Source: American Geophysical Union
URL: http://www.youtube.com/watch?v=RG9SK6NYYss
Date: 10/12/2013
Event: IPCC: The Future of the Assessment
Attribution: American Geophysical Union
People:
Peter Weiss: Good morning. Welcome to our first press conference of today. This one is "IPCC: The Future of the Assessment", and our speakers, in the order that I read them, are Thomas Stocker, Co-Chair of the IPCC Working Group I, Climate and Environmental Physics, at the University of Bern, Switzerland, Dennis Hartmann, Working Group I Chapter 2, Observations: Atmosphere and Surface, he's with the Department of Atmospheric Sciences, at the University of Washington in Seattle, Philippe Ciais, he's Working Group I Chapter 6, Carbon and Other Biogeochemical Cycles, he's with the Laboratoire des Sciences du Climat et de l’Environnement, at the Institut Pierre Simon Laplace, Gif sur Yvette, France, and Olivier Boucher, Working Group I Chapter 7, Clouds & Aerosols, Laboratoire de Meteorologie Dynamique, at the Institut Pierre Simon Laplace in Paris. Thank you.
Thomas Stocker: Thank you very much. Good morning, everybody. I would like to welcome you here to this press conference, on behalf of my co-covenors of the Union session and the town hall meeting here at AGU Fall Meeting 2013. The two colleagues are Qin Dahe, Co-Chair of Working Group I and Kasper Plattner, Director of Science of the Technical Support Unit of Working Group I.
What we want to do is to inform you a little bit about three topics. They will be introduced by my colleagues here on the podium. But, before we go into the science, let me briefly put everybody on the same page. A few numbers as background information - this is the result, this Assessment, of hard work by 259 scientists for the past three and a half years. Over 9,200 peer-reviewed publications were assessed in the process.
We have analysed more than two million gigabytes of numerical data, from which, for example, model projections but also model evaluation and detection attribution studies were performed and assessed during these three and a half years. We also produced more than 1,200 diagrams for this report, including an atlas of global and regional climate projections. Overall, as you know, this is a very complex process, during which several drafts are produced, which go out for worldwide expert review, during which we collected over 54,000 review comments which were addressed by the authors and responded to.
The three major messages by now should be well-known to everybody - I just want to recall them. Warming of the climate system is unequivocal, human influence on the climate system is clear and limiting climate change will require substantial and sustained reductions of greenhouse gas emissions. These are very simple, quotable sentences but their strength lies in the fact that these sentences now are not only agreed by the scientists who have worked on this report but also adopted by the governments of the world participating in the IPCC.
Just reviewing - revisiting the purpose of these Assessments. Our primary customer of these Assessments is the UNFCCC process, policymakers and stakeholders. We provide a snapshot of the current scientific knowledge, but we also serve as a base reference for the scientific community. These reports are eagerly awaited by the community. They are, as you know, very specialised, in 14 chapters, basically comprehensively covering all the content of climate science relevant to the question of anthropogenic climate change.
What is also evident from our Assessment is gaps in knowledge and scientific understanding. For example, the fact that we have, in our current Assessment, a chapter on clouds and aerosols, or a dedicated chapter on sea-level rise, indicated that in AR4 2007 there were these two gaps, at least, that needed to be covered in a comprehensive manner in our process. As that was the case in the past, this also applies to the present report, and with this press conference, we would like to give you a glimpse of open gaps and gaps in the scientific understanding in three important areas.
One is observations, the other one is clouds and the third one is carbon and other biogeochemical cycles, and that's why we have selected three colleagues, coordinating lead authors from Chapters 2, 6 and 7 to share with us their insights out of AR5, plus also their view on uncertainties and gaps in the scientific understanding. With that, I would like to hand over to Dennis Hartmann, Coordinating Lead Author of Chapter 2. We will go through the three presentations, so please keep your questions to the end, and then we will have ample time for questions and answers. So, Dennis, please.
