Theory and Reality- Part 1: DTR

It was two years ago in 2013 that I last posted on the difference between climate scientists’ expectations and reality, so in this series of posts I bring these points up to date, and add a couple of related points.

What the climate scientists tell us:

Dr Karl Braganza in The Conversation on 14/06/2011 lists the “fingerprints” of climate change (my bold).

These fingerprints show the entire climate system has changed in ways that are consistent with increasing greenhouse gases and an enhanced greenhouse effect. They also show that recent, long term changes are inconsistent with a range of natural causes…..
…Patterns of temperature change that are uniquely associated with the enhanced greenhouse effect, and which have been observed in the real world include:
• greater warming in polar regions than tropical regions
• greater warming over the continents than the oceans
• greater warming of night time temperatures than daytime temperatures
• greater warming in winter compared with summer
• a pattern of cooling in the high atmosphere (stratosphere) with simultaneous warming in the lower atmosphere (tropopause).

and later

Similarly, greater global warming at night and during winter is more typical of increased greenhouse gases, rather than an increase in solar radiation.

This post will examine “greater global warming at night” and whether it can be attributed to increased greenhouse gases.

If night time temperatures (minima) increase faster than day time temperatures (maxima), then the difference between these, the Diurnal Temperature Range (DTR) will decrease.

I use BEST global land temperature data,

and annual CO2 concentration data from NOAA.

Fig. 1: Global DTR (derived from BEST Land Tmax and Tmin)

DTR globe

Yes, the long term linear trend shows globally DTR has decreased, at a rate of more than half a degree Celsius per century.

Case closed! That is, if you ignore the sudden turnaround in the early 1980s. Since then DTR has been increasing at +1.1C per 100 years.

The plot showing the relationship with CO2 concentration is even more revealing:

Fig. 2: Global DTR vs CO2 concentration

globe dtr v co2 all

If we break the series in two at the dogleg, we get the following plots:

Fig. 3: Global DTR vs CO2 concentration to 1982

globe dtr v co2 1

Fig. 4: Global DTR vs CO2 concentration 1982 to 2015

globe dtr v co2 2

Calling Global Warming Enthusiasts! I am puzzled:

Is DTR decreasing at 1.14 C/ 100 ppm CO2 or increasing at 0.61 C/ 100 ppm?
Can there be any logical explanation for this distinct turnaround?
Is there a problem with (a) the CO2 concentration data? (b) BEST data? (c) the theory behind decreasing DTR being an indicator of enhanced greenhouse warming? (d) all of these?

I now turn to the Australian context, with Australian surface data.

Fig. 5: Annual DTR Australia (from ACORN)

DTR Aust

While averaged across Australia, DTR has decreased since 1910, there has been a marked increase recently. As well, the pattern is different in different regions.

Fig. 6: DTR North Australia

DTR Aust nth

Fig. 7: DTR Southwest Australia


Fig. 8: DTR South Australia


Fig. 9: DTR Victoria

DTR Aust Vic

Fig. 10: DTR Tasmania

DTR Aust Tas

The effect is strongest in the tropical northwest and northeast, and weakest in the southwest and South Australia, Victoria, and Tasmania.

Moreover, the dominant influence on DTR is rainfall:

Fig. 11: DTR vs Rainfall

DTR Aust vs rain

Definitely not CO2!

Fig. 12: DTR vs CO2 concentration

DTR Aust vs CO2

Assessment of decreased DTR as evidence for the enhanced greenhouse effect: Fail.

Other factors- especially rainfall- overwhelm the enhanced greenhouse effect.


Perhaps I should be more blunt:  If Global Warming Enthusiasts stick to decreasing DTR as an indicator of greenhouse warming, then this shows BEST and ACORN surface data are completely unreliable.  If they stick to claiming ACORN and BEST are “world’s best practice” then they must accept that DTR as an indicator of greenhouse warming is a dead duck.

Tags: , , ,

27 Responses to “Theory and Reality- Part 1: DTR”

  1. Martin Says:

    You can fit a straight line to anything but that doesn’t necessarily mean anything. Putting a straight line through the data on your first chart is clearly meaningless as you point out due to the change in the early 1980s. Have you ever tried using cusums. Apart from being fun to see what effect different “control or target” values can have the cusum is a very good technique for identifying change points. I think of it as showig a rate of change. Unfortunately I don’t have any basic reading but this page looks like a bit of a start.

