Sea Level Comment

Reader “TLM” submitted a comment to this post about sea level rise. Let’s look at some of his (her?) statements.

It starts thus:

Not wishing to defend the original post too much, but I think you have slightly missed the point of the WUWT article. He is not saying that you should measure SLR on the basis of 6 years of data, …

“Measure SLR on the basis of 6 years of data” is exactly what he did. He fit a straight line, treated us to a graph showing it, and reported the trend rate. I quote: “Since December 2009, the sea levels have declined in both Washington DC and The Battery NY, -3.3 mm/year in Washington DC and -10.7 mm/year in The Battery NY.” But you tell us “He is not saying that you should measure SLR on the basis of 6 years of data“?

… he is questioning Sallenger’s method of comparing the first and second halves of periods of 60, 50 and 40 years when there is an apparent sinusoidal periodicity of 30 – 60 years, which is a reasonable point to investigate, but not proved in his article (or even addressed properly to be frank!).

To do so he chooses to “measure SLR on the basis of 6 years of data.” He doesn’t establish any purported “sinusoidal periodicity.” He certainly doesn’t test Sallenger’s result with sea level, but with sea level rate which he gets from 6 years’ data.

If he had compared sea level to what’s expected from Sallenger’s method, say, the 2nd half of the 60-year split interval at New York, we’d be looking at this:


The red dotted line is the extrapolation of the 1980-2009 trend, the big blue dot is the average sea level since 2010. Since then, mean sea level at New York (the Battery) has been 12.7 mm higher than what would have been expected based on Sallenger’s 2nd-half-of-60-year-period analysis.

His 6 years of decline is simply looking at what has happened since the end of the data analysed by Sallenger in 2009. Which is, of course, far too short a period to determine anything!

I have to say, looking at the actual data there does not seem to be any noticeable acceleration in the Tide Gauge data nor the satellite data in any record I have seen. …

The most trusted global sea level record based on tide gauges is that from Church & White. The monthly data look like this:


Annual averages, together with a straight-line fit, look like this:


If we subtract the linear fit to compute residuals, we’ll get a much clearer picture of its departures from linear increase at a constant rate. It’s this:


That’s “noticeable acceleration.”

Taking individual tide gauge records from PSMSL (the Permanent Service for Mean Sea Level), you needn’t look any further than the very first one they list (for Brest, France) to see this:


That too is “noticeable acceleration.”

What should we conclude? That you feel qualified to lecture us on sea level rise, but when it comes to the most reliable global data based on tide gauges or the very first tide gauge data set listed by PSMSL, you haven’t seen them?

They gauges all seem to show a pretty monotonic rise – and usually less than the satellite data seems to show. …

Perhaps you meant something other than “monotonic.”

If we apply changepoint analysis to the Church & White data set, it indicates its most recent slope change about 1999:


The slope of the most recent segment is +4.1 mm/yr. Taking the two data sets used by “Bruno,” the rate at New York over the last 50 years is +3.4 mm/yr, that at Washington DC is +3.3 mm/yr. All of which belies your claim of “less than the satellite data seems to show.”

… Some researchers have tried to infer acceleration based on estimations from proxies for pre-gauge data, but splicing proxy data and direct measurements has a rather dodgy history and probably best avoided!

“Dodgy” describes the strategy of dismissing evidence you don’t want with nothing more that the insult “dodgy.”

Read the literature, conclusions of acceleration based on proxy data don’t rely on combining them with recent direct measurements anyway.

I think it is better to rely on the actual measurements we do have, which globally do not show any significant increase in the rate of rise.

Your saying they do not, don’t make it so.

Sallenger chose to concentrate on a “dynamic hotspot” of acceleration in north east USA but that just seems to be cherry picking locations rather than time periods! I don’t think you can prove “dynamic sea level rise acceleration” from just that data, although it might be consistent with dynamic rise acceleration if it is actually occurring.

Have you even read Sallenger et al? They make it clear at the outset (in the abstract even) that they’re not “looking for a hotspot” — they’re seeking to “verify predicted patterns” which are “forced by dynamic processes, arising from circulation and variations in temperature and/or salinity, and by static equilibrium processes, arising from mass redistributions changing gravity and the Earth’s rotation and shape.”

