Valérie Masson-Delmotte is an internationally renowned climate scientist. She was a lead author for the "paleoclimate" section of the fourth assessment report published by the Intergovernmental Panel on Climate Change (IPCC). She was the lead coordinating author for the "information from paleoclimate archives" section in the IPCC's fifth and most recent assessment report. For the IPCC's upcoming sixth assessment report, she is a lead coordinating author for the entire Working Group One, the Physical Scientific Basis, which oversees all other scientific chapters.
At the recent Second Open Science Conference of the International Partnerships in Ice Core Sciences, in Tasmania, our contributing editor, Benoit Lecavalier, conducted a lengthy interview with Dr. Masson-Delmotte. Later this month, we will publish Part II, on public outreach on the future of the IPCC.
Benoit Lecavalier (BL): The historical climate record covers a period of direct measurements that inform us about past states of the climate system. How was our knowledge of the climate extended into the past?
Valérie Masson-Delmotte (VMD): First, I would like to say that the instrumental period is in fact sometimes rather short. A few areas go back three centuries, like the meteorological records in England, while on the global scale the records extend back approximately 150 years at best. In some tropical regions, the precipitation records only go back 50 years. While measurements of Antarctic sea ice only begin with satellite measurements starting in ~1979. The instrumental records are very short, and so it is important to complement them with the use of either early instrumental measurements from the historical archives - for example historical information such as harvest dates, and dates for freezing over lakes and rivers found in Europe, Japan, or China.
Finally, we have to use indirect information from natural archives, where in fact you have continuous recording of environmental parameters related to the mean climate state and variability. Starting basically at the beginning of the 20th century, people have developed methods to make use of all sorts of archives. First from continents, because they were easier to reach, came tree rings and lake sediments, then deep sea sediments, corals, and ice cores.
BL: Can you elaborate on how tree rings bridge the gap between the instrumental and historical period with natural archives, and specifically how they represent a proxy for climate?
VMD: In trees you can measure tree ring widths or density. What I do myself is the measuring of the relative abundance of heavy to light molecules, the atoms in some molecules. I also use the cellulose in the tree rings. In fact, the information on climate that you get depends on the tree's location. It is mostly related to what would be the strongest stress factor for tree growth. In cold areas at the tree line limit, it is temperature, mostly summer temperature, while in some places it can be the availability of moisture.
The tree ring signal therefore depends on the location, and so when people are focusing on temperature then they will look for high elevation and latitude tree ring records where temperature is the key control. These tree ring records are then only selected to build a chronology. We stack the tree ring records and obtain tree samples from different ages with overlap periods with younger trees and older trees to correct for age effects related to the aging of the trees.
Then the tree ring record must be calibrated, which is hard work. We must relate the parameters measured in the tree ring to a climate parameter and it implies a lot of statistical analysis to yield a robust methodology. Then a number of trees are agglomerated for a given region to produce a regional signal, to remove dependence on a single tree in the event that say, an insect attack compromised a tree. Ultimately it is about playing with a large number of records and then agglomerating them at the regional or even hemispheric scale. Depending on the place, tree ring records provide climatic information that can go back in time at the scale of one to two thousand years.
BL: If we want to extend the record even farther back in time, we might look at ice cores. What types of measurements are conducted on ice cores and how do those proxies relate to environmental conditions?
VMD: An ice core is basically a record of atmospheric deposition, so you have information on what remains of past snowfall after it has been eroded by winds. Furthermore, you have information on temperature through a number of direct and indirect parameters. The direct information arises when you drill an ice core - you are left with a hole in the ice and you can measure a temperature profile down this borehole. Any change in surface temperature will be imprinted but filtered by diffusion processes in this temperature profile. Therefore, borehole temperatures are probably the most direct information. When you have summer temperatures above zero degrees you have melt and refreezing and so melt layers form, a crucial piece of information on past temperature from ice core.
Myself, I work on water stable isotopes, again measuring heavy to light water molecules directly in the water from the ice cores. And these are reflecting past changes in moisture transport from ocean and continental evaporation areas towards the central part of the ice sheets. Since the 1960s, we have greatly expanded our physical understanding of how hydrogen and oxygen isotope fractionation relates to past surface temperatures from theoretical models. We can also simulate it in atmospheric models.
BL: We have temperature reconstructions from borehole temperatures, summer melt layers, and water isotopes from ice core studies, we also have additional information from the air trapped in the ice cores. What can be said of the ancient gases trapped in air bubbles, such as CO2?
