North Carolina Sea Level Rise: Problem Not Solved

If your science gives you a result you don’t like, pass a law saying the result is illegal. Problem solved.

— Stephen Colbert

Back in 2010, North Carolina’s Coastal Resource Commission released their North Carolina Sea Level Rise Assessment Report. It suggested that communities should be prepared for 1 meter (that’s over 3 feet) of sea level rise this century. They didn’t say there would be 1 meter of sea level rise … but it was a distinct possibility, so communities should be prepared.


It caused quite a furor; real estate agents and moguls and coastal developers objected that it could mean higher insurance rates for homeowners, it could re-define flood zones (make them much larger), it could undercut the value of existing property tremendously, and of course it would be a severe obstacle to future development in some areas. Leading the charge against the report was “NC-20”, a group claiming to represent the 20 North Carolina counties which border the Atlantic ocean.

Also in 2010, both chambers of the state legislature flipped from democratic to republican, which made it easy to persuade the state legislature not just to ignore the report, but to pass a law prohibiting the state government from basing policy on its recommendations.

Dave Burton, a member of the board of directors of NC-20, objected that the “science” in the report was speculative, and wasn’t sound. He (and in general, all of NC-20) claimed that the only valid “science” to use was to fit a straight line to the data from the past, and extrapolate that into the future. And, he was quick to point out, the past — at least, in North Carolina — showed a steady rise in sea level, at a snail’s pace, with no hint of acceleration to faster speeds. Here’s his version of a graph of sea level as measured at Wilmington, which has the longest tide gauge record in the state:

Here’s my graph of the same data:

The rate of rise — based on data available at the time — was only 2 mm/yr, which only amounts to about 8 inches in 100 years. That’s a far, far cry from the 39 inches (1 meter) which the report suggested communities prepare for. In the 75+ years that sea level had been monitored in Wilmington, it had only risen about 6 inches. Expect only a little more than that in the next century, said Dave Burton and NC-20.


What has happened to sea level in North Carolina since then?

As it turns out, it didn’t take a century for the tide gauge at Wilmington to measure “a little more than that.” It has happened already.

Problem not solved.


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


65 responses to “North Carolina Sea Level Rise: Problem Not Solved

  1. 50 year trends have also increased – the latest is 2.9 mm/yr: https://tidesandcurrents.noaa.gov/sltrends/plots/8658120_50yr.png

    Given accelerating rates, the latest 30 year trend would be considerably higher than the 50 year trend.

    Two thoughts:
    1) if we were to take the last 30-50 years of data and fit a quadratic (the way that the Colorado sea level research group has started doing: https://sealevel.colorado.edu)… would extrapolating both the linear term and the quadratic term over the next century be a reasonable thing to do? I prefer physics-based projections, but maybe Dave Burton-adjacent people would be more accepting of this kind of simplistic projection than numbers from models they don’t understand.

    2) I’m a little surprised that the 50 year rate is so low – I generally expect eastern seaboard relative SLR to be substantially higher than global (and, just spot checking, I get higher 50 year rates both north in New Jersey – 4.71 at Atlantic City – and south in Georgia – 3.88 at Fort Pulaski).

  2. Tor Ole Klemsdal

    Seriuosly? What kind of ‘cut and paste’ figure is this? Why not present a link for the official NOAA record and graph, which is completely different from the one here presented.
    https://tidesandcurrents.noaa.gov/sltrends/sltrends_station.shtml?id=8658120

    [Response: It’s the exact same data; I got it from NOAA. I prefer my graph, but if you like theirs that’s fine.]

    • Tor should take a closer look at the NOAA version. In particular, note how the single straight line fitted to the data on the NOAA version has a substantially steeper slope than the blue line fit to the data (ending in 2010) in Tamino’s graph It’s the 2011-2021 data that causes that change in slope in NOAA’s version. The part of the data that Tamino shows with the red linear fit in the second graph.

      Seriously, Tor. Can you not see the difference, and understand what it shows? It shows that 2011-2020 sea level is rising faster than it did before. Exactly as Tamino has stated.

      [Response: A minor correction: the data I show in blue is Wilmington sea level prior to 2012 (not 2010), to coincide with Dave Burton’s graph of the same (the report came in 2010, but the law was passed in 2012).]

      • Thanks for the clarification, tamino. I included 2010 in “ending in 2010”, because I could tell it went past Jan 1, 2010. Hard to tell from a quick look at the graph how much past Jan 1 2010.

    • The graph is a comparison of what Burton showed and what happened afterwards. Tamino’s graph shows the clear error in thinking extrapolating from Burton’s graph (and basing any policy on that extrapolation) was worth anything.

      Burton said to expect little more than 6″ in the next 100 years after 2012 — 6″ is about 150mm, about we’ve already seen since 2012.

  3. Nice.

    Mr Burton is a well-known denier activist, and he may be on the payroll of the so-called CO2 Coalition. At least he was a member in 2018 and 2019.

    I wrote a post about the group in 2019, after doing a little digging. And I mentioned his North Carolina SLR connection and his claim that doing proper science depends in part on the political party with which one is associated.

    He actually made comments at my blog, comments which, in fact, were off the mark (חטאה) and impugned Yi, Heki, and Qian’s work from 2017. Naturally, it didn’t end well.

  4. That data looks a whole lot like acceleration to me. By eye, at least.

  5. Tor Ole Klemsdal

    (Response to Bob Loblaw) Well, the point is that handpicking out a very limited period (10 year), that includes a major El Nino, from one particular station and pretending that this represents a systematic change in trend, is not scientific OK. The very latest measurements are now again close to the long term trend, just as it is for another of North Carolinas tide gauge stations, Beaufort, as can be seen here https://tidesandcurrents.noaa.gov/sltrends/sltrends_station.shtml?id=8656483

  6. Oh, my. So 10 years of additional data is not enough to determine that the trend has changed – but “the very latest measurements” (that is, the past few months) is long enough to determine that “everything is normal”.

    You have a very odd concept of statistics, Tor. You are looking at the noise at the end and thinking that is the trend.