The intimation that they’re “cherry picking locations” is both false, and a cheap shot.

I been running my own analysis of SLR based on the CU Satellite Data and discussed it in a brief email exchange with Dallas Masters. The trend rate of rise is a very consistent 3.4mm per year +/- 0.4mm. Most of the excursions from this are associated with ENSO and the rate is currently a bit higher than the past because of the end point in the recent El Nino. This is, of course, with an isostatic adjustment of 0.3mm per year (making it more akin to a rate of growth in sea level volume rather than height).

The average rate of rise in the apparent sea level (what you would actually see at the coast) is 3.1mm +/- 0.4mm per year. Looking at start and end points in ENSO neutral periods, the trend rate is between 2.8mm and 3.0mm a year. The rate of growth has been higher in the west Pacific and Indian Oceans, where tide gauge data is very poor, which probably explains the difference between satellite data and the tide gauges.

Considering your false statements, your mischaracterization of the WUWT post by “Bruno,” and your apparent unfamiliarity the most basic data sets, perhaps you’ll understand that we’ll seek elsewhere for insight about sea level rise.

This blog is made possible by readers like you; join others by donating at Peaseblossom’s Closet.


45 responses to “Sea Level Comment

  1. Ah, but do you deny that sea level has monotonically decreased from $HIGHTIDE to $LOWTIDE?

    • Already addressed. Somebody from the Swiss Climate Bureau recently said Swiss children will never see a low tide.

      The current ENSO surge is huge. Assuming La Nina conditions emerge later this year, what will the no acceleration crowd say if the data fails drop below the trend line?

  2. Wow, not really sure I deserved all of the vitriol poured into that post, but you are never one to pull your punches so I was expecting a good spanking! I prefer my learning opportunities to come without corporal punishment though. Pity about the selective quotation of my post so missing where I dismiss the WUWT post as of no value, but that would have meant you would have looked like you were beating me for agreeing with you.

    I was not aware of the Church & White series so will look into that. I have only ever seen individual tide gauge records and none going back as far as 1800 so an index is useful to know about, thanks. I note that the first acceleration it shows is around 1935, before anthropogenic CO2 could have had any effect. The 1999 apparent change in rate comes just after the 98 El Niño and presumably is reinforced by the 2010 and 2015/6 events, although it is difficult to tell when the chart ends. Will the trend continue? A mere 17 years in so too early to tell.

    Failing to discuss the CU satellite data by dismissing anything from me as coming from an ignoramous so therefore not worth discussing comes across as a bit of a cop out. All I was doing was presenting their data for your comments.

    • With respect TLM, your comments sound like those of a concern troll. If that is not your intention then some self reflection may be in order.

    • I note that the first acceleration it shows is around 1935, before anthropogenic CO2 could have had any effect.

      Just speculation, but if Minnett is correct, and Held’s belief that ACO2 may larger than acknowledged role in early 20th century warming, then there would be an effect. That it happened sort of points in that direction.

    • Greg Simpson

      “I note that the first acceleration it shows is around 1935, before anthropogenic CO2 could have had any effect.”

      Not true. It is believed that Columbus, and similar people, caused a small carbon dioxide drop by wiping out most of the Indians. Then the industrial revolution (1760-1830) raised it a bit by burning most of the forests of Europe. They were small compared to today, but the changes lasted centuries.

    • Chris O'Neill

      the first acceleration it shows is around 1935, before anthropogenic CO2 could have had any effect

      There was enough CO2 forcing in 1935 (0.14 of a doubling) to produce 0.4 degrees C of global warming eventually, so you are sadly misinformed when you claim 1935 was “before anthropogenic CO2 could have had any effect”.

      But being sadly misinformed doesn’t stop you from making false claims of course.

    • “I note that the first acceleration it shows is around 1935, before anthropogenic CO2 could have had any effect.”

      I don’t know why you would think that.

    • Bernard J.

      I was not aware of the Church & White series so will look into that.

      Eh? That’s like saying that you you can’t be held responsible for not knowing the germane facts because you had a third of your brain pulled from your skull by a hook passed up your nose, and the remaining two thirds sucked out with a straw.

