February 07, 2007
Clarifying IPCC AR4 Statements on Sea Level Rise
Posted to Author: Pielke Jr., R. | Climate Change | Risk & Uncertainty
The statements in the IPCC’s AR4 SPM released last week on sea level rise have led to some confusion and conflict over what exactly they said and how it compares to the 2001 IPCC TAR. The IPCC could have made it easier for all of us by presenting the data in a comparable manner. This post reflects my efforts to make sense of this situation. I hope that experts on the subject will weigh in on my initial thoughts.
I conclude that the IPCC has indeed lowered its top end estimates of sea level rise over the 21st century relative to 1990, in contrast to the conclusions at RealClimate which suggest that this has in fact not occurred. For details, please read on.
First, what did the IPCC 2001 TAR say about sea level? It reported:
For the complete range of AOGCMs and SRES scenarios and including uncertainties in land-ice changes, permafrost changes and sediment deposition, global average sea level is projected to rise by 0.09 to 0.88 m over 1990 to 2100, with a central value of 0.48 m
Some information was not included:
In addition, Warrick et al. included an allowance for ice-dynamical changes in the WAIS. The range we have given does not include such changes. The contribution of the WAIS is potentially important on the longer term, but it is now widely agreed that major loss of grounded ice from the WAIS and consequent accelerated sea-level rise are very unlikely during the 21st century.
The range we have given also does not take account of uncertainty in modelling of radiative forcing, the carbon cycle, atmospheric chemistry, or storage of water in the terrestrial environment.
This is quite similar to the just-released IPCC AR4 (PDF) which says:
Models used to date do not include uncertainties in climate-carbon cycle feedback nor do they include the full effects of changes in ice sheet flow, because a basis in published literature is lacking.
The IPCC AR4 does apparently incorporate information from Greenland and Antarctica:
The projections include a contribution due to increased ice flow from Greenland and Antarctica at the rates observed for 1993-2003, but these flow rates could increase or decrease in the future.
It suggests that on the increasing side that:
For example, if this contribution were to grow linearly with global average temperature change, the upper ranges of sea level rise for SRES scenarios shown in Table SPM-3 would increase by 0.1 m to 0.2 m. Larger values cannot be excluded, but understanding of these effects is too limited to assess their likelihood or provide a best estimate or an upper bound for sea level rise.
Over at RealClimate they seem to have added to the confusion by asserting incorrectly:
Note that some media have been comparing apples with pears here: they claimed IPCC has reduced its upper sea level limit from 88 to 59 cm, but the former number from the TAR did include this ice dynamics uncertainty, while the latter from the AR4 does not
As documented above the TAR did not include such uncertainties, writing of its Figure 11.12:
Note that this range does not allow for uncertainty relating to ice-dynamical changes in the West Antarctic ice sheet.
I asked RealClimate about this, and they responded:
The TAR range included mass-balance estimates for the Greenland and Antarctic ice sheets (though did not include dynamical changes - i.e. changes due to changes in ice streams, calving, grounding line movement, etc which were then thought to be small). Recent observations point to the vital importance of such terms in assessing the net mass balance, thus since they are highly uncertain, it was thought more prudent to not include the mass-balance terms this time around. Our statement above should probably state that "the former number from the TAR did include some ice-sheet mass balance uncertainty, while the latter from the AR4 does not"
What RealClimate fails to acknowledge is that because the TAR did not consider dynamical uncertainties, then a similar uncertainty range would have to be added on top of the TAR top end estimate to make it apples-to-apples with the top end uncertainty in the AR4. So in effect they cancel out and are not relevant to this discussion.
Presumably when the IPCC AR4 says "a basis in published literature is lacking" it is indeed prudent not to speculate. I would assume that there is also no basis in the published literature to conclude that sea level rise might stop instantaneously next year, so they didn’t include that either;-)
So what then do we get when comparing the two reports? The following figure shows the TAR and AR4 estimates on the same graph, taken from the TAR with the AR4 values superimposed. The AR4 ranges are delineated using the same color scheme as the TAR, but with rounded ends. The AR4 values are for 2090-2099, which I have presented as 2095. There is, as noted above, some error term on the upper end of the range. But it should be applied to both the TAR and AR4 estimates, so for comparative purposes they basically cancel out.