  7. Sorry, you got it wrong. No, 10 years is not enough to establish a deviation from the general trend, and the most recent data confirm that the deviation was most likely only temporary and related to the 2015-16 extreme Nino event – which is also quite evident in NASA’s global sea level sattelite observations https://climate.nasa.gov/vital-signs/sea-level/

  8. Ah, yes. Double-down on the claim.

    Tell you what. Why don’t you actually do some statistical analysis, and show that the trend tamino shows for the last 10 years is not significantly different from the trend from the previous period.

    I’ll give you a hint. If you do a linear regression from the start of the series (May 1935) to the end of 2011, the slope is 0.001982, with a standard error of 0.000110 (no adjustments for autocorrelation, though). If you do a linear regression from the start of the series to the end, the slope is 0.002601, with a standard error of 0.000098.. The difference in slopes (0.000619) is 5.6 standard errors (using the larger 1937-2011 result).

    The regression of 2012-2021 gives a slope of 0.014850, with a standard deviation of 0.002406. The difference of 0.012868 is 5.3 standard errors, using the larger standard error from the 2012-2021 series.

    Please gives us your expert opinion on why such a large change in slope is “not enough”. Clearly, it has changed, and I seriously doubt that accounting for autocorrelation in the series will change the significance of the result. Saying “but El Nino” is just blowing smoke on the reality of the change. What your claim amounts to is “if I ignore all the high values in the noise, and focus on only the low values in the noise, things are not going up.” There is a word for that.

  9. You again miss the point, unfortunately. A so-called signficant fluctuation in the trend is irrelevant if it is only temporary. You cannot define the start and the length of a period where you want to look for a change retrospectively (after you have observed a deviation) – no significant test is then valid, it remains a post hoc test. Also, you cannot compare a 80 year trend with a 8 year trend from obvious reasons.
    Take a look at the curve, and do the same for the period from 1940 to 1948, as an example, and you may again find a ‘significant’ increase in slope – which of course later have reversed again to the overall trend. Several other 8-10 year periods in the whole observation period have ‘signficant’ different trends – at times even negative trends (f. eks. from 1974-84).
    Now it should be known that there was en extreme Nina i 2010-11, and rainfall in Australia that made global sea level drop (see for instance: https://phys.org/news/2012-10-la-nina-global-sea.html “The 2011 La Niña was so strong that it caused global mean sea level to drop by 5 millimeters (0.2 inches), a new study shows. Since the early 1990s, sea level has been rising by about 3 millimeters (0.1 inches) per year, satellite data show. But between the beginning of 2010 and the middle of 2011, sea level fell by 5 millimeters (0.2 inches).”
    So when you examine a period (from 2012) that starts with an extremely negative enso-affected sea level and include both the recovery from this episode and a markedly positive enso-affected sea level b(y the Nino 2015-16) -so Voila; the rate is in this limited period may of course seem ‘significant’ higher than the overall trend. And the relevance?
    About as important as the pause i global athmosphere temperatures from 2002 to 2012, perhaps?

  10. Tor Ole, I agree that 10 years is too short, but there are other data. See Tamino’s earlier post “The Truth About Sea Level Rise.” Global sea level isn’t just rising, it is accelerating. The last 10 years for North Carolina have shown that sea level can increase much faster there than the NOAA trend projects, and with global sea level rise accelerating, shouldn’t NC officials assume that the most recent reversion to the mean at NC is what is only temporary?

  11. Tor Ole Klemsdal

    Yes, Martin, this would appear so, and Nerem’s often cited publication supports this, and explains the acceleration with increased melting from Greenlabd and Antartica (i.e. ocean mass), while thermosteric effects have remained quite stable in the sattelite era. (see Climate-change–driven accelerated sea-level rise detected in the altimeter era, https://www.pnas.org/content/115/9/2022 , table 1 and table 2 ).
    However, since the end of their analyses (2017), GRACE-2 have given us new and updated data, demonstrating that annual increase in ocean mass is no longer, accelerating, but instead returning to earlier levels. This can most simply be seen in NSAS’s sea level pages https://sealevel.nasa.gov/understanding-sea-level/key-indicators/ocean-mass where the curve is steep from 2011 to 2016, but then markedly reduced. From the interactive figure you will find that the latest value presently is 36,5 mm (sept 2021). During the Nino in 2015 it peaked at 30,8, while dropping to 28,4 in august 2016. The average increase the last 6 years (from the nino top) are hence just (36,5-30,8)/6=0,95 mm/year, compared to 2 mm/year for the whole period. If you choose a less enso-affected time in august 2016, the annual increase can be estimated to (36,5-28.4)/5= 1,6 mm/year.
    Hence, what looked like a systemtic increase in melt from Greenland and Antarctica was moslt likely a temporary situation, considerably influenced by enso events.

    • Tor Ole, I’m not sure what you are saying. Greenland and Antarctica are continuing to melt, but the thermosteric increase has remained constant. If the ocean mass is reducing (“returning to earlier levels”), where is the water going? If it is increased precipitation over land and flooding, that water will return to the oceans.

      But, again, there is other data. Global average surface temperature is increasing, and faster in the polar regions, and most of the energy goes into the oceans. So when you say, “Hence, what looked like a systemic increase in melt from Greenland and Antarctica was most likely a temporary situation, considerably influenced by enso events,” do you mean it’s probably about to get a lot worse?

    • I’m curious. What’s the rationale for assigning causation to ENSO as opposed to other effects? Other effects might be the extraordinary warming the Arctic has seen, in the case of Greenland, and Southern Ocean warning at depth in the case of Antarctica. That is, unless ENSO is the root cause.

      I personally would find that an unsatisfactory explanation, since it seems to me to put the question off. It’s as if once a cause is assigned to something not well understood it’s sufficient to stop there.

      Will respect to ENSO I find the geophysics/oceanography all over the place, and insufficient attention paid to the lunar forced sloshing models of Paul Pukite, influenced by ocean warming, though.

  12. Tamino’s graph shows the rise in sea level for the added period with a linear regression having a greater slope than for the first period (Burton’s graph). Would it make sense to use changepoint analysis to look for a break in slope (and without a discontinuity) in the vicinity of 2012?

    [Response: Yes, it seems sensible to me. Confession: I already did so. Report soon.]

  13. Tor:
    You have nothing in the way of analysis to support your argument that the current rise is temporary. Clearly the change has happened. You are placing your bet that it is temporary on the last few months of data, and some resemblance to previous short-term variations in the data.

