 yng Nghymru Cymru. A gael ychydig ar y Heddlifod y Llywodraeth yn y Llywodraeth Teglionol. Yn ymdill yn Llywodraeth Llywodraeth a'r ysgolwyd yna yn ymddill y gwybodaeth o'r Llywodraeth Cymru, yn y ffordd yw'r ymdill yma. A Peter, gyd-do i'n golygu sy'n tw ymlog ardi? Felly yna'n ddigwch o'n holl i'r pwysig, o'r gyffredig a'r Siadad Pwyd-Y. Gyd-do i'r prwythiau a'r sicr? Gyd-do i'r siadad pwyd-Y o'r sesedlaethau i'r Pwyd-Y? Cymciodd drwy o'r pihaf pwyd... ac mae'r angen i'r syniadau gwneud o'r issue yw'r hyn yn ei ddweud i ddau? ac mae'r ddweud o'r cyffredin. Felly yw'r hynny'n meddwl yw'r cyffredin? Yn ymddangos? Yn ymdweud yw'r hynny'n meddwl. Mae yw'r llawwch, yma'r yma'r ymddangos a'r hyn o'r rhefyd o'r gweithio'r cyffredin yn ychydig o'r 1850, bearing to the fifth assessment report assessment of global mean surface temperature records. Is human nature to be drawn to the end of a time series? The latest greatest is what matters to people. And it's also human nature to extrapolate wildly. So if you've gone 10 o 15 oed, yn cael y gwirio, ac mae gweithio'r ddechrau, ac mae'n cael ei cael eu gwirio. Mae gennych eu cysylltu sy'n ym gweithio. Y cyfnodol, y cyfnodol, wedi ddweud o gweithio'r cyfnodol. Gweithio y ddechrau o'r cymryd o'r ddechrau, o'r cyfnodol, o'r cefnodol, o'r cyfnodol? Efallai oedd y gallwn cymryd o'r cyfnodol? Peyfodol, gallwn cyfnodol o'r cyfnodol? i'r periwyr. You'll see that there are in there are actually longer periods in the past and short periods in the past when if you'd cut the record at that point, there would have been periods of a slowdown or cessation at Haya T, a pause as long if not longer than the one that we were all concerned about. Mae yna'r amgylchedd yma ar y prominent? Felly mae'n gorfodd yma, dyma'r ymwrth fel gyfnodeis, ac yn gwneud i niw yn gweithio lle o mae'r lle o'r gweithio sgol a'r amser o'r chi'n gweithio sgol yn gwneud mae rhael i'r amser o mae'n gweithio sgol am gael mae cerddoli a gweithio, mae pobl fe fawr yn cwestiynau. We know we were increasing the greenhouse gas burdens through emissions of heat trapping gases from combustion of fossil fuels and the other sources, and yet the temperature wasn't rising. Why?anaed What's a really simple narrative to give to a non-expert. It's really intuitive To a non expert i ni ddweud i'r fath ddwynt, dyfodol yn gallu bod wlad gweld i'r gwirdog ac arddangosio'r gwirdog yna yn erbyn yn oedd y rheiddiol yw'r ffaswn. Mae'r ydynt yna sydd gyda'r tarri, mae'r anylion gwahanol i'r cyfeinol o'r rhan o'r brosadau fath. Dyna'r ymchwil yw'r rymian o'r prifsgol, yw'r niedig ar gyfer y lefel, o'r wrth i'r drafodion. Mae'r hyn yn ei wneud bod yn oed yn gweithio'r cyffredinol yn gwleidio'r cyffredinol yn gynnyddiadau. Mae'r hyn yn ymhwyloedd yn cryfiau oherwydd mae'n gwneud, ond mae'r allan iawn i'w cerdyn nhw'n ddarparu gwirio. Felly mae'r ymddindig gwneud yn gweithio'r cyffredinol sy'n 1998. Mae'r gweithio'r cyffredinol yn gweithio'r cyffredinol yn gweithio'r cyffredinol yn gweithio'r cyffredinol. profitable? Who was doing this framing? It was mainly the sceptical of the scientific consensus. There was not a lot of framing being done by the scientific community. There was a lot going on in blogs. In certain media outlets and in political think tanks. There were some early efforts to assess in the literature and I可愛 four papers ond, yn ymweld, ac mae hynny'n ei wneud am y gweithio ar y blog yng Nghymru. Yn ymweld, mae'n ei wneud am ymddiwch ar gyfer y gweithrebu, yw'r gweithrebu ffordd yn ymddiadol yn ymddiadol yn 1997, sy'n ddod o'r cyfnod hynny yn ymddiadol yn 1979, ac mae'n ddod o'r gweithrebu fel hynny'n gweithrebu. Mae'r gweithrebu yn ymddiadol yn y cyfnod i'r cyffredinol. Mae'r amser llygau yng Ngotsion Llyfrgell yn 2006, byw'r Bob Carter, cyfreunau cyflwynoedd. A phemfawr yn i-fawr, ac mae'n gweithio, mae'n mynd i'r llygau ffordd, dyma'n gwell yn ddynol ac yn ond â'r cyflwynoedd, ac i'r hwrs, ac mae'n dod i'r thain, bod o ran o'r blwysig cyflwynoedd. Their oedd yn yno'r mwyafhwynydd ac theoedd yn ychydig iawn. i'w wneud o'r cyffredinol ar ddigon? Mae'n dda i'r llyfr o gael o chyfnodd o'r cymhreif, i'r gael o'r gyflawn o'r llyfr o'r chyfnodd i'r cyflawn o'r cyflawn. A dyna sy'n clywed mewn gwirio yn provenwynt. Felly mae'r snobl yn ymlaen gyda'r gwirio mewn gwirio'r gwirio'r mewn gwirio'r mewn gwirio'r gwirio. A dyna fyddai, ychydig wrth fynd, mae'n ddod yn cyfeirio'r cyffredinol, the economist in nature, saying this is a problem for the science community for climate science. The longer it went on, the more it was in the public forum, the less science had a definitive