WUWT has stepped up their ongoing campaign to downplay the threat of sea level rise. This includes a recent post by Larry Hamblin which indulges in a just-plain-wrong method for pushing the the “no acceleration” meme, and a post by someone calling himself “Giordano Bruno” which disputes the increased sea level rise in the northeast U.S. “hotspot” based on — put your coffee down, please — the “trend” over a whopping five whole years.
What most strikes me about the “Bruno” post is that the terminology is far too reminiscent of Albert “Making Up Stuff” Parker. He’s the fellow who sometimes goes by the name Albert Parker, sometimes Alberto Boretti, and once even submitted two comments on the same paper to a peer-reviewed journal, one under each name. Perhaps now he isn’t satisfied with either name, instead fashioning himself after the famous Italian. Is the post really from Alberto-Giordano Bruno-Boretti-Parker?
The “Bruno” post makes trend pronouncements based on a ridiculous, silly, yea ludicrous short span of time. Yes, a mere 5 years and four months. It would be hilarious if it weren’t so pathetic one is tempted to take pity on Bruno-Boretti. Here’s his illustration for the tide gauge data from the Battery in New York:
The section at the end is his basis for claiming that sea level is now falling. Of course that ignores over 95% of the data … nothing to worry about at all. It reminds me of a similar attempt which ended in embarrassment for Bjorn Lomborg (image courtesy of Greg Laden):
The “Bruno-Lomborg” strategy has become the mainstay of those who deny the reality, human cause, or danger of climate change: find some short time span somewhere — anything, anywhere, no matter how brief — during which the never-ending fluctuations of data mask the underlying trend, then declare those fluctuations to be the trend. Perhaps the fact that their chosen time spans are so embarrassingly short, doesn’t matter to them any more than actual facts matter to Donald Trump.
But when it comes to sea level rise, what are the WUWT boys to do? They can’t deny that sea level is rising (even they aren’t that stupid — maybe), and they can’t deny that the people of south Florida (voters, even) are extremely worried about it (as are a lot of people) and are spending a lot of money to deal with it.
Larry Hamblin retreats to the “no acceleration of sea level rise” meme. Here’s the gist of his argument:
Additionally NOAA performs 95% confidence interval analysis of each location to determine if there are significant changes in the tide gauge measurement trends over time.
…
The 95% confidence interval analysis demonstrates that the linear sea level rise trend is stable and unchanging.
That’s wrong. Maybe it’s “not even wrong.”
Comparing the trend from 1856 to 2006, to that from 1856 to 2015 isn’t going to reveal whether there’s been any acceleration (i.e., a change of the slope of the long-term pattern). Those two time intervals are practically guaranteed to show the same linear trend because they’re based on almost the same data!
Allow me to illustrate. I generated some artificial data with a negative slope from 1895 to 1990 and a positive slope from 1990 to 2015, plus random noise. It looks like this:
The slope change isn’t easy to see, but it’s ridiculously easy to detect and confirm statistically. You could fit a quadratic, or a piecewise linear model, and either would confirm a slope change without doubt.
But if we compute trends from 1895 through 2006, then through 2007, etc., all the way up to “through 2015,” all those calculations shared most of their data in common so we expect them to show similar linear trends in spite of the easily-confirmed change. And indeed they do:
Certainly their “95% confidence intervals overlap.” By a lot.
If you really want to know whether there’s been a slope change in sea level along the U.S. coast, stay tuned here for a post on that very topic soon because there actually are decent methods to look for it. But don’t rely on posts at WUWT; they aren’t likely to accidentally wander into “valid analysis” country any time soon (or should I say “ever”?).
This blog is made possible by readers like you; join others by donating at Peaseblossom’s Closet.
Open Mind 
I read that article last night and promised the first rebuttal would get a donation. Done.
[Response: Thank you! Your donation will be put to good use — funding further rebuttals, as well as some actually relevant analysis.]