    Your original bogus claim was that Tamino’s graph was “completely different” from the NOAA one. Now, you are hauling out the “but climate [sea level] has changed in the past” bogus argument.

    Dave Burton, NC-20 et al claimed that linear extrapolation was the only justifiable way to look at sea level changes for the next 90 years, arguing that past trends will continue and sea level change was nothing to worry about. The data presented by tamino (and accepted by you, even though you were incapable of realizing that it was the same data) clearly shows that linear extrapolation is not working, and that The Burton/NC-20 predictions for 2100 have already been met.

    Tamino (and I) did not pre-select 2012 as a break point – Burton did.

    You are ignoring the long-term trend results I gave – the additional 10 years of data cause a significant increase in the slope for the entire data set.

    You aren’t happy with 10 years at the end? Try the last 14 (from Jan 2008 onward). That period both starts and ends with a local minimum, and the slope is 0.0152. Steeper than 2012-2021, with a longer period of data.

    If you want to focus on the lowest point in the downward spike near the end of the data (the 46mm reading in May 2021), then why do you not compare it to the low point in February 2009 (-170mm). That is over 200mm change in only 12 years. You accuse others of picking inappropriate time periods, but your visual “analysis” is picking features that only support your bias.

    You want to argue about when El Nino kicks in, but you only want to look at when the El Nino cycle brings the curve down at the end, not when it shifts things opposite to your preferred story line.

    Your “it’s returning to the trend” that you see in the May 2021 minimum is no more valid than if I claimed that the rise from May 2021 to July 2021 (where the reading is 176mm) is return to the rapid rise seen from 2012-2020.

    As long as you fail to provide any sort of numerical analysis, and continue to wave your hands (“but El Nino”) you have no credible argument.

    • Apologists for using linear regression really need to address and estimate degree of specification error when applying it to data produced by complex systems. There are a few ways of doing this

      But to the point here, I don’t know why if there’s a solid argument for the kinds of secular variation to which Tor (and others) appeal to argue why a Tamino like approach may not be justified, they don’t offer a quantitative argument based upon residuals. This is another case, to me, of appealing to the “god in every bush” explanation I mentioned in my question about ‘ENSO is the cause of everything’, or other teleconnections, for that matter.

      Yeah, there are non-stationarities, but they don’t all have the same size or dependency structure, and they can’t all have the same size (degree of influence).

      I think there was a recent paper bounding the relative contribution of teleconnections to climatic manifestations, compared to forcing. Don’t recall specifics and don’t have my (digital) library here, unfortunately.

  14. Tor Ole Klemsdal

    Ok, Martin, Greenland and Antartica continue to melt – or more correctly, the mass balance is negative, as both mass gain and mass loss occurs each year, of course. Higher temperature though may increase both melting an snowfall, and changes in cloud cover may be considerable importance. The net effect can be projected by models, but only observations can tell us what actually happens. The mass loss in Greenland is highly variable, but has been generally lower the last 8 years, after an increase up to 2012 http://polarportal.dk/fileadmin/user_upload/PolarPortal/season_report/polarportal_saesonrapport_2020_EN.pdf (se figure 2).
    The question is whether at present the annual total mass loss is rapidly increasing, as this post’s major figure seem to indicate, or only fluctuating for instance due to multidecal oscillations (AMO) and Enso events. Temperatures in Antartica has not increased much the last 20 years https://www.climate4you.com/images/70-90S%20MonthlyAnomaly%20Since2000.gif
    and the temperatures in Greenland appear to have had a peaked around 2012 (se for instance https://rmets.onlinelibrary.wiley.com/doi/10.1002/joc.6771 ,from the abstract; “We provide an updated analysis of instrumental Greenland monthly temperature data to 2019, focusing mainly on coastal stations but also analysing ice-sheet records from Swiss Camp and Summit. Significant summer (winter) coastal warming of ~1.7 (4.4)°C occurred from 1991–2019, but since 2001 overall temperature trends are generally flat and insignificant due to a cooling pattern over the last 6–7 years. Inland and coastal stations show broadly similar temperature trends for summer.”
    So we should lot our expectations of how the future sea level rise may develop trick us into overinterpreting the significance of short-term variations.

  15. Tor Ole Klemsdal

    I think you need to think through the rationale that lies behind Taminos graph, and what you imagine is going on:
    1) The graph suggests that sea level rise was quite stable in the period from 1935 to 2012, at a rate of 2 mm/year, and Bob describes the slope as follows ” If you do a linear regression from the start of the series (May 1935) to the end of 2011, the slope is 0.001982,..”
    2) Then all of a sudden in 2012, the rate of sea level rise seems to more than 7-double, that is 150 mm in 9 years (>16 mm/year) , and Bob writes “The regression of 2012-2021 gives a slope of 0.014850,”
    3) What physical climate factor could explain an abrupt rise like this?? CO2 level in 1935 was approx 310 ppm, and increased to 393 ppm in 2011, apparantly without changing the slope up til then. In december 2012, CO2 level was 393.6, so obviously, there is no direct connection between CO2 levels and slope – it goes through the gradual warming effect of CO2.
    4) Since sea level rise must be caused by either increased ocean mass or steric expansion or a combination of these two, anyone that belives that sea level rise changed dramatically in 2012 must present a plausible corresponding dramatic change in at least one of these underlying factors. As I have pointed out, NASAs data on Ocean mass does not show such changes https://sealevel.nasa.gov/understanding-sea-level/key-indicators/ocean-mass and neither NOAAs observastions on total steric change in sea level https://www.ncei.noaa.gov/data/oceans/woa/DATA_ANALYSIS/3M_HEAT_CONTENT/GRAPHS/totalsteric_2000m.png

    Since there is no evidence for any dramatic upward change in ocean mass or steric height from 2010-2012, one needs to look for other, more temporary causes for the observed change in slope, and the most important are obviously the Enso-events, and cyclic chages in arctic tempertures/Greenland melt.

    • Tor Ole, aren’t you a bit biased in how you use the terms “fluctuations” and “returning to earlier values?”

