Recent Comments

Prev  906  907  908  909  910  911  912  913  914  915  916  917  918  919  920  921  Next

Comments 45651 to 45700:

45651. KR at 05:34 AM on 11 May 2013
       The anthropogenic global warming rate: Is it steady for the last 100 years?
       

       Dr. KK Tung - You stated that "atmospheric CO2 has been measured to increase almost exponentially".

       This is incorrect, CO2 is increasing more than exponentially over the last century, and the CO2 portion of forcings is therefore increasing faster than linearly. 

45652. KR at 05:29 AM on 11 May 2013
       The anthropogenic global warming rate: Is it steady for the last 100 years?
       

       jdixon1980 - What DS said:

       Notice that CO2's radiative forcing increases faster after 1950, because increasing CO2 faster also increases its logarithm faster. That's what makes the forcing "slightly more curvy than linear".

       Now if CO2 were increasing in a polynomial fashion, its log (while still increasing with CO2 - the log of a larger number is larger than the log of a smaller number), would by definition have a positive first derivative, but would have a negative second derivative - less than exponential growth, less than linear forcing increase. If that polynomial grew larger (faster growth curve), the log would increase faster than it would otherwise have, with a higher (although potentially still negative) second derivative. 

       In the case of actual CO2 measures, the growth of CO2 is higher than exponential, the log of CO2 has positive first and second derivatives. DS did skip a step, though, and did not mention that CO2 growth is faster than exponential. 

       But no matter what, if the rate of growth of a value increases, the growth of the log of that value (always with a positive first derivative under growth) will grow faster than it would without that increase. 

45653. jdixon1980 at 05:01 AM on 11 May 2013
       The anthropogenic global warming rate: Is it steady for the last 100 years?
       

       For a reality check, I asked my friend Lucas about it (a math professor and a Climate Reality Leader with the Climate Reality Project).  Some of his reply went a little over my head on a quick reading, but the gist of it is that whether increasing the first derivative of a function will also increase the first derivative of its logarithm depends on what the function is: 

       "J: Just because a function's derivative is increasing doesn't mean the derivative of the log of the function is increasing. For a function's log to have increasing derivative, it must satisfy a certain differential equation: d^2(ln(f))/dt^2 = (f(t)f''(t)-(f'(t))^2)/f^2(t) > 0 <=> f(t)f''(t) > (f'(t))^2. This is never true for lines (f''(t) = 0) and typically not true for polynomials (for a polynomial with leading term x^n, the leading term on the left is n(n-1)x^(2n-2) and the leading term on the right is n^2x^(2n-2), so the right always eventually dominates).

       However, I'm not sure that's what DS meant. DS might mean that if you increase the first derivative of f then you also increase the first derivative of ln(f). More formally, if f(0) = g(0) and f'(t) > g'(t) for t > 0 then (ln(f(t)))' > (ln(g(t)))' for t > 0. Even this might not be true though. For it to be true, we would need the following differential equation to hold: (ln(f(t)))' = f'(t)/f(t) > g'(t)/g(t) = (ln(g(t)))', or equivalently f'(t)g(t) > f(t)g'(t). This now holds true for lines (if f(t) = at+f(0) and g(t) = bt+f(0), the differential equation is abt+af(0) > abt+bf(0), which is true exactly when f'(t) = a > b = g'(t)). It is also true for pure exponential functions (if f(t) = a^t and g(t) = b^t then the diff eq becomes ln(a)a^tb^t > ln(b)b^ta^t, which is true iff a > b iff f'(t) > g'(t)). But for more complicated functions, the exact relationship depends on the functions involved."

45654. jdixon1980 at 04:28 AM on 11 May 2013
       The anthropogenic global warming rate: Is it steady for the last 100 years?
       

       KR @45 "Abrupt or not, if CO2 concentration shows a faster increase over time, the Ln(CO2) will also show a faster increase over time."  This broad mathematical proposition is false, as I showed above by the counterexample in which CO2 = t^2.  But the point is moot because, as you point out, we can see from observation that ln(CO2) actually is increasing faster over time.  I know that now, because you pointed it out to me, though it wasn't clear to me from DS's statement at 7.  The reason why the mathematical proposition holds true in for the particular case of CO2 as a function of time is that CO2 apparently is not proportional to t^2, or for that matter even to e^t (the second derivative of ln(N*e^t) being zero, not positive), but to something even higher order than that.       

        

       SkS is good for people like me who are technically literate (I'm a patent attorney with a mechanical engineering degree, which is admittedly growing stale after ten years) but lack broad climate knowledge, precisely because of instances like this: I read something that confuses me by contradicting what I remember from first-semester calculus, I comment about it, and someone like you can straighten me out by explaining the context that makes sense of the confusing statement, in this case by showing me that the confusing statement wasn't intended as broadly as its literal reading.  

45655. KR at 02:32 AM on 11 May 2013
       The anthropogenic global warming rate: Is it steady for the last 100 years?
       

       jdixon1980 - While I used N and N+1 for illustration, there hasn't been a sudden abrupt shift, but rather (as seen in the CO2 data) an increase in the rate of CO2 increase rising over time, a positive second derivative. And a fairly simple check on this, taking the ln(CO2) and looking at its behavior over time, shows that forcing is actually increasing greater than linearly, that CO2 is increasing greater than exponentially. 

       Abrupt or not, if CO2 concentration shows a faster increase over time, the Ln(CO2) will also show a faster increase over time. 

45656. jdixon1980 at 02:15 AM on 11 May 2013
       The anthropogenic global warming rate: Is it steady for the last 100 years?
       

