Recent Sea Level Change

NOAA provides an excellent website for acquiring and examining sea level data from tide gauges. It includes maps with which one can select individual stations, but which also show the rate of sea level rise based on fitting a linear trend to all the available data. Here’s their map, zoomed in on the USA:

The lengths of the arrows corresponds to the sea level rise rate, and are color-coded as well. Those in green have rates from 0 to 3 mm/yr, those in yellow from 3 to 6 mm/yr, those in red higher than 6 mm/yr, and the blue indicates negative rates.

Note that local rates can be below zero even when global sea level is rising, because tide gauges show local sea level, and that includes both the change in the height of the sea surface and in the height of the land. The land height can change for many reasons. There’s glacial isostatic adjustment: when glaciers and ice sheets melt away the weight of the ice no longer presses down on the land, pushing it down, and the land itself will rebound (but very slowly). Earthquakes can change the height of the land. Extracting groundwater can do so too. All these effects contribute to vertical land movement (VLM).

I’ll create essentially the same map, but I’ll change the color coding so that blue is for rates from 0 to 3 mm/yr, red for rates from 3 to 6 mm/yr, and black for rates above 6 mm/yr. I’ll also plot fewer stations, because I want to focus on just those stations in the U.S. that have enough data to give a reasonable rate estimate recently (since 1990, although I’ll estimate rates using all the data). And here it is:

Clearly, sea level rise has been greater on the east coast of the U.S. than on the west coast. And, northern New England (from about Boston northward) seems to see slower sea level rise than the rest of the east coast.

But there’s quite a lot of difference between stations, even within small regions. On the east coast, for instance, we see a lot of blue arrows (less than 3 mm/yr) and a lot of red ones too (3 to 6 mm/yr), even some black arrows (more than 6 mm/yr). There doesn’t seem to be much rhyme or reason to the differences, and one might be tempted to think that there are lots and lots of differences in vertical land movement, even for nearby stations, which would account for those differences.

But that’s not the only thing that can cause such differences, because these rate estimates come from fitting a straight line to all the available data. This means they represent the average rate of sea level rise during the entire time span of observations. But different tide gauges cover different time spans, so they’re not necessarily estimating the same thing. A station with data from 1950 through 1980 gives us the average rate 1950-1980, while one with data from 1990 to the present gives the average rate 1990-present. Those rates may not be the same (in fact they are not the same), even if two tide gauges are at the same location.

That’s because the rate of sea level rise (of global sea level rise) hasn’t been constant over time. It has shown acceleration (speedup), and at some times deceleration (slowdown) as well.

Climate deniers hate that. They want you to believe that global sea level has risen at a constant rate for the last 80 years or so at least, so they can deny the danger from even faster sea level rise in the near future. Also, the rate over the last 80 years or so has been less than the rate recently, so by claiming that there’s no “real” change they can make the claim that it isn’t even rising now as fast as it really is rising now. Perhaps the most scientifically accurate description of their approach would be: “total bullshit.”

I was interested in comparing the rates at different locations recently, based on a time span these tide gauges all share — at least, much more closely than using the entire time spans. So, I isolated the data from 1990 to the present to estimate the recent sea level rise rate. Stations will still show differences, even very nearby ones, since vertical land movement is a real phenomenon and really affects local sea level. But at least we’ll greatly reduce the differences caused by using (sometimes extremely) different time spans.

Using just the since-1990 data, and insisting on enough data to get a good estimate, I get this map:

Now the differences, especially on the east coast, are much reduced. Northern New England no longer stands out as having especially slower sea level rise than the rest of the eastern seaboard. This not only shows the greater consistency of sea level rise estimates at nearby locations, it also shows the importance of using a common time span for estimation if we want to know what those difference really are.

Differences remain. In particular, there are a few stations which “stick out like a sore thumb” and suggest that those locations are greatly affected by vertical land movement. The most consistent differences are the very high rates in the western Gulf of Mexico, with several stations in Texas and Lousiana showing rates greater than 6 mm/yr. It’s possible that this is due to changes in ocean currents in the western Gulf of Mexico, but I strongly suspect that it’s really a reflection of the impact of vertical land movement in those areas.

All this emphasizes that sea level rise is definitely higher than its rate over the last 80+ years; sea level really has accelerated, and the recent rate is probably the highest seen in many thousands of years. It also warns us about the U.S. east coast, a warning not to be taken lightly because the slowdown of the Gulf Stream system is one of the factors accelerating sea level rise in that region.