      You point out that the last 10 years is too short an interval (a fluctuation) to say the trend has suddenly increased, but you are using even shorter intervals:

      “The very latest measurements are now again close to the long term trend…”
      Why are the very latest measurements a return to the long term trend and not a fluctuation?

      “and the most recent data confirm that the deviation was most likely only temporary and related to the 2015-16 extreme Nino event…”
      Why are the most recent data most likely an indication that the upward deviation was temporary? Why are the most recent data not the fluctuation?

      “However, since the end of their analyses (2017), GRACE-2 have given us new and updated data, demonstrating that annual increase in ocean mass is no longer, accelerating, but instead returning to earlier levels.”
      Why are the 4 latest years a return to earlier values and not a fluctuation?

      “The average increase the last 6 years (from the nino top) are hence just (36,5-30,8)/6=0,95 mm/year, compared to 2 mm/year for the whole period.”
      Why are the last 6 years not the fluctuation?

      “The mass loss in Greenland is highly variable, but has been generally lower the last 8 years, after an increase up to 2012…”
      So why are the last 8 years not the fluctuation?

      “So we should lot our expectations of how the future sea level rise may develop trick us into overinterpreting the significance of short-term variations.”

      But you seem to be doing exactly what you warn us against doing. short term Increases are “fluctuations,” but even shorter term decreases are “returning to earlier values” or “returning to the long term trend.”

      • What Martin Smith said.

        Tor is selecting to interpret the very short term fluctuations that tell Tor the story Tor wants to hear, and rejects the variations that tell a story that Tor does not want to hear.

        Tor is the one that has the imaginative tale of what is going on.

      • Tamino and many others have long noticed and pointed out there is no piece of noise that a denier cannot turn into a “pause”, a “recovery”, or other obfuscating/minimizing term. Noise is their “friend” just as much as uncertainty when of course neither is friendly at all. They are best described as a pair of rotting crutches.

    • 1 – Tamino is not suggesting what you have claim he suggests in your point 1
      2 – Therefore from this point forward your comment is pure garbaggio.

    • Tor:
      Your point 4 is also incorrect. It is not necessary for the mass or volume of the ocean to change for there to be local sea level rise.

      It is my understanding that the thermohaline current results in an increased sea level in Europe compared to North America. This divergence is as much as a meter. Since the thermohaline current has been reported to be slowing, the water can shift back towards North America and cause sea level rise in North Carolina (and a fall in Europe) without the volume of the ocean changing at all. The question then becomes is this a permanent change or only a short time variation. Since the change in the thermohaline circulation is believed to be due to global warming, the local sea level rise North Carolina is due to AGW.

      I note that Tamino has previously shown that the rate of sea level rise globally is increasing. You would expect that it would increase locally in most places since it is increasing globally. The global data set is much bigger so the statistics are better. You need to provide a reason why global rates are increasing while the South Carolina rate is staying the same.

      • Tor Ole Klemsdal

        Michael, I see your point about that the distribution of sea level rise is not uniform across the globe. But a steep change around 2012 is nevertheless quite remarkable and unlikely, and from Taminos latest post (with his “blue line shows the best-fit parabola” ) gives a much more interesting development to discuss.
        The fact is that judged from >1270 tide gauge stations, from Jevrajevas 2014 publication https://core.ac.uk/download/pdf/191605011.pdf
        It seems that an accelerated sea level rise probably was more markedly from around 1990,
        (from the abstract: “We use 1277 tide gauge records since 1807 to provide an improved global sea level reconstruction and analyse the evolution of sea level trend and acceleration. In particular we use new data from the polar regions and remote islands to improve data coverage and extend the reconstruction to 2009. There is a good agreement between the rate of sea level rise (3.2 ± 0.4 mm·yr−1) calculated from satellite altimetry and the rate of 3.1 ± 0.6 mm·yr−1 from tide gauge based reconstruction for the overlapping time period (1993–2009). The new reconstruction suggests a linear trend of 1.9 ± 0.3 mm·yr−1 during the 20th century, with 1.8 ± 0.5 mm·yr−1 since 1970”.)

  16. Tor Ole Klemsdal

    In science , the null-hypothesis,is what needs to be proven false. So if you want do demonstrate a deviation from a trend, the fact that a deviation over a 8 years was present, but sea level rise is back at earlier rates again the last 6 years, cannot be compared. The first observation is the outlier.
    The main problem however, is that there is no support for increased slope for increase in ocean mass nor for thermosteric sea level rise after 2010, and until you can document a change in the underlying causes of sea level rise, you should be sceptic to shortlived fluctuations in local observations.

    • In modern statistics obtaining a small p-value is no guarantee the null should be rejected, since it depends critically upon the threshold chosen and in classical stats, that very typically goes unexamined. And the failure to achieve a small p-value means nothing at all. It could mean the entire calculation is questionable. It certainly does not affirm the null.

      There’s no longer any reason to continue using the Fisherian superstitious magic. Use Bayes. What are the posterior masses assigned to various slopes as supported by data and your model? With the computational engines we have now, failing to do that is inexcusable. See for instance

      Kruschke JK. Bayesian estimation supersedes the t test. J Exp Psychol Gen. 2013 May;142(2):573-603. doi: 10.1037/a0029146. Epub 2012 Jul 9. PMID: 22774788.

      [Response: Despite the fact that I’m an old Fisherian fossil, I have to agree. The future belongs to the Bayesians.]

    • Ha, ha, ha, ha ha! Oh, you really break me up, Tor.

      We should be skeptical about a “too-short” 10-year period with higher rates of sea level rise, because… an even shorter 6-year period conclusively shows a return to previous values? You haven’t even provided a number for that 6-year period, let alone done any analysis or statistical testing.

      The only thing you have shown conclusively is your ability to ignore anything you don’t like. Your Morton’s demon is strong…

      https://rationalwiki.org/wiki/Morton%27s_demon

  17. Tor Ole Klemsdal

    OK Bob, it seems clear that you have no data to contadict the fact there is no basis for a systematic dramatic shift in the slope of sea level rise around 2012, and so your arguments are now replaced by “ha-ha” and a pointless call for irrelevant statistic exercises on 8 vs 6 years periods, which cannot prove or disprove anything of importance anyway.

