Wednesday, December 17, 2014

Bob Tisdale scores a 0/3 FAIL in a quiz on global surface temperatures

Brandon R. Gates has been commenting here and at WUWT lately. Yesterday he put three questions to Bob Tisdale, whose answers gives us more insight into how Bob operates (or doesn't). And since Bob himself suggested something about HotWhopper, let's take him up on his offer - except that I'll do my best not to misrepresent him. In fact I'll paste Bob's answers to the questions in full, thereby removing that risk altogether. Here goes.

Q1: Score = 0/1 Bob can't read his own charts

Brandon Gates  December 16, 2014 at 6:08 am
Perhaps you can shed some light on what’s going on in these graphs:
1) Why do the hiatuses from 1800-1915, 1940-1975 and 1998-present each successively have a more positive slope than the last?

Bob Tisdale replied:
1) I can’t respond to your question 1 because you haven’t illustrated trends on your graphs. Are you aware that you’ll get different results if you use HADCRUT data, because of the modifications to HADSST3 where they eliminated the “discontinuity” around 1945? So the answer to your question depends on the sea surface temperature dataset, now, doesn’t it, Brandon?

Now you might think that first sentence a half-reasonable answer except for the fact that Bob could have looked at his own charts in his own article to see the trend during those periods. Bob himself showed the rates of warming, in the article to which Brandon was responding. I've shown his chart below. Yet Bob said he couldn't respond to the question. Weird, huh. I guess he doesn't understand his own charts. (You'll notice that Bob throws in a reference to his favourite subject - sea surface temperatures.) BTW I think that's a typo in Brandon's question and he meant to write 1900-1915. That doesn't excuse Bob's non-answer.

Q2: Score = 0/2 Bob repeats the question as his answer (circular reasoning)

To Brandon's second question:
2) Why does each hiatus begin and end at a higher temperature than the previous one?

Bob Tisdale did his usual circular reply for which he is now famous (here at any rate) and said it got warmer because it got warmer. He just repeated back the question as his "answer". Or using his words:
2) I’m surprised you haven’t figured that out the answer to your question 2, Brandon. The obvious answer to your question 2 is, because there is a warming period between the hiatus periods.

Q3: Score = 0/3 - the question isn't even answered

Brandon Gates had a third question:
3) Why on earth would I pick a 1986-2005 baseline calculate temperature anomaly?

Bob Tisdale didn't answer Brandon's question. Instead went to great lengths to say what he, Bob, did. Or rather what he, Bob Tisdale, didn't do - rather than figure out why Brandon did what he did. Bob wrote:
3) I didn’t pick 1986-2005 for a baseline. Odd that you should ask about them. I’ve listed the base periods on the graphs, Brandon. Any reason you’re being trollish about base years, Brandon? For Figure 00, I used the base period of 1880 to 2013 so that the base years didn’t bias the difference. The data in Figure 1 and the Supplemental graph that follows it use 1951-1980 (GISS base years). In Figure 2, the base period was 1901-2000 (NCDC base years). The period of 1961-1990 was used in Figure 3 (HADCRUT base years). Figure 4 uses 1981-2010 as a base period (UAH base years), and last but not least, the data in Figure 5 uses 1979-1998 (RSS base years). Those are the base periods selected by the data suppliers. I’m surprised you didn’t know that, too. You’re learning lots today, Brandon. The graphs in Figure 6 and 7 are trend graphs and are not dependent on the base years. And for Figures 8 through 11, I used the base periods of 1981-2010…which are the base years recommended by the WMO, just in case you’re not aware of that also, Brandon.

As Brandon explained in a comment recently, he uses the reference period of 1986-2005 for CMIP5 because 2005 is when observed forcings stop in the CMIP5 models and the RCP's start. Scientists typically use 1986-2005 as the reference period for CMIP5, as did the IPCC in its AR5 report.

Brandon also helpfully gave Bob a clue to help with his first two questions, but they went right over Bob's head - just as his questions 2 and 3 did:
If you need help with (1) and/or (2), you might want to slap a linear regression line over the entire range of your chart here, and ponder the meaning of an upward-sloping 2nd derivative: 

Here's the "your chart" of Bob's that Brandon linked to showing the rates of warming, for which I've shaded the hiatus periods Brandon referred to. The chart plots the linear trends for successive fourteen year periods (or more precisely, successive 167 month periods) in degrees/decade:

Source: WUWT - with my annotation showing the "hiatus" periods to which Brandon referred in Q1.

Here is the upward sloping second derivative that Brandon provided as a clue for Bob:

Data Source: NASA GISS (Note: it's 13 not 14 years - see comment)

So what does an upward sloping second derivative tell us? That it's getting hotter more quickly than it used to. In other words, the rate of change is increasing.

