By Steven Goddard and Anthony Watts
Fort Collins, Colorado is most famous for Balloon Boy, and Boulder, Colorado is most famous for Jon Benet and Ward Churchill.
Both are hotbeds of Climate Science, with familiar names like Roger Pielke (Jr. and Sr.) Walt Meier, William Gray, Kevin Trenberth and Mark Sereeze. Both are of similar size (Boulder 91,000 and Fort Collins 130,000) and located in very similar geographical environments along the Front Range – about 50 miles apart. The big difference is that Fort Collins has tripled in size over the last 40 years, and Boulder has grown much more slowly. Fort Collins population is shown in blue and Boulder in red below.
Sources:
http://en.wikipedia.org/wiki/Fort_Collins,_Colorado
http://en.wikipedia.org/wiki/Boulder,_Colorado
Until the mid-1960s, NCDC temperatures in the two cities tracked each other quite closely, as you can see below. Again, Fort Collins in blue, and Boulder in red – with Fort Collins temperatures shifted upwards by two degrees to normalize the left side of the graph. Since 1965, temperatures in Fort Collins have risen much more quickly than Boulder, paralleling the relative increase in population.
Source: NCDC Boulder Temperatures NCDC Fort Collins Temperatures
The graph below shows the absolute difference between Fort Collins temperatures and Boulder temperatures since 1930. There is some sort of discontinuity around 1940, but the UHI imprint is clearly visible in the Fort Collins record. The Colorado State Climatologist, Nolan Doesken manages the Fort Collins Weather station. He has told me that it has never moved or changed instrumentation. and that he believes the increase in temperature is due to UHI effects.
Roger Pielke Sr. further commented:
“the Fort Collins site did have the introduction of the CSU Transit Center a few years ago, although this is well after the upturn in temperature differences between Boulder and Fort Collins started to increase.”
From the promotional photo on the CSU website, the Fort Collins USHCN weather station (below) seems reasonably sited.
However when you look at the Google Earth street view, you realize that it is surrounded by concrete, asphalt, nearby parking, and a building just 7.5 meters away (By the GE ruler tool). It would rate a CRN4 by the surfacestations rating. It also appears to have been modified since the promo photo was taken as there is a new fence with shrubbery and wood chips surrounding it.
Besides the pressure of CSU expansion, Fort Collins has seen an increase of about two degrees since 1970, corresponding to a population increase of 90,000. This is probably a little higher than Dr. Spencer’s estimates for UHI.
The Boulder weather station is similarly sited since the concrete path is just under 10 meters away.
It is at the campus of NOAA’s and NIST’s headquarters in Boulder. Anthony Watts visited the station in 2007 and took photos for the surfacestations project. Like Fort Collins, it gets similar expansion pressure due to nearby construction as seen in this aerial photo.
Here are the temperature records fro these two USHCN stations:
NCDC Fort Collins Temperatures
There is some UHI effect visible in the Boulder record below, but much less than Fort Collins.
Conclusion:
We have two weather stations in similarly sited urban environments. Until 1965 they tracked each other very closely. Since then, Fort Collins has seen a relative increase in temperature which tracks the relative increase in population. UHI is clearly not dead.
wayne (17:20:45)
I work on a Mac, and I use Excel for the graphs. I take screen shots with SnapZPro. I wrote my own Excel gaussian function.
I can write C, but I don’t use it anymore. I now use R, it’s much more suited for analyzing blocks of data.
w.
Willis Eschenbach (19:26:36) :
R, I suspected. Had already downloaded it but just never installed it. Thanks. However, I can seem to get Excel to have both left and right scales. Is that a Mac feature? I’m on XP Excel 2003, that might be the difference, the 2003.
Willis,
The slope from 1920-1940 is 0.007
The slope from 1895-1940 is 0.029
The slope from 1902-1940 is 0.040
You can get what ever slope you want by cherry picking the start date on the section prior to the discontinuity.
Again, please post your code.
