RealClimate logo

How much will sea level rise?

Filed under: — group @ 4 September 2008 - (Español) (Italian)

… is the question people have been putting a lot of thought into since the IPCC AR4 report came out. We analysed what was in the report quite carefully at the time and pointed out that the allowance for dynamic ice sheet processes was very uncertain, and actually precluded setting a upper limit on what might be expected. The numbers that appeared in some headlines (up to 59 cm by 2100) did not take that uncertainty into account.

In a more recent paper, our own Stefan Rahmstorf used a simple regression model to suggest that sea level rise (SLR) could reach 0.5 to 1.4 meters above 1990 levels by 2100, but this did not consider individual processes like dynamic ice sheet changes, being only based on how global sea level has been linked to global warming over the past 120 years. As Stefan discussed, any non-linear or threshold behavior of ice sheets could lead to sea level rising faster than this estimate. Thus, otherwise quite conservative voices have been stressing the ‘unknown unknown’ nature of this problem and suggesting that, based on paleo-data (for instance), it was really hard to rule out sea level rises measured in feet, and not in inches. (Note too, the SLR is very much a lagging indicator, and will continue for centuries past the time that atmospheric temperatures have stabilised).

The first paper to really try and assess the future limits on dynamic ice sheet loss appeared in Science this week. Pfeffer et al looked at the exit glaciers for Greenland and West Antarctica and made some back of the envelope calculations of how quickly the ice sheets could dynamically drain.

Good news: they rule out more than 2 meters of sea level coming from Greenland alone in the next century. This is however more than anyone has ever suggested and would be comparable to the amount that disappeared at the Eemian (125,000 years ago) (see this post for more on that).

Bad news: they can’t rule out up to 2 meters in total.

In summary, they estimate that including dynamic ice sheet processes gives projected SLR at 2100 somewhere in the 80 cm to 2 meter range, and suggest that 80 cm should be the ‘default’ value. This is remarkable in a number of ways – first, these are the highest estimates of sea level rise by 2100 that has been published in the literature to date, and secondly, while they don’t take into account the full uncertainty in other aspects of sea level rise considered by IPCC, their numbers are significantly higher in any case. And this week the Dutch ‘Delta Commission‘ published its estimate of sea level rise that the Dutch need to plan for (p111): 55 to 110 cm globally and a bit more for Holland, based on a large number of scientists’ input. [Clarifying update: this is meant to be a “high end estimate”.]

Lest readers think this is no big deal, the estimates for the number of people who would be affected by 1 meter of sea level rise is more than 100 million – mainly in Asia. Of some recent relevance is the fact that the storm surge caused by Gustav in New Orleans was within 1 foot of the top of the levees. Another 3 ft caused by global sea level rise would have put a lot more water into the ‘bowl’.

Thus better estimates of sea level rise from ice sheets remain a high priority for the climate community. More sophisticated models and deeper understanding are coming along and hopefully those results will be out soon.

We were going to leave it at that, but we’ve just seen the initial media coverage where this result is being spun as a downgrading of predictions! (exemplified by this Reuters piece, drawing mainly from the U. Colorado press release). This is completely backwards. We stress that no-one (and we mean no-one) has published an informed estimate of more than 2 meters of sea level rise by 2100. Tellingly, the statement in the paper that suggests otherwise has no reference.

There have certainly been incorrect assertions and headlines implying that 20 ft of sea level by 2100 was expected, but they are mostly based on a confusion of a transient rise with the eventual sea level rise which might take hundreds to thousands of years. And before someone gets up to say Al Gore, we’ll point out preemptively that he made no prediction for 2100 or any other timescale. The nearest thing I can find is Jim Hansen who states that “it [is] almost inconceivable that BAU climate change would not yield a sea level change of the order of meters on the century timescale”. But that is neither a specific prediction for 2100, nor necessarily one that is out of line with the Pfeffer et al’s bounds.

Thus, this media reporting stands as a classic example of how scientists get caught up trying to counter supposed myths but end up perpetuating others, and miss an opportunity to actually educate the public. The problem is not that people think that we will get 6 meters of sea level rise this century, it’s that they don’t think there’ll be anything to speak of. Headlines like that in the Reuters piece (or National Geographic) are therefore doing a fundamental disservice to the public understanding of the problem.

Update: Marc Roberts sends along this cartoon illustrating the problem… (click for full size).

386 Responses to “How much will sea level rise?”

  1. 51
    Figen Mekik says:

    You are right, drought is right up there as one of the worst problems associated with AGW. I don’t know. Some RC regulars may know that I am very sensitive about drought around the Med, but sea level rise is a pretty nasty problem too. Maybe our moderators can provide some perspective?

  2. 52
    sidd says:

    Figen Mekik wrote on 5 September 2008 at 2:38 PM:
    “And [2] as they mention in the post, sea level continues to rise for a significant period of time even after temperature rise is stopped or reduced.”