      And if you’re like most of the denialist bent your understanding of the science will be enbalmed by your ideology, so no amount of teaching will see you rise above your wrong-headed defence of the post at WUWT.

  3. TLM,

    Why couldn’t anthropogenic CO2 have had an effect before 1935? The industrial revolution started around 1800.

    • Yes, but only in a very limited way on a tiny island off the European west coast – and then mainly for industrial production. The use of fossil fuels did not really take off big time until after the Second World War. There are plenty of sites out there showing how fossil fuel use and consequent CO2 concentration took off exponentially in 1945. I might find some evidence and post again. Although I am sure you know that already and could cite me several papers on the subject! I think there is a good page at SkS About this.

      There is a bit of a conundrum here that perhaps somebody can explain. An increase in CO2 concentration slows the escape of IR radiation to space causing surface air / sea temperature to rise. This in turn causes expansion of the sea surface and an increase in glacier meltwater runoff. However, in the 1940s, at the same time as the sea increased its rate of rise and CO2 was rapidly increasing, surface temperatures stopped going up and in fact fell slightly between 1940 and 1979. If surface temps were not rising, what was causing sea level rise? My thought is that perhaps surface temperatures may not be as closely linked to SLR as is generally thought? Is the temperature record faulty? Are we measuring temperatures in the right places? I have no idea myself but it does look odd.

      ENSO is driven by SW radiation penetrating the sea surface in the tropics and warming it during a period of low cloud cover (la Nina). IR radiation cannot penetrate the sea surface so direct heating by the sun is the way the sea gets warmer. Of course an enhanced greenhouse effect reduces the sea’s ability to cool at night but that would show up as surface air temperature rise.

      • Chris O'Neill


        only in a very limited way on a tiny island

        I don’t know if you’re trying to be deliberately dense here but as I implied above, the CO2 concentration in 1935 was 309.4 ppm (measured from Law Dome) which causes, as I pointed out above, a forcing of 0.14 of a doubling that would produce about 0.4 degrees C of global warming eventually. We only reached 0.5 of a doubling fairly recently.

        I might find some evidence and post again.

        Don’t bother. You obviously don’t have much interest in the pertinent facts.

      • This is a gimmick argument:

        ENSO is driven by SW radiation penetrating the sea surface in the tropics and warming it during a period of low cloud cover (la Nina). IR radiation cannot penetrate the sea surface so direct heating by the sun is the way the sea gets warmer.

        It is likely additional GHGs almost always cause OHC to go up in all phases of ENSO: La Nina, neutral, and El Nino, so the micron talk is rubbish… straight out of the skeptic playbook.

      • Incorrect, the industrial revolution had already spread across the European continent, and to north America. Fossil fuel use was widespread throughout the European continent in steel production which was one of main concerns before WW1 as Germany caught up and over took British steel production. All that steel being used for all those dreadnoughts and other warships used in WW1 and before.

  4. Further to the ‘early CO2’ issue, two points of interest. One (and as hinted in one or two comments above), there’s a hypothesis out there that emissions from land use changes are detectible on multi-century timescales. The name chiefly associated with that hypothesis is Bill Ruddiman; he had a guest post at Realclimate not too long ago. It is here:

    He sees interest in, and acceptance of, the ‘early CO2’ hypothesis, increasing fairly drastically.

    Two, the first published ‘detection’ of anthropogenic warming in the literature dates from 1938: Guy Callendar’s “The Artificial Production of Carbon Dioxide…”. The prevailing view on that is that Callendar was probably wrong–but he was wrong in such a thoroughly investigated and provocative way that his paper turned out to be enormously fruitful over time. Its synthesis of new data and analysis with an old hypothesis was crucial in reviving the study of CO2’s role in forming climate.

    Better to be ‘wrong’ like that than right in a more mundane fashion!

    • According to this geologist/astronomer, the amount of CO2 and methane released – cumulatively – into the atmosphere by agriculture up until the industrial revolution made about 20 ppm difference to atmospheric concentrations of ghgs, from 260 to 280 ppm. Which was enough to stop temperatures declining in accordance with Milankovitch cycles – then we started burning coal in a big way. Watch this from about 30.10 onwards.