Willis,
Thanks for getting me looking closer at the whole 1895-2008 record. Looking at the early years, it is clear that the other half of the discontinuity is before 1941(not after) in 1920
https://spreadsheets.google.com/oimg?key=0AnKz9p_7fMvBdElxNDA4Vlh2OGhvOUdEX1N0bm1CeWc&oid=2&v=1268456429186
I tried subtracting 0.7 from all years from 1920-1941 and came up with a graph that makes sense.
https://spreadsheets.google.com/oimg?key=0AnKz9p_7fMvBdElxNDA4Vlh2OGhvOUdEX1N0bm1CeWc&oid=6&v=1268456562370
I still have no idea what it is you are doing with your analysis. How you got a straight line in a linear analysis of this mess looks like black magic to me.
https://spreadsheets.google.com/oimg?key=0AnKz9p_7fMvBdElxNDA4Vlh2OGhvOUdEX1N0bm1CeWc&oid=2&v=1268456429186
Steve Goddard (21:05:09)
Steve, the whole record is interesting. However, I don’t find a break in 1920. See my residual analysis Figure W6 here. Once the 1941 discontinuity (a one-month step change of 0.6°C) is removed, the residual analysis does not show any other discontinuities in the record. So while you can claim that there is a 1920 discontinuity, the residual analysis disagrees. A discontinuity shows up as a deep “V”, not a single low value.
I’m not sure what you mean when you say “I still have no idea what it is you are doing with your analysis.” Do you mean the residual analysis? If so, I’m happy to explain it further. Let me know.
w.
Willis,
No need to explain. Please just post your code so that I can see how your graph was generated and attempt to reproduce it.
Steve Goddard (20:07:47)
As I said before, I am working in Excel, so there is no code.
Steve, you keep bringing up irrelevant points, like the trends prior to the 1942 discontinuity. What does that have to do with your claim about alleged changes in the latter half of the 20th century?
Next, you imagine some model of adding 0.7 to the years 1942-1953. I show that all that does is create a new discontinuity … and you ignore that. I show mathematically that there is a single anomaly in January 1942, and you call it arbitrary. So let’s go back to the basics.
You have claimed that the increase in Fort Collins – Boulder is due to the increase in Ft. Collins population post 1970. You say:
I asked you explain why, if that is the case, the trend 1942-1970 is the same as the post 1970 trend.
Or you could pick 1965. Trend 1942-1965 is 0.19, same as 1942-1970.
Trend 1965-present is 0.19, trend 1970 to present is 0.20
So no, Steve, the temperature did not “track each other quite closely until the mid-1960s” as you claim. They diverged at the same rate 1942-1965 as they did after 1965.
Where is the break that you claim to see in the data? I see no change in the divergence of the two datasets anywhere around the 1960s. If you want to postulate a reason that the pre 1965-70 data is different from the post 1965-70 data, first you have to show that they are different. I have shown that they are not different in several ways: by trends in this post, and by a graphic representation here. So where is the change in the trend that you are trying to explain via population?
Answer that question, and we’ll go forward from there.
wayne (19:32:35) : edit
I think it’s in XP 2003. Double-click on the line you want on a different axis, and click on the “Axis” tab.
Willis,
Spreadsheets use code functions very similar to other programming languages. Please make your spreadsheet available online. Why are you avoiding doing this?
Willis,
No doubt you make very pretty graphs and somehow manage to get them embedded in the comment section. I’m interested in how the graphs are generated.
Suppose that the trend from mid 1940s is linear. We seem to be in agreement now on that idea. That is right when the growth in population started to occur.
https://spreadsheets.google.com/oimg?key=0AnKz9p_7fMvBdElxNDA4Vlh2OGhvOUdEX1N0bm1CeWc&oid=3&v=1268464309132
Steve & Willis, I have downloaded Boulder and Fort Collins datasets form here
http://cdiac.ornl.gov/cgi-bin/broker?_PROGRAM=prog.climsite_monthly.sas&_SERVICE=default&id=050848
Using the “download file summarized by Year” I get different results to you, after year 2000 they have almost identical values. Unfortunately I have no idea how to show an Excel chart on here.
I’m out of town for a few days, but here are my final thoughts.
1. The discontinuity is likely to due to some sort of data collection issue at one of the sites between 1920 and 1941 (station move, TOBS, etc) The problem can be seen in the full data set.
https://spreadsheets.google.com/oimg?key=0AnKz9p_7fMvBdElxNDA4Vlh2OGhvOUdEX1N0bm1CeWc&oid=2&v=1268485447023
When all the years between 1929 and 1941 are reduced 0.7 degrees, the full record graph makes sense from a physical point of view.
https://spreadsheets.google.com/oimg?key=0AnKz9p_7fMvBdElxNDA4Vlh2OGhvOUdEX1N0bm1CeWc&oid=6&v=1268485011450
2. The divergence started in the mid-1940s when populations started to grow rapidly in both cities. It appears that the Ft. Collins station was affected more by growth than Boulder. As Tom Moriarty has astutely noticed, the current Boulder site is surrounded by open space.
