GMSL Data Correction: Follow-up

Brief Note by Kip Hansen — 12 February 2020


FEATURED_IMAGEOne month ago I posted a Brief Note here about a correction to the GMSL data and graphics at NOAA’s Laboratory for Satellite Altimetry (LSA).   The problem at that time concerned the data files and their associated graphics, often used here at WUWT and elsewhere as definitive and reliable up-to-date information of Global and Regional Sea Level Rise, found from this web page:



star_nesdisThis is the link:

At the bottom of the page illustrated, you see four columns of links to various forms of the data.  The data is noted to be: “Only altimetry measurements between 66°S and 66°N have been processed. An inverted barometer has been applied to the time series. The estimates of sea level rise do not include glacial isostatic adjustment effects on the geoid, which are modeled to be +0.2 to +0.5 mm/year when globally averaged.”

The different versions of the data are:


Last January, there was a glitch in the data and the graphics that resulted in 2 data points from the month of December to be missing from the files and images. They looked like this:


Up at the upper right hand corner, you can see the one latest data point, and before it, blank space where at least two other data points should have been.

I wrote to the Webmaster of STAR, Eric Leuliette, who was helpful and replied:

“For 2 cycles of Jason-3 data, our database was missing some data due to a script failure. I’ve fixed the database and rerun the mean sea level data. The replacement files are on our web site now.”

Unfortunately, it appeared that the fix broke something else, and the result of the repair was this:


It appears that the missing data points for December have been added in but according to the data file for this graphic, the point you see above at the intersection of the red lines, the first data point for January 2020,  is shown to be just under 70 mm.  The data file says it should be 63.41.

Naturally, I politely notified Eric of the new discrepancy by return email….and never heard back.


Here is the image of the same data set pictured above, download at 1400 hrs 11 February 2020.


Oops, it is exactly the same at it was on 10 January, no new data points have been added, there is no correction for the apparently missing/incorrect data for 2020.022.

In fact, checking all four of the graphics, and all four of the data files (.csv files), NONE of them have been updated since 2020.022 (one on 2020.023).   It is not only the global mean sea level data sets and graphics, clicking through to the Regional sea level time series page, we find the exact same problem — nothing has been update since the first data point in January 2020.  The ninth of February is, in digital date format, 2020.040.

For the year 2019, the .csv (comma-separated values) file shows 36 entries for the year, one approximately every 10 days.  When investigating the previous problem, there were two missing values, each about ten days apart.  This year, the last date for which there is any entry is, any of the sea level rise files (or on their graphics) is 2020.022 (in one case, 2020.023, which represents the same date).

The Official Scoop:

The webmaster at NOAA STAR/NESDIS, Eric Leuliette has been very responsive — the epitome of a public employee fulfilling his responsibilities to the general public.   Truthfully, he has been more than patient with my several emails and many questions, responding and clearing up a less-than-simple situation.

Point 1:   Why have these images and files not been updated since 9 January 2020?

Answer:  Jason-3 has been in safehold this week, so it’s likely that there will be missing data for one or two cycles.”   


CNES is investigating the source of the Jason-3 safehold.”

“The safehold position safeguards instruments by drastically reducing the satellite’s power usage, but it has the perturbing side effect of forcing DSCOVR [in this case, Jason-3 — kh] to stop transmitting data back to Earth.” [ source ]

Point 2:  Why are the data files (.csv files) different from the graphics files which share the same name?

noaa_greenAnswer:  “The plots have been smoothed with a 4-month boxcar filter. I have changed the text on the website to note this. This is the source of the differences that you noticed between the csv files and the plots. The attached figure shows the smoothed values (green) versus the unsmoothed values (red).”

As promised, the text on LSA/Sea Level Rise page has been changed to read:

“Only altimetry measurements between 66°S and 66°N have been processed. An inverted barometer has been applied to the time series. The estimates of sea level rise do not include glacial isostatic adjustment effects on the geoid, which are modeled to be +0.2 to +0.5 mm/year when globally averaged. The plots have been smoothed with a 4-month boxcar filter.” [the bolded text was added –kh]

Point 3:  Why don’t the data files (the .csv files) agree with the plots?

Answer:  “The plots have been smoothed with a 4-month boxcar filter. I have changed the text on the website to note this. This is the source of the differences that you noticed between the csv files and the plots.”

Point 4:  What?  Where are the data files with the smoothed data?

Answer:   “We have not been providing a file with the smoothed values.”

Explanation:  The data files contain data points for each “10-day pass” of the Jason-3 satellite.  Apparently, the script that produces the images (.png files) shown above, the multi-color, multi-satellite graphs, first runs a 4-month boxcar filter on the data and then puts the smoothed data on the graph — not the original data.  This is true, even though the .csv file and the graphics share the same name — they do not share the same data points.

Point 5:  When we will see updated data and images?

Answer:  “I only update the time series sometime around the 15th of each month.”

And, as Jason-3 is/was in safehold, it will depend.

 # # # # #

 Author’s Comment:

 I promised last month that I’d let you know if there was any progress on this, expecting an answer in a few days.  It took longer but now we know.

I normally wouldn’t fuss with an oddity like this, but I write about Sea level Rise and the STAR/NESDIS sites are my mainstay for up to the minute SLR data from the Jason series satellites.

I have been very impressed with NOAA’s response — from Eric Leuliette, the Center for Satellite Applications and Research (STAR) webmaster, who has gone the extra mile to answer my questions.

# # # # #

109 thoughts on “GMSL Data Correction: Follow-up

    • They are waiting until 2021. Then they can start in 2011 and say the decadal rate is now 15 mm a year

    • Years ago I wrote to UC and asked where the data before inverse barometer (IB) adjustment could be found. The said that they did not make it available to the public. How’s that for transparency, falsifiability and validation?

      When I followed up and asked why IB was applied to a global dataset and how could air pressure affect the global average ( this was what I wanted to check ) , they failed to reply and have ignored me ever since.

      I decided that if they are doing questionable adjustments that they were intent on hiding from the public, their dataset had no credibility. Yet another source of potential climate information has been hijacked by activists.

