 I'm Tim Osborne from the Climatic Research Unit at the University of East Anglia and I research into many aspects of climate change including natural changes in the past as well as the instrumental temperature record and future projections of climate change as well. Okay so paleoclimate is when we're trying to understand what happened to the climate prior to the period when we have instruments so before the invention or the widespread deployment of thermometers and rain gauges and things like that and we use indirect estimates of the climate. So things that are living in their environment will gradually record some information about their local environment including the climate. So for example trees, grain will include something about their environment, how good it is to go each year, soil moisture, temperature etc. Ice cores will be developing, will be providing the archive of the snowfall each year but also some of the chemical composition of the snow which is interesting and records like that. So things which are living in the environment and which are accumulating some kind of record of their environment which includes climate but also includes other features of their environment. Well geological proxies go back hundreds of millions of years even but the ones I'm most interested in are ones with a kind of a finer resolution that can tell you about individual years or decades for example truings and those truing records if they're based on living trees obviously they're limited by the lifespan of a tree which might be a few hundred years occasionally a thousand years or more but if you're willing to use or if you're able to use dead material that's been preserved in some way then you can build up a composite record by overlapping modern living trees with trees which died a few hundred years ago and then overlap further back in time and trees which died before that and so some of the longest records are getting on for 10,000 years in length but it's fairly few and far between but there's a growing number especially in the northern hemisphere for you know one to two thousand years length. The oldest trees tend to be the ones that grow slowest so they're often growing in fairly arid regions maybe mountainous areas where there's not much rainfall and the soils aren't able to store much rainfall so the slower they grow the longer they tend to live and the faster growing trees tend to live shorter times in terms of the longest trees some of the north american various different species of pine tree which can be rather long-lived. So in order to understand something about past climate from trees the first thing is to consider where to sample the trees from and so if you want to get a record of past temperature variations then you would sample the trees from a cold region where their growth is limited by how warm it is in the summer so you might go higher mountains or you might go to the high latitudes towards the poles. So if you get a tree growing in such conditions it tends to the amount of growth it puts on each year which is reflected in the thickness of the rings the annual growth ring that will be larger i.e. bigger rings in a year when the conditions are more conducive to good growth so in those type of trees that are limited by temperature a warm summer will give you conditions that are more conducive to greater growth and therefore a thicker tree ring whereas if it's a very cold year the tree will be the tree ring will be rather narrow. The link between climate and trees is really not that strong in individual trees because there are so many other factors that affect trees growth so it's really crucial to get multiple records so you have to get samples from many different trees at one site and ideally samples from different sites in the vicinity as well because only by looking at the common features in those tree ring width records that you can build a picture of the temperature because if some tree is affected by some localized feature it may not be reflected in other trees in the site or in other nearby sites whereas if if a summer is warmer than average it will tend to affect the whole region in one go and they're foresharpen all the trees together. In order to understand past climate we need more and more data we have some data that tells us a hazy picture of what happened in the past but we need more data to get a clearer picture it can never be as accurate as having real data from thermometers of course but you can either improve the accuracy by having more records or you can assess the accuracy by comparing different records so if you've got different tree rings from different locations or if you have different types of climate proxy maybe tree rings and ice cores and maybe other things like tropical corals growing in the oceans you can make comparisons between those different proxy types and see whether they show a consistent picture or not. Some of the difficulties in doing that though are tree rings may respond to summer temperatures, ice cores may respond to some other aspect of climate, corals to ocean temperatures and therefore you don't necessarily expect them all to agree so you have to use our understanding of how these individual proxies behave in order to interpret how much in agreement or how much disagreement there is. The coverage of proxy data across the globe is very variable so tree rings of course are limited to the land but actually the ones that are sensitive to temperature even on the land are limited to the colder parts of the land to the higher latitudes or the higher elevations in the mountains and so we can build up a good picture perhaps in the northern hemisphere land areas we can make some estimates of the global temperature from these but they are less reliable because we're assuming that there's some statistical link between what's happening over the land and the northern hemisphere and what's happening further afield and so we can make those inferences but we have to be careful about the extra uncertainty involved. The best records come from the northern hemisphere land masses. In addition to the width of the individual tree rings the density of the wood that makes up the trees also appears to respond to the climate so with warmer summers we tend to get trees which are denser especially at the end of