It seems that a number of folks have missed one of the more important conclusions from our first paper on the surfacestations project. Co-author Dr. John Nielsen-Gammon has some things to say about it the finding that is the title.
For layman readers that don’t know what diurnal variation is, it is the daily variation of temperature due to the variation of incoming solar radiation from rotation of the earth on its axis.
It looks like this:
Source: http://apollo.lsc.vsc.edu/classes/met130/notes/chapter3/daily_trend4.html
Fall, S., A. Watts, J. Nielsen-Gammon, E. Jones, D. Niyogi, J. Christy, and R.A. Pielke Sr., 2011: Analysis of the impacts of station exposure on the U.S. Historical Climatology Network temperatures and temperature trends. J. Geophys. Res., in press. Copyright (2011) American Geophysical Union.
Fall et al. 2011: What We Learned About the Climate
By Dr. John Nielsen-Gammon
Our paper has a lot of info and analysis about temperatures and temperature trends and their correspondence to siting class. Perhaps the most important question, “Is the mean temperature trend different from previous estimates?” is answered in the negative, albeit with an asterisk associated with the limited scope of the study. While negative results are useful, they’re also boring. So in this post I’ll talk about something we did learn about the climate that’s new and different, and why I think it matters.
This new finding is stated succinctly in the abstract as: “According to the best-sited stations, the diurnal temperature range in the lower 48 states has no century-scale trend.”
The diurnal temperature range is simply the difference between the daily maximum temperature and daily minimum temperature, “diurnal” being a more impressive way of saying “daily”. It’s conventionally abbreviated DTR.
…
The change in global DTR trends roughly coincides with the change in phase of the Pacific Decadal Oscillation, for example, so it’s hard to tell whether the DTR trends were natural or forced.
That’s where Fall et al. 2011 comes in. The figure below shows the change in DTR anomalies over time over the United States, as estimated using data from each of the four classes of station siting. The data goes all the way back to 1895, making this (as far as we know) the longest evaluation of regional DTR trends done anywhere.
Dotted lines represent average values and are plotted 1.5 C apart from each other
All four classes of stations show the decline from the 1950s through the 1970s. But if you take a broader view, you see that the black line, representing the estimate from the best-sited stations, has a long-term positive (!) trend using unadjusted data or time-of-observation adjusted data, and has no trend at all for fully adjusted data (top set of curves). The lower-quality siting classes all show a negative long-term trend, and the worse the siting, the larger the trend.
These results suggest that the DTR in the United States has not decreased due to global warming, and that analyses to the contrary were at least partly contaminated by station siting problems. Indeed, DTR tended to increase when temperatures were fairly stable and tended to decrease when temperatures rose. Maybe DTR really isn’t a robust signal of global warming, and maybe the discrepancy between models and observations is primarily a problem with the observations rather than the models!
I’ve used the words “suggest” and “maybe” here. That’s because I regard our results as tentative. The zero trend estimate is based on only 80 stations, which might be only marginally adequate. The systematic change in trend with station siting quality makes me more confident, but the fact that the closest poorly-sited stations have a weak but positive DTR trend suggests that DTR may be strongly site-specific and makes me less confident. Maybe the best-sited stations have actually improved their siting over time, and maybe the adjustments haven’t fully corrected for this. Because of all this, I think these results need to be confirmed through other means or in other parts of the world before I will wholeheartedly believe that the real DTR has not been decreasing.
Nonetheless, all the ongoing work to understand the consequences of a faster rise in minimum than maximum temperatures for ecosystems and human health might, just might, be misguided.
Read the full post here
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Can I applaud you and your colleagues for your labour of love in trying firstly to identify the best sited, maintained and monitored weather stations and secondly to carry out analyses to show that by determining which set of weather stations you use in your analyses, you may determine what conclusions you are going to draw.
I hope that the next 50 years sees the standards you have set rolled out globally in order that any scientific conclusions drawn about our climatic patterns are based on the best standards of data capture, comparison and standardisation.