      There are other large and very suspicious changes from one version of earlier satellite series as they mangle the various data into one supposedly homogeneous dataset. They are obviously just hacking the numbers to get the result they wish push onto the public.

      Don’t spend too long trying to read sense into any of this fictional “data”.

      • The surface pressure varies throughout the year because air mass in the warmer hemisphere is pushed into the cooler hemisphere. Thus there is a sloshing of water back and forth between hemispheres also. You can see that the data rise to a peak value in the northern hemisphere winter. There is likely a dynamic effect involved as well because the flow of water back and forth between two basins is fundamentally a second order system. I suppose they have found that IB corrects for a large part of the annual variation.

      • The easiest way to get a global sea level is to calculate local anomalies and take an average. To get the local anomaly you need to know what the local sea height should be. The first approximation is the geoid, ie. the Earth’s shape if the Earth were at rest. Then you need to correct for the tides and those are pretty reliable and then you need to correct for IB and a bunch of other effects. Inverse Barometer It’s so simple. LOL

        Unlike adjustments to the temperature record, I suspect that satellite altimetry records are fairly honest. As far as I can tell, the methods are well published, similar to the link above.

    • The correct name for the institution is the “University of Colorado, Boulder” and it’s abbreviated CU, as in CU Buffs.

  1. I am not sure how important accurate data is at this point of the discussion.

    Climate Change activists seem to be simply ignoring any examination of scientific papers in favour of citing press releases, speeches from favourable lobbyists, or ‘studies’ from environmental activists. This means that they have isolated themselves from any discussion of the facts, and can simply proceed with their agendas, manufacturing ‘evidence’ to justify them as they go.

    It is hard to see how to respond to this. They have created a self-defining political movement, which is happy to ‘de-platform’ anyone who moves even slightly off-message….

      • Phil, the problem is that we have given over near complete control to the Progressive Socialists. The elitists have worked tirelessly, since the Wilson administration to get control of the necessary mechanisms to move the masses with misinformation they made great strides in the Protest 60’s and 70’s

        And just look what having control of education, media and Hollywood have done to California Phil. Voting is an exercise in rubber stamping the socialist movement and they are dangerously close to that same control nationwide.

        Dumb down the populous and frighten them with an endless parade of hobgoblins and then bolster and embolden them by teaching them that they can slay the hobgoblin and save the world by voting (D). In order to make your one word response relevant we need to defeat this 3 headed hydra the Progressive Elite created.

        Voting only works when it is given over to people capable of critical thinking and we are in short supply of those people today. Just watch a few “man on the street interviews” with our college educated Borg drones as they march in protest to the many Hobgoblins our Primary media has created. Their responses are cringe worthy Phil and these kids vote.

        Why do you think Democrats keep floating the idea of lowering the voting age even further? They are tired of waiting for the Baby Boomers to die off. They want complete control now and they intend to get it with, in two words, the vote, Damn the Electoral College, full speed ahead.

        • Yes, I was talking specifically about bringing about the death knell of the climate hoax, not the overall fraudulent lying, parasitic phony-elitism. As ever that will require more drastic action, unless the dupes can actually find ways to understand that they’ve been had.

        • Responsible enough to vote at 16, not to own a gun until—NEVER, but mostly 21. What about driving, many states reduce driving privileges below 18.

          I never hear anyone on the news/talk shows bring this up with leftists when they bring up lowering the voting age.

    • Its not just climate change activists. Our favorite “climate scientist” from Penn State resorts to the same tactics. If you watched last years debate that included Judith Currey and Patrick Moore, his entire discussion featured little more than newspaper clippings about wild fires and floods and continual references to his self proclaimed “iconic” hockey stick graph. I kept waiting for a discussion of the science, but that never occurred.

      • Marc, he is in Australia at the moment. I try to keep an open (skeptical) mind and was interested to see he had commented on his website in the last week about RCP 8.5 as reported on WUWT in the article: Climate science does an about-face: dials back the ‘worst case scenario’. His comment was:

        “Finally, let’s not forget that even a 3C warmer world would be catastrophic. Here in Australia, we’re already seeing the catastrophic impacts of less than half that much planetary warming.” I am an Australian citizen, and have spent 6+ decades here – Michael Mann – probably not six months.

        Australia has bad bushfires because it is a very dry country.

        It has always been dry.

        Look at the history of bad bush fires here:

        1851 Black Thursday 12 lives lost
        1926 Black Sunday 60 dead 1,000 bldgs.
        1939 Black Friday 71 dead 5,000 bldgs.
        1967 Black Tuesday 62 dead 1,300 bldgs.
        1983 Ash Wednesday 75 dead 3,000 bldgs.
        2009 Black Saturday 180 dead 3,500 bldgs.

        There were bad bushfires, long before global warming started. Also, in 1960 the population was about 10 million. In 2019, the population is about 24.6 million.

        We have also in the last week had large downpours, with Sydney’s Warragamba Dam filling almost 30% in a week and extinguishing almost all the fires.

        The Australian Bureau of Meteorology reports that it is all down to the Southern Indian Dipole ( a multi-decadal weather phenomenon) They (BOM) state: The IOD affects the climate of Australia and other countries that surround the Indian Ocean Basin, and is a significant contributor to rainfall variability in this region.

        Having observed and noted all that has happened in my lifetime, and then to see a tourist scientist come here and make such an ill-informed proclamation, I must agree, he has seriously discredited himself, in my view.

  2. Interesting comments, Kip, about one type of datasets in common use (occasional misuse?) in the AGW debates. Out of curiosity I searched for some hard data on satellite wave height/sea level altimetry and found several. A simple one to read the abstract of is: Yang and Zhang, Validation of Sentinel 3A/3B Satellite Altimetry Wave Heights… These guys wanted to verify that their Sentinel satellites were as good as the Jason-3 satellite, so they utilized data from 30 minute satellite crossing paths and known floating ocean buoys positions. They were delighted that their Sentinel was as good as the Jason-3! All three satellites, wait for it, are accurate to 0.3 meters (that’s 300 mm). Somehow this translates to data-posting of 1 (one) mm? I readily defer to your knowledge on this issue, but what does the large disparity between satellite accuracy and posted value mean?

