Sea Level Acceleration

Sea level isn’t just rising, it is accelerating. It did so during the 20th century, and has done so even more quite recently. ABC news reported the story, based on just-published research (Nerem et al. 2018), that the latest satellite data now show it plainly. The authors of the new study conclude:

When taken with a rate of sea-level rise of 2.9 ± 0.4 mm/y (epoch 2005.0), the extrapolation of the quadratic gives 654 ± 119 mm of sea-level rise by 2100 relative to 2005, which is similar to the processed-based model projections of sea level for representative concentration pathways 8.5 in the IPCC Fifth Assessment Report. Stated alternatively, the observed acceleration will more than double the amount of sea-level rise by 2100 compared with the current rate of sea-level rise continuing unchanged.

They did more than confirm acceleration of sea level rise; they estimated how much of it is because of man-made climate change. To do so with precision, you have to account for the changes due to other factors, in particular terrestrial water storage (TWS), and volcanic eruptions.

Variations of TWS are largely due to changes in precipitation patterns brought about by the el Niño southern oscillation (ENSO) and the Pacific Decadal Oscillation (PDO). Therefore they estimated, and removed, the influence of ENSO/PDO as well as the major volcanic eruption of Mt. Pinatubo. Without those adjustments, the estimated acceleration is 0.096 mm/y2, but when they are accounted for it reduces to 0.084 mm/y2. They also included in their estimate of the uncertainty changes in decadal variations of the cryosphere, the influence of variations in precipitable water on the satellite data, and the rate of thermosteric (due to temperature) sea level rise. The net result is that the uncertainty in the acceleration is about 0.025 mm/y2. This is about the same as my own estimate of the uncertainty based on the raw data alone, but that’s because my estimate is higher due to modelling the noise as an ARMA(1,1) process rather than an AR(1) process as did the authors.

Here’s the latest from the sea level site at the University of Colorado:

I’ve already removed the seasonal cycle. I also added a smooth fit (modified lowess) to highlight the overall pattern. The acceleration is evident. Here’s my estimate of the rate of sea level rise during the satellite era:

The average rate (from satellite data) is 3.1 mm/y, but the present rate is closer to 4.8 mm/y. That’s a substantial increase — a 50% increase.

My estimates are based on the raw data, and do not remove the estimated influence of ENSO/PDO or the Mt. Pinatubo volcano. That’s why the rate seems to “level off” at the end, a behavior which is due to the 2015/2016 el Niño. With that influence removed, sea level rise is still accelerating.

Another possible source of uncertainty in their estimated rate of sea level rise is that they use a quadratic function to estimate its changes. It’s clear to me that the pattern is more complicated; a quadratic (fit to the raw data) estimates the rate of sea level rise now at 4.3 mm/y, but a more realistic fit gives a higher rate.

Extrapolating the quadratic to the end of the century is fraught with uncertainty and shouldn’t be taken as a realistic forecast, but it does provide a reasonable lower bound on this century’s imminent sea level rise. That’s because, as the authors say,

If sea level begins changing more rapidly, for example due to rapid changes in ice sheet dynamics, then this simple extrapolation will likely represent a conservative lower bound on future sea-level change. In contrast, few potential processes exist to suggest that this estimate is too high.

There are many sources for sea level rise, but the likely dominant causes of the recent acceleration are increased melt from the Greenland and Antarctic ice sheets:

The consequences for humanity, especially the billions of people who live in coastal regions, will be profound. The cost to society, both economic and otherwise, will be astounding. We need to do everything we can to limit future global warming, especially to avoid those “rapid changes in ice sheet dynamics” that could make what is already a terrible situation seem like the “good old days.”

While real scientists, doing their best to give us the most precise estimates possible, are warning us of the danger, climate deniers continue to spread fear, uncertainty, and doubt based on bullshit. Witness a recent post at WUWT in which Philip Lloyd actually expresses doubt that present sea level rise “is partly due to heating of the oceans and their consequent expansion.” He tries to show that the thermal expansion of the ocean is negligible, by characterizing global sea level rise with a single tide gauge record (from Wismar, Germany), comparing straight-line fits for some time spans (1910-1950 and 1980-2015) to the long-term straight line fit, then declaring “… both of which are barely distinguishable from the long-term regression shown by the black line.” It’s only “barely distinguishable” to climate deniers and Stevie Wonder.

Continued denial of sea level rise acceleration can cost more than money and property, it can cost lives. When storms do come, when hurricanes strike, flooding can be deadly, and every inch of sea level rise makes whatever flooding happens much worse, makes the death toll higher. Denial makes us less prepared to reduce the future impact and less prepared to deal with what comes.

That’s why climate-denier politicians in the U.S. like EPA head Scott Pruitt and president Donald “grab ’em by the pussy” Trump have got to go. To all Americans: get out and vote! Vote them into oblivion, and all the climate-denier politicians (almost entirely Republicans) who either conspire with their denial, or tolerate it. The only thing necessary for the triumph of evil is for good men to do nothing.

This blog is made possible by readers like you; join others by donating at My Wee Dragon.