Thus, I conclude that the top end estimate has indeed come down from the TAR to the AR4, and those making this observation are accurately representing the AR4. Why didn't the IPCC just say so?
Posted on February 7, 2007 03:18 PM
I read that differently than you, Roger:
The way I see it, these uncertainties were left out of TAR because they thought they were known to be small, but they were left out of AR4 because they are no longer confident that they are small.
So the reason that they were left out has changed. and the take-home message is that sometimes more research gives you larger preceived error bars, not smaller ones.
So the basic explanation that dynamic effects have moved from the "knowns" to the "known unknowns" catagory seems to be consistent with both your quotes adn the realclimate explanation.
Posted by: Lab Lemming at February 7, 2007 06:34 PM
LL- Thanks much .. but I don't think that the TAR supports this interpretation, it states:
"We emphasise that estimates of the relevant factors are highly uncertain"
Posted by: Roger Pielke, Jr. at February 7, 2007 06:40 PM
Just a layman here, I'm not sure I understood correctly. It seems from your quotes that:
Neither number has the dynamical changes.
So it is left to someone to quantify the effect of these mass balance numbers on the upper bound. (not dynamical changes.)
I hope I'm clarifying and not adding to the confusion. :)
Posted by: mz at February 7, 2007 06:40 PM
Roger, you have not got the story right here. There are a number of reasons why the two sets of numbers are different. Two are purely methodological - the change to 2090-2099 from 2100 for the headline changes, and the reduction to 90% from 95% in the uncertainties. Both of these reduce the headline number but not the substantive change. The physical change is how the ice sheet mass balance is treated. Looking purely at the thermal expansion numbers, the two reports best estimates are within 10% of each other for each scenario.
The uncertainty for the ice sheet change has increased on one side only - people now think there is a greater chance of more rapid melt (due mainly to dynamical effects) than they thought before. Thus the impact of the ice sheet change has not been reported in the same way. The partial inclusion of ice sheet effects in TAR is thus removed, but the potential magnitude of these changes in AR4 is larger. This cannot be described as a reduction in the top range.
However, this is clearly a complicated and confused issue, and so we will try and give a more quantitative analysis in a couple of days.
Minor note, if there is a signed comment on RC, it is not 'RC' talking, it is simply the person that signed it. 'group' posts are discussed among ourselves and can be assigned to the group.
Posted by: Gavin at February 7, 2007 06:48 PM
Thanks Gavin, much appreciated. A few replies:
1. The AR4 SPM says:
"For each scenario, the midpoint of the range in Table SPM-3 is within 10% of the TAR
OK, I'll bite -- when it says "within 10%" is that above or below the TAR model average? Someone obviously knows the answer to this, and I suppose we'll see in May. But I cannot imagine why the IPCC leaves this completely ambiguous -- can you report the actual numbers for us? I am assuming that the AR4 values are below the TAR.
The AR4 SPM suggests that the narrower range is due to "improved information about uncertainties" and doesn't indicate what the quantitative effects of changing the significance level. There is the footnote which says, "The TAR would have had similar ranges to those in Table SPM-2 if it had treated the uncertainties in the same way." Perhaps this is an allusion to the changing confidence level? Why not just tell us the numbers? Why the unclear presentation?
If the IPCC believes that the uncertainty ranges are actually larger than they reported, then why in the world didn't they report them as being larger? It makes no sense to report quantitative numbers indicating a 90% confidence range and then explain that they don't actually mean that.
So yes, I am sure that many people would welcome some clarification. Until then, given what a muddle this is, I do not think it is at all fair to criticize anyone who reads the AR4 SPM and concludes that there has in fact been a decrease in the top range projection of sea level rise. That is absolutely what it looks like to me, but I will await your further quantitative explanation and we'll be happy to prominently post that quantitative information once provided.
On the attribution to RC, we'll try to keep that straight in the future ... Thanks!
Posted by: Roger Pielke, Jr. at February 7, 2007 07:33 PM
Gavin- A further confusion -- the use of the 90% confidence interval isn't exactly clear as none of the sea level rise ranges appear within the square brackets said to denote the 90% range. Thanks!