    • Ron Long

      “All three satellites, wait for it, are accurate to 0.3 meters (that’s 300 mm). Somehow this translates to data-posting of 1 (one) mm?”

      You seem to confound in situ sea level measurement accuracy and its yearly rate of change.

      May I propose to read e.g. the following:

      It (luckily) doesn’t cover the brute statistics behind the job, but it’s a good place to start in better understanding the problems to be solved.

      J.-P. D.

      • Blindidon, I have worked extensively with satellite data, including X-band ground-penetrating radar data from the space shuttle. Basic statistics suggests that an accuracy of 300 mm cannot produce a value of 1 mm difference. For sure it is true that utilizing radar reflection from the waves and tides disrupting the ocean surface, and then calculating a “sea-level” measurement is a statistical nightmare, however, it cannot be defined any closer than the basic variance in the measurement technique.

        • Ron Long

          As usual, WUWT commenters know everything by far better than all people working in the domains they ‘comment’.

          Thanks btw for the ‘Blindidon’.
          Good grief!

          • Bindidon
            Why do you assume that commenters here are less knowledgeable than those currently working in the field(s)? Many of us have formal educations at least equal to those employed, and decades more of applicable experience.

            Since you seem dependent on authority for acceptance, how about this instruction from Pennsylvania State University? Note particularly the instructions for subtracting two numbers.


        • Ron Long

          Thank you for speaking up. The claims involving lots of statistics are being used to baffle brains. What they do is take a large number of measurements and get a distribution. Suppose it is Normal. Then having an uncertainty of 150mm, they can make a set of statements like this:

          Sea level hasn’t changed, confidence nm1% where m1 is a number
          Sea level has changed by x2, confidence nm2% where m2 is a smaller number
          Sea level has changed by x3, confidence nm3% where m3 is a smaller number
          Sea level has changed by x4, confidence nm4% where m4 is a smaller number
          Sea level has changed by x5, confidence nm5% where m5 is a smaller number

          Each statement gets a more and more precise x value as the confidence number gets smaller and smaller until it is a 1mm change with 1% confidence.

          The key to knowing what they really communicate is that they drop the uncertainty and the confidence from the final number. In this manner they get an answer that is impossibly precise given the instrumentation involved.

          I wouldn’t care if they didn’t spend very large sums of money based on these proclamations.

          How precise are the air pressure values used? That uncertainty has to be propagated through the calculation too. It is all very interesting but it is not providing information that is not available from land stations the 3D position of which is known very precisely.

        • “a statistical nightmare”

          You have that right Ron.

          However when one looks at the individual error sources, it is very hard to identify any that aren’t either subject to the law of big numbers or biases or both.

          The law of big numbers things like estimation of midpoint of wave height, are surely mitigated by the size of the big numbers which really are big. It appears that they start off with 187200 observations a day (20Hz*86400Sec). After removing observations over land, ice, heavy rain,passing container ships and such, they will probably have about a million observations a day. They will get substantial improvement when they average all those observations over a few hundred days. Maybe not a full four orders of magnitude — but substantial.

          Biases are constant offsets from “real” values. In theory and one thinks mostly in fact, they cancel out when one subtracts “past” values from “present” values to get the amount of change. Both (hopefully) are offset by identical biases.

          If you’re curious, I’d suggest reading the Jason3 product handbook › Products › documents › hdbk_j3 Chapter 5 goes on for many pages discussing error sources. It includes things I’ve never heard of and you probably haven’t either. e.g. Pole Tide? Turns out that there is a very small planetary scale “tidal” affect on the oceans with a period of either a year or 466 days (I’m unclear which) due to the Earth’s rotational axis wobbling a bit (“Chandler Wobble”). Anyway, these guys clearly have given a lot of thought to what they are doing.

          Where does that leave us? Personally, I suspect there is some sort of systematic problem with long term average satellite sea level estimates. The delta between the satellites and tidal gauges is way bigger than it reasonably ought to be. But I can’t tell you where the problem (if there is one) might lie.

      • Bindidon
        Your suggested link may be a “good place to start,” but it is rather sloppy in its use of terms such as “accuracy” and “precision.”

    • Ron ==> Reading the manuals for these satellite measurements is an enlightening exercise. There are many adjustments and corrections that are made to try to arrive at a sea surface height for every local reading. 300 mm is very close agreement….an order of magnitude less than most adjustments.

      • Kip
        Probably one of the least reliable corrections is the correction for the mass of the geoid, which is an approximate model, subject to changes over time. Further, it is impossible to perform altitude verifications over the oceans and the vast expanses of ice over Greenland and Antarctica. Thus, the altitude of the satellite(s) over land are most accurate, but of lesser interest than the altitudes over water.

        • Clyde ==> Satellite altimetry is extremely complex and there are so ,many variables with so many individual uncertainties – they correct for tides (in the open ocean) yet the known error of tide predictions at Tide Gauges (prediction vs actuality) is shown at every TG station in the Tides and Currents collection.

          Currently, for the Battery in NY, there is a difference in tide prediction and actual water level of 0.8 feet — over 9 inches. (see Water Levels chart)

          I don’t how they correct for the uncertainty of the tidal wave (the tide as it moves across the ocean) if the uncertainty is of that magnitude.

          • Kip, I appreciate your posting this data. I went directly to the NOAA website and read the Jason2/3 Users Manuel. Saying that “satellite altimetry is extremely complex” is right on the money. One of the biggest problems, mentioned by Clyde Spencer above is the reference geoid. NOAA still uses the old WGS84 reference geoid, uncorrected for bulges and distortions due to the molten core sloshing around and discharging heat and magma, etc. The radar altimetry is assisted by retracking, wherein checking on a reference point is utilized to adjust the altitude above the reference geoid and the reflecting surface, which is the sea surface, the intended target. NOAA claims 3.4 cm accuracy (34 MM) which only means that correcting for wave activity, tide effect, velocity of radar signal in dry versus humid atmosphere, etc allow math processing to correct all of this to an accuracy of 3.4 cm. The other persons I cited above actually accept a 30 cm accuracy at any instant and stand-alone data point. If we take NOAA at their word and accept their 3.4 cm accuracy it is still an order of magnitude larger error than cited results.