Again, climate deniers will deny this and claim that sea level has changed dramatically in the last few thousand years, a claim that should also be classified as “total bullshit.” They’re particularly eager to deny the reality of sea level rise because it’s one of the impacts of global warming which will cost trillions of dollars — Miami is already spending about half a billion dollars trying to keep the water out when “king tides” (the highest tides of the year) cause the streets to flood even with no storm, no wind, no rain. They can’t deny the cost without looking like idiots, they can’t deny it’s because of sea level rise without looking like idiots, so they deny the only thing they can: the truth, that it’s happening right now a lot faster than it was in the 20th century, and it’s very, very, very likely to rise even faster during the remainder of the 21st century.

I expect them to continue to deny, to continue to fiddle about, while great cities drown. They reap the profits. We pay the price.

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

9 responses to “Recent Sea Level Change

  1. Having seen talks by an expert in sea level rise it is true it is very hard to see the trends in the details because of so many confounding issues. For example Europe is still rising since it was last pushed down by the last global ice age. The first time I saw this speaker he concluded the trend observations were less than the error bars. Years later he concluded sea level rise was occurring once they refined the data that the trends observed were not inside the error bars. In some strange parody at the same event a denier handed me a magazine where they claimed scientists do not take this or that into account, not realising how foolish they looked because that is exactly what they took into account, and why is was some years before they could state with certainty the sea was rising.

  2. The tide gauge in Galveston is in the old part of the city. I think it is actually close to an offshore drilling museum. In the 1990s Galveston County started restricting well drilling for water. While they no doubt experienced a great deal of historic subsidence, I have found studies that cite either a very low level of recent subsidence or even a low level of recent uplift. The subsidence in Houston (Harris County) is well inland. From my Houston condo, it was 37 miles to the Strand.

  3. Yes, the NOAA site is excellent. Some but not all stations have the option to plot “variations in 50-year trends.” From Boston northward stations with this data option include Boston, Portland and Eastport Maine. Each of these 3 stations shows an increase in the rate of sea level rise in more recent years as you describe.

    Boston’s sea level rise rate data show the 50-year periods centered on 1945 (the beginning of the data set) and 1950 to be a bit higher than the current period. Portland data show the sea level rise to be much higher in 7 consecutive 50-year periods from the beginning of the data set in 1935 thru 1965. The Eastport data show sea level rise to be much higher for the 50-year periods centered on the initial data periods beginning in 1955 thru 1965.

    Any thoughts on the higher sea level rise rates from earlier periods?

    [Response: Perhaps the first thing to bear in mind is that a 50-year trend based on linear regression of data ending now isn’t the rate now, it’s the average rate over the last 50 years. If we’re to assign it a “time” at which it applies, the logical choice is the midpoint of the time span. So the most recent 50-year linear trend is an estimate of the trend at about 1993 (25 years ago). There’s also the sea level *fall* induced by reduced gravity from the Greenland ice sheet, but I’m not sure off the top of my head how strong this effect is in New England.

    Your question motivates me to look at the New England tide gauge station data in more detail. Stay tuned.]

  4. I was teaching in Texas about a decade ago, and watched the degradation of the information about sea level and coastal erosion. When I first started, the Texas General Land Office ( had extensive coverage of coastal erosion and sea level rise, with links to good data sets, e.g. shoreline retreat.

    Then the disappearance of data began–first the links were broken, then the websites that carried the data disappeared, and now there is only a paragraph at the site that mentions coastal erosion, but says nothing about sea level rise. Significant coastal subsidence has occurred from oil and water withdrawal from below ground, but sea level rise in the Gulf of Mexico is still major. Ironically, the subsidence data is being disappeared as well. The first step seems to be removing graphics from the top web page, e.g.

  5. Correct me if I am wrong, but I believe NOAA uses PSMSL data. So far, I am the only person I have seen use this feature on climate blogs. If you go the PSMSL and click on PRODUCTS, and then click on derived trends, you get a similar map. Move the map to the NE coast. Click on an arrow for a gauge. I’m clicking on Boston, gauge 235. 1900 to 2016 is 2.79 mm/yr. Across the bottom there is a blue slider, 1900 to ~2017. Click on the button above 1900 and slide to as far your right as it will go: 1987 to 2016 is 4.58 mm/yr.

    1637: Bergen Point, Staten Island – 4.73 mm/yr
    366: Sandy Hook – 5.39 mm/yr
    360: Washington DC – 4.19 mm/yr

    [Response: Perhaps it’s more accurate to say that PSMSL archives the data. I usually get data from PSMSL myself, because it’s so easy to get the entire data set. But sometimes I use NOAA because they update their data more frequently. Almost all tide gauge stations archived in PSMSL only go through the end of 2016, but from NOAA (for U.S. stations at least) you can get data which includes 2017 and part of 2018.]