  5. @TLM and other commenters:

    Then, of course, there is the Ruddiman hypothesis that human-induced increases in greenhouse gases began some 8000 years ago as the result of agricultural practices of our early agrarian ancestors. Google “ruddiman hypothesis” for further information and references.

  6. E. Swanson

    TLM, in reply to your question from the earlier blog post, I recall a presentation from a researcher who claimed that his data exhibited an oscillation in sea-level with a period of about 50 years over thousands of years. That was back in 1995, when I was beginning to get serious about climate change and I later wrote him several times about his work. While I was looking for the reference, I came across other similar, more recent, studies searching Google. The idea that sea-level (and other climate variables) are subject to periodic variations of natural causation is not new. For example, the so-called “Pacific Decadal Oscillation” or variations in the THC influencing the North Atlantic sub-polar gyre are possible causes of local variations. Of course, what appears in a local record may not pertain on a global scale, which makes such investigations problematic, while also being favored cherry picking for the denialist camp.

  7. The full paper Sallenger et al (2012) is available here. It is plain that the attempt by the post at Wattsupia to brand Sallenger et al. as “bad science misinterpreting the sea level oscillations by cherry picking the time window” is pure nonsense. And I perhaps side with TLM in that the 6-year-trend analysis is a minor part of that nonsense, which may have been his point.
    What the Wattsupian post does fail to report is the acceleration amongst all the wobbles in the data it introduces. The 30-year averages in the Battery NY data wobble from 1mm/yr to 5mm/yr and also show an underlying acceleration of (difficult with the wobbles) somewhere between +1mm to +2mm increase in annual rise over the century. And as NE US is a hotspot for SLR it is right to analyse what is happening there separate from elsewhere.
    I would take issue with TLM when he says “there does not seem to be any noticeable acceleration in the Tide Gauge data nor the satellite data in any record I have seen” (although he does grudgingly accept there is some acceleration in the Church & White data). As shown by the banana-shaped graph in the post “De-Linearized” SLR, how can there not be an acceleration? OLS yields a values of +1.4mm annual increase/century. Such an acceleration would be bad news for sand castle dwellers, adding 6″ to any underlying SLR over a century. This acceleration is statistically significant but that significance relies on including the last 30-years of data.
    That last 30-years of data in C&W shows a far stronger acceleration, as does the satellite data. And why not? This is the period of strong global temperature rise. The ENSO-induced wobbles do not prevent OLS yielding an acceleration of roughly +14mm annual increase/century for both. With a 3mm/yr rise at kick-off, that would see almost 1m SLR over the century which is a serious rise.

  8. “It is likely additional GHGs almost always cause OHC to go up in all phases of ENSO: La Nina, neutral, and El Nino, so the micron talk is rubbish… straight out of the skeptic playbook.

    Because water is an incredibly efficient absorber and emitter of IR radiation, it cannot penetrate the top millimeter of the surface of water. That is why industrial driers all use infra-red heaters.This means IR warming only happens at the very surface resulting in evaporation and some heat mixing with the layer below. The main reason the air gets warmer above water as a result of GHGs is because they trap the IR that the water radiates. It is the SW radiation penetrating deep into the water column that warms it up below the very surface. The skeptic argument is wrong when they argue that because IR cannot penetrate water therefore GHGs cannot warm water, but that is not how the greenhouse effect works. Everything is warmed by the SW radiation from the sun on both the sea and the land. GHGs do not have much of a role in directly warming the sea surface (it is more direct on land), what they do is prevent cooling, they slow the return of IR back into space which reduces the temperature gradient between the sea / land and the air above it and slows down the rate of cooling. That is probably why we are getting step changes in temperature after every El Nino. The warm water is not releasing all of its heat into space before the next EN comes along.

    Everybody here seems to assume I am a “denier” or a “concern troll”. I don’t think anything I have said has implied that. I had some questions and observations. Tamino was very helpful in his rather robust way. Swanson and Rodger have posted interesting stuff as well. Nobody has yet ventured an answer to my question “what was causing the sea level to rise, and even accelerate, during the period of static / cooling surface temperatures between 1940 and 1979”. This is not a sceptic troll post it is a genuine question that I would like an answer to. If there is continued silence on the subject here I will assume nobody does have an answer and I will try and find someone who might elsewhere.