    This cannot be overemphasized. There will be no stable coastline for centuries. We will have to either a)keep building higher levees or b)keep retreating from the coastline for centuries

  3. 53
    gavin says:

    Some people appear to think that we are balming (sic) the media for the misleading headlines on this story. That is certainly not the case since most of the spin originated from the press release (as we stated clearly above). It is unfortunate that these headlines arose because they will mislead, but it is not because of the ‘bad media’ (though many stories did see past that press release). Pielke makes another error in equating thought experiments (as performed by Hansen in his ERL paper) with an informed prediction. Postulating a non-linear doubling process in order to get an order of magnitude estimate is not the same as making a prediction as he makes very clear, and later in that article, he states what he fears may happen (quoted above). Pielke might think that the Pfeffer et al constraints are reduction in the projections, but as we have stressed numerous times, the projections that matter are the ones that come out of the assessment processes, not from the musings of individuals (however their sometime ambiguous pronouncements are interpreted). With many people (linked above), still incorrectly insisting that sea level won’t rise above 10cm in a century, these latest constraints should be sobering. We’ll see.

  4. 54
    Richard C says:

    North and south ice sheets are breaking up. Greenland is a hollow bowl. Has anybody calculated how much sea rise is required to turn Greenland’s interior into a grounded ice sheet?

  5. 55
    Roger Pielke, Jr. says:

    Hi Gavin-

    Now it’s the press release?

    You would be better served by simply correcting your silly claim that “no-one (and we mean no-one) has published an informed estimate of more than 2 meters of sea level rise by 2100”. Anyone can read Hansen and see that he thinks that his estimate was “informed” and likely to be more “better estimate” than Rahmstorf’s approach for 2100.

    I have little confidence in the ability of the scientific community to accurately predict sea level rise — though the constraint approach by Pfeffer et al. helps move beyond the prediction problem. After all the 1990 IPCC grossly overestimated sea level rise to date:

    And in 2007 didn’t really offer any predictions.

    If there is anything to take issue with in the Colorado press release it is the idea that decision makers can optimize a response to sea level rise, they cannot.

    [Response: Roger, my interactions with you are always singularly unproductive and so my patience in pointing things out to you is limited. Having said that, please read the line in the above post (third paragraph from the bottom) that starts “(exemplified by…”. Perhaps you missed that earlier, but in case it still isn’t clear, we were pointing out how the spin in the Reuters piece took a very clear lead from the press release. So it’s not “now” that the press release is an issue it was “then”. Are you even disputing this? And as for IPCC AR4 not making statements about sea level projections, I have no idea where you get that notion (see section 10.6 in WG1). We might be able to agree that they were inadequate or poorly communicated, but not nonexistent. – gavin]

  6. 56
    Alain Goyette says:

    re no 25. I was in Thule AFB in Greenland for 2 weeks around the end of August. In 14 days, we had perhaps 4 days where it did not rain.

    I was speaking with a few local workers at the base, who indicated that the tongue of the glacier visible about 30K west used to be almost at the end of the runway 10 years ago.

    Interestingly enough, we were not able to resupply one arctic location as the runway was too soft for landings (2nd year ever it happened) and had to stop a couple of days at our main location as the top of that runway softened due to the landings and the constant wet snow and drizzle (Alert).

  7. 57
    Roger Pielke, Jr. says:

    Here is another example:

    “Sea-level rise from melting of polar ice sheets is one of the largest potential threats of future climate change. Polar warming by the year 2100 may reach levels similar to those of 130,000 to 127,000 years ago that were associated with sea levels several meters above modern levels; both the Greenland Ice Sheet and portions of the Antarctic Ice Sheet may be vulnerable. . . Corals on tectonically stable coasts from the last interglaciation period (LIG) provided strong evidence that sea level was 4 to >6 m above present levels during a sea-level high stand that likely lasted from 129,000 ± 1000 years ago to at least 118,000 years ago.”

    From the NCAR press release:

    “Ice sheets across both the Arctic and Antarctic could melt more quickly than expected this century, according to two studies that blend computer modeling with paleoclimate records. The studies, led by scientists at the National Center for Atmospheric Research (NCAR) and the University of Arizona, show that Arctic summers by 2100 may be as warm as they were nearly 130,000 years ago, when sea levels eventually rose up to 20 feet (6 meters) higher than today.”

    [Response: You aren’t reading carefully (though both cases could have been much better worded). The polar warming at 2100 was projected to be comparable to Eemian warmth, but the sea level rise was a long term response with an unknown timescale. However, I’m sure it’s comforting to know that you aren’t the only person to get that wrong. – gavin]

  8. 58
    sidd says:

    Richard C. asked at 3:48 pm on 5th Sept 2008:
    “how much sea rise is required to turn Greenland’s interior into a grounded ice sheet?”

    Parts of Greenland’s interior are currently below sea level. There are some pretty deep spots under Petermann , Humboldt, in the NW and Nioghalvfjerdsbrae (how does one pronounce that ?) and Zachariae Istrom in the NE just south of Independence Fjord (where the NEGIS emerges)

    might want to look at and an animation

  9. 59
    David B. Benson says:

    One meter sea level rise in a century will do no good, and potentially some harm, to the near-shore marine life community; shellfish such as clams and limpets are particularly at risk. During the transition for LGM to the Holocene the sea level rose about 120 meters over, say, 12,000 years; an average of a meter per century. During this the near-shore marine community of eastern South America, with a gentle grade to what is now the continental slope, was compeletely (or nearly so) disrupted; it only recovered during the Holocene. In contrast, the Pacific shore of South America, very steep, suffered no ill effects.