       KR @42, 43 - thanks for the clarification that time is a multiplier in the exponent.  I don't think that's a sign of my needing to check my "math" per se, so much as my (lack of) knowledge of the actual relationship of CO2 rise to time.  My math is consistent with what you are saying - in your example where a function that is initially e^(T*N) subsequently becomes e^(T*(N+1)),  it may be that the constant term in the exponent changed from N to N+1 because it is not actually a constant, but a linear function of time, which would give you something that acts like e^(t^2), an example that I mentioned of a situation in which a function "increases faster" as its logarithm also "increases faster."  

       However, I assume you are suggesting that the second term is not actually a linear function of time, but rather a step function that abruptly shifted being a constant N to a constant N+1, at some precise moment or over some short interval, because of some historical event causing an abrupt shift in the emissions trend?  As you illustrate, DS's statement would hold true for that case as well, because as you illustrated the first derivative of the logarithm would change from N to N+1.  

       I hadn't considered this kind of abrupt shift when trying to understand and evaluate DS's statement.  I still would contend that not all abrupt transformations of a function resulting in a transformed function having a larger first derivative would also result in a logarithm of the transformed function having a larger first derivative; for instance it wouldn't hold true if CO2 = (N)t^2 suddenly became CO2 = (N+1)t^2, but that's just a quibble with his wording that I wouldn't have brought up had I initially understood the context, which I do now thanks to your clarification.  

45657. John Hartz at 01:59 AM on 11 May 2013
       What you need to know about climate sensitivity
       

       The results of the study that Albatross cites above are also summarized in:

       New study tells three million-year old story of the Arctic by Roz Pidcock, The Carbon Brief, May 9, 2013

45658. Albatross at 01:28 AM on 11 May 2013
       What you need to know about climate sensitivity
       

       Nice summary Dana, thanks.

       John Mason just alerted SkS to a new paper in Science by Brigham-Grette et al. (2013). They examined sediment cores from a lake in Siberia going back over 3 million years.  Their findings are not reassuring and suggest that fast-feebacks (and by extension equilibrium climate sensitivity) may be more aggressive than previously thought:

       "Evidence from Lake El’gygytgyn, NE Arctic Russia, shows that 3.6-3.4 million years ago, summer temperatures were ~8°C warmer than today when pCO2 was ~400 ppm. Multiproxy evidence suggests extreme warmth and polar amplification during the middle Pliocene, sudden stepped cooling events during the Pliocene-Pleistocene transition, and warmer than present Arctic summers until ~2.2 Ma, after the onset of Northern Hemispheric glaciation."

       These findings suggest that we are very likely in for a whole lot of hurt in the coming decades.  Even more worrying is that we are already at 400 ppmv-- I shudder to think what the consequences for future generations will be should we continue with BAU and reach over 1000 ppmv... 

45659. Doug Hutcheson at 13:17 PM on 10 May 2013
       Climate change will raise the sea level in the Gulf of Finland
       

       Are there corresponding places where SLR is greater than expected from isostatic rebound? I seem to remember that some areas, such as the Missippi delta, are experiencing land lowering, due to isostatic rebound elsewhere on the crustal plate (one side of the dish is rising, the other side is falling). Such places would be subject to greater inundation, due to the double whammy of isostatic fall and SLR. Nasty combination, if I have my facts about right.

45660. Doug Hutcheson at 13:02 PM on 10 May 2013
       Greenland: A Ring of Mountains
       

       Interesting and disturbing, as if I could be any more disturbed ...

        

       <pedantry>"more data has became available" should be "more data has become available"</pedantry>

45661. macoles at 12:20 PM on 10 May 2013
       2013 SkS Weekly News Roundup #19A
       

       gelderon52 @1

       I checked out your site and I generally agree with your advice on how to discuss climate science with contrarians. Personally I find listening to their full opinion on the matter first in order to categorise the extent of their knowledge and motivation of their denial, before I start methodically explaining where they are wrong. There is a great deal of variation in the knowledge and motivation of those who reject the consensus view.

45662. John Hartz at 10:51 AM on 10 May 2013
       2013 SkS News Bulletin #10: Alberta Tar Sands and Keystone XL Pipeline
       

       Mike,

       As fate would have it, the Globe and Mail published an indepth summary of the status of proposed projects to sequester CO2 emissions in Alberta. Two key paragpraphs:

       "Alberta, too, saw two of four planned projects cancelled this year and last, including TransAlta Corp.’s Pioneer project and the Swan Hills Synfuels LP synthetic gas plant. In both cases, the companies said the economics of the projects no longer made sense.

       "But Mr. Hughes said the two remaining projects – the Shell Quest project that will capture CO2 from an oil sands upgrader and the Alberta Carbon Trunk Line – spurred by provincial and some federal dollars, will go ahead. CCS technology is the cornerstone of Alberta’s long-term climate plan, and is expected to contribute 70 per cent of the planned emissions reductions by 2050."

       Keeping the faith in carbon capture and storage by Kelly Cryderman and Shawn McCarthy, Globe & Mail, May 6. 2013 

45663. KR at 10:24 AM on 10 May 2013
       The anthropogenic global warming rate: Is it steady for the last 100 years?
       

       jdixon1980 - Not incidentally, Tamino demonstrated some time ago that CO2 is in fact rising faster than exponentially, hence CO2 forcings are rising faster than linearly. 

       

       [Source]

45664. KR at 10:19 AM on 10 May 2013
       The anthropogenic global warming rate: Is it steady for the last 100 years?
       

       jdixon1980 - The logarithm of a number increases (logarithmicly) as the number increases, the log decreases as a number decreases. That's by the definition of a logarithm.