answer to the problem, the harder it became. The more interest was garnered in the media. There were also some political think tanks and NGOs who were having a say as well. Ac mae gennym ei bod yn ddweud i gael ei ddechrau, y ddweud ar y Ddiwedd Ffandau Llywodraeth, ac y Ffartland Institution, ddweud ddim yn amlwg o hydu hyn. Ac yn ddweud ar gyfer y ddechrau, yma ymlaen y sydd gweithio yma, yma y prifes, ac mae'n cael ei ddaeth. Ydw i'n rhoi'n gwneud y dyfodol y popeth yn ei hanes yn ei bod yn ymdweud yn ei bod yn drwng i'r lluniau, mae'n mynd i chi'n gwneud. Mae'r clynydd ar y cyfrifysgol ran jaeth. Gwylech yn fawr. Felly mae'r acynnol paeth? Felly mae'r ysgol amsôl am y pethau sydd wedi ceisgwyr cyflym. Felly, Gwylech gyflym doncoedd yr unrhyw yn rhan o'r cyflwpatau mi allan yn y dyfodol a'r llef y Llyfriddol. Dylai gydag y twelffydd ni wrth hyn o gerd, neu sydd yn gwybod yn ddifan yn y gallan 34 sio, and in the climate model simulations of both the 20th and the 21st century. so there was nothing really that they believed needed explaining there were periods already readily available in climate model simulations but climate models are imperfect representations of the hugely complex earth, but they're failing having other a parallel earths, efo'r ddweud oedd, a mae'n dweud i'r ffordd sy'n meddwl mwy yn ymddangos. Mae'r pethau hyuatys, i ffathau o'r cyflawn. Yna yw'r wych ar y ddweud o'r ffafolol. Rydym ni'n meddwl â'r ysgol yw'r edrych ar y cyflawn a'r bywydau yma o'r meddwl a'r sefydlu mewn meddwl yma. that the climate models possess internal variability mechanisms capable of explaining a hiatus. And they also highlight a couple of other things. Potential for data biases... …in the surface temperature record and the effect of other force... ... short lived forcing such as solar. So the sun in the recent solar cycle has been quiet... ...and that means there's less radiation coming in at the top of the atmosphere than for a typical solar cycle. Foster and Ramstorth took another approach, so they said, well, we recognise the short-lived forcings and this variability, so we'll back it out. So they removed the effects of El Nino, they tried to remove the effects of volcanoes and solar, and they came up with an estimate after removal of these short-lived effects that showed a continuation of warming. And then Sandra Tell were looking really more at detection and attribution, but they said, you need at least 17 years for identifying human effects on global mean tropospheric temperature. Now, that was spun disingenuously as being 17 years with no trend is a problem. That's not actually what that paper is saying. What that paper is saying is we would need to observe the atmosphere for at least 17 years to be able to detect human influences, which is a fundamentally different problem. So anyone who spins Sandra Tell in 17 years and no warming as being a problem isn't fully understanding that paper, and I was a co-author on that paper, so I do know what I'm talking about. So what happened? Why did it suddenly blow up? Well, this snowball had been going down this mountain, picking up more media interest, but fundamentally it was the fifth assessment report that rose the hiatus to real public prominence. So what happened? Well, I was in the fifth assessment report as an author on chapter two, the observations chapter, so I had some say in the matter. We did cover the hiatus to some extent in the first order draft. So the IPCC has three public drafts or three public versions, the first-order draft, a second-order draft, and the final published version. There's also a zero-order draft, which the public never sees. But that first-order draft, we didn't really cover it in a lot of detail. It was covered in a few chapters in slightly different ways, but there really wasn't literature to assess. And if there's no literature, there can be no assessment. The IPCC is meant to capture the totality of the scientific literature at the time. If no one's written about it, we can't assess it. The expo reviewers raised requests to consider it in more detail. They pointed at a number of papers. We made some changes to go to the second-order draft, but we were really trying to keep it. It didn't seem to us scientifically to be an issue. Now there's a difference between scientifically an issue and we were trying to do a scientific assessment and politically and societally an issue. Now the second-order draft is when you now get the government's coming in and we got a pretty strong three-line party whip from a large number of governments that we absolutely had to cover