Over the past 2,000 years, volcanoes have played a larger role in natural temperature variability than previously thought, and their climatic effects may have contributed to past societal and economic change, an international team of researchers found.

The University of Cambridge-led team, which includes a University of Arizona researcher, used samples from more than 9,000 living and dead trees to obtain a precise yearly record of summer temperatures in North America and Eurasia, dating back to the year A.D. 1. This revealed colder and warmer periods that they then compared with records for very large volcanic eruptions, as well as major historical events.

Crucial to the accuracy of the dataset was the use of the same number of data points across the entire 2,000 years. Previous reconstructions of climate over this extended period have been biased by over-representation of trees from more recent times.

"As we sharpen the focus on our view of the past two millennia by using annually dated records, especially tree rings, features such as the periods of frequent volcanoes affecting climate – and periods with few or none – emerge clearly, said study co-author Malcolm Hughes, Regents Professor Emeritus in the UArizona Laboratory of Tree-Ring Research. "We've been aware of this for some time, but it's really gratifying to see it show up repeatedly and consistently,"

The results, reported in the journal Dendrochronologia, show that the effect of volcanoes on global temperature changes is even greater than had been recognized, although the researchers stress that their work in no way diminishes the significance of human-caused climate change.

Instead, the researchers say, the study contributes to our understanding of the natural causes and societal consequences of summer temperature changes over the past 2,000 years.

"There is so much we can determine about past climate conditions from the information in tree rings, but we have far more information from newer trees than we do for trees which lived a thousand years or more ago," said lead study author Ulf Büntgen from Cambridge's Department of  Geography. "Removing some of the data from the more recent past levels the playing field for the whole 2,000-year period we're looking at, so in the end, we gain a more accurate understanding of natural versus anthropogenic climate change."

Comparing the data from tree rings against evidence from ice cores, the researchers were able to identify the effect of past volcanic eruptions on summer temperatures.

Large volcanic eruptions can lower global average temperatures by fractions of a degree Celsius, with strongest effects in parts of North America and Eurasia. The main factor is the amount of sulphur emitted during the eruption that reaches the stratosphere, where it forms minute particles that block some sunlight from reaching the surface. This can result in shorter growing seasons and cooler temperatures that lead, in turn, to reduced harvests. Conversely, in periods when fewer large eruptions occurred, the Earth is able to absorb more heat from the sun, and temperatures rise.

"Some climate models assume that the effect of volcanoes is punctuated and short," said Büntgen. "However, if you look at the cumulative effect over a whole century, this effect can be much longer. In part, we can explain warm conditions during the third, 10th and 11th centuries through a comparative lack of eruptions."

Reconstructed summer temperatures in the 280s, 990s and 1020s, when cooling due to volcanic activity was low, were comparable to modern conditions until 2010.

Compared with existing large-scale temperature reconstructions of the past 1,200 to 2,000 years, the study reveals a greater preindustrial summer temperature variability, including strong evidence for the Late Antique Little Ice Age in the sixth and seventh centuries.

Then, working with historians, the scientists found that relatively constant warmth during Roman and medieval periods, when large volcanic eruptions were less frequent, often coincided with societal prosperity and political stability in Europe and China. However, the periods characterized by more prolific volcanism often coincided with times of conflict and economic decline.

"This work enriches the historical record by providing environmental context for events in society. The mid-sixth century events in the eastern Mediterranean region – such as the Justinian plague, for example – took place in the context of a drastic cooling felt around the Northern Hemisphere," Hughes said.

"Interpreting history is always challenging," said Clive Oppenheimer, the lead volcanologist of the study at Cambridge. "So many factors come into play – politics, economics, culture. But a big eruption that leads to widespread declines in grain production can hurt millions of people. Hunger can lead to famine, disease, conflict and migration. We see much evidence of this in the historical record. We knew that large eruptions could have these effects, especially when societies were already stressed, but I was surprised to see the opposite effect so clearly in our data – that centuries with rather few eruptions had warmer summers than the long-term average."

The new temperature reconstructions provide deeper insights into historical periods in which climactic changes and their associated environmental responses have had an outsized impact on human history. This has clear implications for the present and future. As climate change accelerates, extreme events, such as floods, drought, storms and wildfires, will become more frequent.

"Humans have no effect on whether or not a volcano erupts, but the warming trend we are seeing right now is certainly related to human activity," said Büntgen. "While nothing about the future is certain, we would do well to learn how climate change has affected human civilization in the past."

"We point out that our estimates of the cooling associated with volcanic eruptions in the first millennium suggest higher sensitivity of the climate to changing radiation balance than currently accepted," Hughes said. "If so, estimates of coming greenhouse warming would have to be increased. We take considerable care to point out that more and different data are needed to check this."