Bob wrote:
The highest recent rate of warming based on its linear trend occurred during the 166-month period that ended about 2004, but warming trends have dropped drastically since then.  There was a similar drop in the 1940s, and as you’ll recall, global surface temperatures remained relatively flat from the mid-1940s to the mid-1970s.  Also note that the mid-1970s was the last time there had been a 167-month period with a global warming rate that low—before recently. 

Bob is being deceptive, and maybe hopeful (as we all are). Let me illustrate. In the 1940s the rate per decade dropped below zero. It hasn't done that again since the 1960s. Each fourteen year period since then has been warmer than the one before (a positive trend). Bob's 167 months takes us back to 2001, which is only 14 years ago. Bob used monthly data, which is 13.92 years. I prefer annual so that the noise of year to year fluctuations doesn't mask the signal, especially over such a short period of time.

Let's look more closely at Bob's linear trend chart. Here is what the rate of change (bottom chart) looks like with GISTemp surface temperatures plotted above. Notice the trend. The trend line at the top is for the most recent 14 13 year period (I deducted a year to centre the 'moving' linear trend chart - prompted by Everett in the comments. I chose 13 rather than 15 to keep the trend closer to Bob's). While lower than it would be for a longer period, it is not negligible and is positive. Look at the zero line on the bottom chart of the linear trend. Not an ice age in sight. Notice also how most of the observations are well below the trend line. That suggests either there is a definite leveling off or the period isn't long enough to properly represent the actual trend. What do you think?

Data Source: NASA GISS (note it's 13 not 13.9 years)

Compare that to the currently accepted climate period of 30 years, which I've made 31 years to get an odd number for the trend chart. Notice the steeper trend over the most recent 31 year period. Notice how the trend line cuts right through the middle of the period since the 1960s. See how sometimes the temperatures drop below the trend line and sometimes they are above. Notice also how this last year (average to November) is almost back touching the 31 year trend line. Which trend line do you think better represents the global surface temperature? The double chart above or the double chart below?

Data Source: NASA GISS (note it's 31 not 30 years)

Here's what happens when there is a longer period - 50 51 years. Again, the most recent 51 year trendline fits well with the observations, this time going back to the early 1960s. Again, see how high the trend has become - look at the bottom half. It's well above zero and shows no sign of dropping below zero.

Data Source: NASA GISS (note it's 51 not 50 years)

I'll show you one more, just for interest. It's a seventeen year chart - with the trend line in the top going back seventeen years to the super El Nino of 1997/98. Bob showed this one, too, in his article. It's the same as his 210 month trend chart. The slope, while not as steep as the 30 year trend, is steeper than for Bob's 14 year trend. What do you think he'll be doing in future? Five year trends? Three years? One year?

Data Source: NASA GISS

The point is that Bob Tisdale doesn't appear to understand what he's doing, as evidenced by his answers to Brandon's questions. As well as that, he's picking shorter periods now than he has in the past, so as to make the slope flatter. That suggests deliberate deception but I'm not sure he's that cluey.

And he's a rather dumb greenhouse effect denier - referring to the "hypothetical impacts of manmade greenhouse gases". That shows that he doesn't accept centuries-old physics.

Will 2014 break the 2010 record?

Before I finish, I'll just make one more observation. The average of this year to date is the hottest on record compared to other full years, and the second hottest on record for the average to November (after 2010). If the December anomaly is greater than 0.55, then in GISTemp, 2014 will be the hottest year on record. The last four months have had anomalies of 0.74, 0.81, 0.76, and 0.65C. Only two months this year have had an anomaly below 0.55C. Even if the anomaly is as low as that of February (a relatively cool month), 2014 will be the second hottest year on record. So whatever happens, 2014 will end up being a very warm year.


  1. Additional points should be taken off for the handful of sentences that Tisdale forgot to end with the word "Brandon".

    1. Not the first time someone has used my name as an expletive.

  2. Have you noticed Tisdale has terminated his inane blog dedicated to bad mouthing HotWhopper? It did not last long at all.

    1. It outlasted all my expectations. It was filled with posts from skilled people in support of Sou. I gave it two days but it lasted four.

      I give Tisdale credit for resilience. I never knew he had it in him.

    2. He claims it was closed due to lack of interest in what Sou has to say. I presume he wrote that with a straight face even though all he seemed to be doing was acting in defense of WUWT. So it seems more likely there are not too many interested in defending WUWT. Or even what Tisdale has to say.