VMD: The instrumental records on atmospheric CO2 go back to the 1960s. We don’t have high quality CO2 records from Greenland ice cores, but we do have them for Antarctica. The issue is that Greenland is closer to the continents. It receives a lot of organic matter and dust from the nearby continents. There are chemical reactions taking place in the Greenland ice that make the preservation of the atmospheric composition of CO2 problematic. Instead, we rely on the analysis of the air trapped in the firn of Antarctic and in the ice cores. However, Greenland ice core remain very useful for methane and other gases, but not for CO2 unfortunately.
BL: Ice cores yield a great deal of information on past climates. What are some novel inferences that have come out of ice cores in the past several years?
VMD: What I like about ice cores is that we have local information, the amount of accumulation/precipitation, local temperature, then you have original information from the chemical analysis of the ice core, transport of sea salt, transport of dust from the continent, biomass burning from the continents, what we call aerosols, and then we can have more remote information. By looking at water isotopes, we have information on the evaporation conditions of water, not local but at the moisture source. We also have information on climate forcing through the deposition of volcanic aerosols, and through the impact of solar activity on cosmogenic aerosols. This allows us to reconstruct the external drivers of climate change.
Finally, we have global information through the analysis of the atmosphere composition. This helps us place our current changes in a broader context, and it facilitates our understanding of how the global biogeochemical cycles work, from the past methane and CO2 cycle. The new information that arose from the long 800,000 year-old Antarctic ice core, which shed light on glacial-interglacial cycles, demonstrated relationships between atmospheric composition and climate, as well as between ice volume, sea level, and climate. We also have a lot of information on abrupt events of the last glacial periods which are characteristic of spontaneous climate variability, not related to the position of the Earth around the sun, which is the first-order driver of glacial-interglacial changes. These abrupt climate events arise from the internal reorganization of the climate system. We describe these events in more and more detail using data and models but we don’t know exactly the cause.
More recently, a lot of information has arisen that compares the current interglacial with earlier and warmer ones. This helps us understand what happens when the climate is warmer than it is now. This research is not really about analogues but about case studies where you know what is the perturbation to the climate system. Based on this, you can test climate feedbacks under similar ranges of actions or "forcings" [climate influences]. Like what happened when climate in the atmosphere was two to four degrees celsius warmer than today, we have information of the past on that.
We can use this data to test climate models, which is I think very important. Climate models developed for today based on process analyses and physics can be further validated by testing them on case studies for extreme cold and extreme warm conditions, to make sure that the models have the right feedbacks. At the moment, it appears we can start testing the models not only for the mean states but also for transient changes, like glacial-interglacial transitions.
Finally, working on the last 2000 years is very important, it is a challenge for paleoclimatologists because we are looking for a small signal in a lot of noisy data. What is emerging is a detailed picture of regional changes around Antarctica and Greenland, changes in atmospheric and ocean circulation, the response to volcanic forcing, and modes of variability. All of these things will also be at play in the future in addition to being superimposed on the human influence to the system.
I feel that scientists hate not to understand things, and there are things that we don’t understand
BL: Ice cores shed light on ancient conditions going back hundreds of thousands of years, and climate scientists are looking for ever-older ice, termed the oldest ice on Earth. What is the relevance of this ice in particular to our overall understanding of the climate system?
VMD: I feel that scientists hate not to understand things, and there are things that we don’t understand. Sea level was higher by 6-9 metres about 125,000 years ago, during the last interglacial period. We still do not know how much ice melted from Greenland and Antarctica and at what rate. This is a very relevant case study for the time response and the variability of ice sheets, and not knowing what happened to the Antarctic ice sheet is extremely frustrating, so that’s a strong motivation. It was the motivation for the quest to find the oldest ice in Greenland, the NEEM ice core project. It is still a motivation for that period in West Antarctica because right now the oldest ice in West Antarctica is 80,000 years old, from the new WAIS ice core.
Moreover, the oldest ice core record from EPICA Dome C, in east Antarctica, demonstrates glacial-interglacial variations in cycles that last as long as 100,000 years, with intense transitions varying by roughly 5°C globally. From deep sea sediments we know that, a million years ago, there was a major climate reorganization with shorter ice age cycles of 40,000 years, with less intense glacial amplitudes, and we don’t know why. The motivation behind the search for older ice from Antarctica, all the way back to 1.5 million years ago, is to understand just why this major reorganization happened, and what happened to the atmospheric composition, particularly regarding greenhouse gases concentrations.
BL: Where do you see ice core research going in the next couple decades?