Dennis Hartmann: Thank you, Thomas. I wanted to start by just reviewing what we have done, in terms of observations and then attributing - here we go - so Thomas already mentioned that warming is unequivocal. We know this from multiple independent observations of the climate system. Now we're also able to attribute a substantial fraction of that warming to human influences - this is well-established from AR5, from our process and from previous studies, but what I want to do is go on now and just talk briefly about three areas of interest for the future, where we think the science will be going, in the near future.
And the first one is one that is very often asked about, and that is the apparent slowing down of the warming rate in the recent decade, or decade and a half - many people have called it the "hiatus" in warming. So, there's been quite a bit of scientific research done since the closing date for publications for AR5, which have indicated that this may be largely a response to natural variability, particularly in the Pacific Ocean. But there's much research that needs to be done, on that. In AR5 we concluded that part of it was due to changes in the forcing, due to solar constants, slight variations, and also some aerosol forcing.
So, in terms of the natural variability, is that Pacific variability that's classically related to the El Nino/La Nina oscillation - is that the major explanation? How long can we expect that to persist? So this will be a very active area of research for the future, to try to understand the hiatus period, and also it relates to another emphasis of research, which is to try to make better predictions, from physical principles, of short-term decadal climate variability, like the next decade, the next decade or two. To what extent can we predict that, on the basis of the knowledge of the current climate system, particularly the ocean?
Another area that I'd like to emphasise is the sea-ice trends in the Arctic and the Antarctic. Here we have the interesting situation that the Arctic sea ice has declined much more rapidly than most of the models predict, and yet in the Antarctic the sea ice has been increasing for quite a long time, and particularly in the last decade or so - there's a very clear signal. So the short-term variations are most likely related to natural internal fluctuations of the climate system. We'd like to understand these better, perhaps be able to predict them and relate them to the longer-term changes that we know are being driven by human activities.
And the final thing I'd like to emphasise is extremes - precipitation extremes in observations and models, not completely consistent and understood. A lot of the climate models don't have the spatial resolution to do a good job of predicting extremes of precipitation. And the observations are also not quite adequate to tell us where extremes in precipitation are going. We know that the humidity in the troposphere is increasing, associated with global warming, and so the latent energy in the surface areas increasing - we expect that to have some impact on extreme weather events, particularly interesting would be tropical storms, but we don't have the observational and the scientific basis to be - to be able to say with certainty that we've observed a change in tropical storms that's related to humans, or to make a confident prediction about the future of those very highly impactful events. So, thank you very much.
Thomas Stocker: Thank you very much, Dennis. I would like to hand over to Philippe Ciais, and I'll just change the presentation.
Philippe Ciais: Thank you, Thomas. So, it was quite challenging to find three slides to summarise the work of many scientists on the carbon cycle during the preparation of this report. And also I must say that our chapter got a record number of comments, which I think denotes a very large interest from a broad range of scientific communities for the carbon cycle, going from terrestrial ecology to ocean acidification and forward coupled system modelling.
Maybe the most important message is that CO2 is a gas that sticks to the atmosphere, once it is emitted, meaning that after its emissions, CO2 is going to be removed by natural things - ocean uptake and terrestrial sink - but a significant fraction of emitted carbon dioxide is going to stay in the atmosphere, and therefore to change the climate, for a very long time after it has been emitted. Typically, 15-40% of carbon dioxide will stay in the atmosphere for about one millennium after its initial emission. Which means that the problem of carbon dioxide emissions of today is really a problem that will remain for a long time in the future. That's the essence of the importance of the carbon cycle for climate change.
Human activities, as you know, they are really destabilised the natural equilibrium of the global biogeochemical cycle of carbon, of methane, of nitrous oxide... Er, today we emit, in 2011, 9.5 billion tonne of carbon dioxide - so we are tracking the highest emission scenarios. To give you an idea of this, about 10 billion tonne of carbon, I think, in the 3rd Assessment Report of the IPCC - this number was somehow forecasted to be reached by something like 2030, and today, in 2011, we are already here. So we are really tracking high on the emissions of carbon dioxide - CO2, I think, increased by 40% above its pre-industrial level from - this is seen from ice core and atmospheric monitoring data.