the grand season kind of the late wood that's put down by the trees at the end of the grand season. The density there is linked to summer temperature even more strongly than the width of the ring widths. The divergence issue shows up most prominently in the density data though. Some tree ring widths there is Sherrod but mostly it's tree ring density and even then it's not all the records and what tree ring divergence is is a separation in the trends of the tree data and the temperature data in recent decades so if you go back to the early part of the 20th century there's quite a good correspondence between the tree data and the temperatures so that warmer summers tend to coincide with wider rings or with denser wood in those rings and colder summers with less dense wood or thinner rings. In recent decades this this correspondence still occurs on a year-to-year basis so you've got individual years that are warmer or colder than it still shows up in the tree ring density data and the tree ring widths but the more gradual changes on the time scales of the last 30 years and there's been a separation with density coming down while temperatures continue to rise so if you've got the time series of tree ring density and the time series of summer temperature even in recent decades they tend to go up and down together but there's a gradual separation of the two curves and so the decade or time scale you've got a continuing upward trend in the temperatures and the densities are flat or in many cases downwards but nevertheless embedded on that are some year-to-year changes so hot years tend to still be years with higher tree ring density and thicker ring widths in the trees and colder years tend to be the reverse. There's also divergence at with chewing width data as well as chewing density data but it's less clear and fewer sites show it so quite a lot of chewing width data sets from different sites in the long hemisphere which don't show the divergence. In terms of what's caused the divergence it's quite hard to determine with a current understanding and current data. There's many different possibilities some can be rather mundane related to the process in the data so for example the tree rings themselves may have changed so in the structure of a tree in the interior the older tree rings are made of what's called heartwood and then nearer the outside which are the more recent rings they're sapwood and it's possible that the density has some systematic change as it converts from sapwood to density as it converts from sapwood to heartwood. Obviously the recent ones haven't changed yet that's a possibility people have looked into this it doesn't seem like it explains certainly not all of the phenomenon but it may be that it contributes to it. There's other factors associated with the sampling of the trees it may be because we tend to sample the biggest trees because we want to get the longest records that our modern sample is somewhat biased towards perhaps either fast growing young trees or long or slow growing old trees and that and that causes a bias compared to the older data and then there's also things that can affect the trees that are not related to climate so it could be linked to changes in the amount of sunlight coming through the atmosphere with pollution and atmosphere altering the amount of sunlight reaching the surface. Other people have suggested that there's other climatic influences separate from temperature so for example maybe an increasing drought effect on the trees I'm not particularly convinced by that because that would affect the individual years as well as the as well as the kind of decadal timescale changes and individual years we still quite see we still see quite a good correlation between the tree ring data and the temperatures so there's many possible explanations for the divergence phenomenon but probably it's a combination of multiple factors but we haven't yet determined what those are. So the paper recently by Stein and Hoibers that suggested that changes in the sunlight reaching the trees could explain the divergence and they're back for that by looking back at the earlier record and identifying a change when volcanic eruptions occur which also put aerosol and dust into the atmosphere which can also affect the amount of sunlight reaching the trees and that seemed to be consistent with the divergent with the locations where the divergence is strongest in recent decades but it's really at this stage just a hypothesis that analysis wasn't sufficient to prove this was the case it demonstrates some consistency but I think it needs to be looked in further because there's some elements of it which you could test further with the current data and it may when you do that it may turn out that the hypothesis is perhaps part of the explanation but perhaps it doesn't explain the full extent of the divergence. I mean it's an interesting problem because if you look at an individual record from the tree ring network you hardly see it at all because it's dominated by the year to year variability which is quite coherent between the temperature of measurements and the tree ring measurements but because it's a small effect but affecting a widespread number of trees when you average them together to produce a large regional average or even an average for the whole of the northern hemisphere because it's a common effect on all these different sites it starts to show up more and more the more averaging you do so it's um to answer the question as to what causes it you would ideally like to look at individual sites and case studies where you can really understand what's going on but at those individual sites it's such a small component of the variability that it's a hard thing to even identify let alone explain and so you need to go to the large scales to see it and once you go to the large scales it's actually quite hard to pin down something that would affect all those sites together which is why our view has been that this is the cause of it is likely to be something fairly unique to the 20th century because in order to have this common effect on many different trees across northern hemisphere you need something you know large scale and and and affecting a you know many regions in one in one time period and therefore some anthropogenic pollution