Fergus Brown has a post which asks the reader to estimate the odds of certain climate-related changes by the year 2040, about 33 years from now. I thought it would be interesting to take a stab at answering his questions.
The questions tend to come in groups of three, which basically amount to “multiple choice” options for outcomes related to various aspects of climate change. But the “survey” also asks the reader to assign odds. So, I’ll treat them in groups, and try to assign odds to the various options. Here’s the first set: what are the odds?
1. The global average surface temperature will be the same in thirty years time as it is today?
2. The temperature will have risen by 0.3-0.5C in 2040?
3. The temperature will have risen by more than 0.5C by 2040?
This set is about future temperature rise. The sustained global temperature rise since 1975 is about 0.018 +/- 0.004 deg.C/year. At that rate, the projected temperature rise over the next 33 years will be 0.594 deg.C — pretty close to 0.6 deg.C. Brown’s options don’t include all possibilities! If the temperature rises by less than 0.3 deg.C, but not zero, or if the temperature actually falls, those possibilities are not included in his list. So I’ll estimate these odds:
- No statistically significant change: less than 1%
- less than 0.3 deg.C: 12%
- 0.3 to 0.5 deg.C: 23%
- More than 0.5 deg.C: 65%
The next set of questions concerns whether or not CO2 is the main driving force behind warming: what are the odds?
4. Something other than CO2 will be found to be the main driving force behind warming?
5. CO2 will be found to be less important than is currently estimated?
6. The relationship between all the forcings and feedbacks in the climate will be better understood and the role of CO2 as principal verified?
The warming effect of CO2 is undeniable: that’s basic radiation physics. Concentrations of CO2, methane, nitrous oxide, ozone, etc. are well measured. Their temperature effect on the climate is already observed. The effects of solar changes, volcanic eruptions, etc. are also well studied. So, I think this is a well-understood phenomenon. I’ll estimate these odds:
- Something other than CO2: less than 1%
- CO2 less important: 10%
- Better understood, CO2 is the principal forcing: 90%
Now come questions about sea level: what are the odds?
7. Sea levels will rise 5-10cm on average by 2040?
8. Sea level will rise 10-20cm by 2040?
9. Sea level will rise 25cm or more by 2040?
I’ve been investigating this lately, in preparation for a post on the topic. Sea level rise averaged over the 20th century is about 1.8 mm/yr. For the last decade, it’s about 3.1 mm/yr. Perhaps the biggest worry of actual climate scientists is the threat of the mechanical (rather than thermal) breakup of ice sheets; they may start to genuinely disintegrate rather than steadily melting away. Even if they don’t, temperatures are likely to go up, so the rate of melting of glaciers and ice sheets is likely to accelerate. Therefore I’ll estimate these odds:
- Less than 10 cm: 10%
- 10-20 cm: 50%
- 20cm or more: 40%
I changed the first option from “5-10cm” to “Less than 10cm,” and the last option from “25cm or more” to “20cm or more” so that the options would include all possibilities.
Now come questions about Arctic sea ice: what are the odds?
10. Annual average Arctic sea ice levels will be the same in 2040 as they are today?
11. Annual average Arctic sea ice levels will decline by 5-8% by 2040?
12. Annual average Arctic sea ice levels will decline by 10% or more, by 2040?
Computer simulations show that there’s a pretty good chance that during summer, arctic sea ice will be entirely gone by 2040. Furthermore, recent observations show it declining faster than the computer models simulate. So, I’ll go with these odds:
- The same, or decline by less than 5%: 5%
- Decline by 5-10%: 10%
- Decline by 10% or more: 85%
Next are questions about tropical cyclones: what are the odds?
13. There will be fewer Tropical Cyclones in the next thirty years than there were in the last thirty?
14. There will be more TCs in the next thirty years than there were in the last thirty?
15. There will be no significant change in the number of TCs in the next thirty years?
The evidence is significant, but not overwhelming, that there has been an increase in the strength of tropical cyclones. But there’s no evidence of an increase in their number. The number could increase with the rise of sea surface temperature, but it could also decrease due to increased vertical wind shear. So, this is a very difficult question for which to estimate odds. I’ll take a guess:
- Fewer: 33%
- More: 34%
- No change: 33%
The next set inquires about atmospheric CO2 levels: what are the odds?