       If CO2 is rising at e^T*N, then speeds up to e^T*(N+1) the natural log rise rate will go from T*N to T*(N+1). In other words, if CO2 rises faster over time, the forcing (the log of CO2 change) will also rise faster over time. 

       I suggest you check your math - time is not the base, it's a multiplier in the exponent. 

45665. mikeh1 at 10:05 AM on 10 May 2013
       Who is Paying for Global Warming?
       

       Justin Guay from the Sierra Club is quite skeptical about whether many of these coal plants will ever be built.

       Here for example - "Indias coal illusion"

       Here is a Greenpeace map of the large Galilee Basin coal development in Australia that Guay refers to.

45666. Mike3267 at 07:33 AM on 10 May 2013
       2013 SkS News Bulletin #10: Alberta Tar Sands and Keystone XL Pipeline
       

       I read that the Alberta tar sands development is to include CO2 sequestering. See:Technology Review
       Anyone know more about this?

45667. jdixon1980 at 07:32 AM on 10 May 2013
       The anthropogenic global warming rate: Is it steady for the last 100 years?
       

       Heh, I just noticed that I applied the chain rule to differentiate ln(t²) instead of just convering it into 2ln(t) and differentiating that - shows my own math is a little rusty...  But it doesn't change the result of the second derivative, -2/t²

45668. jdixon1980 at 06:57 AM on 10 May 2013
       The anthropogenic global warming rate: Is it steady for the last 100 years?
       

       DS @7: You said, "increasing CO2 faster also increases its logarithm faster"  

       I believe this assertion is false, because increasing CO₂ faster does not necessarily increase its logarithm faster. By "increasing CO₂ faster" I take you to mean that as time marches on, the first derivative of CO₂ as a function of time increases, i.e., the second derivative is positive.  Thus, I take your assertion to be that if the the second derivative CO₂ as a function of time is positive, then the second derivative of the logarithm of CO₂ as a function of time must also be positive.  If this understanding of what you are asserting is correct, then what you are asserting is false.  

       As a simple counterexample, let's say CO₂ = t².  

       t² is a function that "increases faster" over time, i.e., its first derivative, 2t, increases as t increases, and its second derivative, 2, is positive.  

       Therefore,  if your statement that "increasing CO₂ faster also increases its logarithm faster" were true, then the first derivative of ln(t²) should also increase as t increases.  

       Let's have a look to see if it does:

       d ln(t²)/dt = d ln(f)/df * d(t²)/dt, where f = t² (applying the chain rule)

       => d ln(t²)/dt = (1/f) * 2t = (1/t²) * 2t = 2/t, which decreases as t increases.  (I.e., differentiating again to get the second derivative, you get -2/t² which is negative for all real values of t)

       So your broadly stated assertion is false. 

       Even if we take the example of CO₂ being equal to (or proportional to) e^t, which has a positive second derivative (the first derivative of e^t is itself, as is the second derivative), the second derivative of ln(e^t) = d²ln(e^t)/dt²= d²t/dt² = d(1)/dt = 0.  In other words, even an exponential relationship of CO₂ to time tends to result in a linear relationship between  temperature and time.  A linear relationship is what the Tung study apparently found, which is not inconsistent with the understanding that emissions are rising exponentially, assuming a simple relationship where temperature change is proportional to the logarithm of CO₂ change.

       This is not to say that there are no examples of f(t) having d²f(t)/dt² > 0 where d²(ln(f(t))/dt² is also > 0.  (For example, if CO₂ = e^{t^2}, read "e to the t squared" then the second derivative of CO₂ and the second derivitive of its logarithm would both be positive.) But that's not the same as saying the former implies the latter, which is how I read your premise, on which you apparently based your conclusion that Tung's interpretation of the data was aphysical.  

       Don't get me wrong, I know next very little about the physics, but based on my understanding of calculus, I suspect you had overlooked this nuance of the math.  

45669. SEAN O at 00:32 AM on 10 May 2013
       Cherrypicking to Deny Continued Ocean and Global Warming
       

       scaddenp'

       Thanks, Fig. 1 in Levitus makes it clear.

45670. John Russell at 23:05 PM on 9 May 2013
       Who is Paying for Global Warming?
       

       Skeptical Wombat @3

       The claim that "China increased its wind power production of electricity be more than it increased its coal fired electricity production last year" is not quite all it seems and should be put in context.  So it's true, but one of those outliers that can distort the actual picture.

       But don't get me wrong; China is the one country that if it has the will, will find a way, and, because of it's political set up, can turn things round rapidly.

45671. gelderon52 at 19:52 PM on 9 May 2013
       2013 SkS Weekly News Roundup #19A
       

       Hello,

       I would like to introduce a modest effort of mine and of some colleagues and friends in the field of disseminating the results of climate science. It is titled "The frog that jumped out"

       http://thefrogthatjumpedout.blogspot.it/

       As I said, it is a modest effort, but the idea is that there are many blogs about climate science, but not so many about communicating climate science. So, we hope to have a bit of an impact in this field and we are trying to do our best.

       If you were interested in collaborating with us with your thoughts, papers, or comments, you are most welcome.

       Ugo Bardi

45672. scaddenp at 14:49 PM on 9 May 2013
       Cherrypicking to Deny Continued Ocean and Global Warming
       

       Sean O - Sparcity of data is reflected in error bars. Note the increase in size of these at 2003. Its especially noticable in 0-2000 data in the Leviticus 2012 paper

45673. sidd at 13:04 PM on 9 May 2013
       Who is Paying for Global Warming?
       