  6. Given the gravity of the whole matter of SLR, the quality of the data from stations that we have is sadly still poor. These days we should be able to deploy GPS corrected and internet connected SLR monitors relatively cheaply (compared certainly with the issue of SLR as such) around much of the coastal towns into a global network. This, I think, would come in rather handy over the next decades.
    I checked on the last decades (2008-2018) SLR data from a few local stations in NZ and they all are above 10mm/yr. Interesting. I guess that the PDO plays a role but I would not be surprised if one of these years trends in many places will react two-digit territory. Certainly, if projections of 1m SLR or more by 2100 are going to materialize then we would see these soon globally. And given that often there is often as it seems a significant lag in the reporting of data, I would not be surprised if are not already in that region in many places but just don’t have the data yet to confirm it.

    [Response: I’m not aware of a PDO effect, but we do know that el Nino has an impact. However, it’s only temporary (until the el Nino fades). It can be surprising how different are the short-term fluctuations in different regions, which makes it harder to pin down the long term *global* rate. I see evidence already (from satellite data) that the present rate is higher than the ~ 3mm/yr most often quoted (see this.]

    • The matter of shorter-term variations came up on RC when one of our resident denialati pointed out a decline in sea level in California over the period of 1992-2009.

      I responded that ENSO wasn’t stationary over that period and that therefore SLR trend would be biased low, going so far (by way of illustration) as to plot the Southern Oscillation Index (SOI) versus the San Francisco tidal gauge data per NOAA:

      To my eye, the graph does indeed suggest that SOI/ENSO affects mean sea level at San Francisco (though clearly it’s not the only factor.) And of course, you’d expect the inverse correlation in, say, an Indonesian tidal gauge. It’s logical from a physical point of view; ENSO affects temps in the top layer of the ocean, as well as atmospheric pressure and–I expect by far the most important factor–prevailing winds. All of those could, and presumably do, affect local sea level.

      I was surprised to see that there didn’t seem to be much written about the effects of ENSO, PDO, et al., on high-frequency variations in SLR trends. It made me wonder if something along the lines of Foster & Ramstorf, but for SLR, might be a useful line of inquiry. I suspect it would be a messier analysis, but still…

  7. Excellent, as nearly always.

    Two points.

    First, in the Northeast, there are other variations, too, having to do with deceleration of the Gulf Stream. This piles water up along the coast. There, can, too, be mesoscale eddies which break off the main Stream flow and persist for months. The slowdown is thought to be related to the time evolution of the AMOC, which is apparently weakening, even though it has year-to-year variability. There may, too, be some relationship between a slowdown the the “cold blob” southeast of Greenland, near or in the Irminger Sea.

    Second, on adaptation, there is a paper which was just published in Earth’s Future which I find very interesting, since I have a front row seat witnessing this around southern New England and especially around Boston: Keeler, A. G., McNamara, D. E., & Irish, J. L.
    (2018). Responding to sea level rise: Does short-term risk reduction inhibit successful long-term adaptation? Earth’s Future, 6. Keeler, McNamara, and Irish argue that to the degree to which , in good faith, local authorities pursue ”let’s plan for 50 years” adaptation strategies rather than seeing the long term trend (*), they are giving the wrong message to the real estate markets along coasts, suggesting that it’ll all be fine, it’ll all be okay, and property values will be protected. This also happens because of things like the Biggert-Waters Act and such, from the federal side, but local measures are exacerbating it. I understand the motives of local authorities: They really don’t want to create disruptive projects, and they have limited budgets. Still, in Boston, the UMass Boston study recently reported how economic exposures are severe enough that adaptation measures quite probably will cost billions of dollars.

    I think the application of Keeler, McNamara, and Irish to Boston is, at least, that the sum of the costs for a series of short term adaptation measures is likely to be higher than a commitment to a solution for the long term problem, especially if there’s something like a harbor barrier pursued. We don’t know yet what the UMass Sustainable Solutions Lab will say about such a proposed barrier, but if you look at the notes of what’s been released thus far, they budget its construction on the order of $10 billion in present dollars. And barriers kind of work only when your predictions are really good. We’ll see: Their report on the barrier is supposed to be out the end of May.

  8. Oh: I left a footnote off related to the “(*)” in the text.

    That was supposed to be a reference to a climate scientist joke which I’ve heard now at least twice in conference settings that “Everyone knows climate change stops in 2100, right?”