20 responses to “Sea Level Acceleration

  1. After seeing their graph of residuals, I guess the 2011-2012 La Nina isn’t really still as pronounced after removing ENSO effects as it seems from the GMSL graph itself. (I wonder if the blue in Figure 2 is really the green in Figure 1 minus the quadratic, or no?) Their method for removing the effect is much more complex than standard multivariable OLS, I presume they indeed were able to pick up a good portion of the spatial pattern in that rainfall, both in Australia and equatorial S. America.

  2. They are scrambling to every hiding place they can find. One: there’s a 60-year cycle, so soon the rate will go way down; two: they violated this from ReaClimate:


    So remember: don’t fit a quadratic to data that do not resemble a quadratic. Instead, look at the time evolution of the rate of sea level rise. And remember there is something called physics: this time evolution must be expected to have something to do with global temperature. And indeed it does.”

  3. Hey folks, anyone here know why the ice age(s) started and ended, my brother dared challenge the orthodoxy (of CAGW) and I do not have a good answer for him, he says it was not C02. ??

    [Response: Slight changes in the tilt of Earth’s axis, and the shape and orientation of its orbit, redistributed how incoming sunlight is distributed around the planet. When more of the sun’s heat reaches extreme latititudes (esp. in the northern hemisphere), it increases melting of the great ice sheets; when less, it increases ice accumulation.

    When the great ice sheets begin to disintegrate, feedbacks kick in. One of the most prominent is ice albedo; replacing highly reflective ice with land or open ocean means Earth absorbs more sunlight, which warms the planet as a whole. This kicks in other feedbacks, such as water vapor (warmer air holds more) which is itself a potent greenhouse gas, and the release of CO2 into the atmosphere (mostly from the oceans because warmer water won’t hold as much dissolved CO2).

    That’s why, during ice age cycles, the change in temperature precedes the buildup of CO2. Climate deniers often point to this fact in an effort to discredit global warming science, but it’s a stupid argument — the observed fact that temperature change happened before CO2 increase was predicted before it was observed, by Claude Lorius and James Hansen and others (some of those scientists who are now warning us of the danger of global warming). Although CO2 isn’t the trigger which paces ice age cycles, it is an important feedback which amplifies them.]

  4. One quibble on your explanation of glacial-interglacial CO2 feedback. The excess CO2 is stored in deep waters, and is released as average deep circulation speeds up during interglacials and deep water resurfaces in the Antarctic and is exposed to the atmosphere. Deep water pCO2 is higher than would be in equilibrium with the atmosphere and outgasses.

  5. Quibble — this doesn’t look to be an “ABC news report” (suggesting that it made TV broadcast) but is ‘simply’ a republishing of an AP story (by the excellent Seth Borenstein).

  6. An explanation of how the various natural influences are estimated in order to remove them from the GMSL (and temperature) data might make for an interesting article. I’ve always wondered exactly how it is done. Thanks.

  7. But wait! Professor emeritus J. Curry just devoted an entire post to the claim that the 1900-1990 globally averaged tide gauge sea level rise was less than stated in IPCC reports. Clearly *somebody* is engaging in spin, but who? [30 microseconds of looking at data later] I guess it’s the individual who appears in front of Republican Congressional committees to give friendly ‘expert’ testimony. What a shock.

  8. Just a statistical question. A linear fit needs 2 parameters, while a quadratic fit needs 3. A I recall there is a test to see if the improved fit gained by adding the extra parameter is sufficient to justify it.

    With the sea level data, is the extension from linear to quadratic statistically justified?

    For what its worth, I expect you are right and its more complex than quadratic, and probably faster. I’d just like to be prepared when the skeptics start with their fake reasoning…

    [Response: If the 2-parameter model is a subset of the 3-parameter model (as in this case), you can apply an “F-test” for direct comparison. If not, you can still compare models with different numbers of parameters using an “information criterion.” Best known is probably the Akaike information criterion (AIC), although also well-known is the Bayesian information criterion (BIC).

    And yes, the quadratic term is statistically justified — that’s perhaps the most important result of Nerem et al.]

  9. Could someone elaborate on why eruptions affect SLR rates? Is it just the same as the atmospheric temp affect i.e sulphates/particulates blocking incoming IR, and thus less ocean heating and thus less expansion for a few years? Pinatubo was 1991 so mostly affects the first few years of this series. Right?

  10. Jeff Masters used your graph of rate of sea level rise in a discussion of acceleration of sea level rise.

    Oh My, he revealed Tamino’s real name! Better put on those dark glasses again to foil the deniers.

    [Response: My real name is no secret. And, Jeff Masters had asked permission before using my graphs.]

  11. Raymond Horstman

    Let I start with saying that I have no idea if sealevel rise is accelerating. It is just a bit odd that you can claim accelaration of just 0,084 mm/Y2 with a dataset that gives sealevel rise of 3 mm/y +- 0,4 mm. It seems to low to tell.

    • RH: Can you show us the calculations behind this what “seems” to be true?

      I don’t mean this meanly, but you simply cannot eyeball data and say a calculated analysis is not meaningful. You can, on the other hand, eyeball data and say it IS meaningful. BF Skinner was famous for doing this…reasonably correctly so in his case I might add (though doing some trend stats would have provided even more information).

      [Response: A caution is in order. In some cases, yes it’s possible to confirm “by eye” — but there are also cases where what “looks” undeniable is actually incorrect.]

  12. You might want to take a look at this comment at C Etc.