          • Kip
            Tidal waves are similar to wind waves in that the amplitude is related to the depth the wave is traveling through. If memory serves me right, the Bay of Fundy has tides of about 50 feet because of the depth and resonance, while tides on islands in the Pacific may be about one foot. Tsunamis hitting coasts may be of a similar multiple-foot amplitude, yet be only 6 inches in the open, deep ocean.

            Incidentally, you might find this link to be interesting:

          • Clyde ==> Yeah — caught that story earlier — being that I have spent half my adult life on ships and boats, these things catch my eye.

            My wife and I have had experience with rogue waves in the Caribbean — not the GIANT rogues, but 10-15 feet rogues in an otherwise flat sea. One time we were napping in our forward cabin, cabin-top hatches wide open to catch the breeze, and ad a rogue brek over the bow and cabin-top dumping about 50 gallons of sea water in through each hatch, directly onto our bunks and in my face. My son was at the helm and never saw it coming until it hit.

            Luckily, it was directly bow on or it would have tossed us violently side-to-side (being a cat…)

          • Kip. When you do the arithmetic, You discover that water doesn’t move fast enough to actually move a water bulge around the Earth at the Equator(40000km circumference/24hr = 1666kph = about 1000mph). Not to mention the affect of the continents blocking the flow. So the offshore tides moves in complex patterns that are predicted by an elaborate equation that looks to be a 3D spherical Fourier series. There’s a long history on that including the work of many big names in 18th and 19th Century physics — Simon Laplace, Lord Kelvin, etc. I’ve never looked at it in depth. Wikipedia has an article.

            Their equation seems to work.

            Inshore tides are even more complicated as they are affected by where the water is, where the shorelines are, water depths, resonances,etc. My impression is that “they” mostly just measure the tides at some location day after day after day and when they have enough data, they fit a curve to it. (Then they adjust for known long term cyclic affects?). Could be wrong about that.

          • Don K ==> Tides are predictable (and oddly chaotic at the same time). It is their magnitudes that are no so predictable, at the shore, predictions are often off by 25%.

            It is not my understanding that tides are caused by moving water . . .

    • Ron, thanks for the link. Here’s the quote:

      The validation of significant wave height (SWH) data measured by the Sentinel-3A/3B SAR Altimeter (SRAL) is essential for the application of the data in ocean wave monitoring, forecasting and wave climate studies. Sentinel-3A/3B SWH data are validated by comparisons with U. S. National Data Buoy Center (NDBC) buoys, using a spatial scale of 25 km and a temporal scale of 30 min, and with Jason-3 data at their crossovers, using a time difference of less than 30 min. The comparisons with NDBC buoy data show that the root-mean-square error (RMSE) of Sentinel-3A SWH is 0.30 m, and that of Sentinel-3B is no more than 0.31 m.

      So the standard deviation of the wave-height errors is about a third of a meter.

      Finding information on the precision of the Poseidon-3 altimeter is surprisingly difficult. I find the following:

      A posteriori analysis of the tracking information from the four sources has demonstrated a satellite position determination capability in the < 5 cm altitude range and the calculation of ocean height within 1-2 cm every 10 days. The multi-source tracking strategy of T/P is a first in high-performance POD analysis. Many other projects followed this hybrid POD approach.

      Measurement             Required accuracy      Achieved accuracy
      Altimeter                           4.0 cm        3.2 cm
      Satellite position                 12.8 cm        2.8 cm
      SSH (Sea Surface Height)           13.4 cm        4.3 cm 


      Orbit determinations with a precision of <5 cm on the radial distance component are achieved on the TOPEX/Poseidon mission.

      OK, so they’re saying that they can measure sea surface height to within 43 mm …

      My main problem in all of this is that the satellite is at an altitude of 1336 km, or 1.3 billion mm. This means that to get ± 1 mm accuracy, it needs to measure the sea surface height to better than one part in a billion. That’s difficult to do in a lab with everything clamped to a bench, much less to a moving surface from a moving sensor through a varying medium.

      Call me crazy, but I’m not seeing it. It’s like saying OK, I’ll build a platform that juts out over the open ocean. It’s 10 metres up, out of the reach of the waves. But it moves up and down randomly ±5 cm. Every ten days, I’ll measure the average wave height to the nearest 300 mm, and I’ll measure the average surface height to the nearest 43 mm … and from that I’ll calculate the true sea level height to ±1mm …

      Yeah, that’s totally legit …

      Heck, even the folks that do it can’t agree. Figure 2 of the Cazenave paper that Bindidon linked to shows the sea level results from University of Colorado and from CSL Aviso.

      I looked at that and said “Huh?”

      So I digitized the graphs, set each of them as anomalies around the mean, and compared the two. Standard deviation of the difference between them is 1.6 mm … and the difference is wildly non-normal (skew = 0.35, kurtosis = -0.16) so we can’t use the law of large numbers.

      So me … I don’t trust it all that much.

      Best to all,



      • w. ==> I trust them — but my impression is that the accuracy/precision claimed is the results of “Over-Mathematic-ation” — computational hubris — getting results that agree with trheir previous results through adjustments that guarantee the agreement.

        CGPS stations require years of data to arrive at reliable VLM estimates in the single digit millimeter range.

      • w. ==> I just emailed Nerem at CU Boulder and asked about their sea level site: “CU Sea Level Research Group”

        It hasn’t been updated in almost two years. He replied:

        “We haven’t updated it in a while because of some server issues, but we are hoping to resolve that by the end of the month. Steve”

        For what is worth……

        • I contacted him awhile ago -around 30 months about essentially the same thing. At the time his global SLR was showing 3.4mm/ yr. He emailed me back that they were changing databases and would have new results next month. A month later it came out as 3.1mm/yr. and held a press conference blaming Mr. Pinatubo volcano for the distortion.