  2. craigm350 Says:

    Reblogged this on WeatherAction News.

  3. kenskingdom Says:

    Perhaps I should be more blunt: If Global Warming Enthusiasts stick to decreasing DTR as an indicator of greenhouse warming, then this shows BEST and ACORN surface data are completely unreliable. If they stick to claiming ACORN and BEST are “world’s best practice” then they must accept that DTR as an indicator of greenhouse warming is a dead duck.

  4. MikeR Says:

    Ken, Your results are intriguing and on prima facie appear to conflict with the climate change models.

    Your figure 11 is interesting and may be a clue as to what is happening. If you correlate DTR with cloud you get a similar but much stronger correlation of 0.72 with a negative slope (see ).

    In essence this indicates that cloud cover which we know traps heat at night is the most significant determinant of the DTR.

    So what we really need to know, if the data exists, the trend in cloud cover globally since 1910. It is possible that since 1980 the global cloud cover has reduced which may explain the turning point in DTR at that time.

    I see the climate4you site ( ) has some information indicating a reduction in cloud cover but it is only starts in about 1983. It does seem to indicate that as temperature has increased cloud cover has correspondingly dropped.

  5. kenskingdom Says:

    Or as cloud cover reduced, temperatures rose. In Australia,Tmax correlates well with rainfall but not Tmin. Cloud cover reflects solar radiation during the day.

  6. MikeR Says:

    Either way reduced cloud causes an increase in Tmax or is caused by an increase of Tmax. In this case causality does not really matter.

    Because of the strong dependence of DTR on cloud cover ,the only way jump you can to any conclusions regarding DTR is to assume the average cloud cover is fixed. Climate4you suggests this not the case at least since 1983.

  7. MikeR Says:

    Sorry for the above,. I meant to have written ‘the only way you can jump to any conclusions’.

    It sounds like Yoda was responsible for the above comment!

    • kenskingdom Says:

      So you are saying the increased DTR is due to increased cloud cover (lowering minima) or decreased cloud (increasing max) as per climate4you? Which? Either way, as I said in my post, “Other factors- especially rainfall- overwhelm the enhanced greenhouse effect.” Add cloud as one of the other factors. What role is left for CO2?

  8. reichforthesky Says:

    No, I am saying that increased DTR is due to decreasing cloud. possibly lowering Tmin and increasing Tmax.

    With regard to Tmin. I would have thought cloud cover at night is likely to raise Tmin. All those frost warnings issued by the BOM associated with clear winter nights suggest this is likely. Night temperatures in desert locations are another indication as well.

    On a personal note, as an amateur astronomer in my more youthful days, I remember waiting for the clouds to recede and as the clouds receded, the views got better, but I can then well remember how rapidly colder it would get.

    However the Australian data does not show a significant correlation between Tmin and cloud cover The correlation coefficient (Rsq=0.05) has a positive correlation but this low coefficient is not that surprising as cloud measurements are only done during the day and Tmin is usually occurs at night.

    Tmax in contrast is correlated inversely to daytime cloud (Rsq =0.29) and as you said previously, this is probably just the surface being heated by direct exposure to the sun in the absence of clouds..

    DTR is highly correlated with cloud cover Rsq=0.72 with a negative slope. This is strong indication that despite the low correlation for Tmin with day time cloud cover it too plays a large role in the DTR trends. Tmax by itself correlates moderately well with cloud cover but DTR correlates significantly better as it includes Tmin as well as Tmax in its calculations.

    Because the confounding factors of cloud (and rain) play such a large part in determining DTR, ignoring these factors weakens your conclusions enormously.

    Finally here is a plot of DTR and Cloud Cover for Australia from 1957 until 2014, see .

    • kenskingdom Says:

      Sorry, got that bit back to front- decreased cloud lowers minima of course.
      Nice graph. More cloud (and rain – you don’t get rain without thick clouds) -> low DTR, and vice versa.
      And that was exactly my conclusion: “Other factors- especially rainfall- overwhelm the enhanced greenhouse effect”. Rain and clouds are entirely natural phenomena. Tell me about CO2 again.

      • reichforthesky Says:

        Yes I am in agreement with most of your comment above. It appears from the Climate4you website that cloud cover and temperature are inversely related (at least since 1983) , so I guess it all comes down to that major sticking point, the relationship between CO2 and temperature. This is where I suspect we differ.