Reblogged this on Don't look now and commented:
Have the BBC dropped Bjorn Lomborg yet? A nice cartoon here showing his technique…
I implore you all not to read the self-congratulatory comments there. Holy fuck.
I don’t suppose they are any different than ever–hence not very tempting, to say the least.
The clown-shoed dissemblers like Bruno need to get back in the clown car and drive it somewhere where it can’t do any more possible harm to the rest of humanity; perhaps off the edge of a cliff, if that’s not too un-PC.
They’d have the rest of us drive over one, after all.
From the paper on the northeast hotspot, NEH:
Here, we present evidence of recently accelerated SLR in a unique 1,000-km-long hotspot on the highly populated North American Atlantic coast north of Cape Hatteras and show that it is consistent with a modelled fingerprint of dynamic SLR. Between 1950–1979 and 1980–2009, SLR rate increases in this northeast hotspot were ~ 3–4 times higher than the global average. …
Regardless, our correlations suggest that should temperatures rise in the twenty-first century as projected, the NEH SLRD will continue to increase. If future sea-level variability is forced by aerosols and/or is part of a cycle, SLR
in the NEH may also alternately fall below and rise above projections
of IPCC scenarios alone.
The denialists have become so few and so desparate
WUWT is digging itself ever deeper into its tornado cellar of irrelevance. Does it even matter who is posting there under the name Giordano Bruno?
Personally, I think a better strategy would be to smack down higher-profile contrarians like Lomborg, Tol or Ridley each and every time one of their error-riddled commentaries is published in a mainstream newspaper or magazine. Who else is left? Pielke Jr.? All the others are slowing down in their dotage or falling off the radar like Soon.
“Mark Twain” had a story about the length of the Mississippi. Too bad that a river boat pilot had better math skills and better writing skills than “Giordano Bruno”! And Twain’s reader’s seem to be more sophisticated than those at WUWT.
I think Pielke Jr has discovered the better part of valour. and withdrawn.
Tol had his legs cut off long ago but he’s still experiencing phantom stimuli that assures him they’re still attached.
Very few people seem to understand that failure of a statistical test does not prove anything – except that the test has insufficient power to detect any effect that is present. Certainly a statistical test can never prove the null hypothesis, no matter what the outcome.
Well, what do you mean by “failure of a statistical test”?
If, say, the linear temperature trend for the surface over the last 50 years was 0.0 +/- 0.01 C/decade, that’d be, uhh, pretty interesting to say the least. In lieu of a solid physical explanation, it would imply a failure of climate science.
But, hey look, you see that I have to add on extra context in order to actually draw scientific conclusions, like whether or not there’s a solid explanatory mechanism behind the “failure” of the test.
Windchaser
At the risk of making a fool of myself on a blog that belongs to someone much more expert than me, I will have a go at responding to your response.
A statistical test normally means taking a null hypothesis, assuming that it is true and calculating what the probability of getting the actually observed data or data that is even further removed from the null hypothesis would be. If that probability is less than some cutoff (often 5%) then the test succeeds and you have a case for rejecting the null hypothesis. If the probability is greater than the cutoff the test fails, but that does not give you a case for accepting the null hypothesis. In fact if the probability was say 30% and you forced to bet your life on guessing whether the null hypothesis is true or false then in the absence of other information you might well choose to bet on it being false.
So it all depends on the null hypothesis. You could set up the test based on the hypothesis that the temperature was increasing at a rate of at least 0.2 degrees per decade and show that under that assumption the probability of getting the actual observations is less than 5%. Then you would have a case for rejecting the hypothesis and concluding that the rate of increase is probably less than 0.2 degrees per decade.
The basic rule is that this type of test can provide a case for rejecting the null hypothesis it cannot make a case for accepting it.
In the example you give I’m not quite sure what your statistical test is. Is your null hypothesis that the IPPCs range of climate sensitivities is correct? In that case the hypothetical observations you posit might well justify rejecting the null hypothesis. But that is totally in accord with my comment.