    Sandy beaches: sandy beaches are the result of along-shore currents and also wave action bringing sand from production areas to dewposition ares. For example, Maine and further north are production areas for the east coast of North America; the sand is transported further south and eventually to the deposition areas of the Carolina Outer Banks and off-shore thre. (I don’t know what happens further south.)

    This action won’t stop due to a 1–2 meter rise; winter storms will continue to pile up sand onto the current beach. Wherever there is sand now there will continue (mostly) to be sand. Wherever there are rocky shores that state also will (mostly) continue.

    So your grandchildren will be able to go to the beach, but there will be no need to bring a spade and a bucket for clamming.

  10. 60
    Roger Pielke, Jr. says:


    Indeed, you suggest that Hansen made an “ambiguous pronouncement” . . . Overpeck could have been “much better worded” . . .

    But perhaps you could point me to the Real Climate admonishments of these scientists for their imprecision/bad writing when their work was widely misinterpreted? Or perhaps to where these scientists sought to correct the egregious public misinterpretations?

    [Response: So now it’s my fault? Well, we spelled out what it meant at the time. As for corrections, it’s definitely happened – the original headline after the 2006 Science papers in the San Francisco Chronicle had “20 feet by 2100” or similar, the authors immediately contacted the paper and they changed the headline to something better (but not perfect) “OCEANS RISING FAST, NEW STUDIES FIND” and corrected the text from what appeared in the print edition. And RC has often criticised press releases for leading the press astray. I’m very much of the opinion that most of the press does a pretty good job overall given the constraints they work with, but there are fault lines that occasionally get tripped – through a combination of bad framing, or uncontextualised press releases, or naive scientists – that end up (predictably) producing vastly misleading headlines. Maybe you’d like to come to my class at the J-school where I go over these examples and more… – gavin]

  11. 61
    sidd says:

    Richard C. asked at 3:48 pm on 5th Sept 2008:
    “how much sea rise is required to turn Greenland’s interior into a grounded ice sheet?”

    In my last comment I should have mentioned the Pfeffer paper, in Fig 1, shows the marine based outlets as including almost all the usual suspects, not just in the NE and NW.

  12. 62
    Slioch says:

    #23 Mike Tabony asked, “Where does Slioch (#18) think the human race will get the energy and material resources to rebuild the world’s ports?”

    I don’t. I was suggesting that coping with a one metre per century change in sea level continued for several centuries might be even worse than a one-off catastrophic change, since at no point would we be in a position to even begin to rebuild coastal communities and would be in a state of constant flux.

    In other words, it is not just how much sea level rises, but how long it takes which is important. The latter is usually not emphasised.

  13. 63
    Aaron Lewis says:

    Gavin, what value for “total sea level rise in the next 100 years” would you give to an engineer or policy maker that needs to be 99.9999% (1 in a million) certain that his design or policy will protect the public safety for 100 years? Not just for today, but a number we can be certain of for 100 years. If you make a mistake and people die, how many deaths can you live with? I am not being dramatic. You are a climate expert, and we need to set public policy. We need a number.

    As you suggest, it rains on the flanks of the GIS. Where does that water go, how much heat does it carry, and where does it release that heat? It is a foundation engineering problem of the first water. The top of the ice sheet can stay cold, strong , and rigid, but, if the foundation fails, the whole structure fails. The flanks act as compressive buttresses to the ice foundation that supports the central massive. (Think of a concrete building tipping over as its permafrost foundation weakens.) Weaken the flanks of the ice sheet, and the ice under the massive is subject to compressive failure. Worse, it is a basin with its outlets plugged with ice, allowing the formation of subglacial and intraglacial lakes large enough to produce significant hydrostatic pressures. The problem today, is that we have both known unknowns and unknown unknowns in this system.

    Pfeffer, et al’s number is not total, and it has no certainty. To be useful in planning, we need a value for total sea level rise. Without good estimates of uncertainty and defining both ‘known unknowns” and unknown unknowns,” Pfeffer et al’s is useless for engineering or public planning. If our best estimate of sea level rise contains substantial “unknown unknowns” then we might want to adopt a procedure similar to the one used by the US EPA for Human Health Risk Assessments a while back, and apply safety factors.

    They would say:

    “The estimated number is 2 meters, the safety factor is 10, so the maximum plausible sea level rise from Greenland is 20 meters by 2100, take precautions accordingly”.

    They might go on to say:

    “The estimated SLR from permafrost melt is 1 meter with no unknowns, and the estimated sea level rise from Antarctica is 2 meters with very large unknowns so the safety factor is 100. Thus, the total maximum sea level rise for public safety purposes is 221 meters. If you do not like that number, go do the research to eliminate the unknowns and uncertainties so we can remove the safety factors.”

    Ahh! Those were the days when federal agencies had more . . . . funding.

