exaggerate the degree of cooling caused by the two largest volcanic eruptions of the Common Era: the ad 1257 Samalas and the ad 1815 Tambora eruptions in Indonesia11. This discrepancy implies that we cannot be sure how bitter a chill would follow a similar eruption in the future.
The familiar maxim that the climate is always changing is certainly true. But even when we push our perspective back to the earliest days of the Roman Empire, we cannot discern any event that is remotely equivalent — either in degree or extent — to the warming over the past few decades. Today’s climate stands apart in its torrid global synchrony.
Nature 571, 483-484 (2019)