          • James ==> They may have just quit including the GIA “adjustment” which does not add to sea surface height at all.

            It is not clear on their old charts.

    • I went back to Yang and Zhang and read their article about Sentinel, which is a European Union Satellite designed to measure wave-height and sea level, and think this whole satellite data set is only accurate in the minds of some true believers. The data is actually spectacular, just not suitable for measuring millimeter wave/level changes. They are thrilled that their Sentinel data, checked by crossing a Jason-3 track where there is a high-resolution ocean buoy, is within 0.3 meters. Then, wait for it, you massage (torture? interpret? subject to the famous algorithim attack? krieg? interpret taro cards?) the data and it ends up very accurate. How accurate? NOAA says 3.4 cm, but reduced to 2.5 cm, as a spot value, by prior listed torture. Then you show yearly sea level rise accurate to 1 (one) mm.

      When I was District Geologist for Pegasus Gold they asked me to evaluate the proposed extension of the Florida Canyon gold mine south into the ASARCO zone. I told them your production onset is wrong, there is no ore on the second bench. They told me “you don’t know the power of krieging”, which I had to admit was true, but reiterated that there was no ore on the second bench because no drill hole showed this. Three months later an alarming production short-fall, because there was no ore on the second bench. Same deal here: if it is not in the real data it is not real.

  3. “The mission objectives call for the provision of the same measurement accuracy of Jason (3.3 cm) with a goal of achieving 2.5 cm, and to maintain the stability of the global mean sea level measurement with a drift less than 1 mm/year over the life of the mission.”

    That is considerable accuracy from orbit to the sea surface.

    I note the orbital period is 9.9 days so it take 99 days before the same sea height measurement can be repeated. Which means each grid cell of sea height is measured 3.6 times per year. Not many measurements for the law of large nu,mbers to kick in.

    However, climate science will manage to get the correct result so we can sleep safe in our beds.

    • Steve Richards

      I note the orbital period is 9.9 days so it take 99 days before the same sea height measurement can be repeated. Which means each grid cell of sea height is measured 3.6 times per year.


      Circular non-sun-synchronous orbit; 1336 km altitude (2 hour period), inclination = 66.038º, 9.9-day repeat orbits/b> (127 revolutions), ground track repeatability = ±1 km cross-track at the equator.
      Jason-2 is scheduled to join Jason-1 in the same orbit
      with a 10 day repeat observation cycle (9.9156 days to be precise, i.e., 10 days minus two hours).

      Your very personal way of capturing and interpreting information is extremely interesting!

      “However, climate science will manage to get the correct result so we can sleep safe in our beds.”

      Indeed! Yes we can, so to speak.

      J.-P. D.

      • Scott ==> I believe that the 3.6 should be 36 … still a small number and a great deal of uncertainty.

        • Kip

          ” … still a small number and a great deal of uncertainty.”

          Did you ever think of the fact that the revisit interval on which satellite temperature data is based, is not so terribly different?


          • J.P. ==> Actually, yes — and it represents the same sort of problems, though I believe there are fewer confounders in the sat temp calculations. In neither case are they actually measuring the thing they report — sat altimetry measures “time-for-signal-to-return” and “signal-scatter” — sat temp == “They measure radiances in various wavelength bands, which must then be mathematically inverted to obtain indirect inferences of temperature.”

            so, yeah, I have thought about it — and have the same sort of reservations.

          • I’m going from memory here, but as I recall, the temperature MSUs have quite a wide field of view — many tens of km. Or more. The satellite Radar Altimeters OTOH have a quite small FOV. “Highest” (“closest” if you want to be picky) point in a small region “directly” under the satellite.

            In any case, “revisit” is perhaps slightly misleading term. Yes, it’s (almost) the same latitude longitude. But the currents, air pressure, air and sea temperatures, prevailing wind, sea conditions, etc have changed. In some sense it’s not quite the same ocean.

    • So with the 3.6 times a year the waves and troughs average out and Gondwanaland with its tide gauges at Port Arthur and Fort Denison are floating up with the accelerating sea level? OK got it.

        • ‘You just need to correctly read the stuff you access’

          I know I have trouble concentrating on this high tech space stuff mate. I put it down to seasickness on the good ship Gondwanaland and continental drift doesn’t help matters.

  4. This whole scam moved from being driven by ‘climate science’ as speculative as it may have been in the early days, to one almost entirely of ‘science communication’ where there is endless potential narratives along themes of catstrophe and so on that can be confected and fed to the media chickens who duly crap it out into the public domain.

    The seeming incompetence in dealing twith sea level data as evidence by this article is just another peek into the workings of the scammers. Proper, rigorous scientific output is no longer assessed by their kpi regime so is no longer attended to except when the odd fig leaf may be required.

  5. Kip Hansen

    A fair head post, appreciated.

    My guess, if you allow: the vast majority of all these collaborators are doing their job in about the same way.

    The fact that so many people disagree with the results of their work (for social / political reasons and not based on scientific criticism) should in turn not lead to them being discredited, as is all too often the case, here and elsewhere.

    J.-P. Dehottay

    • Bindidon, for the record I am a multiple-degreed scientist and have practiced science my whole career and have been paid handsomely for results.

        • MarkW

          “..with those bindidion is paid to support,… ”

          Thanks for this nice little lie I personally consider to be not far from an insult, but I’m aware of your average comment style on this blog, and prefer to ignore it.

          Btw, what about publishing such things under your real name, Sir?

          J.-P. Dehottay

      • Ron Long

        “… for the record I am a multiple-degreed scientist… ”

        1. I hope that this ‘multiple’ included deep knowledge in statistics 🙂

        2. As far as I can recall, a ‘vast majority of persons’ does by no means include all persons.

        My little finger tells me that you in fact perfectly know what I meant in my comment to KH.

        J.-P. Dehottay

    • J.P. ==> I don’t think that the situation is that so many people here disagree with the satellite derived Sea Level data. There is a lot of discussion about the two major issues:

      1. The degree of accuracy and precision claimed by the various satellite sea level groups — what I personally refer to as “errorless sea level”.