    If our federal agencies, said firmly, “The SLR number is 221 m. DOW, Exxon, BP, US DOE, US DOD, and so forth ; What are your plans for dealing with a 221 m rise in sea levels?” We would have the data to resolve the unknowns real fast, and we would be in a low carbon economy, real fast! True, nobody would be happy. Nobody liked CERCLA either, but it saved a lot of lives.

    Are such safety factors good climate science? No! However, they are good risk management science. They are good engineering and very good public policy. Such policies protected the public health and safety from hazardous waste in the environment with unknown characteristics. Are such numbers alarmist? No, they are the basis of good,conservative engineering.

    [Response: I’ve not consulted directly with planners on this issue, but people I know do this on a regular basis. The most questions come up when some new infrastructure is being planned – a recent example was storm drains in New York City – the planners wanted to know how high above max high tide levels they should build the drains. The factors were that the infrastructure was going to last for 50 or so years, and they didn’t need a guarantee or perfect prediction but just an range that would likely keep them out of hot (cold?) water. I think that the advice was that an allowance for 1 meter additional rise over an above what their storm surge estimates gave. That’s pretty conservative advice (as in erring on the safe side), but it will be much clearer in 2050 (when they are planning for the following 50 years) what the trajectory really is. Advice that no allowance for SLR was needed would have been irresponsible, advice in the mid-range of projections would not have been useful either. The point is that you can give useful advice even in the midst of great uncertainty without having to exaggerate the likely effects. – gavin]

  14. 64
    sidd says:

    Aaron Lewis wrote at 5:20 pm on the 5th of September, 2008:
    “[Greenland] is a basin with its outlets plugged with ice, allowing the formation of subglacial and intraglacial lakes”

    I suspect that some of the outlets might melt open. Is there any evidence yet of substantial subglacial lakes in Greenland? I have heard of several in Antarctica, some of them interconnected, changing sometimes on a daily basis.

  15. 65
    Hank Roberts says:

    sidd, click here for your question, as Googled:

  16. 66
    Hank Roberts says:

    “… Eventually all of Greenland’s floating ice could disintegrate. At that point the ice streams may stop accelerating. Then again, they may not, Steffen says. The weight of Greenland’s ice sheet has forced its bedrock down into a vast basin, much of it below sea level. As the glaciers retreat inland, the ocean may follow, prying them off their bed in a runaway process of collapse….”

  17. 67
    Craig Allen says:

    There is a nifty interactive sea level inundation mapper at the website. Just zoom into your part of the world, select a sea level rise amount (up to 7m) and then strap on a snorkel.

  18. 68


    For about a year i’ve been looking at the material on SLR and 3 months ago I began to summarize SLR for a friend who happens to have a house over the water (S. California). So from a real world perspective he is very interested in potentials. I spoke with a friend at Scripps and he said a cm is a cm regarding wave height, but related the problem of potentially stronger storms in our region as ocean temps rise.

    To my friend, I basically started by saying his question is one of the most difficult to answer scientifically. And I did resort to what you referred to as “male up even scarier monsters” with a high concentration on the uncertainty in that realm (I was careful when discussing the dragons). My objective was to show him the scientific analysis IPCC and Stefans paper and then compare that to Hansen, current SLR observed and paleo record as currently understood re. past SLR and estimated temps, GHG levels, length of time at those levels and hypothesized negative feedbacks such as cloud albedo, even though it seems the climate system managed to get past such effects in the past and of course feedbacks and non linear potential.

    Along the lines of the Arron Lewis question in comment #63, if you feel it is plausible, can you characterize the risk potentials based on the following considerations:

    The PDO is supposedly entering a cool phase, or finishing its warm phase, for eastern pacific, and still we had an unusually strong El Nino event with significant coastal erosion.

    Is there a scientific way to discuss the potential or likely effects such as storm damage rather than SLR, since that is the greater immediate danger as far as I can tell pertaining to ocean temp rise increasing storm strength a as opposed to SLR for the eastern pacific, i.e. specifically the west coast. It’s obvious the gulf and Atlantic regions will get stronger storms but the Pacific has a different tempo and being a larger body, slower thermal inertia.

    So then, just as models indicate stronger hurricanes generated in the Atlantic, Is it then reasonable to expect higher potential for a similar El Nino event to the 98 event, especially considering the likelihood of increased solar forcing when the Schwabe cycle gets back into gear? Was the 98 event just an El Nino riding on top of the global warming wave, or an even more significant event? I’m not sure how to talk about this? Because if that event was more signal than noise, so to speak, then the next one could be even stronger. Or is this too much weather talk and not enough climate?

  19. 69
    Hank Roberts says:

    > where these scientists sought to correct the
    > egregious public misinterpretations?

    I recall the original error was in the Associated Press wire story. By the time I contacted the SF Chronicle science writer Perlman, if I recall correctly, he had already heard about the problem (as Gavin points out, someone was trying to catch the error). He’d told his editors — but they hadn’t done anything. He told them again; they changed the headline. They didn’t change the text. We exchanged more email. He told them again. They finally changed the text. Way too late.

    Wasn’t this AP wire story and error correction talked about on Pielke’s blog at the time? Oddly it’s impossible to find mention of it with Google, for me.