16. Atmospheric CO2 levels in 2040 will be less than 425 ppm (current level = 384)?
17. Levels will be between 430-450 ppm?
18. Levels will be greater than 450 ppm?
The current rate of increase of atmospheric CO2 is 1.95 ppmv/yr. Projecting to the year 2040, the levels will be 445-450 ppmv. But this one is tricky, because there are three factors that could change this. One: the rate has been accelerating. Two: we may, as a species, get our act together, and reduce our CO2 emissions enough to reduce the rate of increase. Three: natural feedbacks may kick in. For instance, about half the CO2 we currently emit is absorbed into the ocean and the soil. But we may be near saturating those sinks, and as temperatures rise ocean CO2 solubility decreases, so those sinks may soon become sources. Also, melting of the permafrost can release large reservoirs of CO2 (and methane as well) into the atmosphere. So, I’ll estimate these odds (again changing the choices slightly, so that they cover all possibilities):
- Less than 425 ppmv: 10%
- 425-450 ppmv: 40%
- Greater than 450 ppmv: 50%
Now come questions about the amount of life in the oceans: what are the odds?
19. The world’s oceans will sustain as much or more life in thirty years as they do now?
20. Ocean life will decline by 10-25% by 2040?
21. Ocean life will decline by more than 30% by 2040?
One of the big dangers of CO2 increase is that the ocean is acidifying, and that’s a threat to organisms which rely on calcium carbonate shells for their life cycle. Also, there’s a good bit of research indicating that we’re simply depleting the stock of many marine creatures. So I expect a decline in ocean life. On the other hand, life is pretty tenacious, so I’ll estimate:
- As much or more: 15%
- Decline by up to 25%: 60%
- Decline by more than 25%: 25%
Finally, questions about the extent of forests: what are the odds?
22. There will be as much forest in 2040 as there is today?
23. Forest areas will decrease globally by up to 10%?
24. Forest areas will decrease globally by more than 10%?
We’re definitely deforesting the planet. In particular, we’re ravaging the Amazon rainforest. As if that weren’t bad enough, changes in the hydrological cycle due to climate change threaten to make areas like the Amazon basin inhospitable for the forests they now sustain. So I’ll estimate these odds:
- As much as today: 10%
- Decrease by up to 10%: 30%
- Decrease by more than 10%: 60%
Looking back on my guesstimates, the outlook is rather bleak. But that simply underscores the need to take action now to mitigate global warming. I don’t want a 10% or more loss of forests, a 25% or more loss of ocean life, or a 0.5 deg.C or more temperature increase.
Fergus Brown leaves us with a few more questions, which I’ll leave the reader to ponder:
What do the odds say about what the world is going to look like in 2040? What are the odds that the world will be no worse off in thirty years than it is today? What are the odds that the pattern of change in climate and eco/biosystems will continue or become more severe? What are the chances that the global generations of 2040 will live with an increased risk of loss, suffering, hunger or conflict?
Good questions.
11 responses so far ↓
Andrew Dodds // May 10, 2007 at 1:53 pm |
As far as atmospheric CO2 goes… given the number of coal fired electric plants currently under construction, and the fact that they have an expected lifetime of 50 plus years, I’d put ‘Greater then 450ppm’ as over 70%. There’s also the factor of the transition to more energy intensive forms of oil (i.e. Tar Sands emit a LOT more CO2 per unit oil delivered) and longer distance transport of natural gas; up to 35% can be used just to transport the other 65%.
The only other thing is about the forests – it’s just possible that we end up with greater forest cover, but the problem is we are often replacing old growth forest with tree plantations, which have nothing like the ecological diversity.
George // May 10, 2007 at 3:16 pm |
Thanks for the estimates!
It seems to me that these sorts of basic estimates are seriously lacking from most discussions about what to do about global warming.