    If you actually think that numbers you calculated on sea level rise (you stated): (from May 1935 to the end of 2011 the slope is 0.001982,..”.”. and ” 2012-2021 gives a slope of 0.014850,” that this reflects a true and important change in how sea level rise have developed and will continue, you must have some idea of why this happened, and if it is caused by more ocean mass or more thermosteric effects. Or could it be local land sinking at Wilmington? Enormous local rainfall? Where does the extra water come from?
    Unless you can provide som kind of explanation, I think further discussion is meaningless, and I will leave it at this.
    (Ironicly, Taminos new post quite clearly points to the questionnable ‘analysis’ you so strongly defend: He writes; ” So why, some might wonder, did I choose in my last post to show the data from a single tide gauge station (Wilmington, NC), do no analysis at all, but manipulate the graph to give a distinct visual impression? It wasn’t much of a manipulation — all I did was plot the pre-2012 data in black and the post-2012 data in red — but it did the job.”
    Tamino’s new analysis, with his “blue line shows the best-fit parabola” is of course much more interesting to discuss, and may describe reality in a much more adequate manner, so maybe you can acknowledge that the first figure was never meant to be a serious analysis, allthough it seems to have convinced you)

    • More hand-waving from Tor.

    • Tor Ole wrote: “OK Bob, it seems clear that you have no data to contadict the fact there is no basis for a systematic dramatic shift in the slope of sea level rise around 2012…”

      I don’t think anyone but you is talking about “a systematic dramatic shift” in sea level rise at 2012. I think the claim is that global sea level rise is accelerating and the data for North Carolina are consistent with that. Therefore it is dumb for North Carolina politicians to pass a law that denies it.

      • Tor Ole Klemsdal

        Martin, you write; ” I don’t think anyone but you is talking about “a systematic dramatic shift” in sea level rise at 2012. ”
        But what else is represented by the two regression lines in Tamino’s figure??
        The black line, representing the trend from 1935 to end 2011, starts around -125 mm and ends at some +25 mm, that is 150 mm in 76 years, or 2 mm/year, as also stated in the post.
        The red line, starts at 25 mm in 2012 and ends at near 200 mm in 2021, which is close to 175 mm in less than 10 years, i.e. 17,5 mm/year.
        The graph seems to suggest that sea level increased more the last 10 years than the previous 76 years. This can hardly be described as less than dramatic?
        Tide gauges with long exact measurements are also present in NOAAs database on the other side of the Atlantic, for instance in
        Bergen (Norway), https://tidesandcurrents.noaa.gov/sltrends/sltrends_station.shtml?id=040-221
        Aberdeen (Scotland) https://tidesandcurrents.noaa.gov/sltrends/sltrends_station.shtml?id=170-011
        and Delfzijl (the Netherlands) https://tidesandcurrents.noaa.gov/sltrends/sltrends_station.shtml?id=150-001
        They all show variations and slopes that depend locations and land rising (as Bergen), but no special trend change around 2012.
        Tamino has now in his new post also explained why he made the figure, so I think we can safely discard the one we are here discussing.

      • Tor Ole asks: “But what else is represented by the two regression lines in Tamino’s figure??”

        Don’t you know? You even referred to it in your earlier comment when you wrote: “He [Tamino] writes; ‘So why, some might wonder, did I choose in my last post to show the data from a single tide gauge station (Wilmington, NC), do no analysis at all, but manipulate the graph to give a distinct visual impression? It wasn’t much of a manipulation — all I did was plot the pre-2012 data in black and the post-2012 data in red — but it did the job.’ ”

        When I read Tamino’s post (the one you are complaining about), I, like (I assume) many others here who follow Tamino, thought: “Wait a minute, he can’t just break it at 2012 without showing why it makes sense to break it at 2012. Does it make sense to break it there? There must be more to it than that; Tamino always does the math.”

        Sure enough, there was more to it than that; his next post explained it. Your quote alludes to that, but you truncated the most important part. Tamino also wrote: “I had already ‘done the math.’ ”

        And then he went on to show: “The real result of both tests is: sea level is not rising at a constant rate, rather it has accelerated, at Wilmington NC.”

        So the answer to your question, “But what else is represented by the two regression lines in Tamino’s figure??” is this: “sea level is not rising at a constant rate, rather it has accelerated, at Wilmington NC.”

        Dave Burton is proved wrong, and the NC state legislature is proved stupid.

  18. Tor Ole Klemsdal

    Yes, Martin, it is fine that Tamino at Nov 26 (2 days after my first comment on Nov the 24.) – submitted a new post with a different graph, and explaining the background for the first one. I can hardly be accused for not knowing what was later to be explained – and actually proves my point.
    Further, as you must have seen, Bob Loblaw made several comments with his own exact calculations of sea level rise slopes in the very same 2 periods, describing (“.. a linear regression from the start of the series (May 1935) to the end of 2011, the slope is 0.001982, with a standard error of 0.000110……..The regression of 2012-2021 gives a slope of 0.014850, with a standard deviation of 0.002406…”)
    So what is the point of calculating a slope of 0.001982 and compare it to 0.014850, unless you want to describe a dramatic increase in slope?

    When I tried to argue that such a slope – 15 mm a year -5 times the global average from sattelite measurements – was temporary, and the latest measurements returned to lower levels, the response from both you and Bob was that this normalization was just as/even more likely to be noise and fluctuation – you asked; ” Why are the most recent data most likely an indication that the upward deviation was temporary? Why are the most recent data not the fluctuation?”

    This is exactly was discussion is all about: Why can I say that the upward deviation was temporary?? Simply beacuse you cannot ignore all other observations around the globe from every recognized source (NOAA, NASA, AVISO etc, etc.) that sea level rise the last 10 years is around 3-4 mm/year, so if you for a local tide gauge series, in a short period (< 10 years) manage to calculate a sea level rise that is instead 15 mm/year, 5 times the expected – well this result cannot represent truly whats happening, and sooner or later the measurements will return to what is commonly observed along the coastside of the Atlantic ocean, unless the observations at this tide gauge are biased.
    It as simple as that. Or perhaps you and Bob think that the best prediction for sea level rise in North Carolina the next 10 years is 150 mm?