       The war continues. China is closing older coal plants:

       http://www.chinadaily.com.cn/business/2010-07/26/content_11047808.htm

       Environmentalists knock down another in India

       http://www.huffingtonpost.com/justin-guay/indian-court-strikes-down_b_3203153.html

       
       sidd

45674. Sceptical Wombat at 12:13 PM on 9 May 2013
       Who is Paying for Global Warming?
       

       While it is clear that we must do much better at replacing fossil fuels with other forms of electricity generation, we should recognise that some progress is being made.  For instance last year China increased its wind power production of electricity be more than it increased its coal fired electricity production.  And that is before you count its investment in photovoltaic generation.

       Also many of the new coal powered generators are replacing older, smaller, less efficient and dirtier plants, so it is misleading to simply quote the number of new power plants.

45675. Mark Bahner at 11:07 AM on 9 May 2013
       Global Warming: Not Reversible, But Stoppable
       

       (-snip-)

       Moderator Response:

       [DB] And with that, we are done here, as we have erred well into sloganeering territory.  Anyone wishing to engage Mark further on this may do so at his blog:

       http://markbahner.typepad.com/

       Let us return now to the OP of this thread, Global Warming: Not Reversible, But Stoppable

45676. Riduna at 10:33 AM on 9 May 2013
       Who is Paying for Global Warming?
       

       John – As far as I know, the only means of reducing CO2 emissions from fossil fuelled power plants is CCS technology which, in its present form, requires 40%-50% of power plant output to operate. If more effective and cost-efficient technology is not developed, ff power plants will be replaced by power plants fuelled by renewable sources.

       That will occur. When depends on the speed with which advances are made in production and storage of electricity.

       Thanks for your PS. Quite right.

45677. John Russell at 09:52 AM on 9 May 2013
       Who is Paying for Global Warming?
       

       It's an inevitable fact of life that 'keeping the lights on'—and the population in a state to which they've grown accustomed—will always trump long-term concerns. I believe it's becoming clear that it's only when the costs of 'carrying-on-as-we-are' manifest themselves in the short-term, perhaps as a result of the escalating costs of extreme weather events and their impact on the food supply, will the pendulum swing towards action on climate. 

       This was driven home to me when I read a paragraph on 'discounting' in the Ereskes/Conway book 'Merchants of Doubt' (p180)—quote: 

       "A dollar today is worth more to us than a dollar tomorrow and a lot more than a dollar a century from now, so we can 'discount' faraway costs." 

       This is the mantra of the majority of economists everywhere, and it's this head-in-the-sand, Micawberesque— 'something-will-turn-up'—mindset that drives the economists and lawyers who dominate our political elites.

       Regarding CCS; I also think it's inevitable that it will be viewed as a means to make new fossil-fuel fired plants viable—that's why Shell is so keen on it—and never as a means to reduce CO2 from the atmosphere. So, what with inevitable leakages and failures, the GHG ratchet will always be upwards.  I hope I'm wrong.

       [PS: perhaps the words 'coal-fired' should be added to the title of fig 1?]

45678. SEAN O at 08:07 AM on 9 May 2013
       Cherrypicking to Deny Continued Ocean and Global Warming
       

       In discussing rising OHC, a commenter objected to Levitus etal on the grounds that the data is too sparse, what is the best response?

       Thanks in advance.

45679. John Hartz at 05:40 AM on 9 May 2013
       Global Warming: Not Reversible, But Stoppable
       

       I believe we can all agree that humans require food and water to subsist. In the context of the ongoing discussion of what manmade climate change may wrought in the near future, here is very sobering assessment.

       "Half the world's population—5.2 billion people—could be doomed to an insecure and greenhouse gas-causing reliance on food imports by 2050, according to a new study.

       "In the study appearing in Environmental Research Letters, Marianela Fader of the Potsdam Institute for Climate Impact Research (PIK) and her team looked at water and land constraints affecting national food self-sufficiency, and investigated how cropland expansion and increased productivity could create the potential for greater national food self-sufficiency."

       Source: Are We Doomed to Food Insecurity? New study shows billions more people reliant on food imports in 2050 by Andrea Germanos, Common Dreams. May 8, 2013

         

45680. KR at 04:11 AM on 9 May 2013
       Global Warming: Not Reversible, But Stoppable
       

       My issue with Mark Bahners projections are that they are a combination of cornucopian thinking (all needs for the future can be met by continuing advances in technology, with an underlying assumption of extremely low costs - no limits to growth) and the assumption of numerous black swan events (AI singularity, economic growth an order of magnitude that which has been observed, revamping of all methods of production without economic upheaval, etc).

       These are not reasonable assumptions, or realistic projections - they are just wishful thinking. 

       And they ignore what has been determined to date - that mitigating further greenhouse warming by reducing emissions will be far less expensive than adapting to it later, or diverting the resources/energy to try to sequester CO2 released by continued "Business as Usual" behavior. The economically sensible path is to minimize the GHG problem as much as possible, not to band-aid it later on.

       Cornucopian/black-swan goals like his, to me, represent little more than an avoidance of the issues - an attempt to kick the problems down the road under wishful projections that the costs down the road will be trivial. 

       Personally, I would rather address the issues now - based on supportable economic predictions that current action will cost far less than future re-actions. 

45681. Rob Honeycutt at 03:33 AM on 9 May 2013
       Global Warming: Not Reversible, But Stoppable
       

       There is a fundamental fallacy where Mark states:

       "The value of this information is a simple matter of science. Science consists of one person predicting “a”, and another person predicting “b” (each with explanations for why they are making their predictions) and then observing to see which predictions are correct (or it could be “c” and neither is correct). Or both persons could agree on “a” or “b” in which case there’s not a point in further debate."

