RealClimate logo

Sea level in the Arctic

Filed under: — gavin @ 20 June 2006

A recent conference presentation at AGU (reported here) while confirming that global sea level is indeed rising (in line with other estimates), showed that Arctic sea levels may actually be falling. On the face of it these preliminary results are a little puzzling (though note that this isn’t yet a properly peer reviewed paper, and so may not reflect what ends up in the journal), but it does reveal some of the complexities in analysing sea level in relatively small enclosed basins and so a brief overview of the different factors involved is probably useful.

Firstly, where is this new data from? The ERS-2 satellite has a ‘radar altimeter’ similar to that flown on the TOPEX/ POSEIDON (and now JASON) satellites and it works by beaming down a radar pulse and seeing how long it takes to come back. These techniques are remarkably accurate (down to a few cm’s and can give accuracy in trends to a few tenths of mm per year averaged over large areas). However the Arctic and Antarctic present special problems due to the poorer coverage near the pole (nothing above 82ºN for instance) and, of course, the presence of sea ice. The ‘sea level’ being talked about is that of the water in between the ice floes, and so the data must be very carefully sorted to yield only the open water values (i.e. where there is no floating ice). This clearly removes the majority of the data and must make any Arctic-wide estimates much more uncertain than the global numbers outside of the polar regions). There is also the possibilty that the analysis has still not removed all sources of contamination in the trend – there may be a seasonal bias (since open water has increased more in summer than winter) or the thinning of the ice cover means that any residual ice contamination could make it look incorrectly like ‘sea’ levels were falling. These issues will obviously be considered in greater depth in any actual publication.

But what if the numbers are roughly correct? What would that mean?

Well, sea level in any part of the ocean is a function of the density of the water (which is affected by its salt content and temperature), but also the local currents, pressure and wind patterns. In this analysis, one of the confounding elements in many assessments; the movement of the land due to the recovery from the last ice age or from subsurface geological/volcanic activity – is not a factor since the measurements are taken with respect to the geoid, not any nearby land stations. The Arctic appears to have freshened in recent decades which on its own would lead to a very slight increase in sea level. This is true whether the freshwater comes from land (in which case it is just adding to the total water in the ocean), or whether it is due to sea ice melting (which contrary to popular belief, does – very slightly – add to sea level rise due to the relative freshness of sea ice compared to the (denser) salty ocean water it displaces). Arctic sea water as a whole does not appear to be warming much (since most of the available heat goes into melting ice, rather than raising temperatures), but it is conceivable that changes to the thermal and saline profiles in depth (as a function of changes in, for instance, the Atlantic inflow) might cause some adjustment of sea level. However the bigger effect is most likely to be dynamic, and related to how much water the winds ‘pile up’ in various places. One example of this effect is associated with the dominant westerlies in the mid-latitude regions. There ‘Ekman drift’ causes water to pile up in the sub-tropics rather than the sub-polar gyres leading to about a meter difference in sea level across the Gulf Stream/North Atlantic Drift (higher to the south). Thus a long term increase in the westerlies (as has been seen in recent decades) could cause an increase in that effect, actually lowering the Arctic sea level. Other changes in wind patterns could conceivably have similar effects. Additionally, changes in sea level pressure can have a direct ‘inverse barometer’ effect on the sea level (a 1 mb sustained increase would lead to about a 1 cm decrease in sea level).

All of these mechanisms can (and no doubt will) be assessed for their possible contribution to this result and it may hold up. It is however extremely unlikely to change any results on the global scale – which as I noted above are completely in line with the other satellite and tide gauge estimates.

52 Responses to “Sea level in the Arctic”

  1. 1
    JohnLopresti says:

    It seems to me there is a further way to utilize the precision satellite sealevel measurment: as one dimension in a volumetric computation describing the total quantity of water in the oceans on the globe. As you have described, for measuring sealevel in a small geographical area, a complex array of local and global factors come into play; the same is true of depicting local weather. Yet, there are merits to attempting to acquire a reliable picture which is widely encompassing. Or, is the sealevel measurement sought because it is the most telltale and sensitive indicator that vast increases in total all ocean liquid phase volume are occurring?–in other words, is there an existing computer model for which the sealevel datum is only one input?

  2. 2
    Leonard Evens says:

    In sci.environment, someone—possibly Thomas Palm—has maintained that sea level could be affected by gravitational attraction from nearby masses, including glaciers. He said this was a little known effect going back to the 19th century, and would explain sea level drop in the arctic. Can you comment on this?