Posted by: Roger Pielke, Jr. at February 7, 2007 07:36 PM
Roger, re your concluding comment and question:
"Thus, I conclude that the top end estimate has indeed come down from the TAR to the AR4, and those making this observation are accurately representing the AR4. Why didn't the IPCC just say so?"
It couldn't be, could it, that the IPCC team writing the SPM for AR4 didn't want to have any statements where the level of alarm in TAR is being reduced?
Posted by: bruce at February 7, 2007 09:12 PM
Hi Roger. You tell us that RC concludes that IPCC hasn't lowered the top of the SLR range. I'm still unsure about the details, but we actually say "Note that some media have been comparing apples with pears here: they claimed IPCC has reduced its upper sea level limit from 88 to 59 cm, but the former number from the TAR did include this ice dynamics uncertainty, while the latter from the AR4 does not, precisely because this issue is now considered more uncertain and possibly more serious than before." which asserts that some of the media discussion is wrong (and of course some, from eg the WSJ, has been nonsense) but doesn't say what the real answer is. As far as I can tell, IPCC says the range has decreased, so they are probably right.
Posted by: William Connolley at February 8, 2007 03:34 AM
I can see your reference to the TAR which says that ice sheet dynamics were not included for the WAIS, but there is no mention of Greenland. Are we to assume that it is the dynamics of the Greenland glaciers that cause the difference between the TAR and 4AR estimates? In the absence of statements in the TAR saying that Greenland's glaciers were similarly treated to the WAIS I am inclined to believe Gavin. The section of the SPM which discusses this makes no mention of this being a post 21st Century problem, the implication is that this dynamic instability is connected to the 2100 sea level rise estimates. It says that the 0.1m - 0.2m uncertainty could be added on directly to the upper bounds of the estimates.
Posted by: Paul at February 8, 2007 05:31 AM
Thanks William -- a nice clear statement like "As far as I can tell, IPCC says the range has decreased, so they are probably right" would have been useful in the RC discussion.
Paul- I don't think we can answer this precisely until we see the actual numbers used in the IPCC.
Posted by: Roger Pielke, Jr. at February 8, 2007 07:15 AM
An observation on a minor note. Gavin says:
"if there is a signed comment on RC, it is not 'RC' talking, it is simply the person that signed it. 'group' posts are discussed among ourselves and can be assigned to the group."
But William says:
"You tell us that RC concludes that IPCC hasn't lowered the top of the SLR range. I'm still unsure about the details, but we actually say..."
So I remain confused whether RealClimate speaks for their individual poster or for their group or if comments about RealClimate posts, as the ones seen here, can be addressed by an individual or the group.
Posted by: Jeff Norman at February 8, 2007 08:19 AM
Let's face it. The entertaining hair-splitting about how many millimeters global sea level will rise in the next 100 years is another indication that climate alarmists are concerned about the new mood within the IPCC. Despite lobbying, leaks and antechambering, the SPM has thrown out the more extreme scenarios regarding sea level and temperature rise, polar ice melting, hurricane activity, etc. This has come as a huge disappointment to many interested parties. After all, the extreme scenarios have been carefully developed and advanced by the disaster lobby since the notorious Met Office conference (Avoiding Dangerous Climate Change) exactly 2 years ago.
One only has to compare the SPM with the Stern Review to appreciate just how much the IPCC has softened its assessment and estimates of core issues. No wonder many doom merchants are fuming about the new moderate temper. Others are simply denying that any moderation has actually occurred.
To make matters worse, every single government has now signed on to the IPCC consensus. Some bloggers are seething about President Bush’s conversion as this newfangled consensus deprives campaigners of a natural target in the “science wars.” Now, that no government is disagreeing with the basic science, campaigners are forced to engage in the much more complex issues of climate policy and economic analysis.
Nonetheless, I don’t expect that the prophets of doom will surrender that easily and accept the IPCC consensus. In fact, I expect the stream of disaster predictions, catastrophe scenarios and hyped media alarmism to go on as usual – in the hope that - never mind - at least the next IPCC report will be more alarmist!