       What Mark is describing is a "controlled experiment."  His predictions are not controlled experiments.  He ignores a certain level of randomness in the system.  Just because prediction "a" might have been correct does not necessarily mean that the methods used by "a" were better than those of "b" when the experiment takes place in an uncontrolled environment, like world economics.

       Again, you can use various methodologies to establish a range of outcomes in order to better constrain uncertainties about future events.  You can't use methodologies to predict the future.

       Think of it this way.  There is some young, hyper-brilliant kid recently born who has the capacity to understand the AI issue better than anyone ever has.  That person could grow up to fundamentally change the future of the human race by the year 2040.  But, as fate would have it, that child is killed in an auto accident at age 9.

       This is what bugs me about what Mark is putting forth.  His "predictions" include highly uncertain events occurring at unknown points in the future.  So, relative to how we respond today to the challenges humanity faces down the road, his predictions are completely without value.

45682. Tom Dayton at 03:29 AM on 9 May 2013
       Trenberth on Tracking Earth’s energy: A key to climate variability and change
       

       The links to Trenberth's papers in this post by Trenberth are broken, because all his files were moved to a new spot at UCAR.  Will somebody please update the links? 

       Moderator Response:

       [DB] All links are now updated; thanks!

45683. John Hartz at 01:20 AM on 9 May 2013
       Global Warming: Not Reversible, But Stoppable
       

       In the context of the ononging discussion of what the future portends, the following paragraph jumped-out at me when I read Tom Englehardt's essay, And Then There Was One: Imperial Gigantism and the Decline of Planet Earth posted May 7 on TomDispatch.com.

       "The present capitalist model (the only one available) for a rising power, whether China, India, or Brazil, is also a model for planetary decline, possibly of a precipitous nature. The very definition of success -- more middle-class consumers, more car owners, more shoppers, which means more energy used, more fossil fuels burned, more greenhouse gases entering the atmosphere -- is also, as it never would have been before, the definition of failure. The greater the “success,” the more intense the droughts, the stronger the storms, the more extreme the weather, the higher the rise in sea levels, the hotter the temperatures, the greater the chaos in low-lying or tropical lands, the more profound the failure. The question is: Will this put an end to the previous patterns of history, including the until-now-predictable rise of the next great power, the next empire? On a devolving planet, is it even possible to imagine the next stage in imperial gigantism?"

        

45684. JasonB at 18:51 PM on 8 May 2013
       Global Warming: Not Reversible, But Stoppable
       

       Mark Bahner @ 105,

       You're a "doer" in the field of how AI affects macroeconomic growth?

       No, I'm a "doer" in the field of AI, which gives me a much better idea of the problems that need solving and the likelihood of fantasies about its capabilities in the next few decades coming true than a macroeconomist who has NFI about the subject and thinks it all boils down to FLOPS/$.

       You know your way around basic economic models like the Solow-Swan model?

       As it happens, I actually did some software development for an economics professor to earn extra money while I was a postgrad and my observation at the time was that the problems in economics are either trivial to solve ("You've found an analytic solution! Well done! I'm sure that those simplifying assumptions you made do hold in the real world...") or so intractable that solutions aren't possible (or meaningful). The paper you presented has done nothing to dispel that notion.

       No, the computing power is not there. Using an estimate of say, 20 petaFLOPS for a human brain, and the human population is approximately 7 billion. So that’s 140 billion petaFLOPs.

       Why would an AI's processing power need to be comparable to the combined processing power of the entire human race? We can't network our brains together to achieve higher levels of intelligence (in fact, evidence suggests the opposite...) so the total figure is meaningless. If a computer had more processing power than a single human brain and we knew how to harness that processing power to create an AI then there's no reason to believe that a smarter-than-human AI wouldn't result. And, unlike humans, we can always build more powerful computers (the first exaFLOPS computer is expected by about 2018, which would have the processing power of 50 humans according to your estimate) so there's no real upper limit to intelligence in that regard.

       The problem is that it's still just a very fast calculator.

       If we ever do manage to crack that nut — and this is still an open question — then it wouldn't be suprising for this genius-level AI to set about writing its own AI software that was better than what we could conceive, and lead to the rapid escalation in artificial intelligence known as the technological singularity, at which point all bets are off.

       In the meantime, we have a gradual and hard-fought increase in "intelligence" that really does improve productivity (or else I'd be out of a job), but it's been going on for a long time now and it hasn't seen the magical boost in GDP that your link suggested, as demonstrated by Tom Curtis. Moreover, as we are less and less constrained by the computational abilities of the machines we are using, the more and more obvious it becomes that the constraints are us, and so the less impact each successive improvement in hardware has. (Some examples to ponder: how much better are word processors today than they were five years ago? Ten years ago? 15 years ago? 20 years ago? During the 80s and 90s improvements were obvious; now the only way you can tell which version you're using is the arbitrary cosmetic changes introduced to differentiate the products. What about games, or film CGI? In the beginning there were massive improvements in hardware and software every few years that allowed more and more realism to be introduced; now each generation is much like the last. This tailing off in advancement is normal as technology matures, and completely the opposite of what your forecast assumes.)

       The bottom line is that there's no evidence that AI will be the magic bullet that makes out-of-this-world growth rates of GDP possible, and if AI really did make such a massive breakthrough, there's a good chance it would be the end of us anyway so it's perhaps not something to hope for.

       The value of this information is a simple matter of science. Science consists of one person predicting “a”, and another person predicting “b” (each with explanations for why they are making their predictions) and then observing to see which predictions are correct (or it could be “c” and neither is correct). Or both persons could agree on “a” or “b” in which case there’s not a point in further debate.

       The problem is that (a) the predictions are utterly worthless and (b) the timeframes are so long that we wouldn't even know who turned out to be correct.

