Sea level rise and the “urban sinking” effect

Most of us are familiar with the “urban warming” effect that can cause temperature stations in and around urban areas to measure warming gradients that aren’t representative of the surrounding areas. Here I present evidence for the existence of an “urban sinking” effect that can cause tide gauge records in and around urban and other populated areas to give similarly non-representative results.

I begin with a plot of three tide gauge records on the east coast of the Malaysian Peninsula, none of them in or close to urban areas. The data aren’t as complete as I would like but I feel confident enough to use them to estimate an average for this particular stretch of coastline (black line):

Figure 1: Three tide gauge records from Malaysia-Thailand (Ko Lak, Ko Sichang, Cendering). Black line shows the average.

Now I superimpose the tide gauge record for Bangkok (Fort Phrachula Chomklao), a couple of hundred kilometers north of Ko Lak at the head of the Gulf of Thailand:

Figure 2: Figure 1 data with Bangkok (Fort Phrachula Chomklao) tide gauge record superimposed in green.


As noted in an earlier post Bangkok is sinking below the waves, but not because of man-made sea level rise. It’s sinking because of excessive man-made groundwater extraction.

Calcutta (Diamond Harbour) has had the same problem, although not quite so extreme:

Figure 3: Calcutta tide gauge record vs. average of eight other Indian Peninsula tide gauge records (Vishakhpatam, Paradip, Madras, Cochin, Bombay, Karachi, Okha, Marmagao)

But Manila is about as bad, although there are no adjacent tide gauge records to compare it with:

Figure 4: Manila tide gauge record

These are of course extreme cases, easy to identify and weed out. But we see the same thing on a smaller scale in other places. An example is Venice:

Figure 5: Venice tide gauge record vs. average of nine other Mediterranean records (Malaga, Ceuta, Alicante, Marseilles, Genoa, Trieste, Bakar, Split, Dubrovnik)

Another example is Chesapeake Bay in the eastern USA. Figure 5 compares Washington D.C. at the head of the Bay with Hampton Roads at the entrance. Since 1960 sea level at Hampton Roads has risen by almost 100mm more than it has in Washington, again as a result of subsidence induced by groundwater extraction:

Figure 6: Washington DC vs. Hampton Roads tide gauge records

And then there’s this example from California. I don’t claim that this is necessarily a case of “urban sinking” because it could be caused by differential tectonic movement. Nevertheless it’s an interesting result:

Figure 7: Mean of four urban tide gauge records in California (San Francisco, Santa Monica, La Jolla, San Diego) vs. mean of four non-urban tide gauge records (Port Reyes, Monterey, Port San Luis, Ensenada)

It would be good to have more examples but there just aren’t that many tide gauge records. However, the above results document the existence of a man-made “urban sinking” effect that impacts only limited stretches of coastline and gives inflated estimates of sea level rise that aren’t regionally representative. The question is, how widespread is it?

Well, a disproportionate number of tide gauge records, and the bulk of the longer and more continuous ones, are located in large port cities where subsidence caused by groundwater extraction might be expected to be occurring or to have occurred. Added to that is the possibility that gauges mounted on piles driven into unconsolidated sediments in busy ports might subside over time because of vibration from ships and heavy machinery, although I have no hard evidence for this. The problem is not so much the easily-identified cases like those shown above but those where the subsidence is on the order of tenths of a millimeter a year, or where it has occurred over limited periods, or where a number of records in the same area are all distorted by about the same amount. These cases would be difficult if not impossible to identify, but enough of them could result in significant overestimation of both relative and absolute sea level rise.

Another question is exactly what level of overestimation we might be talking about. This question is presently impossible to answer, but regardless of how widespread urban sinking is it will result in some overestimation of sea level rise. There are many places along the world’s coastlines where water and other fluids are being pumped out of the ground in large quantities but few if any where they are being injected.

This entry was posted in Climate change and tagged , . Bookmark the permalink.

8 Responses to Sea level rise and the “urban sinking” effect

  1. singletonengineer says:

    A couple of years back, when I was living in Beijing, I read of surface subsidence of up to 180mm due to dewatering compressible sediments. Figures above 50mm per year seemed to be common.

    Surely there are records for the mouths of the Nile, the Ganges and the mouth of the Mississippi (ie New Orleans). If they have not subsided substantially I will be very surprised.

    Other factors on deltas: Lack of sedimemt deposition. Erosion due to wind and waves. Erosion due to concentration of river flows between levees, thus increasing flow velocities in some places and starving other areas of water and silt.

    This is much more than just the odd millimetre or two of nicely worded “urban sinking”. It is compaction and erosion on a massive scale.

    • The Nile delta is an interesting case. Alexandria is indeed subsiding and the tide gauge record shows it, but according to the study below the subsidence was caused by compaction of unconsolidated delta sediments, not groundwater extraction.

      The entire US/Mexico Gulf Coast shows anomalously high sea level rise because of subsidence induced by a combination of sediment compaction and fluid (water and petroleum) extraction. Sorting out how much of the subsidence is natural and how much man-made is not easy, however.

  2. Euan Mearns says:

    Roger interesting post, a couple of questions. 1) Do you have documentation of ground water withdrawal being cause of urban subsidence or is this an assumption you have made? 2) Is it possible that urban loading can cause isostatic subsidence given right geology? i.e. the mass of built infrastructure. and 3) Are these extreme cases removed from data sets used by the IPCC? How does the IPCC manage all the variables.

    I liked the way the data are managed in this chart. Any thoughts?

    • Euan: Here are a couple of refs on sea level and subsidence. They both say that rising sea levels makes things worse but SLR effects are negligible compared to the subsidence.

      Urban loading can cause local subsidence, as it has in Shanghai, but it won’t affect a tide gauge unless the gauge is close to the urban loads.

      The IPCC presents a number of tide gauge reconstructions done by others, like Church & White and Jeverejeva et al. It doesn’t actually do anything itself. I can’t find a list of C&W’s stations so I don’t know if they removed (or corrected) stations like Bangkok or not.

      The graph is interesting but more data are needed to make anything of it. Which stations are included and which aren’t? Are the data corrected? Are the corrections based on GPS or GIA? What time interval is used? And so on and so forth.

    • Peter Shaw says:

      I too like this data presentation. I’d be comfortable with the median of the whole set (1mm?). Is there a lurking statistician willing to comment?

  3. oldfossil says:

    So the urban heat islands are sinking, just as the iPCC predicted…

  4. johndroz says:


    Good post on subsidence. My main critique would be that you should make clear that subsidence can be natural as well as man made.

    For example, you section on the Chesapeake area says “…again as a result of subsidence induced by groundwater extraction” where it would be more accurate to say “…again as a result of subsidence partially induced by groundwater extraction.”

    • Roger Andrews says:

      Hi John: Yes, I probably should have slipped in a word like “differential”. The east coast of the US is indeed sinking naturally, I believe mostly because of the compaction of unconsolidated sediments in coastal areas. The groundwater pumping in Chesapeake Bay just adds to it.

      But it’s a different story on the other side of the continent. The west coasts of North and South America are rising tectonically as the American plates ride west over the Pacific plates. According to the tide gauge records there’s been over 100mm of differential movement between the west and east coasts of the Americas since 1930.

Comments are closed.