 Gall gwnaeth bod yna dwy predictor yn gyda ei wneud ffyrdd o'r llei. Yn ymddeithasod sail yw'r cylles, ond mae holl o'n mynd yn cyflig o'r cyflig o'n mynd. A beth yna y 1,5 bilynyddiadodd ymddeithasod sail. Yn ymddeithasod sail o'n mynd yn gweithio meddorol. Ond mae holl o'n mynd yn gweithio meddorol. Ac mae hynny'n ei bwysig i'n ymddeithasod sefydlaeth o'r cyflig o'r cyflig o'r cyflig o'n mynd yn ymddeithasod sail. We don't know how much fossil fuel is going to be burnt in the next 50 years, but the augur is here rather poor. The projections from Shell and others say that 20 years from now we're putting out about 25% more per year, whereas of course the aim is to halve the emissions of CO2 by mid-century. Second, we don't know just how sensitive the temperature is to carbon dioxide. If nothing else happened, then doubling carbon dioxide would raise the temperature by about 1.2 degrees. But there are feedback effects due to water vapor and others, and it is thought, and in the best models it is found, that the amplification factor is about two or three. This is called the climate sensitivity. But that's the biggest uncertainty still in climate projections. But in 20 years from now, there's a very good chance that we'll have pinned down the climate sensitivity for two reasons. First, we'll have a longer time base of data so we will know how much the world is warming where we can average over a longer period than we can now. And secondly, we'll have much better models of the climate about how the climate should change, how cloud cover should change, et cetera. Well, if 20 years from now the sensitivity turns out to be low and the warming quite gradual, then we can relax a bit and stick with fossil fuels for a bit longer. But if it's high and the climate consequently seems on a rapidly warming trajectory, then if international efforts to cut CO2 emissions still haven't worked, there may be pressure for some kind of panic measures. And these panic measures would have to involve a plan B, being fatalistic about continuing dependence on fossil fuels, but trying to combat its effect by some form of what's called geoengineering. Geoengineering can be of two kinds. First, sucking the CO2 out of the atmosphere. This would be benign. It would be like having lots of artificial trees, essentially undoing the changes that humans have made already. But this is sadly not really feasible on the necessary scale, according to present technology. But the second kind is artificial volcanoes, as it were, putting stuff in the upper atmosphere to stop all the sunlight getting down to the ground, reflecting aerosols or something like that. Or by seeding more low clouds, which is what that would do so that more sunlight is reflected back into space. Well, in principle, it's fairly clear that some of these are feasible. It's already clear that it's affordable to throw enough material into the stratosphere to change the world's climate. Indeed, what is scary is that this kind of geoengineering might be within the resources of a single nation, even a single corporation or individual. The political problems of this geoengineering would be overwhelming. Not all nations would want to turn down the thermostat equally. And there would be unintended side effects. Climate change, of course, isn't just an overall warming. It's a shifting of weather patterns on a continental scale. Very elaborate climatic modelling would be needed in order to calculate the regional impacts of an artificial intervention. Moreover, the warming will return with the vengeance if these countermeasures were ever discontinued. And furthermore, rising CO2 has other desilirious effects. Ocean acidification, for instance, and they will continue unchecked if we merely cool down the earth by putting stuff in the upper atmosphere. This second kind of geoengineering would be a real nightmare. Except for one thing, it would be a bonanza for international lawyers. Because if nations can blame each other for the weather, there's endless scope for disputes and litigation. Well, it seems to me right at least to study geoengineering to clarify which options make sense and perhaps damp down our undue optimism about the technical quick fix of our climate. And that's been done in a small way, as I've shown by studying tether balloons and by studying clouds and trying to see what changes cloud cover. And it's crucial, I think, to sort out the complex legal and governance issues raised by this so-called solar radiation management. And to do this soon and calmly well before urgent pressures for action might build up. But this will be a fallback. So we should surely redouble efforts to decarbonise the economy by using energy more efficiently and by a massive effort to develop and deploy clean energy. In fact, speaking as a chemist professor, it's hard to think of anything more likely to inspire the brightest young people to become engineers than a global commitment on the scale of NASA's Apollo programme to provide clean energy for the developed and the developing world. And to do this quickly enough so we don't need to engage in geoengineering. But I think the question really is, should we nonetheless think about geoengineering to see if it could be done in principle or to perhaps persuade people that it's not an acceptable option to aiming to control CO2? Thank you very much.