    Anyone know where the Associated Press keeps their errata and corrections list, presuming they have such a thing? Andy Revkin might know.

  20. 70
    GlenFergus says:

    Gavin’s inline at #63:

    Good response. Seems you have learnt a little engineering. Lots more similar understanding may be needed. Hint: Setting levee crest levels is tougher. Overtopping is (literally?) a “tipping point”. It tends to kill people.


  21. 71
    Andrew W says:

    The effects of a 1 or 2 metre SLR over a century is not a big deal compared to the changes we’ve made to coastlines over the last century, much of New Orleans was below sea level when it was built, and Man’s engineering abilities are far greater now than they’ve ever been.

    As for 100 million displaced, how many people have been displaced as a result of human factors over the last century, I’ll bet more than 100 million.

    Not disputing than AGW’s happening, or that the effects of some of the changes may be dire, but can we please be realistic about what the numbers actually mean?

  22. 72
    sidd says:

    I took Mr. Roberts most excellent advice, and discovered these references to subglacial lakes in Greenland:

    this one is rather good i thought:

    and this one is good too

    and some nice models by Schoof, Evatt and many others. For some reason i feel compelled to add this last link

    lastly: are the proceedings of this:

    to be published soon ? in the Journal of Glaciology, perhaps ?

  23. 73
    Aaron Lewis says:

    Gavin inline comment on #63

    For storm drains in NYC, a SLR of 1 meter is a cost effective design criteria.

    Do not get me wrong, I love good storm drains, but there is no law saying NYC storm drains must be designed to withstand 200-year return event storms. Thus, PE liability and bonding issues are minor, and the value of enhanced storm drains is limited by the vulnerability of legacy infrastructure.

    The question stands. What is the value for SLR that you are 99.9999% sure will not be exceeded in the next 100 years?

  24. 74
    Mark Bahner says:

    “We stress that no-one (and we mean no-one) has published an informed estimate of more than 2 meters of sea level rise by 2100.”

    Here’s James Hansen’s presentation on the 20th anniversary of his 1988 testimony to Congress:

    “Debate among scientists is only about how much sea level would rise by a given date. In my opinion, if emissions follow a business-as-usual scenario, sea level rise of at least two meters is likely this century.”

    Or is James Hansen’s estimate not “informed?”

  25. 75
    Ryan T says:

    Maybe RC needs a media contacts address book, so corrections can be e-mailed to science reporters. Maybe they’ll eventually start viewing press releases with a more critical eye, or even wait a day for some third-party analysis. :-)

  26. 76
    Alan Millar says:

    This discussion seems to drift between Artic ice conditions and sea level rise.

    Well global ice is maimly concentrated in the SH not the NH! Not that you would realise that from most of the posts here.

    As far as sea level rise is concerned there is absolutely no evidence of an accelerating trend, which would be necessary, for some of the dire predictions, stated here, to come to pass. Indeed sea levels have been stable for 2 or three years now. how come?


  27. 77
    Ryan T says:

    I just noticed something about the press release. It notes that “Some scientists have theorized that continuing warming trends in Greenland and Antarctica could warm the Earth by 4 degrees F over the present by 2100. The last time that happened, roughly 125,000 years ago during the last interglacial period, glacier changes raised sea level by 12 to 20 feet or more. But the time scale is poorly constrained and may have required millennia, Pfeffer said.”

    I thought 4 degrees wasn’t exactly at the extreme end of the IPCC projection range, and that the IPCC would constitute more than just “some scientists”. Am I mistaken? And do we have a reliable estimate of the rate of temperature change during the LIG, and how it compares to projections for this century?

  28. 78
    Guy says:

    Roger Pielke, Jr in #55 states that the IPCC’s first assessment report “grossly over-estimated” sea level rise to date. I see a link to his own blog, but I cannot check the original report as it is not online (and the Wikipedia summary a rather hopeless stub, which I hope can be improved upon). As a layman, I’d expect to see a range presented, but on Roger’s blog it is a single trajectory for each report (with the first report over-estimating, and the following two underestimating).

    Could someone expand on what was actually in the original 1990 report? Is Roger’s summary fair? Specifically – what is each line on his chart – the average forecast? The maximum? And was there a projection for the century? Obviously the first report was based on much less research than followed even four years later (let alone right now), but just interested in getting a better historical conext for this latest paper following Gavin’s thoughts – thanks.

    [Response: We’ve had this discussion before with Pielke, about the temperature projections of the first IPCC report. Gavin can probably quickly find the link. The key point is that the first report included only the warming effect of greenhouse gases, but not the cooling effects of anthropogenic aerosol pollution. You have to remember that scenarios are not predictions, but “what…if” type calculations. What would happen if greenhouse gases rise by that much? Thus, as a calculation of the warming and sea level rise that these greenhouse gas emissions would produce (by themselves), this may well have been correct. But there was an additional effect that IPCC did not include in the scenarios at the time, namely the aerosol pollution. Hence these scenarios were not realistic, even if technically the estimates were correct. -stefan]

  29. 79
    Peter Ellis says:

    As the glaciers retreat inland, the ocean may follow, prying them off their bed in a runaway process of collapse

    I must confess I’ve never understood this particular prediction of doom. If the ocean is able to “pry them off their beds”, it follows that they must already be sufficiently far below sea level that they would float. And if they’re floating, then the sea level rise from melting them is zero. Surely it’s exactly the other way round – the deeper below sea level the Greenland bedrock is, the better?