The other thing that is lacking are the attendant costs/benefits associated with each potential future, which is a whole other set of questions, of course.
The probabilities, combined with the attendant costs/benefits can be used to calculate an overall “expected cost” — as is done in actuarial analysis — which can then be used to decide what to do, if anything.
But if one has no idea what the probabilities are for the potential scenarios (futures) and no idea about the attendant cost of each potential scenario, it’s difficult (if not impossible) to figure out the “best” approach to take to minimize the risk.
I’m not familiar enough with some of the global warming cost studies (eg, Stern) to know if this kind of expectation analysis is the sort of thing that they have done.
I’m only familiar with the basic conclusion of Stern that it is cheaper to act sooner rather than later — which really seems to be common sense to me, at any rate. Prevention is usually cheaper than crisis management (eg, with diseases).
Of course, there are lots of other things that need to be included in such an expectation analysis:
What are the probabilities and attendant costs (to humans) associated with species loss throughout the world? (not just in the oceans)
What are the probabilities and attendant costs/benefits (to humans) associated with changes in availability of fresh water for drinking and agriculture?
What are the probabilities and attendant costs/benefits (to humans) associated with changes in availability of arable land?
What are the probabilities and attendant costs/benefits (to humans) associated with changes in precipitation?
What are the probabilities and attendant costs/benefits (to humans) associated with loss of traditional ways of life (eg, in the case of the Inuit, for example)?
Also, there remains the basic problem with the mainstream economic approach that is geared simply toward costs to humans and that leaves out other species entirely (and not simply because we don’t know how species might benefit us now or in the future — though we don’t in most cases).
How do you put a value on being able to watch and listen to birds in your back yard? On being able to see other animals in the wild that may disappear if their habitat does? Or, on simply knowing that creatures and plants exist outside zoos and arboretums?
In some regards, the expectation analysis helps us to decide what to do, but with regard to the latter issue, it actually may cloud our thinking.
I believe that’s why people like Al Gore have framed this as a moral issue rather than strictly a cost/benefit one.
That leads some to dismiss Gore as an emotional windbag (or worse). But let’s face it, there are issues here that are not purely (or even primarily) objective. In fact, the whole issue of what to do is at its core a subjective one — though it can and should be (again subjective) informed by objective reality (as opposed to the virtual kind, which seems to be so prevalent these days).
Those who try to argue otherwise — and criticize people like Gore and environmentalists for emotionalism — are simply not being honest.
Brian Schmidt // May 10, 2007 at 3:39 pm |
Left a comment on Fergus’ website too, but the odds for temperatures depend on whether you’re discussing annual average or a smoothed average.
[Response: I was thinking of the smoothed average.]
fergusbrown // May 10, 2007 at 4:16 pm |
Thanks for the namecheck. I’ll alter my numbers appropriately, to make them a bit more ’scientific, as you have done.
George: I have one or two problems with your comment. Michael Tobis may have others. My first concern, is that your questions are fundamentally unanswerable because they are highly dependent on decisions made between now and 2040. They also interrelate in ways which make a systematic evaluation extremely difficult, at best.
For example, Roger Pielke Jr has a point when he says that mitigating to avoid hurricane damage is unnecessarily costly; Tamino’s odds explain very simply why this is true, but it is not helpful, because this is neither the aim nor the main benefit of mitigation. Mitigating to do anything by 2040 is probably a sideshow, because its impacts will be further down the line; but this does not mean there is no value in mitigation.
Secondly, the obvious question, is why does this have to be defined in terms of cost/benefit at all? This is the accountant in us trying to justify decisions which may be unpalatable to shareholders. Unless we redefine ‘cost’ and ‘benefit’ to have broader than fiscal terms. One of the things that this might bring into question is how we place value on the future, compared to the present, other than in monetary terms.
George // May 10, 2007 at 5:24 pm |
How about the odds for temp increase due to the other green house gases (methane, CFC’s etc).
From what I understand, CO2 accounts for about 60% of the forcing.
I understand that the non-CO2 contribution to warming has decreased in recent times (largely because of CFC reductions), but what can we expect for the future?