related influence or maybe some climate warming related influence could be um part of the explanation some people ask why we have persevered with the data because if it has this problem why not discard it and and move on to something else but actually we don't have as much information as we want about past climate change and so we don't want to throw away something that could be potentially be useful and if you look at this data set of truing density across the northern hemisphere in terms of the correlation between it and between the density and the summer temperatures there aren't any other proxies that have such a strong correlation across hundreds of sites in the northern hemisphere so it's actually even despite the divergence problem it's actually one of our strongest datasets to understand past climate changes so from the paleoclimate data variations of climate in the last thousand years have been quite large in individual regions um but not necessarily coherent so some regions are warm and um while others are perhaps cooler if you take a large scale picture what you see is as generally warm conditions at the beginning of the last thousand years often called the medieval warm period or some name like that and then coming down to a cooler period called the little ice age um perhaps between 1450 and 1850 as it's peak and then after 1850 coming back up to um kind of warmer conditions in the 20th century now into the 21st century and what we can see quite clearly is is that relative to the whole thousand years the beginning and the relatively warm and the little ice age 1450 to 1850 fairly cold um we have some estimates of the degree of warming and cooling between these periods but actually comparing the warmth during the medieval period and the warmth during the um the modern period they're actually really quite similar given our rather wide uncertainty errors you know modern temperatures are well known from the thermometer record but our estimates of the medieval temperatures from these indirect proxies like truings the uncertainty is much wider and so making this comparison is quite tricky so the assessment in the most recent IPCC report was that if you look at these different records these various different published attempts at reconstructing northern hemisphere temperatures for the last thousand years in most of those it's quite likely that the modern temperatures are warmer than the medieval temperatures if you take their uncertainty estimates or their error estimates as published but in the assessment we consider those error estimates were under estimates those error those errors were underestimating the true error and actually the uncertainty is probably larger than we think in the medieval period and therefore we although we um the assessment was that the modern temperatures were likely to be warmer than the medieval temperatures um only at a moderate level of confidence could we make that statement and there are limitations in the proxy data as you go further back in time and there's also fewer of them so many of them may take us back 300 years 400 years 500 years but the number of records that go back the full thousand or more years is fewer and the number of independent studies is less and so they're the reasons why we had happiness confidence in that comparison but nevertheless on the basis of what's published it's more likely that the modern temperatures are warmer than the medieval ones but there's not something we'd be certain about so our best estimates for the little ice age is that there was a cooling effect in most parts of the world of a moderate level but then superimposed on that are regional fluctuations if there's changes in atmosphere or the ocean circulation they can transport heat around and make some areas even cooler and some areas less cool and so the the phenomenon is probably a northern hemisphere wide event maybe a global event depends exactly how you define it but within that there'll be many regional differences in terms of the amount of cooling and whether it's strongest cooling in the summertime or the strongest cooling in the wintertime for example but the difficulty in defining it is is what is normal climate I mean is that at last age the normal climate in which case you don't need anything to bring you in it because that's where it was and what's brought us out of it is increased greenhouse gases from human activity on the other hand if it was if it is clear that that little ice age is cooler than what we would expect then you also need to have some idea of some forcing that got us into it whether it's increased volcanic activity which is certainly the case for certain decadal periods within the little ice age there were some decades where the number of volcanic events was higher than the average and that is associated with the cooling effect but in terms of a multi-century cooling period you know you need to know whether that is a normal climate or whether that was caused by something if it's a normal climate you don't need a cause okay so we know that climate is very naturally but that doesn't mean to say that nothing else can influence it so natural variations don't preclude the fact that humans can have altered the climate equally even if we attribute recent changes to human activity doesn't mean that natural climate change has suddenly stopped and can't either have influenced the climate in the last 100 years or won't influence it in the next 100 years in fact in making future climate projections we should include the contributions from natural climate variability as well as projections of greenhouse gas induced climate change so the temperature changes in the last the last half century or so have been quite interesting because there's been warming from the 1950s up until the late 1990s and then since then the warming has slowed down and that actually is a good indicator that we mustn't forget the role of natural variability there are possible explanations for the slowdown that don't include natural variability but very few of them could explain the the length of the period during which this temperature slowed down as last said now which is in around about 15 years and so there's likely to be some combination of of ongoing climate warming which we can see in the oceans the oceans have still