Methane has increased by even a larger amount - 150%. It has been stable in the atmosphere from 1995 to 2005, and now we see a - a resuming of the methane increase, with a relatively high growth rate since 2006. And nitrous oxide has increased by 20%, mainly in response of the human use of fabrication of nitrogen fertilisers.
So, how to try to get a - a better estimates and forecast of how, er, concentration of greenhouse gases will evolve in the future, and they will change the climate. The approach that has been followed in this Assessment Report has been to force complex coupled carbon cycle climate models - also called Earth system models, also called CMIP5 models - by a range of prescribed concentration trajectories, So when you prescribe the concentration, you use your Earth system model to calculate the ocean sink, the land sink, how climate is impacting those two quantities - and the result is called "compatible emissions". Essentially, it's the emissions you are allowed to put into the atmosphere in the future, if you want to respect, to follow a given CO2 concentration trajectory.
So behind these curves, which show the compatible fossil-fuel emissions, different colours for different scenarios, lies the work of a very large community of Earth system modelling, who crank up [?] and develop very complex models that try to account for a number of interactions between climate and carbon-cycle processes.
The main message is that if we want to keep climate below 2-degree warming, we can emit about a thousand billion tonne of carbon dioxide. We have already emitted about half of this amount, so the amount which is ahead of us is about 500. If we want to follow the RCP 2.6 scenario - which is most climate-friendly scenario - it's a fact that by 2050 we will have to reduce our carbon dioxide emissions by roughly a factor of two, compared to the present-day conditions. So we have 37 years to cut the emissions, which are on an upgoing trajectory, by almost a factor of two. And even by the end of the century, as you can see in the figure, emissions will have to approach zero, and, in some of the model results, to be negative - negative emission means that we need to remove carbon dioxide from the atmosphere.
Thomas Stocker: Thank you very much, Philippe, and we come to the last presentation, by Olivier Boucher.
Olivier Boucher: Well, thank you for coming, and thank you, Thomas, to take care of the slides. I'm going to discuss, very briefly, about the roles of clouds and aerosols in the changing climate, the main findings of the IPCC Assessment, but also the research gaps and future research directions which we don't discuss explicitly in the report.
So, cloud and aerosols are really critical to understanding the Earth's changing energy budget. Aerosols are responsible for radiative forcing, the emissions and concentrations have changed. They interact with radiation, they interact with clouds and they have contributed to cool the climate system, so far. Clouds, on the other hand, are responsible for feedbacks, so they respond to climate change and they can amplify and they can dampen climate change.
So, first of all, we observe clouds and aerosols much better than before, much better than at the time of the 4th Assessment Report, and this is thanks to a number of satellite instruments and new instruments which give us a global coverage but also the vertical profile of aerosols, clouds and precipitation. We also have analysis of previous datasets, we also have improved models from the finer scale to the global scale.
Yet there is something we don't have a clear vision on, which is trends - aerosol trends, cloud trends and cloud properties are still poorly characterised, and this is because of a number of issues with the sampling, and issues also on satellite instrument calibration. So expect more focus on this in the future, and this is also - this is very important, there are real scientific issues there, but sometimes they're being overlooked, somehow.
So one thing we've done in the IPCC report was to assess the aerosol radiative forcing, and we've done that with all the evidence we had, from in situ observations, satellite observations, the fine-scale model, the global-scale model. And you can see on this chart here that we've revised the aerosol radiative forcing upwards, so it is less negative than it was at the time of AR4, and that's the - you can see that with the blue bar, in comparison to the green bar, on the upper left corner of the plot.