  30. 80
    Bruce Tabor says:


    To quote from above, “…no-one has published an informed estimate of more than 2 meters of sea level rise by 2100.”

    Firstly, Hansen is on record as expecting a SLR of more than 2 metres under BAU – see my post #30. I accept this does not appear in a peer reviewed publication.

    Secondly and more importantly, when it comes to the cryosphere, informed estimates have so far proved to be gross underestimates. Prior to 2007 the informed consensus was for artic sea ice to be gone in mid summer between 2050 and 2100. Now some scientists have “informed” expectations that this will happen within 5-10 years and 30 years is a reasonable consensus.

    As you well know, projections of global temperature rise are underpinned by sound physical models that have proven reasonably accurate since Hansen’s 1988 paper (if not earlier). But projections of SLR is not underpinned by sound understanding of the physics of ice sheet collapse.

    Until models can reproduce the gyrations in SL over the last 100,000 years or so – including the 5 metre per century rise over 4-5 centuries during “Meltwater Pulse 1A”, one wonders what value peer-reviewed “informed” estimates have. (Peak rates were probably greater than 0.05m/yr.)

    Until then surely we can really only attempt to “constrain” the range of expected SLR. (Pfeffer’s paper appears to be an attempt to do this – I can’t access it to check.) Using the IPCC lower bound would be a reasonable, although highly unlikely, lower bound SLR. The maximum might best be determined by paleoclimatic-geologic estimates – at least they *really* happened.

    Meltwater Pulse 1A occurred under substantially less forcing than we currently have, but it almost certainly was preceded by millennia of “build-up”. So, on balance, 5 metres by 2100 is a reasonable, perhaps equally-unlikely, upper bound.

    This gives a range of SLR from 0.2 to 5 metres by 2100.

  31. 81

    David B. Benson posts:

    One meter sea level rise in a century will do no good

    Note, also, that one meter of sea level rise is enough to make many, many coastal cities uninhabitable. All the seawater has to do is back up sewers and seep into aquifers, and bingo — you’ve got a city nobody can live in, unless they take special equipment with them.

  32. 82

    Andrew W. posts:

    The effects of a 1 or 2 metre SLR over a century is not a big deal

    It’s enough to make Miami, Jacksonville, and many other coastal communities around the world uninhabitable, along with much of Florida and Bangladesh. See my earlier post here.

  33. 83
    Bruce Tabor says:

    Re 45 Rod B:
    I suspect 3.4mm/yr/degree should be interpreted as follows: for every degree C global mean temperature rise we predict a sea level rise of 3.4 mm per year or 0.34 metres per century.

  34. 84
    Hugh says:

    For those who are interested, the ongoing Thames Estuary 2100 project has involved intensive research into how London’s resilience against flooding can be maintained/improved over the next century (and somewhat beyond). The Consortium have indicated confidence that, through the implementation of a portfolio of measures, it is possible that the greater part of London could be protected against 5.25m of sea level rise; contingent on the development of effective decision-making, financial and construction plans (e.g. Thames Barrier 2.0) and processes within an adequate timescale.

    Paper is not online but full citation is:

    REEDER, T., WICKS, J., LOVELL, L. & TARRANT, O. (2008) Protecting London from Tidal Flooding: Limits to Engineering Adaptation Paper presented at the Tyndall Centre for Climate Change Research Conference “Living with climate change: are there limits to adaptation?” 7th & 8th February 2008. RGS, London.

  35. 85
    E. Somanathan says:

    I recall reading somewhere that it would take decades for sea level
    rise occasioned by ice-sheet melt to reach distant coastlines like
    those of South Asia. Is this true? It seems implausible, but if true,
    it gives valuable time.

    [Response: No. Most of the sea level rise travels as a gravity wave (like a tsunami) which would go around the world in a day or two. There certainly are longer term adjustment processes that work slower (and a paper recently by Stammer discussed this). – gavin]

  36. 86
    Jeffrey Davis says:

    Don’t wait too long to see Venice.

  37. 87
    Ricki says:

    Ref 63, 73 on probabilities and risk factors for use in design.

    Thanks for raising this point. It is critical to the response of policy makers.

    Factors of 10 or 100 appear to be a bit extreme, particularly when dealing with a 50 or 100 year life (as would be the case for infrastructure. Normally, in the world we knew a couple of decades ago where the environmant was basically stable, we would take the bell curve of data such as wind speeds (extreme values of these above some cut-off) and say that for safety we will assume the 95th percentile value is suitable for design. Then a design event would be chosen that would typify this level of probability of the wind (or other event) being exceeded within the life of the structure.

    In practice due to the conservatisms used in the design and construction process the true safety factor for structures has been of the order of 5 or more above this event. In fact the real safety factor is often not known as we don’t have enough data on building collapses to determine it. Buildings often fail not due to the design being too low, but the details of construction being lacking in some way.