    [Response: It would probably be a noticable effect should the GIS completely disappear, but for the small changes so far (certainly much less than 1% of the total mass), I doubt any effect would be detectable. There are a couple of papers by Mitrovic and Clarke where they did similar calculations in order to estimate sea level fingerprints of various ice sheet changes, so they might be a good place to start to quantify the issue. – gavin]

  3. 3
    Coby says:

    In anticipation of the nattering nabobs seizing apon this, I have written a preemptive entry in the How to Talk to a Climate Sceptic Guide:

    It won’t be long…

  4. 4
    S Molnar says:

    On an ever so slightly tangential topic, the article on beach erosion caused by rising sea levels by Cornelia Dean in today’s New York Times (referenced by another poster on another thread) mentions the IPCC prediction for the amount of global rise in the next century, and adds that there is some evidence the prediction is too low. She then goes on to interview Robert G, Dean, a coastal engineer at the University of Florida, who says he expects the rise to be about the same this century as last rather than the IPCC prediction, and suggests that those who predict a rise of a meter or more are engaging in scientific fraud to prolong their funding. Does the reporter ask Dean to provide a basis for his beliefs? No. Does she point out to the reader that he is not an expert in the field? No. Does she ask how it is that someone could think that publishing blatantly fabricated results is more likely to result in continued funding than publishing honest results? No. Does she ask a climatologist who believes the rate will be that high to justify his belief or respond to Dean’s accusation? No. Did I cancel my New York Times subscription some time ago because of this sort of reporting? Yes.

    (BTW, Gavin, “it’s” is a contraction; “its” is possessive.)

  5. 5
    cwmagee says:

    Does isostatic rebound on the Canadian and European shields have any effect, or are these measurements from the ocean basin? To what extent is the reference geoid dependent on mantle flow anyway?

  6. 6
    Lee Disbury says:

    In reference to Dr. Dean’s interview by N.Y.Times correspondent Cornelia Dean. Is she related? It has been allgeded Dr.Dean has been known to take a more poltical rather than scientific outlook on a number of issues.He appears to favor off shore dredging and filling of beaches in lieu of any type of progressive erosion mitigation technolgy. I reference his insistence of the removal of The Town of Palm Beach’s first attempt at erosion control using P.E.P. barrier reefs, without any substanstitive evidense of the failure of the product. His apparent association with the dredge and fill industry may have clouded his judgement concering this issue.I do not seek to impugn his motives,quailifications or character, however he does seem quick to assume that others, with whom he disagrees, are finacially motivated. “…suggests that those who predict a rise of a meter or more are engaging in scientific fraud to prolong their funding.”

  7. 7
    Lee Disbury says:

    Well that will teach me to actually check facts before posting. I thought it was a direct quote of Dr. Dean that I had included in my last post. The actual quote is as follows:
    “Even Dr. Dean believes that a rise of three feet or more will mean trouble.’I don’t believe that’s going to happen,’ he said, pointing out that people who study sea level rise ‘may have somewhat of a vested interest in keeping the project alive.’ But if it did happen, he conceded, ‘that would not be good at all.'”
    (Speaking of a sea level rise of 3-4 feet at Florida’s coast.)

  8. 8
    Russell Robinson says:

    Anybody have a comment on this? In the full article they suggest that
    glacier retreat similar to now probably happened ~1930 as well.

    GEOPHYSICAL RESEARCH LETTERS, VOL. 33, L11707, doi:10.1029/2006GL026510, 2006

    Greenland warming of 1920-1930 and 1995-2005

    Petr Chylek and others

    Los Alamos National Laboratory, Space and Remote Sensing Sciences, Los Alamos, New Mexico, USA


    We provide an analysis of Greenland temperature records to compare the current (1995-2005) warming period with the previous (1920-1930) Greenland warming. We find that the current Greenland warming is not unprecedented in recent Greenland history. Temperature increases in the two warming periods are of a similar magnitude, however, the rate of warming in 1920-1930 was about 50% higher than that in 1995-2005.

  9. 9
    Steve Sadlov says:

    It’s a tough problem. The Earth itself is not constant in shape. Tectonics play a role. As mentioned there are pile ups due to winds and ocean currents. So much we still don’t fully understand. Still so much to learn.

  10. 10
    Mark A. York says:

    In his takedown of Gore, Ronald Bailey in Reason cites IPCC page here: for a mean rise of 19 inches. He claims Gore is hyping sea level rise, just as Crichton downplayed it. Nothing will flood if this is the case, but I think is potentially more dire than that. This is the correct prediction?