And this is mostly for two reasons. One is: we've re-evaluated aerosol absorption, so aerosol absorption from black carbon is more important than we thought. And also we've revised downwards aerosol-cloud interactions. Yet aerosols continue to represent the largest uncertainty for the anthropogenic radiative forcing, so if you look at the black curve, which is the probability distribution function for the radiative forcing, a large fraction of the uncertainty is coming from aerosols, rather than from greenhouse gases.
The two reasons why we changed the aerosol forcing are actually the two major issues for the future. Aerosol absorption continues to be not very well observed, not very well modelled, and that's clearly a focus for future research, and aerosol-cloud interactions are also a stumbling block in climate models, even though we've made a lot of progress with the small-scale models, so now the challenge is really to use all the new tools that we have, to extrapolate that to the global scale. Um, yeah, there is one more, I think, research direction for the future, which is to go more into the role of aerosols and the response of the climate change to aerosols, especially in terms of circulation changes.
Turning now to cloud feedbacks, we've analysed what we call robust mechanisms contributing to cloud feedbacks, and we've identified three robust mechanisms contributing positive cloud feedbacks. One is the rising of the top of high-level clouds, and this is a mechanism which was put forward by Dennis Hartmann a few years ago. By "robust mechanism" we mean mechanisms which not only we see in the global models but which we also have some kind of independent support from observations or from physical understanding.
So the rising of high-level cloud tops is the first one. The broadening of the Hadley Cell, whereby you have a larger region with little clouds, with less clouds, is another positive, robust feedback, and the third one is the poleward shifts in the storm track, whereby you are pushing the clouds - low-level clouds - towards regions with less solar illumination, so they have less of a negative cloud radiative effect.
So we have these three positive cloud feedbacks, and we couldn't identify any robust negat- any robust mechanism contributing a negative feedback. It doesn't mean that there aren't any, and clearly low-level clouds are the wild card here, so I expect more research in this direction, in the future, to really constrain the feedbacks involving the low-level clouds.
The new frontier, in terms of cloud, is also to look at all the relationships between clouds, circulation changes on climate sensitivity, and there is a new run challenge from the Climate Research Programme on this, and there are new tools as well, with a hierarchy of models and new observations, so I really expect more research on this.
Thomas Stocker: Thank you very much. Thank you to all the speakers for their presentations and keeping in time, so I hand over to the moderator.
Peter Weiss: Thank you. Okay, we have some questions?
David Appell: Hi, my name's David, David Appell - I'm a freelancer. Will there be a 6th AR? Some people have talked about, maybe ARs have outlived their usefulness, in terms of coming out every seven years and things happening in between, and...
Thomas Stocker: I think it's too early to respond to that question, and that question would have to be asked to the Intergovernmental Panel on Climate Change. That is within decision of the panel - the panel has certainly initiated a process to look into the future of the IPCC, in terms of what would be needed for the next few years from the stakeholders and from the policymakers. Certainly we are looking at a continual development, but it's too early to say what the possibilities are. It's clear that these comprehensive reports are extremely useful, but at the same time it's also becoming evident that they are an increasing burden to the scientific community. Just to remind you, at the few numbers that I gave in my introduction that should illustrate really that the burden of the scientific community - to the scientific community - in carrying out these Assessments, has become extremely large.
Michael White: I'm Mike White from Nature magazine, and this is a question for Olivier, I think. So I know that for global cloud-resolving models, you can still only run them for maybe a month or so, and I'm wondering: between that extreme and what we have now, do you see a potential for convergence on things like tropical cloud convection, in terms of how they're parameterised? Or is that going to remain an area of multiple approaches and huge uncertainties for the next few years?
Olivier Boucher: So, you're right - we've got more computing power, now, and some of the cloud-resolving model and large-eddy simulation model, which before were only running for a few hours or maximum a few days, of a small domain - now can be run on a larger domain and maybe up to a month. So it's, sort of, starting to bridge the gap, with the global models, but there's still some way to go. And I think, as you've said, there will be a multiple number of approaches. I mean, I mentioned new parameterisations that really make sense of the subgrid-scale information in climate models, while you can look at - yeah, variations in the dynamics of the subgrid scale. They are combination of climate models with large-eddy simulation in 2D - we call this super-parameterisation - so this is the range of tools now being available, so we can expect progress to bridge the gap between the fine scale and the global scale.