       The fact that you are willing to give people $20 to just Make Stuff Up is consistent with the value you place on other worthless projections, I suppose.

       I think it's also worth pointing out that the difference between these "predictions" and the IPCC's scenario-based projections is that the latter allow us to evaluate the likely effects of decisions that we make today to aid our decision making. The equivalent in your case would be a scenario where the AI Silver Bullet hits, malaria is cured, etc., etc. — in other words, those are your inputs, your assumptions for your scenario, not a prediction of the future.

       Additionally, in the case of my asking for your predictions for all those parameters (not just GDP) is to see whether or not we agree regarding some parameters that I think reflect "quality of life" (e.g. life expectancy at birth, number of malaria deaths, people with access to adequate sanitation, etc.).

       Suppose, for the sake of argument, that the number of malaria deaths reaches zero because e.g. a cure has been found by then. How does that justify pushing the CO2 drawdown problem onto them? It's as if you are saying "Hey, they will be better off than us because they won't have to deal with problem X, therefore we can burden them with problem Y".

       You also seem to be forgetting that there are plenty of countries in the world today that don't have malaria. Doesn't that mean that those countries should start reducing CO2 on that basis?

45685. nuclear_is_good at 18:18 PM on 8 May 2013
       The anthropogenic global warming rate: Is it steady for the last 100 years? Part 2.
       

       While it is great to have scientists explaining their published work to the public, I find a little surprising that people still try to do various forms of regression based on surface temperatures and completly neglect the amount of energy going in and out of the ocean - that amount is easily dwarfing the amount of "AMO corrections" and there is a passing mention on how that could be part of the explanation but as far as I can see there is no attempt at including that data into the model and calculations!

45686. K.a.r.S.t.e.N at 11:05 AM on 8 May 2013
       The anthropogenic global warming rate: Is it steady for the last 100 years? Part 2.
       

       Thank you very much for the posting! I appreciate your effort to elaborate on the AMO issue in such detail. Many points to discuss in much more depth, which I'll do at a later stage (as soon as time allows). For the time being, I would like to clarify and stress that Fig.5 was taken from Tamino! Some may misread that part. It is not mine! While the fluctuations are the true AMO signal (amplifying the external sulfate aerosol forcing), the negative trend is most likely due to the delayed response to the very strong warming once the effect of the Clean Air Act was detectable in the sulfate aerosol loading. It merely reflects the thermal inertia. It's not the aerosols as far as the trend is concerned. The ups and downs are an amplified aerosol response (+ other forcings of course).

       As a teaser: I strongly disagree that debate is about whether the observed AMO is anthropogenically forced or not. The debate is about how strong the Atlantic ocean (namely the AMOC) responds to the external forcing. The external (anthropogenic) sulfate forcing is there. We know it! We measured it! It's therefore a well established fact which I won't further discuss. I pointed to the respective literature already. Your reference to the CET time series is therefore rather pointless in my point of view. It just shows how sensitive European climate (and the North Atlantic at that matter) is to external forcing, i.e. volcanic eruptions before the 20th century.

       More to come later ...

45687. Daniel Bailey at 10:30 AM on 8 May 2013
       Leave It in the Ground, Climate Activists Demand
       

       @ Mark Bahner:

       You have posted 2 comments thus far on this thread that, rather than adding to the signal of participation and dialogue on this thread that the other contributors work hard to maintain, do nothing more than detract from that discussion. As such, they serve no useful purpose, being noise among the signal. To that end, if you do not wish to have your noise filtered out, please attenuate said noise in favor of greater signal.

       Capiche?

45688. Rob Honeycutt at 07:59 AM on 8 May 2013
       Global Warming: Not Reversible, But Stoppable
       

       And Mark...  I just don't think you get the point to economic predictions.  Everyone understands they're wrong.  I'll say it again, you can't predict the future.  

       You can explore potential outcomes.  You can discuss market influence that might or might not come about.  You can discuss trends, opportunities, etc.  

       They way people use this information is so they can mitigate risk.

       The problem with what you're doing is, you're taking your assumptions as fact.  I'm sorry but you just don't have any more insight than the next person on future facts.

45689. Rob Honeycutt at 07:49 AM on 8 May 2013
       Global Warming: Not Reversible, But Stoppable
       

       Mark...  You have not "predicted the future" for the coming century.  You have put forth predictions about the future.  Not only that, you've made predictions about things that you will never know.  

       I can make a prediction that pigs will develop the capacity to vocalize their thoughts in 2150, and I can make grand bets on the accuracy of the prediction, but none of us will be there to see it.  

       Making bets you can never collect on borders on delusional.

       You say, " Based on that calculation, I think only the 2020 prediction is potentially too high. I think the others will be way too low."  And then you add, "(Barring takeover by Terminators or global thermonuclear war, of course.)"

       So, even you are already saying that you are likely going to be wrong.

       (Psst.  Also, you might take strong notice of the fact that China is ramping down it's red-hot economic growth of the past 30 years.)

45690. Mark Bahner at 07:34 AM on 8 May 2013
       Global Warming: Not Reversible, But Stoppable
       

       Rob Honeycutt makes this remarkable comment:

       Mark seems to not be aware of the purpose of economic forecasting. You, literally, can not predict the future.

       Rob, I have predicted the future:

       “The world per-capita GDP in the year 2000 was approximately $7,200. The world per-capita GDP (in year 2000 dollars) will exceed $13,000 in the year 2020, $31,000 in 2040, $130,000 in 2060, $1,000,000 in 2080, and $10,000,000 in 2100.”