  11. 11
    D.D. Trent says:

    Kevin, Michael et al:

    Have you seen this one from Canada? It deserves a rebuttal by you Real Climate folks who really know what they’re talking about.

    D. Trent

  12. 12
    Lee says:

    re. 10:
    A sea level rise of 19″ WILL cause flooding. Among things locally to where I live, it will put the entire California Delta, some of the richest farm land in the world and the key transmission center of the California water system, at serious risk. It would also flood salt marsh up to urbanized boudaries (nearly everything 1’5 feet above high water is already built on), and likely remove almost the entirety of the remainder of the SF Bay salt marsh ecosystems.

    Just off the top of my head.

  13. 13
    Zeke Hausfather says:

    Re. 10:
    As Lee said in #11, 19 inches is pretty significant. Also bear in mind that the 19 inch figure is the central (median?) value of the 35 evaluated SRES scenarios. In a world where fossil fuel use is higher (say, A1F or A2), the predicted sea level rise is closer to a meter.

  14. 14
    Thomas Palm says:

    Leonard, I mentioned melting of glaciers as a counterintuitive factor that *could* explain a local drop in sea level, but I made it clear that without knowing anything more about the result than a very vague statement in sci.environment I was in no position to tell if this was the case here. As is described here in RealClimate there any many possible reasons for local sea level fluctuations.

  15. 15
    Felix Landerer says:

    Since the issue was brought up by various comments: regional sea level changes can be related to the mass balance of polar ice sheets and mountain glaciers by the so-called approach of ‘fingerprinting ice sheet fluxes’. These models look at regional sea level changes in a gravitationally consistent way (includig self-gravitaion in the surface load). When an ice sheet melts, it will leave a distinct geographical fingerprint pattern in sea level change. Qualitatively, as far as sea level fall in the Arctic region is concerned, shrinking mountain glaciers or a shrinking Greenland ice sheet would be consistent with a fall of sea level in parts of the Arctic Ocean (and a rise everywhere else). One would need to check if the reported fall is consistent with changes in ice sheet volume during this period, or if other mechanisms (as mentioned in Gavin’s post) are responsible. Clearly, ocean warming is highly non-uniform, and decadal variability of ocean heat uptake is very significant.

    The sea level fingerprint concept is described in detail in:

    1. Mitrovica et al., 2001, Recent mass balance of polar ice sheets inferred from patterns of global sea-level change, Nature, 409

    2. Tamisiea et al., 2003, Long wavelength sea level and solid surface perturbations driven by polar ice mass variations: fingerprinting Greenland and Antarctic ice sheet flux, Space Science Reviews, 108

    (you need a subscription for these articles; however, you can find a graph illustrating the distinct fingerprints here.

    I hope this helps a bit.

  16. 16
    Leonard Evens says:

    I sent email to Cornelia Dean at the New York Times saying the following.

    “First let me congratulate you on your excellent article about the effect of sea level rise on beaches. But I do have one objection. In the interests of balance, you quoted Robert Dean, whom you describe as a prominent coastal engineer. He doubts that sea level will rise as much as IPCC predictions, and at one point you quote him as attributing such predictions to a vested interest in keeping the project alive. I hope you are aware that the arguments for sea level rise are based on solid scientific arguments. If average temperature increase as predicted, glaciers will continue to melt, thermal expansion will occur, etc. So the physics is not that complicated. If Mr. Dean questions the IPCC predictions, and he understands basic physics, it can only be that he questions the predictions about temperature rise. He is of course free to do that, but he has no particular qualifications to do so. His background in coastal engineering gives him no more expertise in climatology than you or I, so his opinion on that matter is not relevant. On the other hand, his opinions on the effects of a given level of sea level rise on coasts may be very relevant.

    As a science reporter, it is your job not to present some sort of abstract journalistic balance, but to report the science as accurately as you can. In this case, the reliable source of expertise is the IPCC, not Mr. Dean.”

  17. 17
    jhm says:

    FYI: Wikipedia has what looks like a great resource for all things GW (like anyone should be surprised).

  18. 18
    pat neuman says:

    re: 16

    I think a reliable source of expertise on sea level rise is Jim Hansen.