Woman 1: Yes, we have a question from the web-streaming chat. This is a question from Seth Borenstein from the Associated Press - this question's for Dr. Hartmann. Between the increasing Antarctic sea ice and still cold - and according to Ted Scambos, yesterday, record cold - temperatures, how do you reconcile what is happening in the Antarctic with models and climate-change science? Why is melting sea ice in the Arctic due to climate change but growing sea ice in the Antarctic due to local conditions?
Dennis Hartmann: That's a really good question. Er... I think that modelling analysis has indicated that the sea ice in the Antarctic is quite sensitive. There's feedbacks between the ocean, the atmosphere and the clouds, and its variability is bigger than we expected in the past. The models don't produce that, because we think it's naturally caused, by natural internal variability. The models produce, over a longer span of time, actual reduction in sea ice around Antarctica, particularly in the summer season.
In the Arctic, we do have a strong relationship between what we observe and what the models produce. So the models will produce a reduction in sea ice in the Arctic, which is consistent with the physical understanding that we have, so we have assessed that it's - I forget - at least likely, I don't remember the exact probability level, that that decline in the Arctic is related to human activities.
That being said, the actual decline that we see is greater than we've predicted, so far. So we don't know to what extent that greater melting of Arctic sea ice is a result of natural variability that is not forced by humans, or whether there's something about the sensitivity of the Arctic ice that we don't understand.
Mark Fischetti: Another - Mark Fischetti from Scientific American - another question about clouds. You mentioned that feedbacks for low cloud cover - I'm wondering if you could describe an example or two of what the feedbacks may be?
Olivier Boucher: Yeah, well, they're mostly feedbacks relating to low-level cloud properties - I mean, are the clouds getting thicker, or because they are shaped from ice to liquid clouds, what is the implication in terms of radiation, so that's one. And then there is, of course, feedbacks on the amount of clouds in different regions, and then some regions may be more special, like the Arctic region - if you change the sea ice, then you may also change low-level clouds.
Randy Showstack: Yes, Randy Showstack, reporter with Eos Newspaper, the American Geophysical Union. I have a multi-part question. Philippe, you noted that we have about 37 years to get emissions down to meet favourable scenarios. A question to you and to the other speakers: do you have confidence that we can reach that difficult goal? Do you think that your message is getting across sufficiently to the public and decision-makers? And, partly as a follow-up to an earlier question, is the IPCC nimble enough to do the job of a) getting the information out quickly enough, and b) dealing with and confronting sceptics on these issues, such as those who were associated with the NIPCC report that came out a few days before the IPCC recent report?
Thomas Stocker: Yes, I think what we can say that in AR5 we have worked hard to translate the rather technical and complicated language that is usually used in the Summary for Policymakers, including all the numbers and ranges, into summary statements that are short, quotable and understandable. These are the so-called "highlighted statements" in the Summary for Policymakers which, when you filter them out of the Summary for Policymakers, they give you the complete storyline of the Working Group I Assessment.
We found that this way of additional communication - not only of technical numbers and scientific ranges, et cetera, but also summary statements - is very effective with the public. Many media have taken over verbatim such statements, and we are also noting with some satisfaction that in the UNFCCC process, some of these statements were actually finding their way into the adopted documents of the most recent Conference of the Parties.
So I think the scientists have actually carried out their task, the task to inform, to provide the scientific assessment to the policymakers, and it seems to me, from the short analysis that we have carried out, that these messages have arrived at the customer, so to say.
Now, regarding the confidence of what will happen, that's really in the hands of the policymakers, and I should say that the essence of the information and the essence of the urgency of the problem, and also the fact that options will not increase with waiting but they will decrease significantly with continuing emissions - that information has been around for many years, so I think we, what we now have is clear numbers based on the latest scenarios.