the hiatus. So that was from the European Union, the US, and a number of national governments, including a number of European governments who participated in the review process. Several other governments, they took the contrary view that we put too much emphasis already on the hiatus in the second-order draft. So even within governments there is this dichotomy as to what should we do about the hiatus in the IPCC. So the final lead author meeting was in Hobart in Australia and we convened a very large cross-chapter meeting to discuss the hiatus. When I say large, I mean basically every chapter pretty much emptied out and took part in this. It may as well have been a plenary, even though it was meant to be one or two members from every chapter. Everyone was interested in this issue. Everyone had a say in what was causing it and how the IPCC should handle it. But there's still no literature at this point. So there's still nothing really fundamentally to assess. We did agree to address by convening a new box. So a box is a rhetorical mechanism, if you like of IPCC, for pulling together lots of distinct strands of information from across many chapters into a single location where you can do an in-depth assessment of that particular aspect. It allows for cross-chapter referencing and a degree of consistency across the report. So box 9.2 was drafted entirely, and I mean entirely after that final lead author's meeting. So there were several web-based meetings of experts at weird and wonderful hours because we had authors from across every time, say. So sometimes you'd be up at 2am, sometimes you'd be up in the middle of the day, and sometimes you'd be, goodness knows what, it was really difficult to convene. And we basically drafted text with each chapter bringing its own perspective on it. But it was a community effort of about 15 to 20 of us with expertise in observations, in forcing, in models, in projections, huge number of experts from across the spectrum of working group one. But the lack of literature meant this assessment was largely an expert judgment-based assessment. We didn't have the papers to assess fundamentally. And it's important to note that when you add something like this, this late in the process, there is no opportunity for a formal review. So this final draft doesn't get reviewed by anyone other than if you want to by friendly reviewers you put out to, and that's what we did for box 9.2. But it has not got the same level of rigor as some of the other bits that have been through first-order draft, second-order draft, through two very, very in-depth substantive reviews. Trust me, when you open an Excel spreadsheet and see a thousand review comments, peer review will never hold a fear for you again. So what did it say? It's due in roughly equal measure to a reduced trend in radiative forcing and a cooling contribution from natural internal variability, medium confidence. There is low confidence in quantifying the role of changes in radiative forcing and causing the reduced warming trend. Those are the key bits in the statement in the summary for policymakers. Why have I stressed the confidence bits? Well, in IPCC, confidence statements have a very specific meaning, and it's basically a combination of the amount of evidence and the agreement of that evidence. And if you're in the lower left quadrant here, so, in other words, if your evidence poor or if your evidence disagrees substantively, then you only have low or medium confidence. So that's where we were. We put a box out. It was the only thing that was followed up by the media when the fifth assessment report working group one was unveiled in Stockholm. The only thing of interest to the media. IPCC can't explain the warming hiatus. Climate science can't explain the warming hiatus. Well, what we actually want is peer reviewed evidence-based science and we wanted after peer review, and the problem was that we hadn't got the peer review basis. But IPCC recognising the hiatus as a thing, if you like, is effectively a clarion call to the global scientific community and you can guess what happened next. Everyone dropped what they were doing and started doing analysis on the hiatus. So there are tens of papers if not hundreds and nature climate change for a while should have probably been named nature temperature hiatus because that was pretty much all they published. And there were hundreds and hundreds of papers, tens if not hundreds of papers. And these were breathlessly reported by the media. So you had one week. It's the Atlantic. You had the next week. It's Arctic sea ice. You had