    A preview from the July 13, 2006 issue
    Volume 53, Number 12 · July 13, 2006
    The Threat to the Planet
    By Jim Hansen

    “The business-as-usual scenario, with five degrees Fahrenheit global warming and ten degrees Fahrenheit at the ice sheets, certainly would cause the disintegration of ice sheets. The only question is when the collapse of these sheets would begin. The business-as-usual scenario, which could lead to an eventual sea level rise of eighty feet, with twenty feet or more per century, could produce global chaos, leaving fewer resources with which to mitigate the change in climate.” …

  19. 19


    You mention the little known fact that global sea level will rise as a result of the melting of sea ice. The calculations for this effect were performed by Robert (Bob) Grumbine and are included in his FAQ on sea level rise, well worth reading :-)

    The relevant section reads:

    “Sea ice cannot change sea level much. That it can do so at all is because sea ice is not made of quite the same material as the ocean. Sea ice is much fresher than sea water (5 parts per thousand instead of about 35). When the ice melts (pretend for the moment that it does so instantly and retains its shape), the resultant melt water is still slightly less dense than the original sea water. So the meltwater still ‘stands’ a little higher than the local sea level. The amount of extra height depends on the salinity difference between ice and ocean, and corresponds to about 2% of the thickness of the original ice floe. For 30 million square kilometers of ice (global maximum extent) and average thickness of 2 meters (the Arctic ice is about 3 meters, the Antarctic is about 1), the corresponding change in global sea level would be 2 (meters) * 0.02 (salinity effect) * 0.10 (fraction of ocean covered by ice), or 4 mm.”

    However, that is a calculation of the global effect. Combining the law of conservation of mass, and the assumption of a constant oceanic density, then the global increase in sea level must lead to a local decrease in some regions, which will be in the areas where the sea ice has melted.

    [Response: Constant ocean density? Where did that come from? This effect raises sea level everywhere precisely because the density of the ocean becomes less on average. -gavin]

    Using Bob’s model, we can calculate how much the sea level will drop in the Arctic Ocean as the ice there melts. When the Arctic sea ice has completely melted the sea level there will be unchanged but the top 3 meters will be covered with fresh water. When that fresh water (salinity = 5/1000) disperses through the Fram Straight, it will be replaced by N. Atlantic salt water (salinity = 35/1000.) This additional weight will depress the Arctic sea level surface.

    Rothrock et al. reported in 1999 that for the years 1993, 1996 and 1997 the trend in mean draft of the Arctic sea ice for all regions was decreasing by 0.1 m. See Table 2 in If we apply Bob’s salinity effect to Rothrock’s change in draft, we get a drop in Arctic sea level of 0.02 * 0.1 m yr^-1 = 2 mm per year.

    The BBC report you cite states that:

    “Dr Scharroo and colleagues have now established seasonal and yearly sea-level trends in the Arctic (from 60 to 82 degrees latitude) for the period 1995 to 2003. The analysis reveals an average 2.17mm fall per annum.”

    If the agreement of these two figures is not just a coincidence then it is rather worrying :-( Rothrock also reported that that the mean draft for all regions in 1997 was 1.6m. If the draft has continued to decrease at 0.1 m/yr, which is what the decrease in sea level implies, then the mean draft will now be only 0.7m, and the ice is hardly sustainable. Estimates of the rate of melting of the ice pack have been based on its extent, but that has remained fairly stable [IMHO misleadingly] especially in winter.

    CEP Brooks pointed out as long ago as the 1920’s that the Arctic ice pack is self sustaining. It seems that now the extent of the multiyear ice is insufficient for the winter ice to grow to its full extent. Further reduction in the ice thickness will lead to a further decrease in the extent of the multiyear ice, and its disappearance may lead to that of the winter ice as well.

  20. 20

    I really would like to know how the radar altimetry from the satellites achieves an (instantaneous?) sea-level measurement with an accuracy of a couple of cm’s. Is the sea-level the mean between height and low of the waves, and is this mean measurable by radar? How is the problem of changing wave amplitudes during the seasons and weather conditions solved? Can we take a trend accuracy of some tenth of mm derived from the much coarser measurements as granted? I read some years ago (but forgot the reference..) that TOPS could not do measurements more precise than about 1 m, and that all the fine results had to be extracted by mathematical work; may be this is not true anymore…

  21. 21
    Fernando Magyar says:

    Re 19
    As an ex commercial deep sea diver and scuba instructor I recall learning about thermoclines, haloclines and pycnoclines. The only constant with regards the density of ocean water seems to be the variation. I also vaguely recall reading somewhere that if it were not for the compression of salt water with increasing depth the surface of the oceans would be quite a few meters higher than they currently are.