    [Response: It takes a long time for the oceans to warm up (large thermal inertia). So, just because surface temperature plateaus that doesn’t mean the ocean has yet reached equilibrium.

    You *rarely* ask questions, instead you make pronouncements, which are wrong with alarming frequency. Perhaps you need to recognize that when it comes to this subject you’re just a beginner. I will also express my opinion that your overall attitude fits the definition of “smarmy.”]

    • Where are sunlight wiggles?

      • “Where are sunlight wiggles?”

        I guess you will have to look with time-series analysis.

        I took some obsolete length-of-day (LOD) data and ran machine learning on it and it discovered some precise long-period tidal variations. That’s new, AFAICT.

        The more recent “official” data has been more heavily filtered and so I presume that is why the variations are not observed.

        And then there is the case of QBO, which also shows some very interesting seasonal aliasing that is obvious with novel analysis.

      • Greg Simpson

        Apparently, QBO is quasi-biennial oscillation.

  9. TLM – have you considered that no one has answered your question because it’s based on a faulty premise?

    “what was causing the sea level to rise, and even accelerate, during the period of static / cooling surface temperatures between 1940 and 1979”

    I don’t know why you believe there is a mystery. Sea Level Change as a function of temperature, Fig. 2 from Vermeer, 2009 doesn’t show any real discrepancies.

  10. Thanks for the link. I have to say that the Figure 2 does not seem to tally well with the Church & White data that Tamino has pointed to. In the graph the “observed sea level change [rate]” seems to level off and go into slight reverse in the early 1940s whereas the C&W data shows an acceleration. Also the rate shown in 2000 is 4mm per year and apparently accelerating whereas the satellites are currently showing 3mm a year and that is up and down a lot in the last 16 years.

    I will do my own analysis of the C&W data to extract rate of change and put it up here for your consideration.

  11. libertador

    TLM – actually the referred paper, Vermeer 2009, uses C&W. It would be quite interesting, if you got another result. Did you estimate the acceleration of SLR in C&W simply by visual perception?

  12. Libertador: My original post was just based on Tamino’s comment of a 1930s acceleration (confirmed by eyeball) so I went away and compared C&W against HadCrut4 (as I had that to hand) which showed up no noticeable correlation, so I tried again with HadSST but the result was almost identical (not terribly surprising as one is a sub-set of the other). To smooth out the wrinkles I used a 10 year linear correlation at each data point (centred on year 5) and plotted using Excel. The following is the result.

    Not sure what to make of it really. It certainly looks nothing like the Vermeer 2009 chart. Clearly that was much more heavily ‘smoothed’ so I might try some other sort of filter. What is clear though is that the trend rate in the C&W data was around 1 – 2 mm a year for most of the period 1880 – 2000 with odd spikes above 2mm and below 1mm. In 2000 at it rockets up to 4.5mm at the same time as SST change was dropping down to zero, so the two were going in opposite directions.
    This is not what I was expecting at all and is a genuine puzzle to me. Perhaps it is my choice of Linear Regression to smooth? Off to do more work on those numbers.

    • Bernard J.

      …so I went away and compared C&W against HadCrut4 (as I had that to hand) which showed up no noticeable correlation…

      And how do you define “correlation”…?

    • TLM,
      Absent graphics aside, perhaps a good move would be to examine how Vermeer & Rahmstorf (2009) approached the problem.

    • I haven’t looked deeply into it, but I assume that some area weighting would have to be involved. SLR as a function of temperature also needs to account for OHC since each ocean layer has an effect – not just the surface. It may also be necessary to use actual temperature values and not anomalies – though the difference should be minimal

    • Chris O'Neill

      showed up no noticeable correlation

      Rate of sea level rise would be related to global temperature anomaly rather than rate of change of global temperature anomaly. You need to pay attention to physics before you start looking for correlations.