As a green campaigner recently stressed: “one of the big stories around this Friday's release by the Intergovernmental Panel on Climate Change (IPCC) is the conservative edge to the final product ... That means public awareness needs to be raised about two things: the missing pieces, and future IPCC reports that include models of disintegrating ice sheets…”
Posted by: Benny Peiser at February 8, 2007 08:36 AM
Guess what? I've just arrived at home - and what do I find in the post? Another New Scientist shocker: Their front page headline reads as if it's right out of a conspiracy fruit cake:
"GOOD BYE COOL WORLD: WHY OUR FUTURE WILL BE HOTTER THAN WE'VE BEEN TOLD"
So what's the IPCC secret?
Here is there revelation:
"But here's what they didn't tell us: If the official verdict on climate change seems bad enough, the real story looks far worse...."
No wonder the editorial is headlined: "Consensus is not enough"
Posted by: Benny Peiser at February 8, 2007 10:32 AM
Another non scientist, (but not ignorant!). Reading the IPCC summary for policymakers, (yes, I did read through it, turgid and poorly written though it was), I was struck by the table SPM0 on page 5 of the on-line pdf offering.
Posted by: Tony Edwards at February 8, 2007 11:13 AM
Post script. The 18 page pdf file that I have just checked on at http://ipcc-wg1.ucar.edu/wg1/docs/WG1AR4_SPM_Approved_05Feb.pdf
differs markedly from the one earlier released through Junkscience which, apart from anything else, was 21 pages long and had things in a different order. Even the table numbers are different, the one I referred to before is now tablespm1, the numbers have been changed, but the addition is still odd in the errors column. This does not bode well for the final report, particularly when you realise that the man made part of the total co2 load is only about 1/6th anyway.
Posted by: Tony Edwards at February 8, 2007 11:43 AM
A related comment about the "Summary for Policy Makers". My thought throughout reading the SPM was "a policy maker is supposed to read this!!?". I think the report would more accurately be called "Summary of briefing for climate advisors to give to Policy Makers".
Posted by: Cortlandt at February 8, 2007 11:55 AM
Long story short: It looks like a units conversion error.
Posted by: Sean D at February 8, 2007 12:17 PM
I posted a corrected version here:
I actually think that Table SPM-0 made more intuitive sense in meters per century, but apparently the British per mm/yr.
Posted by: Don Thieme at February 8, 2007 06:19 PM
On another thread Ian Castles raises a point worth thinking about and following up on:
"Questions such as whether or not the Panel reduced the upper end of its range for sea level rise between the Third and Fourth Assessment Reports are of minor importance compared with the key issue of whether or not the higher emissions projections used in both Assessments are plausible."
Posted by: Roger Pielke, Jr. at February 8, 2007 08:27 PM
Posted by: Lab Lemming at February 9, 2007 12:58 AM
Off topic question. How was one chosen to be a member of IPCC? was there an effort made to include varying viewpoints?
Posted by: Tom at February 9, 2007 02:38 PM
New Scientist isn't a science magazine. I got a reply years back to a complaint about content in which the person answering me made exactly that statement. He wrote further: New Scientist is an 'Entertainment' magazine.
They said it, I believe it.
Posted by: hank at February 9, 2007 09:30 PM
Here is another good example why climate alarmists and green campaign journalists are losing credibility fast:
Remember the extensive lobbying and media campaign against the IPCC's handling of the sea-level issue in the run-up to the AR4 release?
"Early and changeable drafts of their upcoming authoritative report on climate change foresee smaller sea level rises than were projected in 2001 in the last report. Many top U.S. scientists reject these rosier numbers.
A few days ago, Science published a new paper that shows that this kind of gorilla-hysteria is misplaced. The paper reports that two of the largest Greenland glaciers have suddenly slowed. At one of them observations revealed “average thinning over the glacier during the summer of 2006 declined to near zero, with some apparent thickening in areas on the main trunk.”
I commend science writer John Tierney for concluding that the current moderation on this issue is more than appropriate:
"But for now, with the glaciers moving in fits and starts, it’s wise not to make any sweeping predictions based on a few measurements. Although the Intergovernmental Panel on Climate Change was criticized for not incorporating the recent scary data from Greenland into its long-range projections, these new results seem to vindicate its caution...."
As yet another climate scare has fallen flat, it is instructive but not surprising that alarmists and their groupies in the green media are simply ignoring an inconvenient truth.
Posted by: Benny Peiser at February 10, 2007 06:22 AM