      2. The disagreement between the Tide Gauge record (now, the TG record corrected by CGPS) and the satellite record — satellite S:R pegged at 3+ mm/ye, while TG corrected running at 2 mm/yr. This issue remains unresolved.

      • Kip Hansen

        – Concerning (1), that is the real problem here.

        – one trusts in other people’s work, or
        – one REALLY falsifies it (I mean with the same degree of data evaluation and argument intensity as shown by the source, and not simply a little guest post quickie).

        Anything between these two alternatives has imho zero value.

        – Concerning (2), I can’t agree with your numbers.

        We have to compare apples with apples in a temporal sense.
        It is correct that sat altim gives 3+ mm / year. But… for 1993-2018.

        Let us please compare the data for the period common to all evaluations known to me

        – Church & White
        – Jevrejeva
        – Dangendorf
        – Foster
        – my own layman’s job

        and which I have on disk in spreadsheet files: this is Jan 1993 – Dec 2009.
        Of course because sat altim data starts with Jan 1993, and Jevrejewa’s data ends with Dec 2009.

        My rough evaluation job excepted, all time series integrate vertical land movement data; the most impressive job was done by Sönke Dangendorf & colleagues.

        { I don’t have tide gauge data exclusively restricted to stations with GPS ‘on board’.

        Feel free to provide me with a link to such data, I’ll integrate it into my little collection. }

        Here are the linear estimates for this period, im mm/yr.

        1. Sat altim
        – NASA: 2.63 ± 0.06
        – NOAA: 2.58 ± 0.04

        2. PMSL
        – C&W: 3.20 ± 0.07
        – Jev: 3.12 ± 0.09
        – Bin: 2.73 ± 0.17
        – Dang: 2.61 ± 0.03
        – Foster: 2.33 ± 0.12

        C&W and Jev clearly are the outliers.

        Here you see the match sat altim vs. tide gauges for 1993-2015 (end of Dang):

        On request I add links to all data sources.

        J.-P. D.

        • J. P. ==> If you’d like to publish a comprehensive essay here on the topic, I’d be glad to forward it to the editors here for consideration. You can email it to me at my first name at the domain

          NOAA STAR currently gives 2.9 +/- 0.4m/yr for TOPEX/Poseidon, Jason-1, Jason-2, and Jason-3. Media quotes NOAA/NASA at 3.2mm/yr (taking numbers which have GIA added in – 2.9 + GIA) — and calls this a doubling of SLR rate over that of the 2oth century.

          Tide Gauge records are of Relative Sea Level and are not suitable for determining Global SLR unless they have been fully corrected for VLM by CGPS at the Tide Gauge (mounted on the same structure).

        • Kip Hansen

          Did you not understand what I wrote?

          The goal was to compare apples with apples in a temporal manner, i.e. based on the same reference period, here: 1993-2009.

          Moreover, you write:

          “Tide Gauge records are of Relative Sea Level and are not suitable for determining Global SLR unless they have been fully corrected for VLM…”

          This is more than 100% evident for all the people who evaluated PMSL, me included.

          “… by CGPS at the Tide Gauge (mounted on the same structure).”

          This is your claim. It must be proved by worldwide data comparisons.

          J.-P. D.

          • J.P. ==> take your time and read up on it — why the TG data MUST be corrected with CGPS@TG to provide any useful data about absolute sea levels.

      • Just so, Kip! Regardless of the disdainful and ad hom responses of a few, the issues you so clearly illuminate remain unresolved.

    • What arrogance! You know all our backgrounds and knowledge, and why we hold our views.

      The simple fact that you believe that utterly discredits you, not us.

  6. Kip,

    The four month box-car filter is interesting. Do you know, is it applied to the current data point using the previous four months (with the effect of a phase shift), or are the four months centered on the current point (with the effect that points at the end of the series have to be treated differently)? By four months, do they actually mean four lunar months or something else?

    Inquiring minds want to know.

    • Kevin ==> I only have the information in the essay, taken as direct quotes from Eric at NOAA. I assume (safely I hope) that it is “the current data point using the previous four months” — using future data is tricky and invalid in almost all cases.

      I can add this question to my developing list of things at ask — once I think Eric has re-filled his patience tank.

  7. Re: “the first data point for January 2020, is shown to be just under 70 mm. The data file says it should be 63.41.”
    and “The plots have been smoothed with a 4-month boxcar filter.”

    If the plot has really been smoothed with a four-month (nine data point) boxcar filter, there should be no data plotted for the last two months, because a 4-month boxcar filter requires data from the the following two months, which are obviously unavailable. Yet their graph clearly shows data after 1/1/2020.

    I loaded their data into an Excel spreadsheet, added a column to calculate the 9-point boxcar smoothed values, and plotted them, here:

    It looks a bit different from their plot.

    • Dave ==> I did something similar when communicating with eric at NOAA, but if one rescales everything, they work out to be “close enough for CliSci.”

      I think Eric using the “previous four months including final point” for the boxcar.

      My question is “WHY smooth at all?” I will ask, when I think Eric has caught his breath.

      • Then, following Dave’s question, and in relation to my prior question, there is a phase shift resulting from this filter, and in fact, one might imagine there is a two month shift in peak values in the smoothed series, but it would be nice to know for sure.

        • Kevin ==> The two plots, all data and smoothed, are so similar that it is hard to tell them apart except by very close examination. That’s why I thought it was just a script error.

          So I don’t know for phase shifting — but given how close the plots are to one another — it hardly makes a difference.

          In fact, it makes SO LITTLE difference, I don;t know why Eric does it at all.

  8. Kip, there is no shortage of data sets of all kinds that aren’t being updated at all. Sea ice, Enso, etc. Look at the sea ice for Antarctica. It was expanding rapidly (NSID) recovering to its growth mode of the 1979 ~2015 period. All sets from Japan, Norway and US are in limbo. Ozzie BOM used to report ENSO like clockwork every Monday and now weeks go by.