Is 80 stations better than 1 tree?
I do not believe that temperature, due to your discovered difficulties of measurement, and the fudging which seems to continue despite the arguments against such questionable practice, is a good indicator of human effect on climate. Despite this, Anthony, your painstaking research on the siting of the surface stations has opened many peoples eyes to the many problems of data gathering. Well Done!
If the trend is that the worse the siting – the more negative the trend, I guess this must be because sitings such as next to buildings would provide shade, change the wind conditions, etc.
But could it also have to do with the age of the stations – that the worst sited tend to be older, with aging instrumentation? Or even different instrumentation, different huts, etc?
And if the best sited stations are the newest – could it be that these to a larger extent are located in areas more exposed to urbanization, thus being more affected by UHI?
I’m surprised at the usage you’ve made of the surface station data. The siting data is only a snapshot of the history of the sites. What is now a 1 may have been a 5 or a 5 a 1 historically due to either changes in the site location or site relocation. Your analysis seems to imply a static situation which I don’t think is supportable. Consider Washington National airport. When built in the 1940’s Washington was a sleepy southern town that shut down in the summer. Now the airport is surrounded by concrete and asphalt and is basically in the middle of an urban area. I think there is too much uncertainty in the historical siting to make the inferences you’re making.
Re: BarryW
Try reading the post before commenting:
I wonder could the 3 hour lag between maximum solar radiation and maximum temperature be because the humidity is generally lower in the afternoon than it is in the morning?
You forgot to adjust the most recent temperatures up and the adjust temperatures further in the past down!
This is a key scientific principle when measuring global warming. Don’t you understand science?
In response to the ‘why lag?’ comment – as long as incoming radiation exceeds outgoing radiation (as per the second graph), temperature will rise.
What interests me in the best sited stations is:
1. What is the trend for maximum daytime temperatures (both annually and month by month).
2. What is the trend for minimum night time temperatures (both annually and month by month).
3. How does this correspond with their counterparts in the entire station series.
Any feedback on this would be interesting.
TerryS says: May 19, 2011 at 5:15 am
Re: BarryW
Try reading the post before commenting:
“Maybe the best-sited stations have actually improved their siting over time”
Yes, airports are obviously the best placement:
http://chiefio.wordpress.com/2009/12/08/ncdc-ghcn-airports-by-year-by-latitude/
These posts were published in 2009 and reflect the same thinking:
http://hallofrecord.blogspot.com/2009/11/critique-of-october-2009-ncar-study.html
http://hallofrecord.blogspot.com/2009/11/non-linear-perspective-of-climate.html
http://hallofrecord.blogspot.com/2009/11/more-critique-of-ncar-cherry-picking.html
Guest Post “European Diurnal Temperature Range (DTR) Is Increasing, Not Decreasing by Francis Massen
Can someone remind me again why the tax-funded IPCC, ‘Climate Scientists’ or GISS cannot or did not do such a study. The Dutch had to get a private firm invovled before anyone took notice, in NZ it was said that actually their temp series wasn’t really a temp series and then over in the US you have to have individuals give up their free time to do such studies. I will not even comment on the MET/CRU here in blighty.
I guess when people are just simply handing you money hand over fist it’s too much bother to actually do an in-depth study before you have politico’s declaring that they will cut CO2 emissions by 80% in 20/30 years and all the economic hardships this puts on the poorest in our society.
Dear Shevva:
“……hardships on the poorest” and everyone else, too – please don’t forget!
So what? I read the post and stand by my comment. Using present conditions to extrapolate that the sites have the same conditions in the past is wrong. Especially when the location is not the same historically. Assuming that the site location has remained static you might assume that the microsite state is constant but only during that timeframe taking the data back to 1900 assuming that the CRN value is constant is making a big leap in my opinion especially if the site location has changed. Showing the data for the static siting timeframe would have more validity.