Regarding your question of an NIPCC report, et cetera, if you actually go into that report, nobody can seriously claim that this is a scientific report. So, I think I should not comment further on that.
Stephanie Ogburn: Stephanie Ogburn, ClimateWire. This is for Dr. Hartmann. I was just curious if you could see, in the next seven years or so, more observational networks coming online that would support additional observations in the Poles and the oceans, to look at the melting of the sea ice or the ice sheets in the Antarctic, and also to support more science around the hiatus and what's happening in the oceans?
Dennis Hartmann: Well, we do - related to the oceans, we do now have floats that measure the heat content at depth in the oceans - it's a relatively short time series. SO I'm hoping that a combination of using those observations of where the heat content is changing in the ocean and an understanding of how the circulation works from climate models, that we'll be able to sort out where the heat is going and why the surface warming has slowed down. There's quite a bit of evidence about how that's working, already.
In terms of observations, I think it's pretty challenging, from multiple perspectives. A lot of the information we get from remote areas and high latitudes, it's from satellites and remote sensing. And of course it's very challenging to keep those up there, those data sets continuing to be taken, and also to introduce new, improved instruments that will give us greater information.
There's some disparity among the estimates of the amount of global warming, because of the way the Poles are treated. Some data sets do more extrapolation to the Arctic, they give more warming than models do - than observational analyses that do a smaller degree of extrapolation, so better observations of the polar regions and of the oceans would be extremely helpful. It would be great if we could get to sustain those and improve the observations that we do have. Thank you.
Thomas Stocker: I may add, perhaps, one item that is also important, also with respect to impact studies of Working Group II, is better observations of precipitation - I think you mentioned that also in your presentation - and I think we need to emphasise that, because that impacts a primary resource of ecosystems and human systems.
Stephanie Ogburn: So, just to clarify, are you referring to the NASA GISS surface temperature data, certain extrapolations that they do, versus the EUMETSAT surface temperature data set, when you says there's different data sets?
Dennis Hartmann: Well, there are three primary data sets of global surface temperature. There's one from the Goddard Institute of Space Studies, where they do more interpolation into polar regions, and they get a slightly different, slightly greater warming trend, because there has been, we believe, more warming in the Arctic, versus the HADCRUT4 data set, which does basically no interpolation, and only constructs the global mean from where they actually have observations - that's what I was referring to, there.
Peter Weiss: Are there any other questions, from anyone who hasn't asked one yet?
Man 1: I'm wondering if the utility of climate models is peaking, in terms of their usefulness to policymakers? I mean, the error bars on equilibrium climate sensitivity haven't dropped much, over the decades, and it seems that the more factors are put into a model, each brings a certain uncertainty with it, and the uncertainties compound in a way that the error bar overall doesn't get much smaller.
Thomas Stocker: Well I would disagree with that assessment. Clearly, what we have heard, from the presentation by Olivier Boucher, that
there is still a large - a wide range of possible improvements for the global models. But the models become increasingly important for impact studies, so, going down to the smaller scales, having dynamical models available, being able to simulate the water cycle at a local and regional level is extremely important for impact studies. Also, when it comes to extreme events, as something that has been looked at in depth in a special report published by IPCC in November 2011, it's clear and obvious that extreme events are very important when it comes to adaptation but also vulnerability of societies around the world. There the models can do a good service but there is still a far way to - in terms of improvement and evaluation of these models, when it comes to extreme events.
Man 1: Can I ask one more? Um... for the first time, the 5AR talked about geoengineering - there was a paragraph in the Summary, and I don't know what's going to be in the main volumes, but I was wondering if geoengineering is going to be something that will be increasingly looked at, from the scientific point of view, in this AR and in future ones?
Thomas Stocker: Olivier?
