Andy,
Just adding my $0.02 regarding tipping points: I think that some thought regarding what we mean by tipping point is useful. Typically people think about these as a step change in the slope of some response variable to an input. This does not necessarily imply any hysteretic or irreversible behaviour. Some of the things that people worry about for tipping points (e.g. Arctic sea ice, though don't recall the reference) don't show any evidence of hysteresis in climate models if you subsequently reduced radiative forcing beyond the tipping point (depending on how long you hold the radiative forcing high). Extinction of polar bears would be a very clear irreversibility, collapse of an ice shelf (like Larson-B) is also effectively irreversible. That is, I'm using "irreversibility" here in the sense that if you increase radiative forcing beyond some point and subsequently reduced radiative forcing, do you get back roughly where you started from. This is an important distinction in terms of risk; if there really was irreversible behaviour, then there is an enormous risk to crossing the tipping point in the first place. Time-scales also matter. If we held the radiative forcing elevated long enough to melt all of Greenland (which might be a long time, but is certainly not decades), and then reduce it to todays levels, the ice sheet might not return. If we increase radiative forcing for 50 years, and slowly reduced it back to todays levels, maybe we would get back to something close to where we are today, even though we were forcing the system at a level that if sustained sufficiently long would cause major irreversibility. I'm not at all trying to be technically precise here, just pointing out that the perception of risk from crossing a hypothetical tipping point depends on what the consequences are of crossing it for 5 years or 10 years or 50 years. Because that matters for how "rapidly" one might consider ramping up SRM. Ultimately one would have to trade the risk associated with tipping point with the risk associated with ramping up SRM (which includes unknowns, at least). If you're worried about Greenland ice sheet loss, I don't think there's any argument that this ramp would need to happen on some very short time scale (you didn't define "rapid" below, neither will I). From what I've read, I would say ditto for permafrost, though I understand some people will disagree with me (and I'd be happy to be pointed to scientific references arguing how rapidly an intervention would be required). Arctic sea ice itself I think is reversible, though the affected ecosystems might not be. (Nor the change in shipping and mining interests.) And, of course, whether or not we've already crossed tipping points or we're simply talking about avoiding crossing them in the future will also impact perceptions of risk, and whether SRM would need to reduce radiative forcing below today's levels or simply prevent radiative forcing from increasing beyond some value (assuming that the goal was to avoid tipping points). Sorry, got long-winded. All to say that the subject of tipping points and SRM and rates of change thereof is not as trivial as a sound bite. doug From: [email protected] [mailto:[email protected]] On Behalf Of Keith, David Sent: Wednesday, August 07, 2013 5:32 AM To: Andrew Lockley; geoengineering Subject: [geo] RE: Response to D Keith lecture at Harvard Andy Sorry but I do not find this an effective way to debate these topics, particularly in this case when I am responding to a summary that does not accurately reflect what I said. I would rather put the energy into writing a paper or discussing these topics in smaller groups. I would be very happy to chat more over the next couple days. I do think there is some utility in this blog but I do not now look at it routinely. It may be that I am just old-fashioned, but I find that the 24-hour blogging culture tends to overemphasize emphasize spurious disagreement fueled by jumbled factoids in a way that makes it hard to identify substantive disagreements about facts and values. This is particularly true in a single-threaded format like this. But this may simply be my old-fashioned taste: many folks do contribute smart, relevant and thoughtful posts and are clearly having a lot of fun doing it. Cheers, David From: Andrew Lockley [mailto:[email protected]] Sent: Tuesday, August 6, 2013 10:38 AM To: Keith, David; geoengineering Subject: Response to D Keith lecture at Harvard David Below are a few questions and comments from your interesting lecture today. I'm copying to the Google group to invite wider comment on the important issues your raised. It would be great to hear any comments you have, if you are able to respond. Thanks A Ramp up You considered a linear and shallow ramp up of SRM intervention. This perhaps works particularly well for a linear climate response. Should we not instead be looking at more rapid temperature reductions, to reduce risks of crossing tipping points (Greenland, permafrost, etc)? Health impacts You assume quantifiable health impacts from particulate rain out. However, the aerosol rain out would be well distributed, with much occurring in depopulated areas. Are the health impacts of distributed particulates comparable to those from concentrated loading patterns (eg focussed on harbours and cities)? I assume the calculated risks were derived from these concentrated loading patterns. Will the health risks be reduced because rain out will likely mix with, or condensate, raindrops - hence diluting them to destruction. Delivery mechanism Previously you considered homogeneous condensation of H2SO4, and today you discussed in-situ high-altitude combustion of solid S to release SO2. In response to my verbal question, you stated that particle size distribution issues only kick in around 'a few' MT. However, having a constant delivery mechanism reduces the risk of 'nasty surprises' on switching, despite increased lofting costs in earlier stages. Is there an argument for 'starting as you mean to carry on'? Distribution pattern Particularly with high aerosol loads, there may be advantages to a temporally bound injection regime (when using precursor gases instead of direct particulate injection). This is due to the benefit of constrained particle growth. (See reference below) Does a potential requirement for a temporally-concentrated injection regime mean that the use of aircraft becomes problematic from a cost point of view? EG to do all the injection in a month per year, you'd likely need up to 12x as many aircraft. You also considered the spatial distribution pattern. A 30N to 30S injection regime may take advantage of the Brewer Dobson circulation. However, is there not a risk of particle size growth and rain out. See Heckendorn et al ( http://m.iopscience.iop.org/1748-9326/4/4/045108 ), who discussed particle growth, and I recall this paper may also have discussed spatial (vertical and latitude) and temporal distribution patterns, and advocated a more spatially varied injection regime. Smart particles You touched on the concept of 'smart particles'. Is there a risk that these can be weaponised to make a solar concentrator, capable of burning buildings on the ground? -- You received this message because you are subscribed to the Google Groups "geoengineering" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/geoengineering. For more options, visit https://groups.google.com/groups/opt_out. -- You received this message because you are subscribed to the Google Groups "geoengineering" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/geoengineering. For more options, visit https://groups.google.com/groups/opt_out.