VMD: For ice core research I see things becoming faster, more efficient, and consuming less fuel for operation which is also part of the climate change issue. There is the development of fast drills which allow drilling in areas with difficult access, such as ice streams, fast flowing parts of glaciers, and ice sheets. This would provide important information on the deformation of the ice to better represent its flow. Regarding ice core analysis, we have seen the development of continuous analysis methods that yield data more quickly and provide highly resolved records. It is amazing to see students today that have more data over the course of their PhD than people could accumulate in sometimes 20 years in earlier periods.
There is work on providing new high resolution results, new methods to measure isotopes and trace elements that were not possible before, better dating techniques especially for glaciers, and mountain glaciers. Also there are efforts towards large syntheses of data. It is not simply about going in the field and doing measurement, it is also about having a broader view. I often describe climate science as putting together a giant jigsaw puzzle, each element is like a piece of the jigsaw puzzle, one ice core, one new piece for the puzzle but we need to put all the pieces together and there are many missing. To put the story of past climate change together we must combine different observations and place them on the same timeline, that involves synchronizing the age scale (timeline) for ice cores, lake sediment cores, speleothems, deep sea cores and so on, to develop a clear picture of the ocean and atmosphere through time, the transient climate variation from the deep sea to the highest mountains and above. The goal is to have a 4D movie of climate evolution through space and time.
I’m quite uncomfortable with the notion of a tipping point, but I think it is important to explain to people that in the climate system not everything is linear
BL: We often hear that the rate of change in the climate system is currently unprecedented, however, paleoclimatologists speak of periods of abrupt climate change thousands of years ago. How do these periods in the present and distant past compare to one another?
VMD: I think it is the media that always speaks about unprecedented changes, sometimes on very short timescales. One issue with climate change is what we call detection: do we detect something abnormal that may be attributed to a systemic perturbation, such as human greenhouse gas emissions, and of course we have been testing this for any and all parameters we can measure. Ice cores reveal that the atmospheric CO2 concentration is unprecedented over the past 800,000 years, but then it is also at a level that was encountered in the Pliocene three million years ago, so it is not unprecedented on geological timescales.
It is the same for warming. The current warming, by level and rate of change, is unprecedented for the past 1500 years. It is the same for sea level. We have recent reconstructions of sea level that contribute to that. Still, if you go beyond that in the past, due to natural drivers there are periods that are warmer than today, maybe in the Early Holocene 12 to 8,000 years ago, while sea level was higher than today during the last interglaciation 125,000 years ago. So the issue goes beyond the level, it is also the rate of change and the cause of the process. And when we look at glacial-interglacial changes, we do not often have a resolution of less than 100 years in many records, especially ocean records. So it is still difficult to compare 100-year warming at present with an average of several hundred years in the past.
In the last IPCC report, in the paleoclimate chapter, we relied on a single global temperature reconstruction from the glacial time to today, there is only one. It is just one study and we showed [that] deglaciations are the fastest climate changes that we know of globally. On that scale, we estimate the fastest warming is about 1°C per thousand years globally, but locally there are more intense changes. However, it is important to compare global to global or local to local. Abrupt events near Greenland were particularly large, 8-16°C in a couple of decades to centuries. Today’s changes in Greenland are not unprecedented, but if we go on with greenhouse gas emissions, in the high end business as usual scenarios, we would reach or go above the fastest changes of the last glacial period.
BL: Do you think there might be a tipping point in the system, particularly for Greenland, which could lead to the reorganization of the system, heading towards the last interglacial conditions where we had sea levels 6-9 metres higher than at present?
VMD: I’m quite uncomfortable with the notion of a tipping point, but I think it is important to explain to people that in the climate system not everything is linear. For example, melt of ice has a threshold of zero degrees celsius and that clearly is not linear. From what I understand now, for Arctic snow and sea ice, there is not really a tipping point. If it gets warmer you get less and less sea ice, and then if it stops warming or it gets colder it can recover quickly, so there is no real tipping point. If you look at the Greenland ice sheet there is clearly an amount of warming at which you can trigger a sustained long lasting deglaciation, that is a self-sustained process because when the ice sheet gets thinner then the surface is warmer because temperature decreases with elevation and then the Greenland deglaciation is self-sustained on the long run. The range of warming at which you can trigger this, I would not say a tipping point but a sensitivity threshold is something between 1 and 4 degrees celsius above pre-industrial values. So we are more or less at the lower end of this threshold.
This interview has been lightly edited for readability. Stay tuned for Part II, later this month.