3. The correct way to fix the discontinuity is by subtracting from 1920-1941, rather than adding after 1941 – an adjustment for which there is no theoretical basis.
Willis,
One last item. The slope from 1920-1941 is 0.007. There was no upwards trend before 1942.
Another Colorado whacko hits the front page:
http://www.foxnews.com/story/0,2933,589147,00.html
Is there a Climategate arm to Norway, too?
Norwegian cooler stations skipped, says Tom Segalstad;
Excerp from the newspaper article;
“The panel’s work (IPCC) can not be taken seriously until they change their objects clause, says Segalstad.
He served as an expert by reading the climate panel’s third report which was published in 2001. Then he pointed out that several monitoring stations in Norway was taken out from the scientific basis. Three of the four monitoring stations, which had been in the Second Assessment Report, was now gone. These measurement stations showed that temperature had fallen, not risen so the main conclusion of the second major report said.”
Here;
(Its a google translation, so…..but I think one can decode the meaning of it);
http://translate.google.no/translate?hl=no&sl=no&tl=en&u=http%3A%2F%2Fwww.bt.no%2Fnyheter%2Fklima%2FIkke-alle-vil-vaere-med-i-klimapanel-1046688.html
Same tactics as elsewere?
I strongly agree with Goddard and Watts that the UHI effect is a factor that is underrated, and that there is a remakable correlation between population rates and temperature rates of Boulder and Fort Collins.
I have examined temperature series (NASA/GISS) of Fort Collins and the rural station Chugwater, which is in Wyoming, 41.76N, 104.82W; 1617 m, about 78 miles N of Fort Collins. Fort Collins is 1525 m above sea level, Chugwater 1617m above sea level.
Here are the temperature series NASA/GISS from both staions from 1950-2009:
http://www.klimaatgek.nl/klimaatimg/fort1.jpg
And here is the difference between Fort Collins and Chugwater:
http://www.klimaatgek.nl/klimaatimg/fort2.jpg
There is hardly any trend visible. Can anyone explain this? Is there another factor that is overlooked?
Steve Goddard (23:03:13) : edit
Avoiding doing this? You asked for my code. I said I don’t have code, it’s a spreadsheet. You asked more my code again. I said again, I don’t have code, it’s a spreadsheet.
Now you ask for my spreadsheet, and insult me with a scurrilous accusation of bad faith that I am avoiding posting it.
That’s not nice. I have been operating in good faith all along. You have refused to answer any of my questions. Now you want to attack my actions, as though I were your servant and I’m avoiding your orders.
You made the claim, which I see you are now changing the goal posts on. You started by saying that the reason for the difference in the temperature records post 1970s was that Ft. Collins grew faster than Boulder. That was your hypothesis. I said it was not true.
I pointed out that
a) up until 1970 Boulder grew faster than Fort Collins, and
b) there was no change in the Ft. Collins minus Boulder trend anywhere in the post-1942 record.
Since there was no change in the trend, there was nothing to explain, so your hypothesis was false on the face of it. You had invented a reason to explain a non-existent phenomenon.
I invited you, a number of times, to calculate and report on the following trends:
1897-December 1941
February 1942 – present
February 1942 – 1970
1970 – present
I get ~ 0.2 for all of those. That alone means that your hypothesis is nonsense.
You seem to think these are randomly chosen dates to favor a point of view. They are from the start of the dataset to the discontinuity, from after the discontinuity to your 1970 date, from 1970 to the end of the dataset, and from the discontinuity to the end of the dataset.
Instead of reporting those trends, instead of just admitting your hypothesis was wrong and moving on, you want to argue about trivialities, and make up other ways to adjust the data without mathematical support, and waffle and wave your hands, and attack me for not knowing that when you say “post your code” you have some other meaning entirely.
So no, I won’t post my spreadsheet. It only concerns my analysis of the discontinuity, which I’m sorry I ever mentioned. The discontinuity is immaterial to the question of whether your hypothesis is correct.
Your hypothesis is wrong. If someone wants my spreadsheet, email me, I’m more than happy to share it, that’s not the issue. But I’m not happy to have you use it to distract people from the fact that your hypothesis was incorrect.
Once you accept the fact that your first explanation was wrong, however, you would notice that you have lit upon a fascinating question. Because once the one-month step-change discontinuity in the difference dataset is corrected, we see that the trends of the two cities have diverged at a remarkably constant rate for the last century. I show this above.
This is a curious and interesting finding. I don’t know why this should be, but it certainly bears further investigation.