the next week. It's the Pacific. You had the next week. It's stratospheric water vapor. All this going on. So to some extent we called in the climate sharp shooters. We gave them nice silver bullets. Silver bullets are nice. They're things that help us explain things. They kill nasty things like the hiatus in this case. Everybody wrote a different silver bullet and everyone silver bullet said something different. We could explain the hiatus several times over. In fact, if the globe started cooling at about 0.5 degrees centigrade per decade from the combination of these papers, we could have explained that as well if we wanted to. But that's not sensible. So we need to do another assessment. We need to weed out these papers. And in fact, just this morning, I saw an advanced talk of a talk that we're going to have at the Six Assessment Report scoping meeting in Addis Ababa next week and it's making much the same points I'm about to make. So what's the new knowledge? Firstly, on the observational basis, there are three things all of which are new since AR5 and all of which push up the global temperature record in the most recent 15, 20 years. So the new analyses of marine records which were not in relation to the hiatus actually, they were driven by the need for NOAA to revisit its marine surface temperature record had an effect on the most recent 15 years. And the effect basically relates to the fact that NOAA, which was used also by NASA, treated ships and boys as the same. Ships are not boys and in fact there is a systematic offset between ship and boy based measurements of the order of 0.12 degrees centigrade with the boys reading cold. Now, if ships and boys have been forever more in a given instrumental mix sampling a given area in a given way, it wouldn't be a problem. But since 1990, we've gone from 90% ships, 10% boys to 90% boys, 10% ships. And where we sample has also changed. So if you don't account for the systematic bias between ships and boys, you bias the record and that bias the record because boys read colder than ships that without accounting for that you diminished the rate of warming recently because you're not accounting for a known systematic effect in the instrumental record. We've also made substantial efforts through the International Surface Temperature Initiative which I chair to improve global holdings. So we've now gone from about 7,000 stations to 35 days since stations that we have available to make global long-term temperature records from. And they tend to sample areas that have warmed a little bit more than the old stations did. So there's more stations in the high optic Canada, Siberia. There's somewhat more in a few other areas that have been warming more than where we were already sampling. And then there's another one using re-analyses which are numerical weather prediction forecast systems but run historically and also interpolation techniques. So this is trying to spread information from where you have an observation over a broader region. And both of those say that we underestimated the warming primarily because we weren't capturing the Arctic and some other areas that were warming faster. There have been a whole host of papers about internal variability mechanisms. Primarily around ocean modes of variability. So you have the North Atlantic Oscillation which is the one hopefully everyone in this room knows about because it's whether there's a very strong westerly flow and it's raining all the time on Ireland or whether, as in this past winter and spring, it's a negative and it's a more easterly or southerly flow which bangs out less rain on Ireland. And by the way, we are in drag by ostensible measures in Ireland, I know. There's also longer term Atlantic ones and the Atlantic multi-decadal oscillation which would have a time scale about that of the Hiatus. There's the Pacific Decadal Oscillation and the El Nino Southern Oscillation which are Pacific modes and then there's the Indian Ocean Dipole. And there are high quality analyses all of which say and have been published in places like Nature that that's one of those is the answer. All plausibly saying one of those is the answer. Some papers have also posited a role for sea ice reductions in the Arctic which could cause a reorganisation of the circulation and winter time cooling in Eurasia which is a big fingerprint of the Hiatus. But all of these have some reliance on models and or sparse observations of the phenomena. They cannot all be correct. If we summed those all up we would be moving rapidly towards an ice age so they cannot all be correct. And