    • Tor Ole writes: “When I tried to argue that such a slope – 15 mm a year -5 times the global average from sattelite measurements – was temporary, and the latest measurements returned to lower levels, the response from both you and Bob was that this normalization was just as/even more likely to be noise and fluctuation – you asked; ” Why are the most recent data most likely an indication that the upward deviation was temporary? Why are the most recent data not the fluctuation?”
      This is exactly was discussion is all about: Why can I say that the upward deviation was temporary??”

      No, it is not what the discussion is about. It is what you tried to make the discussion about. You repeatedly referred to short term increases in the sea level rise trend as fluctuations, and to even shorter term decreases as returning to the earlier trend. Without the analysis Tamino has now done (had already done), they are both noise, but you write about the increases as being anomalistic and the decreases as being returns to normal.

      Tor Ole continued: “Simply beacuse you cannot ignore all other observations around the globe from every recognized source (NOAA, NASA, AVISO etc, etc.) that sea level rise the last 10 years is around 3-4 mm/year,…”

      But we have NOT ignored them. Our replies to your comments are based, in part, on those observations, because 3-4 mm/y is an acceleration of the global sea level rise. And, the global average surface temperature is rising. And, the ice sheets and glaciers are melting. Burton and the NC state legislature are the ones who are ignoring all those other observations. They are the ones behaving as if sea level rise on NC shores will forever be 2 mm/y, no matter what is happening globally. And, Tamino has done the analysis. Sea level rise at Wilmington has accelerated too!

      That is what the discussion is about.

      • Tor Ole Klemsdal

        OK, Martin, I understand what you are saying here, but I have never argued to support Burton’s claim or taken part in that discussion. So I cannot agree that “this is what the discussion is about,” allthough this may be what you want the discussion to be about.
        As I wrote in an earlier comment (response to Michael, see above) , there are solid data from Jevrajeva 2014 from 1277 tide gauge stations (not one or two) that documents accelleration of sea level rise, but this trend shift was already present in the period from 1993 to 2009.
        From the abstract: “There is a good agreement between the rate of sea level rise (3.2 ± 0.4 mm·yr−1) calculated from satellite altimetry and the rate of 3.1 ± 0.6 mm·yr−1 from tide gauge based reconstruction for the overlapping time period (1993–2009). The new reconstruction suggests a linear trend of 1.9 ± 0.3 mm·yr−1 during the 20th century, with 1.8 ± 0.5 mm·yr−1 since 1970”.)
        So, what I have argued against is NOT that sea level rise have been accelerated from some 2 mm/year prior to 1990 to at present around 3-4 mm/year (on the contraray, I even presented a well recognized paper to support this). But what i HAVE argued against is that a dramatic shift in slope occurred in 2012 and that this represents more than a temporary local fluctuation, combined with enso effects (the period started with Nina-related extreme rainfall and drop in global sea level, and continued in the super 2015-2016 Nino episode).
        So perhaps we can agree that 15 mm/year was a fluctuation, and that 3-4 mm/year is the current ‘normal’ level of sea level rise’, and not the other way around?

        [Response: There are several reconstructions of GMSL (global sea level} based on tide gauges, and that from of Jevrejeva et al. is among the worst (see this). For a decade or so, the most trusted was from Church & White, then along came Dangendorf et al., and the most recent (and in my opinion, the best) is from Frederikse et al.

        GMSL is not rising at a 3-4 mm/yr, it’s currently 4-5. The satellite record alone shows undeniable acceleration (no, it’s not just fluctuations from el Nino) whether you use the data reduction from JPL, from NOAA/STAR, from U.Colorado, or from CSIRO.

        As for the rest of the century, I doubt that analyzing current trends will be of that much use. I take my cue from the experts working on ice sheet dynamics, and they say that the *good luck* case is: we’ll get about 1 meter (GMSL) from 2000 to 2100, the *bad luck* case is 3 meters or more. In particular, the disappearance of “ice shelves” removes the buttress which keeps so much ice land-locked, that it dramatically (and rather suddenly) increase ice discharge to the ocean (and hence sea level rise). The scientists who study this, in particular, seem to me to be the most worried.]

    • Actually, the first thing Tor tried to make the discussion about was the following:

      “Seriuosly? [sic] What kind of ‘cut and paste’ figure is this? Why not present a link for the official NOAA record and graph, which is completely different from the one here presented.”

      Since then, he’s tried several “Look! Squirrel!” diversions, to cover up the bogosity of his initial claim.

      At least he was half right when he said “I think further discussion is meaningless, and I will leave it at this.”

  19. Oh, the strawmanity.

    None of us ever said the trend over the past 10 (or 14) years would extend into the future for the next 10 years.

    You are the one that has claimed – and have never provided any sort of statistical analysis, – that the “trend” over the next 10 years will see sea level dropping to the old (ending in 2012) trend line.

    We keep pointing out that the recent data does not support that. [Your, and Burton’s, expectation that the linear trend ending in 2012 is the best estimate of future sea level.)

    Stop making stuff up.

    • Tor Ole Klemsdal

      “None of us ever said the trend over the past 10 (or 14) years would extend into the future for the next 10 years.”
      OK, is that so??
      So what was the point of constructing the red regression line for this period, and calculating the slope ” ..The regression of 2012-2021 gives a slope of 0.014850, with a standard deviation of 0.002406…”.? You never thought that it would contunue this way? In other words – you considered that it was a temporary fluctuation? You just did’nt want someone else to point out that this 15 mm/y trend had to be temporary?
      Then it seems that you can continue this debate with yourself without ny help.

      • Yes, Tor, some of us area actually capable of thinking that sea level rise will not continue at the same rate as in the past. Some of us are capable of thinking that a change in the rate, confirmed by actually doing the math (instead of just waving our hands), is not actually “temporary”, and will be followed by additional changes of the rate in the future. Some of us are capable of understanding that just because we can fit a straight line does not mean that every departure from a straight line is a fluctuation that will always return back to that straight line.

        Some of us are capable of understanding the various factors that affect sea level rise. Some of us are capable of understanding that some factors are short-term and will cause fluctuations that do not alter the long-term trends. Some of us are capable of understanding that some factors will cause longer-term changes and therefore alter the overall trend. Some of us are capable of understanding that one possible – no, likely – change in the trend is acceleration.