  22. 22

    Total integrated density of the entire Arctic Ocean may be one other reason, must disagree with shallow sea surface densities as the air immediately above has been much warmer. This leaves us with the known THC slowing down by 30% measurements, in effect this should cool down the entire Arctic Ocean.

  23. 23

    ….. Water temperature at maximum density varies with pressure, the higher the presssure the lower the maximum density temperature. However I have not read of any significant cooling of the Arctic Ocean subsurface, it may be the other way around, whichever approaches the maximum density temperature, on any given Ocean depth vs temperature profile, either an increase in sea water temerature or a decrease in sea water temperature can do it. Its a matter of finding which one.

  24. 24

    Let me try to explain again why the sea level will fall in the Arctic if the ice there melts.

    Following Grumbine we can argue “When the ice melts (pretend for the moment that it does so instantly and retains its shape), the resultant melt water is still slightly less dense than the original sea water. So the meltwater still ‘stands’ a little higher than the local sea level. The amount of extra height depends on the salinity difference between ice and ocean, and corresponds to about 2% of the thickness of the original ice floe.”

    When you remove that 2% thickness and spread it over the oceans the global sea level will rise. However you are left with a column of fresh water below where the 2% has been extracted. When that is replaced by denser saline water of the same weight, due to Archimedes Law, then its height will be less, hence the sea level there will drop.

    There is an alternative way of viewing this, which I tried to describe using conservation of mass and a fixed density. What I was trying to say is that (in Bob’s model) after the sea ice has melted the volume of the ocean does not change. It remains with its mix of fixed volumes of fresh and saline water. Thus when the fresh water which was in the Arctic is spread throughout all the oceans, and replaced with denser saline water, the surface of the Arctic will fall.

    Am I the only person who can see this, or is there someone who can explain it more clearly?

    [Response: I’m still confused at why you don’t appear to get Bob’s point. If you imagine the ice melting instantaneously it will fill the hole left, but because it’s less dense there will be a little extra which will stand up above the pre-existing level. Thus locally sea level will rise. At the speed of surface graviy waves, the local sea level bump will spread globally giving a (smaller) net rise everywhere. There is no source of ‘extra’ saline water and there is nowhere where sea level will fall. The volume of the ocean will not be constant, it will be greater. -gavin]

  25. 25
    Wolfgang Flamme says:

    “It is however extremely unlikely to change any results on the global scale – which as I noted above are completely in line with the other satellite and tide gauge estimates.”

    This of course is to be continued, isn’t it?

  26. 26
    Steve Sadlov says:

    #20 – I had not thought about that. You do appear to have a point, based on what I know about radar. I cannot imagine the frequency components necessary to discriminate on the order of centimeters not getting attenuated rather quickly. Certainly, over a distance of several miles, they might attenuate so much as to be non measurable.

  27. 27
    daniel says:

    I am hoping that someone may be able to direct me to more information in regards to the lag time of a potential 5m global sea level rise, if say the Ross Sea Ice Shelf where to suddenly collapse as fast of Larsen B.

  28. 28


    I fully understand and accept Bob’s argument. My point is what happens next?

    There is a column of fresh water in the Artic where the ice used to be. It will be less dense than the surrounding sea water, and so will float up and out over the surrounding ocean, and then be dispersed via gravity waves. When that dispersal is complete, the column that was filled with fresh water is now filled with saline water. As a result the density of the Arctic ocean has now increased. A more dense ocean means a shallower ocean. QED.

  29. 29
    John Hunter says:

    > In this analysis, one of the confounding elements in many assessments;
    > the movement of the land due to the recovery from the last ice age or
    > from subsurface geological/volcanic activity – is not a factor since
    > the measurements are taken with respect to the geoid, not any nearby
    > land stations.

    This is a fallacy that is, I’m afraid, still being widely promulgated. I just hope that it has been fully expunged from the IPCC AR4.

    The “anthropogenic” effect on sea-level rise is best estimated by making real observations (a) (e.g. by a tide gauge or a satellite altimeter) and comparing these with the values that these observations would have taken without anthropogenic influences (b). We generally use models of glacial isostatic adjustment (GIA), driven by “natural” land-ice changes (i.e. excluding those that we may consider anthropogenic, such as the recent melting of mountain glaciers), to estimate (b). Contrary to common belief, and to the statement at the head of this posting, (b) is significantly non-zero for satellite altimeter measurements. In fact IT IS OF THE SAME ORDER AS THE (b) APPLICABLE TO TIDE GAUGE DATA — the GIA adjustment made to estimates of present global-average sea-level rise is about 0.3 mm/year FOR BOTH TIDE GAUGE DATA AND ALTIMETER DATA (although the details of the processes involved are of course different). The point is that GIA does not only show itself simply by tide gauges being moved up and down by the Earth’s crust at the coast, but also by an overall warping of the ocean basins, changes in gravitational attraction of the displaced water (in both solid and liquid phases), resultant changes in the angular rotation of the Earth …… and probably a few factors which I have omitted because I am not an expert in these things!