       And Nobel Laureate Robert Lucas Jr. predicted the future in his Journal of Economic Perspectives article. The difference between our two predictions is that I think that the development of artificial intelligence will greatly increase economic growth, whereas Robert Lucas Jr. either did not consider the development of artificial intelligence and its effects on economic growth (most likely) or considered the development of artificial intelligence, but thought it would not affect economic growth (possible--I've never had any contact with him--but I think highly unlikely).

       Further, the Long Bets site alone has many other economic predictions, such as #626: "The United States will not experience hyperinflation (defined by 3 consecutive months of 6% Month over Month inflation according to the Billion Prices Project measurement of MoM inflation) from April 17th 2012 to April 17th 2017. ”

       The thing about forecasting is, no matter what, you will be wrong.

       No, I will be right if the world per-capita GDP (in year 2000 dollars) exceeds $13,000 in the year 2020, $31,000 in 2040, $130,000 in 2060, $1,000,000 in 2080, and $10,000,000 in 2100.

       "...you can't build a rosy scenario..."

       Those predictions are not a "rosy scenario." At the time I made the Long Bets #194 I hadn't done my calculation of the number of human brain equivalents added every year. Based on that calculation, I think only the 2020 prediction is potentially too high. I think the others will be way too low. So if I make it past the 2020 prediction of $13,000 (in year 2000 dollars), I think there's smooth sailing to the end of the century. (Barring takeover by Terminators or global thermonuclear war, of course.)

45691. Kevin C at 06:32 AM on 8 May 2013
       The anthropogenic global warming rate: Is it steady for the last 100 years? Part 2.
       

       Dr Tung:

       The spatial temperature patterns you've shown here are very interesting, but I am concerned that you have attributed a significant amount of temperature variation to the AMO which is already well explained by other causes. In particular I draw your attention to the volcanic record, e.g. here for the recent data, and here for a longer record. The short record is shown below, and shows two clusters of volcanoes with the intervening quiet spells.

       

       The volcanic record explains the apparent oscillation in the temperature record (your figure 4) rather well, with the dips in temperature corresponding to eruptions as follows:

       1809-1835: Tambora, Consiguina and two unknown erruptions

       1883-1912: Krakatau, Novarupta, Santa Maria

       1960-1990: Agung, El Chicon, Pinatubo

       This is illustrated by the fit to the data obtained by the Berkeley team using just volcanoes and CO2.

45692. John Chapman at 06:20 AM on 8 May 2013
       Distinguishing Between Short-Term Variability and Long-Term Trends
       

       Dana, now that I hunt it down I see the year is more like 2015 rather than 2013, though there would have to be a transition towards 2015.  The source is www.skepticalscience.com/david-archibald-exaggerates-solar-influence-on-future-climate-c   see figure 3.

45693. IanC at 04:38 AM on 8 May 2013
       The anthropogenic global warming rate: Is it steady for the last 100 years? Part 2.
       

       Dr. Tung,

       Thank you for taking the time to write this series of posts.

       I'm wondering, how sensitive are the results of Zhou and Tung 2013 to the choice of AMO index? i.e. what will the results look like if the index as defined by van Oldenborgh et al. 2009 or Trenberth and Shea 2006 were used instead?

        

45694. Bob Tisdale at 03:19 AM on 8 May 2013
       The anthropogenic global warming rate: Is it steady for the last 100 years? Part 2.
       

       KK Tung:  Thanks.  The post was easy to read and understand.

        

       Regards

45695. Dikran Marsupial at 02:54 AM on 8 May 2013
       Global Warming: Not Reversible, But Stoppable
       

       Mark Bahner, I think you have missed the point, regardless of the PFLOPS available, we don't know how to write the software, so the point is moot.

       Secondly, making projections based on speculation is not science, it is futurology.  In science whether a projection is worth paying attention to (or for that matter making) rather depends on the strength of the supporting argument/evidence.  Merely challenging someone to make a prediction so we can wait and see who is right is not science, and comes across as a rhetorical device to avoid addressing the factual basis for your position.

       At the end of the day, physics is a lot more predictable than economics, which suggests physics is a more reliable solution to the problem than deligating the problem to others in the hopes that they can fix it more cheaply.

45696. Mark Bahner at 02:50 AM on 8 May 2013
       Global Warming: Not Reversible, But Stoppable
       

       One final thing about the predictions I requested JasonB to make. (Offered him $20 to make! I challenge anyone to make me similar offers!) 

       The table with the various parameters is located here. That table has the parameter values circa 2012 (and the units, such as for PM10 concentration), for greater ease in making predictions. For example, malaria deaths are estimated at 770,000 in the table. But I've seen a Lancet article up to 1.2 million. Similarly, the estimate of 37% of the world without proper sanitation circa 2012 is just one estimate. So the idea would be to assume that the number I gave in the table was correct, and go off of that.

45697. Mark Bahner at 02:40 AM on 8 May 2013
       Global Warming: Not Reversible, But Stoppable
       

       Oops. Here is a repeat of that post, with JasonB's words as quotations (it may be confusing the way I first did it):

       JasonB writes: "I can't speak for Tom but I have published (a long time ago) in the field of AI and I make a living today creating technology that greatly enhances productivity so I guess I'm one of the "doers" (as opposed to "thinkers") in that area."

       You're a "doer" in the field of how AI affects macroeconomic growth? So...you've had some courses in macroeconomics? Or even an advanced degree, perhaps? You know your way around basic economic models like the Solow-Swan model?

       Estimates vary on the processing power of the human brain, from about 10 petaFLOPS up to about 100 petaFLOPS, but somewhere in the middle seems reasonable.

       The estimates I've see run from 0.5 petaFLOPs (Moravec) to 20 petaFLOPs (Kurzweil). So yes, "somewhere in the middle...towards the lower end" seems reasonable.