With regard to CFC’s, what impact will China’s growing economy have on future levels?
Also, what about estimates of how much total methane is stored in frozen arctic permafrost, for example, and how much will be released for every degree C rise and how much further temperature rise this might cause? And what about the probability of various scenarios involving such a release?
That seems to have the potential to throw a monkey wrench into the works.
What are the odds? « Old man in a cave // May 10, 2007 at 6:03 pm |
[...] Normally, the Old Man avoids circular blogging, but in this case an exception is worthy: Tamino has made a scientist’s stab at playing the odds game, and there are the comments of others, [...]
George // May 10, 2007 at 7:08 pm |
Fergus said: “My first concern, is that your questions are fundamentally unanswerable because they are highly dependent on decisions made between now and 2040.”
Obviously there are a lot of uncertainties involved, but expectation is definitely one of the better ways of minimizing risk in such cases.
That’s the purpose for basing things on “expectation”. One is able to combine different scenarios (based on different assumptions with different associated probabilities and costs) to come up with an expected cost” — ie, without knowing how the future will play out.
The insurance industry does this all the time to assess risk. They multiply the cost of each possible “future” by its associated probability and sum these to get an “expected cost”.
Expected cost can also include non-monetary costs — human deaths, for example.
If we include only the scenarios for the future in which we do nothing to mitigate, then the expected cost will essentially be expected damages that we will have to pay for in the future. That calculation alone is very valuable because it combines both the cost of the most likely business as usual outcome with the considerably higher cost(s) of the less likely outcomes (eg, multimeter sea level rise). That gives you an idea of what it will cost if you do nothing — which also gives you and idea of how much money it might be worth spending to reduce the future risk.
As to your second point:
“Secondly, the obvious question, is why does this have to be defined in terms of cost/benefit at all? ”
I agree, which I brought up in the the last part of my post
“there remains the basic problem with the mainstream economic approach that is geared simply toward costs to humans … How do you put a value on being able to watch and listen to birds in your back yard?”
Actually, I would advocate a combination approach because the cost/benefit analysis still provides valuable information regardless of what one decides to do.
Heiko Gerhauser // May 10, 2007 at 7:51 pm |
Pretty thorough treatment, so I think I am giving it another go.
Firstly, there’s the question of how to deal with things we can’t predict or reasonably put probabilities on, eg development of a low cost low carbon energy source, or of a catalyst that enables cheap weathering of silicates, or the outbreak of nuclear war, or a major disease killing 90% of humanity etc….
One way to look at this is to compare the time period in question (2007 to 2040) with past time periods. We haven’t had a nuclear war over the last sixty years in spite of nuclear weapons being around, but 33 years is only about half of sixty years. On the other hand, when we’ve been lucky for sixty years, it seems a fair bet, we’ll be lucky for another six.
The IPCC handles this issue by not looking at such unpredictable scenarios, and they don’t assign probabilities to the scenarios they do look at either.
As a first step, I’d therefore exclude unpredictable scenarios, where over time periods of 33, let alone 100, years we can’t put probabilities on their likelihood of occurring.
I therefore assume that we are in the IPCC scenario envelope, and any mitigation action doesn’t take us out of there.
Looking at the report, you’ll find a density probability for the period 2020-2029:
http://ipcc-wg1.ucar.edu/wg1/Report/AR4WG1_Ch10.pdf
That’s on page 808.
And on page 763 they give global mean warming between 2011 and 2030.
Given the IPCC constraints, I’d therefore reguestimate probabilities for 2040 (average of 2035 to 2045) as
less than 5% probability of less than 0.3C, 10% probability of 0.3-0.5C, and 85% probability of more than 0.5C.
But there is also the question of aerosols.
On page 803, there’s a good selection of graphs indicating that aerosols are assumed to move in a similar fashion as GHG’s. They don’t rise as fast, but they show similar behaviour.
And that indicates potential for higher temperatures still, because, in fact, I can imagine scenarios where sulphur falls due to much stricter pollution controls, while CO2 does not. It’s much easier to clean up SO2 than to sequester CO2.