    3. Maybe they just found that defending WUWT wasn't possible.

  3. On trends: as far as I can make out the trend in GISS is now statistically significant since April 1996: 0.103 ±0.102 °C/decade (2σ) Ref:

    That's 18 years 8 months. Between that, plus at least one very likely new warmest global land and SST record (NOAA); almost certain new global ocean temperature record (HadSST3); likely new warmest year for the oldest instrument temperature record in the world (CET), etc....

    Oh dear. 2014 has a very 'bad' look if you're trying to sell 'no warming for 18 years'; even to some of the WUWT crowd. No wonder Bob's a bit tetchy.

  4. I forgot to comment on the R^2 values. At less than 0.1 it's so low as to make the 14 year trend line to 2014 pretty well meaningless. A bit higher for the 17 year trend line. Very high for the 30 and 50 year trend lines.

    1. I hadn't even gotten so far as to plot the actual trendlines much less look at the R^2 values. But now you have done it, saving me the trouble. Something to think about. [mutter mutter]

      Thanks for the much more complete writeup of all that was very wrong with Bob's original post and his total fumbling of the answers to my pointy little questions. He exceeded my wildest expectations.

      As well, thanks for your recognition of my ... work. I do actually see it that way. I'm beaming right now, perhaps a bit too much. :) Cheers.

  5. Sou,

    On your 17, 30 and 50 year moving average trend lines, I would shift all to the left, half the span length or 8.5, 15 and 25 years, respectively. I always like my moving trend plots to be centered wrt the underlying time series, not the end of the underlying time series. The relative phasing of these annual/moving average plots, as is, is visually apparent.


    1. You're absolutely right, Everett. Thanks - my mistake and it's not the first time I've done that. I was sloppy, trying to do too much too quickly. I'll have a shot at fixing it.

    2. I fixed the trend charts but didn't bother with the top charts in the pair except for the 13/14 year chart, so the other trends don't match exactly - the previously even numbered ones are out by a year. I'd say well within any error margin, but I haven't worked it out. It's hopefully near enough for demonstration purposes.

      With the 14 year chart the extra year (15 years) did make quite a difference, so I dropped it to 13 to be closer to Bob's.

  6. Bobs current hypothesis for ENSO contributing to long-term warming is through ENSO interacting with cloud-cover (

    has anyone attempted to disprove this? surely it would be fairly easy?

    1. Does Bob say what caused his clouds to suddenly act differently out of the blue, around the turn of the century or whenever? Why would clouds be any different this past several decades to how they've been for thousands of years or more, if all else stayed the same?

      I mean clouds would be different now. There's a lot more moisture in the air for one thing. But that's because of greenhouse warming - not magical leaping ENSOs. (I reckon ENSO would be acting differently too, these days. All weather is being affected by global/greenhouse warming.)

    2. Answer the question yourself, Greig. Why is ENSO now a "ratchet" on increasing air temperature, but it wasn't throughout the rest of the Holocene?

  7. Bob has now gone from one extreme to the other. In the above he's relying on a very shaky trend of fourteen years. Today he's punched a rough straight line trendline (or a straight line would be more correct) through 134 years to make a different point about something or the other (the two degree target).

    If you look behind the big red line and follow the recent trend up on his chart, he'll probably hit his 2 degrees by 2050!

  8. Sou, in the article you say:

    If the December anomaly is greater than 0.55, then in GISTemp, 2014 will be the hottest year on record.

    Funnily enough, I just happened to do the calculations when the .65 value for November was posted the other day. So yes, you are technically correct that if December's anomaly is 0.56, the figures for Jan - Dec (using the same scale as on the chart):

    would add up to 790, which gives 790 / 12 = 65.83. Whereas 2010 had a value of 65.75. However... this would still appear on the chart as a tie with 2010, because we know they round half up to the nearest whole number. So if you want to clinch it by a full .01 deg C to be outright hottest, then I calculate that it has to be at least 0.64.

    If you add the figures from Jan - Nov for 2014, you get 734. Since 2010 was the hottest year from Jan - Dec so far at 66, to beat it you'd need to get to 66.5, because they would round that half up to give 67. And so:

    (734 + X) / 12 = 66.5

    734 + X = 798

    X = 64

    If you want to get clean to 67 with no rounding up, then December's anomaly needs to be 70. Bit pedantic, and I realise you already know this stuff, but hey, nothing else better to do at the moment.


Instead of commenting as "Anonymous", please comment using "Name/URL" and your name, initials or pseudonym or whatever. You can leave the "URL" box blank. This isn't mandatory. You can also sign in using your Google ID, Wordpress ID etc as indicated. NOTE: Some Wordpress users are having trouble signing in. If that's you, try signing in using Name/URL. Details here.

Click here to read the HotWhopper comment policy.