    Re isostatic rebound, why is this reported as a “correction” to actual sealevel. This is a volume component. We now have sealevel results that are somewhere above sealevel and growing .2-.5mm a year into the atmosphere.

    • Gary ==> NOAA STAR no longer adds in the GIA adjustment: “The estimates of sea level rise do not include glacial isostatic adjustment effects on the geoid, which are modeled to be +0.2 to +0.5 mm/year when globally averaged. ”

      This is proper, since the GIA Adjustment does not actually affect sea surface height.

      I’ll have to check some of the other non-updating online data sites and try to figure out what the deal is.

      All that depend on JAson-3 will be lagging a week or two or three behind if Jason-3 is in safehold.

  9. Why we think a satellite can measure sea levels to millimeters from 1336 km up is beyond me anyway.

    Why not create a tide gauge index from some number of reliable tide gauges with history around the globe, each with a means to measure isostatic adjustments and call it a day? It will certainly be more accurate that satellite measurements, we won’t have to worry about incomplete data and it would be believable (and cost less)

    • rbabcock

      “Why we think a satellite can measure sea levels to millimeters from 1336 km up is beyond me anyway. ”

      Not only you have to live with that. Me too.
      But… it is not a reason for me to doubt about the correctness of other people’s work.

      If you don’t trust in satellite data, maybe you look at this corner…

      Anyway, I downloaded last year the entire PMSL data, generated an anomaly time series out of that (using a simple layman’s approach) and was surprised to detect how near the trends were between NOAA’s satellite data, Dangendorf’s tide gauge evaluation in 2016 and my layman’s evaluation for the period 1993-2018.

      The most amazing detail was that at that time, I didn’t integrate vertical movement data: it seems that all that averaged out worldwide. Maybe one day I finish that work.

      J.-P. D.

        • Malcolm Chapman

          Yes, approximatively: in fact, I should write ‘Bendidon’ like ‘Ben dis donc!’ but I preferred the ‘i’ variant.

          Thanks for the comment, appreciated.

          J.-P. D.

      • J.P. ==> “The most amazing detail was that at that time, I didn’t integrate vertical movement data: it seems that all that averaged out worldwide. ”

        The problem with that idea is that MOST VLM at tide gauges is not continental movement but movement of the structures on which tide gauges are mounted and their very local environments. Piers have a known VLM of – 1mm/year if built on wooden pilings. Most docks at seaports are built on filled areas, thus subside over time. Continental movement either adds or subtracts to that very local movement.

        This is why PSMSl and others are pushing the CGPS@TG (same structure) program — be be able to correct for that very local VLM.

        When we are dealing with measurements as small as single-digit millimeters per year, it is the local details that count the most.

      • Not only you have to live with that. Me too.

        I don’t have to live with it, I have a remote sensing degree, and the claims of accuracy are ridiculous.

        The first rule of remote sensing on Earth is checking your results in situ. The satellites are an accessory, a shiny object, and very few of those who discuss satellite data accuracy have any idea of what they speak. But clearly that does not stop them.

        • Gator

          “… and very few of those who discuss satellite data accuracy have any idea of what they speak. ”

          Such arrogant comments I simply discard, Sah.

          Come up with something really scientific akin to your degree, with valuable data sustaining what you pretend behind you pseudonym, and we’ll see.

          J.-P. Dehottay

          • We are talking measuring millimeters from over 1300 km on an ocean surface that is rarely flat and might have 15 meter swells and 10 meter waves. Throw in drift in electronics, satellite drift, temperature extremes in space affecting sensors, high and low pressure systems over the oceans, winds pushing water here and there and who knows what else and there is no way you can measure a rise or fall of an ocean surface of 1 millimeter. No way.

            So please get real. With all the variables I don’t have to prove it is wrong, those running the satellites have to prove it’s right.

          • Such arrogant comments I simply discard, Sah.

            Do you even read what you type? LOL

            Sorry JP, but extraordinary claims require extraordinary evidence. My claim is ordinary.

            I would tell you to go get stuffed, but apparently you are already full of yourself.

            God this is fun!

          • rbabcock

            “So please get real. With all the variables I don’t have to prove it is wrong, those running the satellites have to prove it’s right.”

            You very well have to prove it is wrong: simply because when I read stuff like

            “We are talking measuring millimeters from over 1300 km on an ocean surface that is rarely flat and might have 15 meter swells and 10 meter waves. Throw in drift in electronics, satellite drift, temperature extremes in space affecting sensors, high and low pressure systems over the oceans, winds pushing water here and there and who knows what else…”

            I begin to believe, with all due respect, that you probably never have read any paper about satellite altimetry (I don’t mean here: scanning its abstract).

            The problems to solve there are of totally other dimension, e.g. gravity.

            And when I read

            “… and there is no way you can measure a rise or fall of an ocean surface of 1 millimeter. No way.”

            I definitely understand that you in turn don’t understand the basic difference between

            – 30 meter high ups and downs in intervals measured in a couple of seconds, and
            – a linear estimate calculated over a period of say 20 years.

            But I know: ‘skeptics’ are ALWAYS right, even without having taken care of really collecting information before criticising.

            Please, rbabcock: google for sat altimetry, and manage to read some papers.


            I sincerely hope you understand that viewed from space, the ups and downs at sea surface are very, very similar to the ups and downs of 60 GHz O2 emissions within e.g. the lower troposphere, where you have to cope, among hundreds of other things, with e.g. extremely strong convection and advection streams?

            If you reject the ability of people to calculate sea level trends down to a few mm/year, how then can you accept that other people are able to calculate the temperature trends for four different atmospheric layers, down to 0.01 °C?

            J.-P. Dehottay

    • rbabcock ==> To be fair, the whole Poseidon/Jason satellite project has a far broader purpose than the simple measuring of sea surface height and sea level rise and fall.

      There are international projects afoot to measure global sea level change using tide gauge data corrected for Vertical Land Movement — but there are not enough “Continuous GPS stations on same structure as Tide Gauges” (CGPS@TG(ss) to derive global estimates. CGPS requires several years of data to make a good determination of VLM. In a few years, we may have something to compare to the satellite record.

