Whilst we have much to learn about how trees affect climate, the trees themselves are also extremely useful for examining our recent climate history. By looking at how much a tree’s rings grow year on year, you can get a good idea of how hot the summer periods were in that part of the world in that year. Cambridge University’s Ulf Büntgen has been tapping into just such a resource, and I went to meet him at his lab…

Ulf - There are different treeing parameters, there are different tree species, there are also different sites. All this has to be taken into account and then depending on our question, we have to basically adapt the methods, the archive, and the approach. So let's start from a simple example. We are interested in reconstructing past temperature variations. So all these trees, they record something that happens during their growing season, which is roughly the summer from spring into autumn. We have to go to trees that grow under conditions where summer temperatures or changes in summer temperatures are becoming most relevant for the growth. So the northern treeline in Siberia or in Scandinavia where small changes in summer temperature are already reflected in the annual growth rings. These trees would function almost like a meteorological thermometer for us. If it's a little bit warmer during these summer months or weeks, we would expect a wider ring. We also would expect the higher wood density. So how dense is the wood in a certain year? And with density we actually mean the ratio between cell size, that is a lumen, and cell wall thickness. And then another parameter beside ring widths and wood density would be wood anatomical parameter. So we can find really fine grain, go into a single year and then look at the cells and they maybe show us some anatomical features that also refer to a frost event or an insect affiliation. So things that are really happening within one of these growing seasons and are reflected or basically are reflected in anatomical anomalies.

Will - When you get this information, when you have a look at the rings that have grown on a tree year on year, and you see the difference in width. And if you get, say, a wider one after a narrow one, you go, 'we must have had higher temperatures this summer.' Are you able to discern the difference in degrees? How good is your resolution?

Ulf - The first important thing is that it's never the absolute width, but it's a relative change from year to year. So we would measure these different parameters, ring width, density and so on for hundreds of trees for hundreds or thousands of years, right? And then we build tree ring chronologies. Each of the trees that we consider will show slightly different growth variations because each tree is an individual and influenced by different things. If we merge the information, if we average the information from many hundreds of trees, we cancel out the individual 'noise,' so these individual growth variations. And we emphasise the common signal. So then we already get a very strong time series of what this forest basically tells us and what we need. Now that is the answer to your question. We need an overlapping period to instrumental measurements of temperature. So we know, for instance, for the last a hundred or 150 years, in some places, like in England, we even have longer meteorological records, right? And we need the overlap between the proxy and the measured temperatures. And in this period of overlap, which is usually the 20th century, we are building a model. And then we can transfer wide narrow, wide narrow into degree Celsius or millimetre precipitation or a drought index. And then we actually see what is worse that we are taking out of the trees.

Will - A massive advantage, as you've already said, is the fact that you have access to so many samples so close to each other. Because say for instance, you could only look at one tree and it may have had a blight one year, that affected growth. The ability to then look at 400 other trees and go, 'no, that was actually pretty good growth that year' seemingly gives you a big leg up over say drilling a hole in the Antarctic ice sheet.

Ulf - Yes, that's correct. And obviously the different climate proxy archives or records like ice cores, tree rings, they all have their advantages and disadvantages. So ideally the paleoclimatologist at the end wants to create something that we call a multiproxy record or approach, where we basically take the benefits from each of the records and disregard where they are less good.