    The lesson is: altimeter observations are, on average, affected just as much by GIA as are tide gauge observations.

    [Response: Maybe I wasn’t clear, but I was not referring to the overall euststatic change due to changes in the shape of ocean basins, but the differences between relative sea level change (due to local effects) and the global sea level change. The GIA effect is still a change in global sea level – though not an anthropogenic one of course – and so is measured by all techniques. It is only in the attribution calculation that it needs to be removed. Altimeter measurements are not however affected by local uplift corrections that need to be applied to tide gauge measurements, for instance in Scandanavia. – gavin]

  30. 30
    llewelly says:

    John, thank you for that comment. I was wondering why it was that measurements w.r.t. the geoid would not be affected by GIA – and now I understand that they are.

  31. 31
    Felix Landerer says:

    RE 24 and RE 19

    Alastair, where is your source for the increased salinity in the fresh water block from the melted ice? For this 1st order kind of calculation, it is reasonable to assume that the ‘block’ of fresh water (the now liquid ice) mixes completely horizontally with the rest of the ocean layer (if you want after that extra little bump has spread over the world ocean at gravity wave speed). You then decrease salinity (and hence raise sea level) in the previously ice-free regions (by a very small amount), and you do increase salinity in the previously ice occupied region, where that rigid block of fresh water was sitting. BUT: you cannot state that this decreases sea level in the now ice free region. This statement implies that you define sea level in the ice covered region as the water column height as is (to the base of the ice) + the liquid phase equivalent of the floating ice. However, one would generally define sea level right next to the floating ice (like the altimeter does). Or how would you define sea level in ice covered regions? It appears to me that this is the crucial point in this discussion.

    Felix Landerer

  32. 32
    Wacki says:

    I really love this site but do you think you could add a community forum where we can create threads? I get into debates all the time about links such as this one:

    Not only is it inappropriate for me to post questions about such articles here but it’s surely too taxing for you to answer every single question one of us has. This is especially true when many of the questions have probably been asked before. I’m sure many of your readers can lessen the burden via community discussion.

  33. 33
    Brian Gordon says:


    Initial posts by Coby, followed by discussion.

  34. 34
    John Hunter says:

    Gavin (your response to #29): Thanks for that (I appreciate that you probably have more important things to think about right now!) — I find these kind of discussions very helpful and this one has forced me to delve a bit deeper into the subject. However, I do not agree with your assertion that “altimeter measurements are not however affected by local uplift corrections that need to be applied to tide gauge measurements”, which presumably suggests that the GIA correction for satellite observations is approximately spatially uniform and only represents a “global sea level change” (i.e. a global-average change). Unfortunately there is not much in the literature to help here. However, Douglas and Peltier (2002), Physics Today, Vol. 55, No. 3, 35-40, provide a map of the GIA correction for altimeter data (their Figure 6 c) which shows a range of about -0.3 to +0.8 mm/year over the oceans (I have changed the sign so that this represents a CORRECTION to be added to the altimeter observation). The only other information I have is from a paper that is in press by other authors, which shows rather smaller corrections and a very different spatial pattern over the globe. However, it seems that the spatial variations of the GIA correction to satellite altimeter data ARE significant (of order tenths of a mm/year).

    Incidentally, I would also take issue with another statement in your original posting that “the measurements are taken with respect to the geoid, not any nearby land stations”. As far as I understand, satellite altimeter measurements are relative to the centre of the Earth, and not to the “geoid” (whatever that is — I could also take issue with the often casual use of the term “geoid”, but I won’t do it here!).

  35. 35
    Coby says:

    Here is an interesting notion: if this result holds, there will need to be a bit of SLR attribution shuffling. Some of the observed rise elsewhere would be coming from a falling in the arctic.

  36. 36
    John says:

    Well if the sea level is rising then why the Artic level is falling?That’s really strange. The two quotes are actually really worrying!!