more recently there have been some analyses using very large ensembles of single models. So this is an ensemble of 100 plus members and they show that it will be very difficult to apportion responsibility for an internal variability mechanism. So you can have the Atlantic or the Pacific or the Indian Ocean be in a condition that could drive a Hiatus but it is not a given that the atmospheric dynamical response will play ball with that. And so you get this issue that that you cannot, we will never know if it's internal variability we will probably never know which ocean, which mode is responsible. It's just not possible. There are also other things so there are short lived falls things, volcanoes in particular and a problem that we have is that we don't include in the historical climate model runs the small to medium sized volcanoes which isn't a problem if they go off infrequently but over the last 15 years there have been a large number of volcanoes that are not your pinotibos, your crackatoas but are still climatically potentially important. It's an order of 0.1 watts per metre squared which doesn't sound much but that's about comparable to a couple of years, three years increase in the greenhouse gas burden from carbon dioxide. And the high latitude volcanoes which many of these have been you all will remember the Iceland volcano that shut down European airspace for four or five days. These have tended to be very much underestimated in their impact because they put it to because the climate model forcing ancillaries don't catch the low trochipause height in the high latitudes. So they don't catch the burden of volcanic aerosols there. The net effect of the volcanoes could at most explain about half the observed slowdown or hiatus. Then there's a whole bunch of other forcing that have been posited in single papers. So one thing is that we have been incredibly successful in the Montreal Protocol to remove ozone-depleting substances which were causing the ozone hole in the stratosphere comparing contrast perhaps with our efforts at mitigation on climate change. It's probably because it's somewhat easier to change your refrigerant and change your hair your hairspray than it is to change your way of life fundamentally. But CFCs are not just ozone-depleting substances they are also incredibly potent greenhouse gases. So the fact that we've been able to face them out more aggressively than was hoped means that there is a reduced greenhouse gas burden compared to what we thought there would be. There's also been this quiet solar cycle. On the flip side some greenhouse gases may have increased more than thought and in particular the atmospheric radiative efficiency of methane has been increased a little bit, bumped up a little bit and that wasn't reflected in the climate models. There's been quite a lot of changes in stratospheric water vapour which isn't a forcing per se it's a feedback but it is very radiatively efficient and the climate models don't get that stratospheric water vapour. And then the real wild card in climate science is aerosols and their effects. So we don't observe aerosols well we don't understand aerosol lifetimes well and we don't understand in particular the indirect effects of aerosols on both cloud condensation nuclear density and cloud lifetime which are huge uncertainties. So the net effect of all of those is uncertain but they certainly in part cancel each other. So there's some fortuitous cancellation and errors between what forcings actually existed in the real world and what forcings were in the climate models. But it's probably a non-zero cancellation there is probably an error in the climate forcing that was given to the models and if there's an error in the climate forcing that was given to the models it would be alarming if there wasn't then an error in what the climate models showed. That would be far more alarming than if the climate models agreed with the observations because then we'd have to work out why they agreed with the observations despite the forcings being wrong. So if you want to understand the hiatus you really need to have very good inventory of forcings and we fundamentally do not have that. Particularly for some of the shorter-lived harder forcings to understand such as aerosols. So if we were to create a new assessment today what would in my opinion we say and I think this is going to be very consistent with what I'm going to hear in Addis Ababa next week. I don't think it would be a million miles away from what was actually said in AR5 but now instead of saying low confidence, medium confidence