2011-05-18 06:49:05Trend in Diurnal Temperature Range
John Cook


Saw in an email that DTR has not changed from 1979 to 2004 as both day- and night time temperature have risen at about the same rate. IPCC AR4 report this. Am I overselling nights warming faster than days?


2011-05-18 07:16:38
Dana Nuccitelli

Well if you look at the Braganza paper, models only expected to see a 0.1–0.2°C decrease in DTR from 1951 to 2000.  So it's a pretty small effect.  If you're just talking 1985 to 2004, you'd only expect to see in the ballpark of a 0.05°C decrease in DTR.  And the full DTR change since 1951 is still much larger than the models predicted.

2011-05-18 07:22:31


The DTR has decreased globally, there's plenty of evidence and lots of papers. If anything, you oversold the effect of GHG on DTR, there are other possible causes for this reduction.

2011-05-18 08:09:51
Julian Brimelow


And S posted a link to this paper by Zhou et al. (2010).  Dai, Dirmeyer and Dickinson (authors) are exceptionally reputable names, so it is almost certainly a good paper.

I know...models are not reality, but FWIW, they are suggesting that the reduction in DTR is not going to be huge (at least globally), and their data suggest that perhaps the observed DTR has been overestimated. Their results also suggest that  the reduction in DTR is not uniform and is likely most marked at higher latitudes continental areas.

"When anthropogenic and natural forcings are included, the models generally reproduce observed major features of the warming of Tmax and Tmin and the reduction of DTR. As expected the greenhouse gases enhanced surface downward longwave radiation (DLW) explains most of the warming of Tmax and Tmin while decreased surface downward shortwave radiation (DSW) due to increasing aerosols and water vapor contributes most to the decreases in DTR in the models. When only natural forcings are used, none of the observed trends are simulated. The simulated DTR decreases are much smaller than the observed (mainly due to the small simulated Tmin trend) but still outside the range of natural internal variability estimated from the models."

So, are you over selling the DTR?  If Zhou et al. are anything to go by...probably not. 

I was just looking at some of the figures.  The modelled decrease in DTR over the contiguous USA between 1950 and 1999 are not significant at the 95%confidence level... but note other than southern Africa,and Europe those are pretty much the only large terrestrial areas where that is predicted to be true.  And grid cells where the trends in observed DTR is negative are spotty over the contiguous USA. So perhaps the data over the USA are outlier....

Also, say the DTR is being overestimated b/c of siting issues in the USA, that would bring the models in closer alignment with reality there-- remember the models are not predicting statistically significant decrease in DTR there between 1950 and 1999. Hardly a win for the contrarians.

Zhou et al. did not look at seasonal trends, but I suspect that looking at annual change sin DTR weakens the signal, b/c I expect that the greatest decline in DTR will be during the winter months.  But now I am getting close to talking through my hat, so I'll stop there... ;)

2011-05-18 15:57:31
Ari Jokimäki


Here are some papers, and there are more in the peer-reviewed database.

2011-05-18 16:11:16
Rob Painting

Conceptually, won't the DTR change as large parts of the world dry out, and reduced evapotranspiration leads to enhanced daytime surface temperatures?. Well, that, amongst other things.

2011-05-18 16:20:30Yes, a lot of influences affect DTR
John Cook


When I talked to Karl Braganza about it, asking if DTR was a robust signature of a human fingerprint (he's da man when it comes to human fingerprints), he said the annual cycle was a more robust indicator than the diurnal cycle, because so many factors affect the diurnal cycle. So it was okay to mention it but don't pin your hat on it in isolation - it's all the different lines of evidence taken together. And probably best to give annual cycle more prominence than the diurnal cycle. Which is why I wince when I see people just grab 'nights warming faster than days' as the poster boy proof of a human fingerprint.

2011-05-18 18:04:53Don't forget aerosols
Glenn Tamblyn


We also need to consider the impact of Aerosols. If the argument is that 'It can't be the Sun' because nights warm as well, then aerosols are like a cooling sun, day time only impact. Doesn't this imply that Nights warming faster than days may be evidence of the daytime cooling effect (relative to what would happen withou them) of aerosols.


I must admit I have never entirely understood why nights would be expected to warm faster than days (discounting aerosol effects). Nights warming as well as days yes, thats obvious, since the GH Effect is 24/7 while the Sun is 12/7. But why would nights warm even more than days?

Is there a basis in here to start leading into some posts on Aerosols? That seems to get missed from the big picture a bit.

2011-05-18 21:24:05
Mark Richardson

Glenn: Maybe something to do with the constant radiative effects and nights being cooler?



Nighttime in Europe is ~6 C average, and daytime ~14 C. A 3 W m-2 radiative heating would warm days by 0.56 C and nights by 0.65 C.

I think I dropped something in my calculations, and most of the ultimate heating comes from feedbacks anyway, so to get a real calculation you'd need a far more complex model. But it might be the source?

2011-05-19 02:21:19
Julian Brimelow

John @18 May 2011, 4:20 PM,

Karl's suggestions seem perfectly reasonable, and as you said, you cannot go wrong quoting him.  DTR is but one small piece in a complex, yet compelling puzzle of evidence.