Steve Goddard (05:35:57) : edit
You’re starting to hallucinate. Look at the chart here. From the start of the dataset in 1897 to 1941, temperatures went up by about a degree. A full degree upwards in about fifty years, how is that “no upwards trend before 1942”.
As shown by the graph, the trend from the 1920-1941 was smaller than the overall trend. But there is a very clear upward trend of 0.2°C/decade from the start of the dataset to 1941, and the 1920-1941 trend was still upwards.
Steve Goddard (05:05:40)
I say again. You can’t just eyeball a dataset and say “there’s a discontinuity”. There are recognized mathematical procedures to find them, one of which I detailed above. Using the procedure, we don’t have to handwave and say “between 1920 and 1941”. The procedure identifies the exact month of the discontinuity, January 1942.
Just like your claim about doing the same thing to the period 1942-1953, all this does is displace the discontinuity. In this case, rather than displace it to to 1953, it displaces it to 1929.
Steve, the math is clear. There is one and only one discontinuity in the Ft. Collins minus Boulder dataset. You keep claiming there are two … if you believe that, you must show it mathematically, not just squint at the graphs and make declarations.
We’ve gone over this many times. The divergence has been there since the records began in the late 1800’s. This is visible even in the uncorrected dataset. You keep repeating this claim as if repetition would make it true, but we can look at the graph. You’re like Groucho Marx saying “Who are you gonna believe, me or your own lying eyes?”
For our purposes, it doesn’t matter whether you add or subtract, because it makes no difference to the resulting trends. At present there is no theoretical basis favoring either, because we don’t know which dataset contains the error.
Willis Eschenbach (22:10:56) :
I think it’s in XP 2003. Double-click on the line you want on a different axis, and click on the “Axis” tab.
Willis, I’m indebted! After many years I had never stumbled across that simple tie. On the other hand, I have always been a rather light user of Excel, usually programming from the ground up, including charting. Thanks again!
wayne (17:10:58)
After twenty five years or so of driving Excel, and having written some fairly outrageous macros and systems, I know most of its foibles.
Speaking of Excel, here’s how I do the residual analysis to find the discontinuity:
Figure W7. Formula used to analyze the residuals to determine discontinuities in temperature difference datasets (e.g Ft. Collins minus Boulder)
The LINEST function with the variable set to TRUE (note the two commas) gives a five row by two column matrix containing the analysis statistics. The INDEX(somerange,5,2) function selects the residual sum of squares, which is in row 5 column 2 of what the LINEST function returns.
Note that the first function is anchored (with the dollar sign $) to the start of the dataset, and the second one is anchored to the end. Also note that the function leaves out the data point for the month being calculated.
To use this, you’ll need to remove the blanks from the dataset. The Excel LINEST function is dumb, it chokes on blanks. So just cut out the lines (both date and missing difference data) where one dataset or the other is missing temperature data and therefore the difference is also missing data. Makes no difference to the analysis.
Like I said, this is peripheral to the discussion of when the divergence began between Ft. Collins and Boulder, but it is a useful tool.
Other Excel questions related to analysis and display of temperature datasets? I’m happy to answer them.
w.
http://www.climate4you.com/LongyearbyenUHI%2020070202.htm
Henry chance (10:09:04) :
Boulder is morally superior. They have removed autos from downtown and other endeavors. The geography is also different. Boulder is adjacent to the slopes and mountains and Fort Collins is 15 miles east of the mountains.
I live in Boulder. It definitely sees itself as “morally superior” and “cool”. Hey, check out the Tesla dealership on the West end of the downtown mall sometime.
In reality Boulder lost its “coolness” to outsiders around 2001.
I can tell you Boulder has not removed autos from downtown. True, they converted part of it to a pedestrian mall in the mid-1970s, but if you consider the whole downtown area, no, there are still cars there. They make it very difficult for too many autos to park, though.
As to the population, I recall it getting up to 96,000 or 98,000 in the early part of this decade and then declining for the last several years to its current level.
Meant to also say, there is definitely a UHI in Boulder. I remember one really hot summer day in the late 1990s the power grid for the whole western part of the U.S. went down for a day (a power overload happened at a hydroelectric plant in CA, as I recall, which caused a cascade). Power was knocked out at my apartment in town, and it was stifling hot. A friend of mine lived in the countryside, 3 miles outside of the city limits. She had electricity. So I drove out there. Once I got out of my car I could immediately feel the difference. It was perceptibly cooler outdoors there than it was outside in town. This was before I had about the idea of UHI.
Correction:
“This was before I had about the idea of UHI.”
This was before I had heard about the idea of UHI.