we'd probably be saying high confidence or very high confidence and we may even be using likelihood statements such as very likely or virtually certain. I would say it would give slightly more weight to the role of internal variability. Internal variability probably explains most of the hiatus. There would be a greater recognition of the role of the oceans and potentially sea ice. And there would certainly be a far more acknowledgement of the role of both spatial and instrumental biases in global mean surface temperature in the most recent 20 years. And there would be a more nuanced discussion I think of the multiple potential pathways and mechanisms of decadal variations in climate and a greater recognition that 21st century climate will not be linear even in the global mean and sure as hell won't be locally at the level that matters for planning. We need a plan for variability. Climate change does not negate climate variability. It acts in concert with it. Sometimes they act together in which case you may get very rapid change globally or locally and sometimes they will act against each other in which case you may not get much change. Well, you may even get periods of cooling still in the 21st century even under an increase in greenhouse gas burden. Now since AR5, so we stopped in AR5 we stopped just about there in that last trough. Since AR5 the last three years according to all of the estimates have each been record warm. And 2017 as it stands has about 50-50 chance of exceeding even 2016. So if it was variability variability can act against it for a while. It's like walking a dog up a hill and the dog sees the sheep down the hill and it strains on the leash for a while and then it sees the sheep up the hill and it strains on the leash the other way. So the flip side of hiatuses or hiati is surges. And it may be that by the time we're actually drafting AR6 finalising the drafting of AR6 which is just three years time what the government will be asking is not a damn thing about this hiatus which is yesterday's news. They'll be worried about is this surge going to continue? What does it mean? And rather than having the skeptical side pulling us we'll have the alarmist side pulling us. And we're the scientists in the middle trying to make sense of this. My view is that this is probably as much variability as the other. Indeed, if you put a linear trend plus some autoregressive moving average process on top it would look very like this. This is the superposition of two things of some variability with some kind of long-term memory component and a much more linear. So the dog walker walking up the hill is the forcing and the dog is the observations. It's going all over the place because it's getting pulled by the natural variability made in the climate system. So what's the take-homes? Did a hiatus ever occur? Ever even occur? Well, statistically no. There is no robust evidence that there was a change in behaviour. There is no robust statistic that I could use that would show a change point in the observational time series behaviour then particularly after taking into account the new observational insights in the most recent three years. But practically yes. Your eyes don't lie that badly. Okay, fundamentally there was a period in the early 21st century in which global mean surface temperature changed less than it had done before and less as it turns out than it has done since. Did it matter or cast data on the central tenets of the science? Publicly yes, absolutely. Scientifically no, which is why we were so against such a high profile inclusion in AR5 because we knew it didn't fundamentally call into question aspects of the science. Why did it occur? Mainly natural variability with some contribution from both long-lived and primarily short-lived forcings. So volcano, solar, some from errors in what we've given to the models in terms of the long-lived forcings. Did any good come of it? Well, I would hope we don't forget anytime soon that decadal scale change and variability is important and that climate change is not some linear process even at the global mean scale that there will be periods when we have surges. There will be periods when we have hiodeses. And those matter on a policy basis and they matter on a mitigation and particularly an adaptation basis. They matter hugely because you need to adapt for that variability. The climate in the future is not some deterministic value we can know absolutely. We need to understand and plan for the range of possible futures which depend upon climate variability as well as mitigation choices we make.