    • You are correct on all counts, rbabcock.

      The advantages of tide gauges over satellite alimentary are that their measurement records are much longer, their data is much more reliable, they measure sea-level where it matters (at the coasts) instead of where it doesn’t matter (mid-ocean), and they cost a lot less than satellite altimeters.

      From alarmists’ perspective, though, the satellite altimeters have two advantages: their high cost and complexity makes them much cooler toys, and their data are infinitely malleable, so it can be made to support whatever narrative you wish.

    • rbabcock, “Why not create a tide gauge index from some number of reliable tide gauges with history around the globe”

      “They” thought they’d done exactly that. Except when they started getting reasonably credible satellite based sea level data, it was coming in at 50% or more higher than the tidal gauge data. That called attention to possible problems with the tidal gauges. There are, as it turns out, a LOT of potential problems with tidal gauges. The tidal gauges were, by and large intended to make sure that ships didn’t bang their keels on the bottom. They were not intended to be high precision devices. But some have a long time span so they should be a good cross check on the shorter term readings from the satellites. So now “they” are trying to improve the tidal gauge results in hopes that once corrected for local tectonics, the numbers will be closer.


      BTW — I wouldn’t get too hung up on the satellites being in 1300km orbits. As I understand it, that was done to minimize drag and thus reduce the need for station-keeping orbit adjustments which would complicate the data analysis. It doesn’t need any complicating. Once one gets above the atmosphere, the speed of light is effectively constant and distance adjustment is (in concept anyway) just a matter of adding or subtracting a constant times the distance from the satellite to the top of the atmosphere.

      • Don K

        “Except when they started getting reasonably credible satellite based sea level data, it was coming in at 50% or more higher than the tidal gauge data.”

        Sorry, Don K: this is simply not correct.

        The people who pretend that made the mistake of comparing the tide gauge trends for their average lifetime (less than 2 mm/year) with the trends of the satellite data during the satellite era (currently over 3 mm/yr).

        As you certainly know, we have to compare trends of data coming from different sources for the same reference period, whenever the data is presented in anomaly form.

        The difference then disappears.

        Here are the linear estimates for the reference period 1993-2009, in mm/yr.

        1. Sat altim
        – NASA: 2.63 ± 0.06
        – NOAA: 2.58 ± 0.04

        2. PMSL
        – Church & White: 3.20 ± 0.07 (outlier)
        – Jevrejeva: 3.12 ± 0.09 (outlier)

        – Dangendorf: 2.61 ± 0.03
        – Foster: 2.33 ± 0.12
        – Bin: 2.73 ± 0.17 (own evaluation, very rough layman’s work)

        Here you see the match: satellite altimetry (NOAA) vs. PMSL tide gauges, this time for the reference period 1993-2015 (2015 was the end of the Dangendorf evauation):

        J.-P. D.

        • Bindidon

          You have a point. To an extent.

          But I don’t agree. In the case of C&W and Jevrejeva you are comparing subsets of tidal gauges that appear to be cherry picked to get the answer establishment “climate science” wants to see. But, if your thesis were correct, EVERY tidal gauge that doesn’t have some special consideration like altered rate of underground fluid (oil or hydrocarbon) removal. would show a similar modest change in SLR starting about 25 years (?) (or more) ago.

          I just don’t see that in the data.

          It’s true that the data is noisy. And it also seems to be true that there is no reliable math for detecting small changes in rate of change in noisy data. But would I think that a 20%-50% change in SLR rate would be easily visible.

          I suppose I could be wrong. Maybe some day I’ll get around to running some simulations to confirm my gut feeling.

          In the meantime, I’m not buying it. But, as Kip argues, once they get local tectonics at the tide gauges sorted out (in the next decade hopefully?), the nature of the values and extent of the uncertainties should be much clearer.

    • Mosher ==> I almost never understand what you are saying….it has become a point of amusement to me — part of the pleasure of dealing with the comment section. Thanks for brightening my day.

      • Kip
        You are assuming that Mosher understands what he is trying to say. His terse, unintelligible remarks may have nothing to do with his presumed mastery of the English language.

      • It’s climate science mumble rap .

        You got to get into your climate gangster persona and do those drive by’s.

        Heaven help us if he gets into climate emo rap.

  10. Where’s Stokes, Moser, & some of the regular trolls? Remember how they criticized the UAH satellite temp record for a minor error or two (which were immediately corrected)? And now this gmsl measurement is apparently riddled w/errors & problems….

    Rhetorically waiting for their criticisms.

    • beng135

      I agree ith you concerning Steven Mosher’s style, keeping silent would be better than what he writes. Sometimes he writes really meaningful things, but… hmmmh.

      Concerning Nick Stokes, I don’t.

      Simply because his criticism about UAH had nothing to do with such microscopic details, but with the difference between UAH5.6 and UAH6.0 {beta5 at that time}, which was far bigger than the alleged ‘huge adjustments’ made on GISTEMP.

      Here is his graph comparing the two adjustment contexts:

      Both UAH and GISTEMP will have had their valuable reasons for adjusting.

      J.-P. D.

      • If that different versions graph wasn’t meant to mislead, it would at least have done the same thing with both data sets: either before and after plots of each or difference plots of each. The best presentation would have been three plots for both: before, after, difference.

        As is one can only guess what significance there might be in the proposition.

    • beng ==> Mosher weight in with “Man bugging a website master about science data”.

      If you can translate Mosher-isms, I’d appreciate it.

  11. As just an interested (but not at all knowledgeable) observer in the proceedings here, I’m thinking that the next development in climate ‘science’ might be that plastic surgeons will feel they are well qualified to apply their skills to raw data sets to ensure they have the right look before being presented to the media.

    (Well, if psychologists can claim expertise applicable to climate ‘science’, why not plastic surgeons?)

  12. Most of us deplorables casually follow coastal RE prices and figure no tipping point yet for the movers shakers and omniscient overlords.

Comments are closed.