        Some of us understand that the mathematical term “acceleration” means that the slope is not constant, and the next 10 years will probably be different again. Some of use are capable of understanding that the future is not limited to the binary choice of “continue on this straight line, or go back to the previous straight line”. Some of us are capable of understanding that over a period of 10 years, a different slope can be the result of a combination of temporary fluctuations plus a real change in the long-term trend.

        Some of us area capable of actually doing some mathematics and statistics to evaluate what is happening.

      • Yes, Tor, please. Stop helping.

  20. Pretty definitive:

    Hamlington, B. D., Frederikse, T., Nerem, R. S., Fasullo, J. T., & Adhikari, S. (2020). Investigating the acceleration of regional sea level rise during the satellite altimeter era. Geophysical Research Letters, 47, e2019GL086528. https://doi.org/10.1029/2019GL086528

    • Definitive?
      The abstract you linked to concludes as follows ” While the dominant features in the acceleration pattern can be attributed to internal variability, there is an indication that the forced acceleration pattern may emerge as the record continues to lengthen.”

      “an indication.. that may emerge “.
      Not ny idea of something definitive

      • It’s entirely possible for a research result to be definitive without affirming a hypothesis as certain, whatever that means in geophysical work. You are equivocating “definitive” in the popular sense of the term with how it is used in scholarship.Thwy are not the same.

        The result is not rejecting the hypothesis, indeed, offering evidence in its support.

      • To get an idea of why I tag this study as “definitive”, consider, first, the approach:

        In this study, we followed the approach of Royston et al. [2018], although we did not assess the most appropriate noise model, as reproducing the results of the referenced paper was beyond the scope of this study. Instead, the uncertainty in the acceleration is estimated at each location using a generalized Gauss Markov noise model, which in cases where an AR(1) or other noise model may be most appropriate, would lead to a likely overestimate of the uncertainty, rather than an underestimate. We wanted to err, however, towards a more complete assessment of the uncertainty rather than the alternative, which could lead to overinterpretation of the resulting acceleration. Furthermore, an important part of the approach here is to include the uncertainty associated with the technique for removing the modes of natural variability. This was done using the results of Hamlington et al. [2019] that were obtained by applying the CSEOF analysis to the individual ensemble members of the CESM that include both natural variability and the forced response, and determining the extent to which the description of these modes changed in the presence of that forced response. Although not strictly independent, the uncertainty from the removal of natural variability is combined with the uncertainty from the acceleration estimation by squaring the values, summing and then taking the square root. When including this uncertainty in the estimate of the acceleration, the impact of the choice of noise model became less apparent. When including this uncertainty in the estimate of the acceleration, the impact of the choice of noise model became less apparent.

        (from the text’s Supplement)

        and then consider, secondly, from the Discussion:

        For example, there has been a notable shift in the trends measured by the altimeters in recent years, made clear by the trend patterns in Fig. 2 but quantified by the acceleration pattern in Fig. 3b. The perception of the significance of ongoing sea-level rise can be influenced by these short-term changes and trends (e.g. internal signals suppressing sea-level rise for an extended period in a particular location), but it is important to emphasize that such variability generally oscillates on top of a background trend and acceleration. Examining the acceleration in the altimeter record allows us to both highlight the influence of internal variability, while assessing the degree to which we can uncover the forced, underlying acceleration by attempting to remove it.

        (from the main of the paper, in its Discussion, with emphasis added, and see the paper itself for context)

        Figure 2 from paper:

        https://hypergeometric.files.wordpress.com/2021/12/fig2-2021-12-01_111215.jpg?resize=446%2C446

        Figure 3 from paper:

        https://hypergeometric.files.wordpress.com/2021/12/fig3b-2021-12-01_111309.jpg?resize=446%2C446

      • Sorry, WordPress cropped those two figures. Here they are again:

        Figure 2:

        https://photos.app.goo.gl/QRbkZhduQYtKC93bA

        Figure 3:

        https://photos.app.goo.gl/GDrhELM3VYy7imhk6

  21. Tor Ole Klemsdal

    To Bob.
    It might be possible to have a more fruitful discussion if you could back your arguments with published papers, in stead of thinking that your own analyses says it all. The problem with statistics, and especially linear regression, is to know when not to use it.
    I have repeatedly shown that NASA, NOAA and other recognized sources demonstrate no marked upward shift in the underlying causes of sea level rise in 2012 -i.e. there is no change in the rate of increased ocen mass or steric expansion at this time point (se also https://iopscience.iop.org/article/10.1088/1748-9326/abceea
    and Jevrajeva and others suggest that the acceleration in general took place well ahead of 2012 – probably starting around 1990.
    The fact that you can show a different slope in one tide gauge station from 2012 and onward is trivial, and now it is obvious that you realize that the 15 mm/year was indeed a fluctuation, However as you wrote ” Some of us are capable of understanding that over a period of 10 years, a different slope can be the result of a combination of temporary fluctuations plus a real change in the long-term trend.”
    Fully agreed. But your primitive analysis from one single station cannot tell us how much is temporary fluctuation and how much might be a real change in the long-term trend. In this respect it’s worthless.
    If in deed it is correct, as Tamino writes above “GMSL is not rising at a 3-4 mm/yr, it’s currently 4-5 mm/y”, then still at least 10 of the 15 mm/y you estimated from 2012 on represents a temporary fluctuation, and perhaps 1-2 mm/y (above the background slope) the possible ‘real change in the long-term trend’.
    Obviously, the temptation to use the observed steep increase in sea level at this station in 2012 as a ‘proof’ for that ‘dangrous sea level rise already being present’ was so high, that the red regression line starting in 2012 was presented without any reservations, for instance that perhaps 80-90% of the increase in rate stems from temporary fluctuations. Seems like another example of ‘Never let the facts get in the way for a good story’.

    • You just keep digging, don’t you Tor?

      The problem with statistics, and especially linear regression, is to know when not to use it.

      This comes from someone who is certain, without any analysis at all, that “…that perhaps 80-90% of the increase in rate stems from temporary fluctuations”. Coming from the person that claims “The very latest measurements are now again close to the long term trend”.

      And now you are back to claiming that NOAA analysis supports your case, when you started your comment here with “Why not present a link for the official NOAA record and graph, which is completely different from the one here presented.” – even though the data used in the two graphs are exactly the same.