       The computing power is there.

       No, the computing power is not there. Using an estimate of say, 20 petaFLOPS for a human brain, and the human population is approximately 7 billion. So that’s 140 billion petaFLOPs. What was the total number of petaFLOPS added by all computers in the world in 2012? How many will be added by computers in 2022? 2032? 2042?

       "Why? You've made that offer a couple of times now and I can't see what possible value there would be in any of our predictions of future GDP growth,..."

       I offered $20 to Tom Curtis to predict world per-capita economic growth in each decade of the 21st century. I offered you $20 to predict world per-capita GDP in 2050 and 2100, but also many other parameters for 2050 and 2100: 1) world life expectancy at birth, 2) global CO2 emissions, 3) atmospheric CO2 concentration, 4) global surface temperature anomaly (NASA GISTEMP), 5) global malaria deaths, 6) Beijing annual average PM10 concentration (that’s particulate matter less than 10 microns), 7) percent of world population without access to proper sanitation, 8) global sea level rise (relative to 2013), 9) percentage of species in the 2010 IUCN “Red List” listed as “vulnerable,” “endangered,” “critically endangered,” and “extinct in the wild” that will be declared extinct, and 10) the cost to desalinate 1000 gallons of seawater.

       The value of this information is a simple matter of science. Science consists of one person predicting “a”, and another person predicting “b” (each with explanations for why they are making their predictions) and then observing to see which predictions are correct (or it could be “c” and neither is correct). Or both persons could agree on “a” or “b” in which case there’s not a point in further debate.

       Additionally, in the case of my asking for your predictions for all those parameters (not just GDP) is to see whether or not we agree regarding some parameters that I think reflect "quality of life" (e.g. life expectancy at birth, number of malaria deaths, people with access to adequate sanitation, etc.).

45698. Mark Bahner at 02:36 AM on 8 May 2013
       Global Warming: Not Reversible, But Stoppable
       

       JasonB writes: "I can't speak for Tom but I have published (a long time ago) in the field of AI and I make a living today creating technology that greatly enhances productivity so I guess I'm one of the "doers" (as opposed to "thinkers") in that area."

       You're a "doer" in the field of how AI affects macroeconomic growth? So...you've had some courses in macroeconomics? Or even an advanced degree, perhaps? You know your way around basic economic models like the Solow-Swan model?

       Estimates vary on the processing power of the human brain, from about 10 petaFLOPS up to about 100 petaFLOPS, but somewhere in the middle seems reasonable.

       The estimates I've see run from 0.5 petaFLOPs (Moravec) to 20 petaFLOPs (Kurzweil). So yes, "somewhere in the middle...towards the lower end" seems reasonable.

       The computing power is there.

       No, the computing power is not there. Using an estimate of say, 20 petaFLOPS for a human brain, and the human population is approximately 7 billion. So that’s 140 billion petaFLOPs. What was the total number of petaFLOPS added by all computers in the world in 2012? How many will be added by computers in 2022? 2032? 2042?

       "Why? You've made that offer a couple of times now and I can't see what possible value there would be in any of our predictions of future GDP growth,..."

       I offered $20 to Tom Curtis to predict world per-capita economic growth in each decade of the 21st century. I offered you $20 to predict world per-capita GDP in 2050 and 2100, but also many other parameters for 2050 and 2100: 1) world life expectancy at birth, 2) global CO2 emissions, 3) atmospheric CO2 concentration, 4) global surface temperature anomaly (NASA GISTEMP), 5) global malaria deaths, 6) Beijing annual average PM10 concentration (that’s particulate matter less than 10 microns), 7) percent of world population without access to proper sanitation, 8) global sea level rise (relative to 2013), 9) percentage of species in the 2010 IUCN “Red List” listed as “vulnerable,” “endangered,” “critically endangered,” and “extinct in the wild” that will be declared extinct, and 10) the cost to desalinate 1000 gallons of seawater.

       The value of this information is a simple matter of science. Science consists of one person predicting “a”, and another person predicting “b” (each with explanations for why they are making their predictions) and then observing to see which predictions are correct (or it could be “c” and neither is correct). Or both persons could agree on “a” or “b” in which case there’s not a point in further debate.

       Additionally, in the case of my asking for your predictions for all those parameters (not just GDP) is to see whether or not we agree regarding some parameters that I think reflect "quality of life" (e.g. life expectancy at birth, number of malaria deaths, people with access to adequate sanitation, etc.).

        

45699. Djon at 02:32 AM on 8 May 2013
       Distinguishing Between Short-Term Variability and Long-Term Trends
       

       Dana - I suspect John had http://wattsupwiththat.com/2012/08/13/when-will-it-start-cooling/ in mind, though it's nowhere near as dramatic a drop as McLean was foolish enough to predict and it looks to me as though Archibald left himself a fair bit of wiggle room to claim, at least for a few more years, that the failure of cooling to start when he said it was likely to doesn't invalidate his longer term prediction of major cooling over the course of the next few solar cycles.

45700. Rob Honeycutt at 00:54 AM on 8 May 2013
       Global Warming: Not Reversible, But Stoppable
       

       Mark seems to not be aware of the purpose of economic forecasting.  You, literally, can not predict the future.  The point of forecasting is to establish potential outcomes in order to mitigate risk.  No one is going to take the best case scenario and base their plans on that.  

       The thing about forecasting is, no matter what, you will be wrong.  You can create boundary conditions and base plans on that, but you can't build a rosy scenario where all problems are magically solved and expect anything like that is going to happen.

Prev  906  907  908  909  910  911  912  913  914  915  916  917  918  919  920  921  Next