  37. 37
    C. W. Magee says:

    Re #34
    The geoid is the equipotential surface of the Earth; the mathematical surface where the gravitational attraction is constant. It is close to, but not the same as, the reference ellipsoid, which is the theoretical ellipsoid that says where the equipotential surface should be, based on the density and rotation of the planet. I don’t know why they are different, but the biggest difference is a -80 meter “hole” in the Indian Ocean south of Sri Lanka.

  38. 38

    Re 37 It is the ellipsoid not the geoid which is mathematical. In other words the ellipsoid is theoretical and the geoid is the actual graviational sea level surface. See for a bref explanation and for a picture of the difference.

  39. 39
    mark schneeweiss says:

    hiya all,

    been lovin ur site 4 about 6 months now. when our grand children ask us what we did when global warming/heating first became obvious, the moderators/authors of this site will be safe. the rest of us may well be shot for crimes against gaia!

    thank u gavin, raypierre, coby et al, on behalf of the yet unborn.

    I’ve been waitin for a good few weeks now (i appreciate u all have important work to be gettin on with), for some one to pick up on the 3 papers in nature on artic paleoclimate, using data from seabed sampling.

    any chance of a response?

    keep up the desperately important good work

    [Response: Thanks for the compliments, and sorry for the delay on the arctic papers. I’m in charge in writing those up, but my day job is getting in the way. Hopefully soon! –raypierre]

  40. 40
    John Hunter says:

    C. W. Magee (#37): A geoid is a surface of constant gravitational potential — it is NOT the “surface where the gravitational attraction is constant” — the “gravitational attraction” presumably means the gravitational field strength (commonly labelled “g”), which is the GRADIENT of this potential — “g” is NOT constant on a given geopotential surface.

    Sorry, I don’t think I’m being pedantic here — the literature contains much confusion concerning the meaning of the term “geoid” — a confusion which in turn confuses those who are new to the subject. It isn’t helped by those who do GIA modelling using the term “geoid” when what they really mean is “mean sea level under conditions of prescribed ocean volume and no baroclinic or surface forcing”.

  41. 41
    Jim Crabtree says:


    Your response for # 24: If I remember correctly from a graduate level course in physical oceanography back in 1971, when sea ice forms, the salt molecules are pushed back into the surrounding water making the surrounding water more dense. When the sea ice melts, then the salt water is added back to the fresh water resulting from the melting process. This would result in a zero change in the sea level, would it not? Now if one considers other actions on the sea ice (sublimation and snowfall) then this will change the equation. But by how much? Where does most of the moisture come from for the snow in the Arctic? Sublimation of sea ice and evaporation of salt water? Unless the moisture comes from further south, it will still result in a balanced result. If I remember correctly, the Arctic does not get a lot of snow (until recently when more water has been exposed), correct?

    I have been visiting RealClimate on a regular basis for over six months. I read my first paper on global warming back about 1968. I have an MS in Marine Science and we talked about global warming in two courses: chemical oceanography and physical oceanography. Since then I have been keeping up on the science of global warming. Your site has helped to answer some question that I have had. Thanks much and keep up the good work.

    BTW: I have been running the climate modelling program from the BBC on two workstations and am now up to 2018. Results are becoming scary.


    [Response: Your point is correct but not relevant. The issue here is not the seasonal cycle of melting and freezing – which as you note has no net impact on sea level, but a sustained melting of sea ice which will have an effect over longer term scales. – gavin]

  42. 42
    pat neuman says:

    re 41.

    Until the most recent decades, annual losses to evaporation from the Upper Great Lakes were largest in fall and early winter, before freeze-up, when water temperatures were warmer than air temperatures. Now that the Upper Great Lakes don’t freeze-up anymore, the level of Lake Michigan-Huron is near the 100 year historical record low level. Regarding the Arctic, an increase in the amount of open water would mean greater evaporation over the combined water/sea ice area than in past decades and less local contribution to the local Arctic sea level.

  43. 43
    John L. McCormick says:

    Pat, your #42 comment is the most plausible Arctic sea level drop I’ve read thus far. And, increased evaporation would likely contribute to increased snow fall in the interior of Greenland. I believe that is being measured lately.

  44. 44

    Re #31 – Thank you Felix for your constructive criticism. I think you are correct to point out that we have to be concerned about which sea level is actually being measured here. There is a great temptation to believe it is that of the geoid, in other words global sea level, but the drop is a local not a global effect. Thus although one would expect areas of sea surface near the ice edge to follow global trends, the results show that the sea there is following a local trend. That is opposite to what is happening globally. Moreover, the measurements were made of the sea level in the leads within the ice pack, a local measurement.