Olivier Boucher: Yeah, that's correct, we have assessed geoegineering - both the carbon dioxide removal techniques, in Philippe's chapter, and the solar radiation management techniques, in my chapter. It is clearly an emerging field - I mean, the literature is still thin but there is a lot of studies going on now, and there is a little bit of funding - not that much, actually - a little bit of funding going into that research. Then - I think I want to make one comment here, is not - it's a field in itself, but at the same time there's a lot of connection with climate science. I mean, we've learned a lot on the climate system just by looking at some of the geoengineering technique, on this sort of - goes two way, in terms of how it interacts with climate science.
Thomas Stocker: Philippe, do you want to... add?
Philippe Ciais: Yeah, the carbon cycle - free-moving carbon dioxide has been studied in our chapter, mostly in terms of what are the profound implication, for carbon-cycle mechanisms, of voluntary and large-scale removal of CO2. We didn't look at the technological feasibility of carrying out sustained and large-scale removal of CO2, but, for sure, if you want CDR to be significant as a so-called geoengineering practice, the scale at which it must be deployed will have very deep and very profound implication, compared to the current functioning of the carbon cycle.
Thomas Stocker: I think it's also important to realise that the paragraph on geoengineering that was put into the Summary for Policymakers by the scientists, in conclusion of that scientific assessment that was carried out in the two chapters, Clouds & Aerosols but also Carbon and Biogeochemical Cycles, also contains some important indications of risks associated with these methodologies.
Woman 1: Are there any more questions from the room?
Gabriel Popkin: Gabriel Popkin with Science News. I think this question is mostly for Olivier, but I was wondering if - um, kind of - computational resources and [inaudible], the expected growth of those resources is a limitation in climate models, and cloud aerosol models in particular?
Olivier Boucher: Well, it is a limitation, like it is for climate science in general, but in terms of clouds and in terms of aerosol-cloud interactions, I mean, we shouldn't underestimate also the limitations in our physical understanding. So if, by chance, by a miracle, we had a tenfold increase or a hundredfold increase in computer time, or computing power, I mean, it would still take time to make some progress. So I would say it's not the only limitation to our progress.
Thomas Stocker: We've already also realised in the 5th Assessment Report that there's a third limitation, actually more serious - and that is the human factor. The sheer amount of data that has to be analysed, extracted from data repositories - it's just a big task and big burden to the scientific community. And that has come really to the limits. There is also a question of timing between the CMIP programmes and the Assessment - this time around, it was about as short and timely as possible, but it's clear that there is a lot of numerical data out there that still awaits analysis, and there's a big opportunity for data-mining of these 2 million gigabytes of numerical data.
Michael White: Mike White from Nature magazine, again. And this is a question for Philippe. I'm wondering about - if you think back about the work on climate sensitivity, the numbers haven't changed much, over the decades. If you think about CO2 fertilisation as another important factor, I'm wondering, if you think back about the early days of greenhouse studies, going into eddy co-variance, and the face experiments now with coupled ESMs, where you have nutrient constraints, how have the numbers on CO2 fertilisation changed and where are we now, with that concept?
Philippe Ciais: That's a very good question. This concept is embedded into all the carbon cycle models, and it is the main process by which land carbon storage is increased. And we still have an increasing effect of CO2 on land carbon storage in the AR5 models, but the fundamental discovery is that the two models that include nitrogen limitation have a much smaller effect of CO2 fertilisation. So it means that if we start, in the next Assessment, to account for nitrogen, and perhaps for phosphorus limitation, I believe that the effect of CO2 fertilisation is going to be reduced at [inaudible], in particular for nutrient-limited ecosystems - nitrogen-limited ecosystems in the north and phosphorus-limited forests in the tropical regions. And the problem is that if you look at the measurement side of it, at the moment we have only two experiments where forests have been grown under elevated CO2 atmospheres for a significant amount of years. So we have two data points and - you're right, this is a key process by which our system model increase carbon storage in the future. So it give you an idea of where we are with this.
Woman 1: Any more questions for our panel? Okay, thank you very much to our panel. Our next press conference, on improved marine systems, will start at 9 o'clock.