continued to warm but at the surface the warming has slowed down and maybe superimposed on this gradual warming over the last 60 70 years there are periods when natural variability has enhanced the warming at the surface and periods where it's suppressed the warming which is one explanation for the recent changes there's a clear dependency between the currents of sea ice in the Arctic and polar bears and so the fact that climate change is altering the sea ice in the Arctic and will continue to do so in the future means we need to consider what effect that might have on polar bears and some of our research looked at the seasonality of sea ice and it it seems that polar bears in in some regions at least thrive when the sea ice are part of the year and open water other parts of the year but when it's completely sea ice covered or when it's completely open water they seem to do less well and so our research was looking at how this might influence polar bear populations in the in the future and so the more southerly regions where polar bears currently exist you may move from an area where there's some seasonal sea ice to no sea ice all year around and that may be detrimental to the polar bear population but further north in the high Arctic it may be that you improve the conditions for polar bears at least initially because there may be more ice all year round than they would perhaps would ideally like melting of the ice and some of them actually be a beneficial effect of course if the melting continued and and the sea ice covered even in the higher latitudes became less and less maybe that'd be a detrimental effect but now it's fairly uncertain how they'll respond but this is some indications that we found out by talking to experts in the field so one of the things we tried to do with the paper was to separate out the projections of future sea ice from the interpretation of what that means for polar bears because some people like to pull it all together and there may be a polar bear expert and they may have their own view about future sea ice so you ask polar bears and they're concatenating what they think would happen to sea ice together with what their expert opinion is about how that affect polar bears and we wanted to present them with a single scenario and said what happens if this is what happens to sea ice whether you believe that would happen or not what would that what in your opinion would that do to the polar bear populations that was that was the purpose of the paper to separate these things out and to use an expert elicitation study to understand expectation polar bear response and for a large extent it worked quite well there was one respondent who looked at our sea ice projections and said they're wrong I'm going to ignore them I'm going to tell you what I think will happen anyway but apart from that one person the experts were able to give us the you know focus their response on the thing they're experts in which is the polar bear populations okay so one of the really interesting problems at the moment is the slowdown in warming at the surface and it's interesting because it brings in many different aspects of the climate system virtually everything we do has some bearing on it so our understanding of natural climate variability which we can get from the instrumental temperature record but also by going before the instrumental temperature record to look at paleoclimate data through proxies we need that to better understand the level of natural variability that might have contributed to the slowdown but it also brings into into the picture things like the climate sensitivity so how sensitive is the climate to external forceings like the increase in greenhouse gases because one way of estimating climate sensitivity is to look at the real world and see how well how much warming there's been in response to forceings that we've applied to the climate system up until now and obviously the slowdown has affected that because had it not slowed down and the warming continued then we would now be a warmer position than we are it affects climate future climate projections then as well because if the slowdown affects our estimate of climate sensitivity then it will affect our estimates of how sensitive the climate is to future changes as well and so all these all these features are kind of rolled together into this one issue about understanding the the temperature record over the last hundred years or more but also over the last decade or so where it's where it's rather different than the warming in the previous decades well I got into it because I was interested in oceanography initially and I then did a PhD on the role of the oceans in in the climate system and how ocean currents can alter our climate after my PhD I kind of developed into all different realms of climate science so I moved away from the oceans a bit and into atmospheric change and temperature records paleo climate a little bit about climate models and future climate so I cover a bit of everything really climate change is a really difficult problem involving all elements of the climate system from humans through to chemistry and biology and the physics of atmosphere so it's a really difficult problem to solve we understand many aspects of it at least at a global scale but making regional scale projections is still rather fraught with uncertainty and knowing how to respond to it in terms of what global society might do about climate change problem is also difficult in some extent dependent upon those regional projections of climate change but of course if you're more risk averse and you want to avoid the possibility of very adverse impacts then your opinion might be different than if you're willing to take more risks the the experiment we're conducting with the climate system involving moving a large quantity of carbon from stores underground into the atmospheric reservoir such that there's more carbon in the atmosphere than there's been for well hundreds of thousand years actually probably more like millions of years is really quite a large experiment we're doing with the climate system it's a system that we know from past records is rather sensitive to changes and sometimes it can be rather abrupt in how it responds and so therefore there's quite a risk in doing that experiment