On the next set of 3 questions, I’d summarise it as >90% probability that our understanding will improve, but no fundamental change to our understanding of the role of CO2 compared to what we know now will occur, make that 99%, if fundamental is defined fairly restrictively.
On sea level change, glaciers outside of Greenland and Antarctica are due to run out at some stage, and that’s one reason for a potential decelaration of sea level change. Looking at page 820 of the report I conclude that 10 cm or more are 50% likely and 10 to 20 cm 40% likely and more than 20 cm is 10% likely.
Sea ice cover is dealt with on page 771, and isn’t expressed in percentages. Eyeballing the numbers 10% or more loss by 2040 looks about 50% likely. Complete melting 0% probable.
Looking at page 790 it looks to me like without mitigation >450 ppm is >90% likely by 2040.
The discussion on tropical cyclones in the AR4 has managed to surprise me:
“A synthesis of the model results to date indicates that, for
a future warmer climate, coarse-resolution models show few
consistent changes in tropical cyclones, with results dependent
on the model, although those models do show a consistent
increase in precipitation intensity in future storms. Higherresolution
models that more credibly simulate tropical cyclones
project some consistent increase in peak wind intensities,
but a more consistent projected increase in mean and peak
precipitation intensities in future tropical cyclones. There is
also a less certain possibility of a decrease in the number of
relatively weak tropical cyclones, increased numbers of intense
tropical cyclones and a global decrease in total numbers of
tropical cyclones.”
I think this means that for number of storms, I’d have to put in a slightly higher probability for a decrease than for an increase.
On ocean life, I’d need a definition of the measure, is that by photosynthetic activity?
I don’t think there’s much evidence ocean life will decline by this measure, or that it has declined.
Having read Lomborg’s Sceptical Environmentalists and “rebuttals”, I’d note that global forest cover hasn’t declined, if memory serves right, because afforestation in developed countries has made up for rainforest felling, well, at least as measured by hectares.
fergusbrown // May 10, 2007 at 7:58 pm |
George: on your first post, about which I know something, the numbers for different GHGs vary; methane has been relatively stable for a while, whilst the others are still rising as fast as CO2. I am not sure, but you appear confused about the relative forcings of CFCs and HFCs. The odds of the net emission of non-CO2 GHGs rising in line with CO2 itself are probably reasonably high.
The permafrost question is more problematic; on the one hand, there is the paper from late last year which said that methane release from Siberian peat may be as much as 5 times greater than previously estimated. On the other hand, a new paper – it may be forthcoming in the AGU, in fact – suggests that permafrost damage might be lower than estimated, because of a stabilisation process in the deeper layers. As far as a clathrate release from a sub-oceanic source, this looks about as predictable as a volcanic eruption, with our current level of knowledge. So, you are right that it might throw a spanner into the works, but putting odds on it might (currently) be near impossible.
On your second comment: sorry if you thought I was implying that you weren’t concerned about cost/benefit as an evaluative tool; I did understand you have your doubts and should have been clearer in my response. I would suggest though, that there is a distinction to be made (which I am sure you understand and probably agree with), between cost/benefit and/or risk analysis as a toll to aid decision-making, and the same as a means of defining the parameters under which decisions should be made.
Regards,
George // May 11, 2007 at 4:59 pm |
Heiko noted above: “The IPCC handles this issue by not looking at such unpredictable scenarios, and they don’t assign probabilities to the scenarios they do look at either.”
I would just like to point out that, while it may be true that IPCC has not assigned probabilities to scenarios (which I believe is unfortunate, at least in the case of the more likely scenarios), any attempt to assign probabilities to future warming (as Heiko and others have done above) does effectively assign probabilities to future scenarios (which make certain assumptions about future emissions) — even if implicitly.
More odds, less time « Old man in a cave // May 23, 2007 at 9:05 am |
[...] Science and Weather, Global Warming, climate change When the Old Man asked what the odds were, Tamino very helpfully put a scientific gloss on answers which were no great surprise at the cave. To help [...]