    The most likely candidate to cause a local fall in sea level would be a change in atmospheric pressure, but according to the BBC report “The data is also corrected to take account of ocean tides, wave heights, air pressure, and atmospheric effects that might bias the signal.” Thus it cannot be a trend of higher atmospheric pressures that is causing the sea level fall.

    One way of looking at the problem is to consider a beaker of water, and a test tube containing a few ball bearings which allow the test tube to float upright in the beaker. The water in the beaker will be at a certain level. If we now remove a few of the ball bearings, the test tube will float higher out of the water, but the level of the water in the beaker will fall because less water is being displaced.

    If we now think of the Arctic Ocean as the beaker, then the Arctic ice is the ball bearings in the test tube. When some of the ice, as Bob Grumbine explained, flows away to raise the sea level in the global ocean, it has a similar effect as removing ball bearings from the test tube, and the Arctic sea level falls. The argument against this idea is that the Arctic is part of the global oceans, and is not a beaker of water enclosed around the edges. However, we know that the Arctic is not responding to global effects, and is nearly an enclosed sea. The measurements were made of sea level in leads. Won’t the sea level in a lead behave just like the water in the beaker, falling as the ice surrounding it rises when the weight of the ice is reduced, due to melting at its base?

    In effect I am saying the same as Coby who wrote in #35 “Here is an interesting notion: if this result holds, there will need to be a bit of SLR attribution shuffling. Some of the observed rise elsewhere would be coming from a falling in the arctic.”

    Confirmation that this explanation is correct lies in the fact that the decrease in height of the sea level in the Arctic corresponds with the calculation by Robert Grumbine of the mass of water lost :-) My excitement at discovering that fact is my excuse for having proposed this idea before I had fully thought it out. OTOH, I am sure that facts, even with this fuller explanation, will not prevent sceptics from rejecting it out of hand :-(

    It is tempting to extend this idea to the Great Lakes, where the water is indeed in an enclosed space and water level is falling in an iceless regime. But in that case the ice is not like a test tube or like the Arctic sea-ice where there is a net loss of mass from the system. Hence, for the Great Lakes, the explanation of greater evaporation from an ice-less surface seems more likely to be correct. However, loss by evaporation is unlikely to be the reason for sea levle drop in hte Arctic, since it is not yet iceless.

    This should set another set of alarm bell ringing. If an ice-less surface can produce more evaporation, then it is also producing more water vapour. Water vapour is the main greenhouse gas. Thus with the loss of ice cover we have a strong positive feedback situation, and the possibility of a runaway warming. When Arctic becomes iceless, then rapid climate change seems very likely!

  45. 45
    bill says:

    may I have a glass of water with ice, and please add the ice first.

  46. 46
    S K McIntee says:

    To the question to arctic sea level drop- I have a general principle question for the experts from a non-scientist.

    If the earth’s ice, overall, was melting and the volume of liquid water increasing around the earth and the earth is spinning (as it usually does). I have this image of increasing volume around the girth (equator) with liguid flowing from the poles to the equator leaving lower sea level at the poles. By this same reasoning, the earth, with increase weight at the equator, might even slow down from the extra water weight. (I guess this is why earth is a “she.”)

  47. 47
    satmd says:

    The results aren’t puzzling me at all.
    In short: the additional water will concentrate around the equator because of the centrifugal power of our rotating earth.

    [Response: Nope. The mass distribution of water vapour is much more closely tied to the temperature than the latitude. -gavin]

  48. 48
    Hank Roberts says:

    Distinguish water vapor from liquid water — and the tides more than Earth’s rotation move the water around, on a line between Earth and Moon that varies over the year (note how high the Moon is in the N. Hemisphere winter sky, and how low in summer sky).

  49. 49
    Steven Mann says:

    Here’s a thought. Bryden et al. (2005) have described the slowing of thermohaline circulation at the 25 degree North latitude by about 30% from 1957 to 2004. What effects on Arctic sea surface elevations would a weakened thermohaline circulation have? Would you not expect a lessening of frictional drag primarily on the East Greenland Current, and to a lessor degree, the Labrador an East Spitsbergen Currents, if the Irminger and Norwegian Atlantic Currents, whose force is largely derived from the North Atlantic Current, the northernmost extension of the upper limb of the thermohaline circulation, were weakened. Could this play a role in the decrease in sea surface elevations in the Arctic?

  50. 50
    missy says:

    what are the great lakes at in sea level