      You obviously can’t see anything in the data that does not fit your preconceptions. You “know not to use it” when it does not fit your story.

    • Re. linear fits…

      Question for Tor: How “explanatory” would you say a linear fit to the first 30 squares–{1,4,9,16…900} is?

      Well solving the linear regression gives us the nice straight line Y = 31X – 165.33 where the R-squared is .94, so it’s a GREAT fit, AND it “explains” the series behavior sufficiently well we need look at nothing else in the data. Right? (Well it actually does predict OK around the 7th and 25th values of the series, not so hot elsewhere, though.)

  22. Tor Ole Klemsdal

    Bob, you write; “This comes from someone who is certain, without any analysis at all, that “…that perhaps 80-90% of the increase in rate stems from temporary fluctuations”.
    (I’m not native English, but I never realized that the word ‘perhaps’ was used when a person was certain of anything??).

    But OK. Back at Nov. 25, you wrote:” Tor:
    You have nothing in the way of analysis to support your argument that the current rise is temporary. Clearly the change has happened. You are placing your bet that it is temporary on the last few months of data, and some resemblance to previous short-term variations in the data.”

    Today, you sound a bit different, and write, first; ” None of us ever said the trend over the past 10 (or 14) years would extend into the future for the next 10 years.” and then
    “Some of us are capable of understanding that over a period of 10 years, a different slope can be the result of a combination of temporary fluctuations plus a real change in the long-term trend.”

    Good, so what is your best estimate for how much of this change in slope is caused by temporary fluctuations and how much is related to a real change in the long-term trend?
    You have calculated a trend in this period after 2012 of 15 mm/y, while Tamino states that GMSL is currently 4-5 mm/y. NOAAs long term trend for this station is 2,56 mm/y https://tidesandcurrents.noaa.gov/sltrends/sltrends_station.shtml?id=8658120
    So what is your best prediction for sea level rise the next decade, and what kind of analysis and reasoning do you suggest to make an estimate?

    • “Look Squirrel!

      Once again, you provide no analysis. Just hand waving.

      Your “temporary” argument repeatedly claims that sea level will return to the old linear trend. The recent data, the linear trend from my quick calculation, and from Tamino’s more detailed analysis on the other blog post you won’t comment on, is strong evidence that things will not return to the old linear trend. The data already fall well outside the extrapolation of that trend.

      Either intentionally, or because you can’t conceive of any other possibilities, you are back to only considering a binary choice: either “it’s temporary (implying it will go back to the old trend), or I must think it is going to continue at the same trend as the last 10 years.

      …and your “perhaps” is placed in the context of your “80-90%” confidence statement.

      I’m with the experts on this: the long-term trend will not be linear. It will accelerate. How much depends on how rapidly we warm the planet, and how rapidly land ice melts and sea water warms. And fluctuations such as El Nino and other local factors will make local trends quite noisy over periods of years.

      The next 10 years? Only a fool tries to predict the next 10 years in a system with a lot of noise. We have already seen your prediction: it will return to the linear trend. We already have a lot of evidence that your prediction is wrong.

      • Tor Ole Klemsdal

        OK, so you strongly disagree to my statement that perhaps 80-90% of the dramatic increase in slope of sea level rise at Wilmington in the period 2012-21 must be temporary, but you will not come up with your own (presumably more solid) estimation.
        As regards predicting the next 10 years I totally agree that it is difficult beacuse it is a system with a lot of noise.
        But I’m pretty confident that using NOAAs linear trend of 2,56 mm/y will come much closer to reality than a prediction following the red line in Taminos graph or your corresponding “regression of 2012-2021 gives a slope of 0.014850”. The graph and your calculation is misleading, and I think Tamino’s follow-up comment in his next post acknowledged this

        Finally, I would like to have your opinion on the following 2 simple questions.
        1) Do you agree that global sea level rise is determined by only two factors, increased ocean mass and thermal expansion?
        2) Since you are convinced that sea level rise already have, and will coninue to accelerate, which of these factors (or how much of each) is/will be responsible for the acceleration?

      • Oh, i think I am starting to see your problem, Tor. You actually think that “it’s temporary…”… “…because El NIno” actually represents a scientific explanation of sea level patterns.

        And you are back to creating straw men: pretending that I used the slope from the past 10 years as a prediction into the future. If I had actually done that, yes it would be misleading, but I didn’t. Maybe it’s that your reading comprehension is pretty poor, but I suspect you know exactly what you are doing. Let’s say I’m 80-90% confident.

        And your “perhaps” has returned to “pretty confident that using NOAAs linear trend…”, in spite of the statistical analysis that has shown that the trend is not linear (cf. other tamino posts). Since you have yet to provide even a single slope/statistical analysis of your own, I am beginning to think that you actually can’t do it.

        You want a comprehensive explanation/estimation of future sea level rise? Try here:

        https://www.ipcc.ch/report/sixth-assessment-report-working-group-i/

        If that is too tough for you to read, there is a nice summary here:

        https://www.realclimate.org/index.php/archives/2021/08/sea-level-in-the-ipcc-6th-assessment-report-ar6/#comment-796393

      • @Tor
        “As regards predicting the next 10 years I totally agree that it is difficult beacuse it is a system with a lot of noise.
        But I’m pretty confident that using NOAAs linear trend of 2,56 mm/y will come much closer to reality than a prediction following the red line in Taminos graph or your corresponding “regression of 2012-2021 gives a slope of 0.014850”. The graph and your calculation is misleading, and I think Tamino’s follow-up comment in his next post acknowledged this”

        1 – Strawman if I ever saw one…
        2 – Tamino’s graph IS NOT a prediction for the next 10 years. Repeating false assertions in different guises does not make them less stupid.
        3 – Tamino’s graph is only misleading to a moron who doesn’t bother to read the whole post and follow links. Again – it is NOT a prediction, it is REBUTTAL of a prediction.

  23. Tor Ole Klemsdal

    To Ecoquant.
    Thanks for the link and figures. I read the paper, and it is interesting, and I see your point.
    Nevertheless, the authors chose to write their conclusion in the abstract as I cited it;” the dominant features in the acceleration pattern can be attributed to internal variability,..”
    I think that is rather relevant for the discussion we have had here.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s