    A method has been derived that sets a ‘reliability’ level for the structure and this is used to determine the probability values used for design. There is a lot of mathematics in this method and not relevant here in this blog. However, for the simplest approach, structures are designed for an ‘event’. Drainage might be designed for a 1 in 20 or 1 in 50 year event (based of course on the “stable” climate we all know). Buildings are dsigned for winds of the order of 1 in 500 year event and earthquake is similar (with an additional look over the sholder at the 1 in 2000 year earthquake in high risk areas such as California).

    If we start to look at SLR, we might say that we should take the 95th percentile of the upper extremes of the peer reviewed ranges provided in the last say 2 year period (rounded to a logical value). It would seem appropriate to then make some assessment of the rate of change in such predictions over the last 5 years to see if there was some adjustment to be made. On the basis of this there may be an additional factor put on the 95th percenbtile value as a sort of “uncertainty” factor for safety reasons.

    A review process would be necessary to adjust the “design value” given emerging new data and better predictions. Such adjustment may allow for a reduction in the uncertainty factor while raising the 95th percentile number, thus leading to a best approximation of the likely value as understood at the time it was set.

    From a very limited understanding of the current range of predictions (please correct me if you know of further info), I would say the value should be of the order of 0.5m for 2050 and 1.0 for 2100 with an additional factor of 2 for 2050 and say, 3 for 2100; giving

    1m for the 50 year design
    3m for the 100 year design

    These design values would have to be revised on a 6 mth or yearly basis at the moment, as the information available is changing so rapidly. Personally, the IPCC should be giving us a report every 1 or 2 years, not every 4. Knowledge is changing to quickly in this area.

  38. 88
    Count Iblis says:

    There are a few places on Earth that are below sea level where inland seas have evaporated away, like e.g. Death Valley. If we flood these areas again and recreate the inland seas, we can lower sea levels.

    [Response: Yes. but do the math – to get rid of 1 m of sea level from 70% of the globe is a lot of water. You would have to flood a lot more than Death Valley and the Dead Sea. Something more like Montana/Ontario and Quebec to recreate the paleo-Lake Agassiz for instance. – gavin]

  39. 89
    Guy says:

    #78 inline response – thanks Stefan, that helps enormously. Am I right in thinking, then, that really the IPCC have only made two sets of outright all-encompasing predictions, then – in 1995 and 2001? 1990 was a scenario, and 2007 left out dynamic ice sheet processes completely because they were so uncertain.

    From my reading of Roger’s blog, I guess he’d interpret the “scenairo” 1990 explanation as spin. I’ve no reason to doubt that the 1990 and 1996 processes were quite different, and exactly as you say, although I don’t have access to the reports. I’d hope that those writing reports – especially IPCC ones – would flag up very clearly important distinctions such as this one (I didn’t realise that the critical significance of the different methodology in 2007 until Real Climate pointed it out). And I’d hope that all climate scientists would help get these sorts of messages across to clarify to the public, rather than point score based on comparisions which should not legitimately be made.

  40. 90
    Hank Roberts says:

    Aaron, your question
    > What is the value for SLR that you are 99.9999% sure will not be exceeded in the next 100 years?
    indicates you haven’t taken a statistics course. Look up ‘long tail’ for pop science articles that may be helpful.

    Peter Ellis,
    William Connolley would have fixed “pry them up off their beds” if that had been from Wikipedia (where someone had the glaciers dragging themselves out of their beds).
    What the National Geographic writer was waxing (or waning) poetic about is this: the Greenland ice, like the Antarctic ice, has piled up quite high above ground and above sea level —- look at the elevation of the top of the ice in both cases.
    Yes, ice floats. But if you take a cubic foot block of ice and put it in water an inch deep, it doesn’t float. It rests firmly on the bottom of the container. You can look up “basal temperature” and “subglacial lakes” for more — at the bottom of these huge thick piles of ice there are often voids or layers of meltwater just from the pressure, but the pressure is still enormous. Maybe enough to ‘float’ the ice or maybe enough to force any fluid there to move ‘uphill’ if a route is available. It’s not just sitting there inert.

  41. 91
    jose says:

    when the sea level rises and all the ice melts whats gonna happen to the water?? if global warming was to slow down and cease would their be a refreeze?? what does the moon have to do with this??

  42. 92
    Mark says:

    Odd is’t it.

    “Sea rises may be lower than some predicted”.

    But the corollary is that the sea rise WILL be higher than some predicted.

    Odd is that they word it so as to push “we’re A-OK, people!”

  43. 93
    Jeffrey Davis says:

    Hansen is on record as expecting a SLR of more than 2 metres under BAU – see my post #30. I accept this does not appear in a peer reviewed publication.

    Hansen has, if I remember correctly, acknowledged as much. I think his phrase was that his predictions aren’t matched by any model except the geological record.

  44. 94
    Rod B says:

    Thanks, Martin (50) and Bruce (83), but my question remains. It sounds like Stefan is saying if global surface temperature increases one degree (and then stays there forever) oceans will then rise 3.4mm each and every year hence. Doesn’t sound right. Actually seems impossible. What does he (the units) really mean?