      • MikeR Says:

        Ok i have changed it back to Miker. I think old age is catching up with me and I am dementing!

  9. reichforthesky Says:

    Sorry my name keeps reverting from MikeR to my old name from about 3 years ago Reichforthe sky. I need to keep checking each time which account I am logged into.

  10. MikeR Says:

    Hi Ken,

    There is a lot of stuff around about cloud cover on a global level.

    Interesting reading but there appears to be little agreement as to whether there has been significant changes in cloud cover globally from the 1980s onwards.

    One of the papers you will clearly love is by a Ph.D student from James Cook University

    but it gets numerous criticism from others see-

    Click to access Evan_etal_GL028083.pdf

    Click to access OSGC-000-000-000-521.pdf

    and particularly adverse comments at the web site-

    There is also a lot discussion regarding cloud cover and temperatures. In the above they are referring to global mean temperature so the situation is very different to DTR and Tmax and Tmin.

    Despite my evidence that, at least for Australia, cloud cover is the dominant factor determining DTR, this still leaves the puzzle regarding the 1980s change in direction for global DTR.

    Now Ken you think rain is the major factor. That is interesting because if you plot global rain as a function of time then it shows an increasing trend of 0.23 mm/year from 1900 to 2014 – see .

    Interestingly for the period 1900-1979 the trend is only 0.43 mm/year while from 1980-2014 it is 1.01 mm/year . So for the period from 1980 onwards the rainfall more than doubled . Ken you are clearly right, rainfall does have an impact. The only problem is increasing rainfall would be expected to decrease DTR not increase it!

    If it is not changes in clouds or rain for which there appears to be little available evidence then what could be behind the puzzle?

    I am not claiming increasing CO2 is responsible but I am open to suggestions.

    Thanks again Ken for providing a blog with loads of thought provoking information and importantly allowing open debate to occur.

    • kenskingdom Says:

      A couple of points. 1. Cloud is very difficult to describe objectively- height, thickness, are not mentioned in ‘oktas’. Also based on subjective observation mostly unless by satellite. This will improve. 2. Rain and cloud are obviously linked. 3. If as you say DTR is gong the wrong way for increased global rainfall, then this brings us back to whether the BEST DTR data are reliable. BEST just might be unreliable just like other severely homogenised surface datasets, which is one of the alternatives I suggested above. Actually we know BEST is rubbish, this just further proof.

  11. MikeR Says:


    If the Best data is rubbish and cannot be relied upon then what was the whole point of the above blog? Especially figures 1 to 4 which rely on this data.

    Is there any other data set that has global values for Tmax and Tmin?.

    With regard to oktas you make a good point that it cannot provide information about cloud types but the relationship between rain, oktas and DTR was highly significant in the Australian context so this part was a useful exercise.

    Ken, you do excellent and perceptive work when you tease out relationships that others have missed. Your earlier blog regarding the relationship between UAH, Tmax and rain is a prime example.

    I am eagerly looking forward to Part 2.

    • kenskingdom Says:

      Either decreasing DTR as an indicator of enhanced greenhouse warming is nonsense or BEST is nonsense. Or both.
      A couple of posts coming- Pt 2 to this and a follow up to Trends- maybe this weekend.

      • MikeR Says:

        WIth respect to your last comment, you have not produced any evidence , in fact you have painted yourself into a corner with your rejection of the BEST data.

        I will remind you of your comments in the body of your blog (between figures 4 and 5) i.e.

        Is there a problem with (a) the CO2 concentration data? (b) BEST data? (c) the theory behind decreasing DTR being an indicator of enhanced greenhouse warming? (d) all of these?”

        With regards to each problem

        (a) You provide no evidence about problems with the CO2 concentration data? Is any one disputing this data?

        (b) You claim that BEST data is rubbish, an evidence free assertion, but if you say so, then

        (c) It follows your BEST based evidence needs to be disregarded and cannot be used to either to disprove (or even provide evidence against) the proposition with regard to the theory that decreasing DTR is an indicator of enhanced greenhouse warming?

        (d) From either (a), and/or (b) and/or (c), the answer is – there is clearly not a problem with all of these and possibly there is not a problem with any of them.

        Bring on Pt 2.