I spent four summers planting trees in Northern Ontario (Canada) – a completely forested area the size of France (maybe bigger). We lived and laboured outdoors 24/7 so we were very much in touch with the elements. We worked in huge clear-cuts, which were surrounded by the remaining forest. Late afternoon was the time I looked forward to with dread. It was almost always the hottest time of the day. The wind would die down, the sun would still blaze (albeit at a higher angle of incidence), and we toiled in the latent heat that accumulated throughout the day. Not a very scientific explanation, I realize, but the lag in temp. between noon and 3 – 5 P.M makes sense to me.
“Why lag?”
Heat inertia in the heated ground?
Yes, heat inertia in ground. Once did an experiment related to radon release mechanisms near a uranium mine, by using long stem thermometers set at different depths near Darwin Australia. There is a pronounced lag after the surface warms, before it warms at 20 cm depth (IIRC, the deepest we went) and then a later lag as it cools, of about the same duration. Unfortunately, I think the data were lost in a corporate takeover.
One question/observation regarding high temp of the day has a correlation in the low temp of the day. The lowest temp of a day is typically 30 – 45 minutes after sunrise, sometimes an hour. My basic understanding of this has to do with atmospheric heating and the change in wind direction. This causes the cooler air to blow from where the sun has not risen yet, blowing from cooler air masses to warmer air masses I wonder if there is a correlation in that at some point after noon, perhaps there is sufficient heating to cause the winds to blow from warmer air masses to cooler air masses until 5 pm’ish.
Anyway, just my $.02 worth…
We not only have a lag during the day but a lag for each season. The hottest days of summer are not around 21 June nor are the coldest days around 21 December.
It takes time to heat/cool a large mass of humid air. (daily) It takes time to heat/cool a large mass of water. (seasonal)
Using DTR as a single analog for global warming is fraught with potential problems. There are other influences that must be accounted for before reaching any conclusions. For example:
1. There could be global warming (or cooling) and the DTR remain constant if the driving force(s) affected both night and daytime temperatures equally.
2. The daytime temperature could go up (or down) or the nighttime temperature could go up (or down).
3. Natural cycles such as ENSO affect humidity which plays a role in heat transfer.
4. Natural cycles affect winds which play a role in heat transfer.
5. Shielding may affect daytime temperatures more than nighttime temperatures.
6. From a harm to earth perspective, increasing daytime temperatures are more critical than increasing nighttime temperatures. DTR does not capture this aspect.
While one always tries to make answers as simple as possible, DTR seems too simple to me to fully explain changes in the temperature of the atmosphere. Less I sound too critical, earlier I expressed appreciation to the whole team who did the work and published the paper on the impact of measurement stations on climate change. I repeat my thanks.
REPLY: No worries. I think you may not have read the entire post on his website. The IPCC made significant use of DTR as a metric for AGW, and John N-G talked about the history of that in context. Here’s his full essay. Also this recent guest post on Pielke Sr. ‘s website
Guest Post “European Diurnal Temperature Range (DTR) Is Increasing, Not Decreasing by Francis Massen is instructive, and also talks about the IPCC using it. – Anthony
BarryW has a reasonable point so what sort of result would the following give?
Take the current best 80 and the current worst 80. I’d suggest that the best 80 had tended to get better not worse on average over time and the worst 80 now would have tended to get worse on average over time. Reasonable?
Then plot the 2 sets of data and see how they differ. That would give an insight into how improvements and degradation affect the trends if at all.
Who would have thought something as simple as measuring and tracking temperatures would prove to be quite challenging?
Just defining the standards across a network of weather stations is a major task.
Lots of work and sort of boring after a while. No wonder a lot of guys would prefer to do all this temperature tracking stuff from a computer.
For the diurnal charts at the beginning of the post, just noting that the temperature change is exaggerated by quite a bit in these charts compared to a real-world station. There is far less change in the actual energy levels/temperature at the surface over a 24 hour period (between Max and Min), than should be expected.
http://img140.imageshack.us/img140/4109/tablemountainall.png
For the large differences in diurnal patterns that is found in Fall et al., there has to be quite large energy changes in terms of Watts/m2.