    [Response: See my response above (#45) Stefan]

  45. 95

    jose posts:

    when the sea level rises and all the ice melts whats gonna happen to the water??

    It will become slightly more fresh and less salt. Ocean currents which depend on temperature and salinity may change.

    if global warming was to slow down and cease would their be a refreeze??

    Not for a long time.

    what does the moon have to do with this??

    Nothing that I can think of. The sea level rise is far too small to affect the tides noticeably.

  46. 96
    Andrew W says:

    Paul Barton Levenson, a link for you:

    “It’s enough to make Miami, Jacksonville, and many other coastal communities around the world uninhabitable, along with much of Florida and Bangladesh.”

    Only if nobody does anything for a hundred years.

  47. 97
    Rod B says:

    Stefan, thanks. Makes considerably more sense!

  48. 98

    Rod B #94:

    It sounds like Stefan is saying if global surface temperature increases one degree (and then stays there forever) oceans will then rise 3.4mm each and every year hence.

    Yes it does, doesn’t it. But that’s precisely the misunderstanding you shouldn’t make as I tried to explain. Caveat, I cannot access the article, but am somewhat familiar with the theory.

    In the general (not: 20th century exponential) situation the proportionality that should exist is not between sea level rate and anomalous temperature (your one deg), but between sea level rate and disequilibrium, say, the warming that is still “in the pipeline”.

    If warming stops at +1 deg, then the warming “in the pipeline” will start decreasing at that point, going asymptotically to zero. And so will the sea level rise rate do.

    This is actually compatible with the paleo observation that every forcing situation has not only its own equilibrium mean temperature, but also its own equilibrium continental ice situation and equilibrium sea level. Change the forcng, and all of these will start changing to a new equilibrium, which will not depend on the history of getting there, only on the amount of forcing. (Ideally; Snowball Earth is an exception.)

    I know this is tricky. Try thinking house, furnace, thermostat, temperatures in different rooms. Sea level is but a special proxy of global mean temperature.

  49. 99
    SecularAnimist says:

    Barton Paul Levenson wrote: “… one meter of sea level rise is enough to make many, many coastal cities uninhabitable … enough to make Miami, Jacksonville, and many other coastal communities around the world uninhabitable, along with much of Florida and Bangladesh.”

    I’d note that coastal areas may become economically uninhabitable long before they become, literally, physically uninhabitable. I am thinking for example of the American southeast — Florida and the Gulf coast region — which will be impacted not only by rising sea levels but also by more large, powerful hurricanes. Well before it becomes physically impossible for people to inhabit these places, it can become economically impossible to inhabit them, given the enormous cost of not only holding back or retreating from the gradually rising sea level, but of repeatedly and frequently rebuilding in the aftermath of huge, powerful, massively destructive storms.

  50. 100
    Peter Ellis says:

    Yes, ice floats. But if you take a cubic foot block of ice and put it in water an inch deep, it doesn’t float. It rests firmly on the bottom of the container.

    Well yes, that’s kind of my point. And adding another inch of water doesn’t pry it off the floor either. It doesn’t actually rise off the floor until you’ve added about 9/10 of a foot of water, at which point of course it will begin to float. And at that point, if the block melts, it won’t make the water rise any further.

    To turn the same situation on its head, say you have a nice flat sheet of water, and a 1 foot cube of ice sitting on a table at sea level. Melt the ice, the water level will rise. Now say you have a nice flat sheet of water, and a cuboid 1x1x10 feet of ice floating vertically in it. You have the same 1 foot cube of ice exposed above sea level, but this time when you melt it there is no sea level rise because it’s already floating.

    I can’t see why the same doesn’t apply to Greenland. The more of the ice there that’s already below sea level, because of crustal depression, the better: it means less sea level if the ice does melt out. And yet the fact that the centre of Greenland is below sea level is often raised – on this blog and other sites – as a concerning factor in regard to sea level rise.

    Is it simply that the rate of sea level rise may be higher if the glaciers/ice sheet are grounded below sea level? I can see that in that circumstance, a rise in sea level would increase buoyant forces – and reduce friction between the ice and the bedrock, allowing for faster outflow?

    Captcha fortune cookie: Rickman Valera. An odd couple that might have revolutionised the course of Irish independence. Or maybe not.

    [Response: The issue is different. For glaciers that are grounded below sea level, there is generally a cavity below the glacier tongue and a point at the back of the cavity (the grounding line) where there is a transition between floating and grounded ice. As I understand it (and this could be corrected by someone who actual knows something), the ability of the ice stream to sustain shear changes radically at this point – floating ice (I think) has no shear (velocity differential in height), while grounded ice does (it moves faster at the top). Therefore, as the water warms and the grounding line retreats, the part of the ice that was grounded now starts moving much faster delivering more ice to the ocean. Grounding line stability is a big unknown and for some glaciers, once the old one is released, a new might not form until the glacier has retreated kilometers up the fjord. The issue with WAIS and for a significant chunk of Greenland is that in extremis, it might not stop at all (until all of that ice is spread out and floating). (Please someone correct me if I have this wrong!). – gavin]