  13. Image did not make it through, sorry. Not sure of the code that is necessary. A straightforward hyperlink here:

  14. OK, here is another stab at it.

    30 year linear regression smoothing on both the HadSST and C&W GMSL. Starting to make a bit more sense now. There have been significant variations in both factors, possibly looking like some kind of periodicity, but with only one cycle it is too early to tell. However it looks like a roughly 20 year lag between a rise in sea surface temperatures and a rise in sea level. I am not sure why that might be, anybody venture a mechanism / hypothesis?

    • Maybe sea level rise is causing surface temperatures to increase.

    • Bernard J.

      However it looks like a roughly 20 year lag between a rise in sea surface temperatures and a rise in sea level.

      I weep.

      • Bernard J.

        Bah, I missed the next sentence, but Libertador’s comment below addresses the issue…

    • libertador

      Your results are not surprising regarding the Vermeer 2009 paper mentioned above. Have you read the Vermeer paper. It does not seem like you have read it. The basic idea of this paper is also explained in

      They test the hypothesis, that the sea level rate depend upon an temperature anomaly term(relative to a temperature, where sea level and temperature are in equilibrium) and an instantaneous term, directly connecting rate of change in sea level with rate of change in surface temperature(In Rahmstorf 2007 only the first term). The first term lead to the lagged behavior, between change of rate in temperature and sea level. If you want to see this in your graphics, maybe you plot surface temperature (not the change rate) and change rate of GMSL and look into both mentioned papers.

      • Thanks for the links I will read them when I get the time. I can understand how there might be a lag, but just need to get my head round how it works.

  15. So what we are looking for is some physical mechanism that would get the tidal gauges to show a maximum rate of SL a little after 1940, then for that rate to drop back before starting to increase again a little after 1980.
    I wonder! How long do you think the time lag would be between knowing what we’re looking for here and twigging what the integral of d(HadSST)/dt actually represents?

  16. Brian Dodge

    The massive Greenland and Antarctic ice sheets respond slowly to changes in earth’s temperature; as far as they’re concerned, the temperature rise over the last fifty years is no different from a

    step change. The changes in ice sheet balance would continue for hundreds of years even if temperatures miraculously stabilized – and those changes are growing exponentially.
    An improved mass budget for the Greenland ice sheet; Ellyn M. Enderlin, Ian M. Howat, Seongsu Jeong, Myoung-Jong Noh, Jan H. van Angelen, and Michiel R. van den Broeke4; Geophys. Res. Lett., 41, 866–872, doi:10.1002/2013GL059010
    “The rate of loss increased from 153 ± 33 Gt/a over the period 2000–2005 to 265 ± 18 Gt/a from 2005 to 2009 and 378 ± 50 Gt/a between 2009 and 2012, giving a total acceleration of 27.0 ± 9.0 Gt/a2 since 2000. This acceleration is in good agreement with the 2003–2012 acceleration of 25 ± 9 Gt/a2 detected by GRACE [Wouters et al., 2013]”
    The conversion from ice mass loss to sea level rise can be derive from Velicogna who showed “The combined contribution of Greenland and Antarctica to global sea level rise is accelerating at a rate of 56 ± 17 Gt/yr2 during April 2002–February 2009, which corresponds to an equivalent acceleration in sea level rise of 0.17 ± 0.05 mm/yr2 during this time” and “The F-test show that the improvement obtained with the quadratic fit is statistically significant at a very high confidence level.” doi:10.1029/2009GL040222
    That was in 2009; a more accurate answer should include recent findings from – “We use satellite altimetry and gravity observations to show that a major portion of the region has, since 2009, destabilized. Ice mass loss of the marine-terminating glaciers has rapidly accelerated from close to balance in the 2000s to a sustained rate of –56 ± 8 gigatons per year, constituting a major fraction of Antarctica’s contribution to rising sea level. The widespread, simultaneous nature of the acceleration, in the absence of a persistent atmospheric forcing, points to an oceanic driving mechanism.”

    One wonders where the SLR denialists, (yes, using a linear fit is just a trickier form of denial), think all that ice melt will go; drained away to Middle Earth, perhaps? Can they set up a spreadsheet which will project how high mean sea level will be, and what the expected rate of rise will be every decade between now and 2100, if the observed 0.17 mm/yr^2 acceleration continues? What happens when Totten glacier starts accelerating too?