[edit]
[Response: I’ve decided to make a post about your comment. Perhaps it will be an opportunity for learning.]
I am glad you think you can learn from it! ;-)
Happy to post you my workings (Excel sheet) if it helps. My address is [email address redacted]. I note that you pre-moderate all postings so please do not reveal the address here. Ta.
[Response: Consider the possibility that you might learn something.]
OK, that was a true LOL!–or maybe a true LOL(DK).
Looking forward to that post; I’m sure that *I* at least will learn something.
Probably a neat ROFL (leaving off the rude ending)
Next time I will put [just pulling your leg] after every joke, I appreciate that English banter and joshing is not always comprehensible to foreigners!
OK then how about some charts?
Basic sea level time series 12 month moving average and trends.
http://tinyurl.com/z9phtgh
Rate of change in sea level rise
http://tinyurl.com/j2a5msd
This the rate of change as measured by CU. They subsequently add 0.3mm a year to their measurement to estimate the effect of movements in the earth’s crust.
Although they don’t publish the unadjusted series any longer, in an email exchange Dallas Masters at CU confirmed that they simply add 0.3 mm to the measured figure each year and that I was using the correct method of removing this adjustment to arrive at the actual measured SLR.
CU Sea Level Research Group chart showing correlation between rate of sea level rise and Multivariate ENSO Index:
http://tinyurl.com/jcd292a
This shows the close correlation between “changes in the rate of change” and ENSO.
Your links don’t work for me.
Not sure why. I have checked on phone and laptop and are OK here. But as Mr T says, I am a worthless ignoramous so don’t worry you aren’t missing anything.
The CU Sea Level Research Group can be found here: http://sealevel.colorado.edu/
TLM – You claim that the recent rise is simply due to ENSO, but please explain the 1998 ENSO and accompanying SLR compared to the 2015 ENSO and its accompanying SLR.
It sure looks to me like we had less SLR from a stronger El Nino in 1998 compared to 2015.
Reading the paper by Sallenger, et al., 2012, it would appear that their analysis splits the SL time series into different periods, then looked at the trends. For example, the 60 year period was from 1950 thru 2009 and the 40 year window was from 1970 thru 2009. “Bruno” appears to have applied trailing averages to the time series, for example, comparing a 60 year with a 30 year filtered result, etc. His plots show all his filtered results ending with April 2016, which demonstrates that he doesn’t understand that a trailing (or moving) average average results in shifting dates for the resulting filtered series. Also, a 60 year centered moving average cuts off 29.5 years at each end of the filtered result, but using a trailing average cuts off 59 years at the start. What he gets is an apples and oranges comparison because of the date shifting. He can’t even see the last 10 years in any of his filtered results.
There’s no question that Sir Bruno’s conclusions are flat out wrong.
Agreed. I would, however, be interested in seeing if there really is periodicity in the SLR figures but suspect the record is not long enough yet
Is there a 60-year oscillation in global mean sea level?
That is so last century. Up in the North Atlantic, the big blue blob is stuck in a standoff with the warm ocean that surrounds it. People keep talking about the negative phase of the AMO being on the verge of commencing. My question is, if the big blue blob can’t drag down the AMO index, what is left that can?
TLM – it’s obvious that SLR is highly correlated with the ENSO index. Andto see the various frequencies in the ENSO index you may wish to visit WHUT’s blog: Deterministically Locked on the ENSO Model
Thnaks KO. Can’t forget that ENSO causes something called the inverted barometer effect whereby low pressure raises sea level and vice versa.
For example, a two-year differencing on Sydney tidal gauge readings will remove any annual and biannual cycles, which are also strong. The ENSO signal is clearly evident. Worse yet is that in places such as the low-elevation interior of Australia, the depressed regions will pool excess water, which subtracts from the ocean SLH, at least until that all evaporates over time.