  12. kenskingdom Says:

    No, the opposite. BEST shows DTR is NOT an indicator of greenhouse warming. If BEST is good, greenhouse warming is problemmatic. If it is true that greenhouse warming produces less DTR, then BEST is no good. Either / or.

  13. MikeR Says:

    No Ken no. This time with feeling.

    If the BEST data is crap then it cannot be used to confirm or deny anything.

    If the BEST data is good then the mysterious up swing from the 1980s ,as I pointed out could be related to decreased global cloud cover, as evidenced by the Australian data which demonstrates a clear link between DTR and cloud . It appears that the global cloud cover data is unreliable so it cannot be used to rule out or rule in a decrease in global cloud cover from the 1980s.

    So if BEST is OK then the data you presented in your blog cannot be used to confirm or deny a link between DTR and CO2 unless all other variables that affect DTR, such as cloud cover, are assumed to be fixed. This is one humongous assumption.

  14. kenskingdom Says:

    Let’s assume BEST and ACORN are excellent- we’ll never know for sure- and I have no doubt that NOAA CO2 data are good too (I was merely canvassing possibilities on this one). Forget cloud- that has never been my concern for reasons I outlined above. ACORN shows DTR with a lousy fit with CO2 and a much better fit with rain. And BEST clearly shows the enhanced greenhouse effect of lowering DTR being completely overwhelmed by some other factor(s) for more than 30 years. If you cannot see this point we will just have to agree to disagree.

  15. MikeR Says:

    Ken, I have just had an epiphany!

    I now realize that your argument has been a strawman argument all along. I am wondering now why it has taken me so long to realize this. Maybe I am really dementing .

    Braganza and others may have claimed the DTR is a signature of global warming over the long term, but have they claimed that the only cause of a change in DTR is via CO2?

    I don’t think so, but correct me if I am wrong.

    Actually Braganza stated in his abstract of his 2004 paper ”This difference is likely attributed to increases in cloud cover that are observed over the same period and are absent in model simulations.”

    Now I understand why you seemed to be obsessed with rain ,even when I displayed above, that cloud cover was more highly correlated with DTR than rain. I think your obsession is more with Braganza.

    Again in the conclusion of Braganza’s paper he states the following-

    “Changes to water vapour, soil moisture and precipitation, not considered here, are also likely to have lesser but important impacts on simulated trends in Tmax and Tmin. [19] Since DTR is highly sensitive to small changes in maximum and minimum temperatures it is important that we fully understand the causes of model inconsistencies with observations. This study does not consider simulated changes in DTR due to changes in solar irradiance or volcanic aerosols or forcing due to other aerosols such as black carbon, the indirect radiative effect of aerosols or land use changes, all of which may be expected to impact on DTR. [20] ”

    I have spent an inordinate amount time unnecessarily huffing and huffing to blow this straw man down.

    So I hope Part 2 will have less straw content and more mortar and cement.

    [My apologies for not approving this earlier when I replied: as you know all your comments go into moderation but I approve before commenting (usually!). This one slipped through, my mind was elsewhere yesterday with trying to get rid of malware from my computer which slowed down all processes as well. My mistake.KS.]

  16. kenskingdom Says:

    Whatever floats your boat. Time to move on.

  17. DaveR Says:

    Moving on… as a geologist I am always intrigued what causes the flip from warm global temperatures to ices ages – with four major ice ages in the last 400,00 years. The flip seems to be (relatively) rapid, with the globe some 10 degC colder in the ice age period.

    It appears like an equilibrium breaks down, with a rapid move to another equilibrium state, then it breaks down …..and on it goes.

    I once asked Ian Plimer what he thought an ice-age day would be like (in Melbourne) if you could be there – his answer was – cold dark (totally overcast) and very windy.

    Clearly Tmax, Tmin, DTR, cloud cover, CO2, wind levels and rainfall (water vapor?) are all related, ie not independent variables.

  18. These items caught my eye – 12 February 2016 | grumpydenier Says:

    […] global warming at night” and whether it can be attributed to increased greenhouse gases. – Click here to read the full article […]

  19. BEST Adjustments | kenskingdom Says:

    […] years ago I wrote a post about changes in Diurnal Temperature Range (DTR) and whether these were a “Fingerprint of […]

Comments are closed.

%d bloggers like this: