Guest essay by Jim Steele, Director emeritus Sierra Nevada Field Campus, San Francisco State University
For researchers like myself examining the effect of local microclimates on the ecology of local wildlife, the change in the global average is an absolutely useless measure. Although it is wise to think globally, wildlife only responds to local climate change. To understand how local climate change had affected wildlife in California’s Sierra Nevada and Cascade Mountains, I had examined data from stations that make up the US Historical Climate Network (USHCN).
I was quickly faced with a huge dilemma that began my personal journey toward climate skepticism. Do I trust the raw data, or do I trust the USHCN’s adjusted data?
For example the raw data for minimum temperatures at Mt Shasta suggested a slight cooling trend since the 1930s. In contrast the adjusted data suggested a 1 to 2°F warming trend. What to believe? The confusion resulting from skewing trends is summarized in a recent study that concluded their “results cast some doubts in the use of homogenization procedures and tend to indicate that the global temperature increase during the last century is between 0.4°C and 0.7°C, where these two values are the estimates derived from raw and adjusted data, respectively.” 13.
I began exploring data at other USHCN stations from around the country and realized that a very large percentage of the stations had been adjusted very similarly. The warm peaks from the 1930s and 40s had been adjusted downward by 3 to 4°F and these adjustments created dubious local warming trends as seen in examples from other USHCN stations at Reading, Massachusetts and Socorro, New Mexico.
Because these adjustments were so widespread, many skeptics have suspected there has been some sort of conspiracy. Although scientific papers are often retracted for fraudulent data, I found it very hard to believe climate scientists would allow such blatant falsification. Data correction in all scientific disciplines is often needed and well justified. Wherever there are documented changes to a weather station such as a change in instrumentation, then an adjustment is justified. However unwitting systematic biases in their adjustment procedure could readily fabricate such a trend, and these dramatic adjustments were typically based on “undocumented changes” when climate scientists attempted to “homogenize” the regional data. The rationale for homogenization is based on the dubious assumption that all neighboring weather stations should display the same climate trends. However due to the effects of landscape changes and differently vegetated surfaces,1,2 local temperatures often respond very differently and the minimum temperatures are especially sensitive to different surface conditions.
For example even in relatively undisturbed regions, Yosemite’s varied landscapes respond in very contrary ways to a weakening of the westerly winds. Over a 10-year period, one section of Yosemite National Park cooled by 1.1°F, another rose by 0.72°F, while in a third location temperatures did not change at all.16 Depending on the location of a weather station, very different trends are generated. The homogenization process blends neighboring data and obliterates local differences and then fabricates an artificial trend.
Ecologists and scientists who assess regional climate variability must only use data that has been quality controlled but not homogenized. In a climate variability study, scientists computed the non-homogenized changes in maximum and minimum temperatures for the contiguous United States.12 The results seen in Figure A (their figure 1b) suggest recent climate change has been more cyclical. Those cyclical changes parallel the Pacific Decadal Oscillation (PDO). When climate scientists first began homogenizing temperature data, the PDO had yet to be named, so I would like to suggest instead of a deliberate climate science conspiracy, it was their ignorance of the PDO coupled with overwhelming urbanization effects that caused the unwarranted adjustments by causing “natural change points” that climate scientists had yet to comprehend. Let me explain.
Homogenizing Contrasting Urban and Natural Landscape Trends
The closest USHCN weather station to my research was Tahoe City (below). Based on the trend in maximum temperatures, the region was not overheating nor accumulating heat. Otherwise the annual maximum temperature would be higher than the 1930s. My first question was why such a contrasting rise in minimum temperature? Here changing cloud cover was not an issue. Dr. Thomas Karl, who now serves as the director of the NOAA’s National Climatic Data Center partially answered the question when he reported that in over half of North America “the rise of the minimum temperature has occurred at a rate three times that of the maximum temperature during the period 1951-90 (1.5°F versus 0.5°F).”3 Rising minimum temperatures were driving the average but Karl never addressed the higher temperatures in the 1930s. Karl simply demonstrated as populations increased, so did minimum temperatures even though the maximums did not. A town of two million people experienced a whopping increase of 4.5°F in the minimum and was the sole cause of the 2.25°F increase in average temperature.4
Although urban heat islands are undeniable, many CO2 advocates argue that growing urbanization has not contributed to recent climate trends because both urban and rural communities have experienced similar warming trends. However, those studies failed to account for the fact that even small population increases in designated rural areas generate high rates of warming. For example, in 1967 Columbia, Maryland was a newly established, planned community designed to end racial and social segregation. Climate researchers following the city’s development found that over a period of just three years, a heat island of up to 8.1°F appeared as the land filled with 10,000 residents.5 Although Columbia would be classified as a rural town, that small population raised local temperatures five times greater than a century’s worth of global warming. If we extrapolated that trend as so many climate studies do, growing populations in rural areas would cause a whopping warming trend of 26°F per decade.
CO2 advocates also downplay urbanization, arguing it only represents a small fraction of the earth’s land surface and therefore urbanization contributes very little to the overall warming. However arbitrary designations of urban versus rural does not address the effects of growing population on the landscape. California climatologist James Goodridge found the average rate of 20th century warming for weather stations located in a whole county that exceeded one million people was 3.14°F per century, which is twice the rate of the global average. In contrast, the average warming rate for stations situated in a county with less than 100,000 people was a paltry 0.04°F per century.6 The warming rate of sparsely populated counties was 35 times less than the global average.
Furthermore results similar to Goodridge’s have been suggested by tree ring studies far from urban areas. Tree ring temperatures are better indicators of “natural climate trends” and can help disentangle distortions caused by increasing human populations. Not surprisingly, most tree-ring studies reveal lower temperatures than the urbanized instrumental data. A 2007 paper by 10 leading tree-ring scientists reported, “No current tree ring based reconstruction of extratropical Northern Hemisphere temperatures that extends into the 1990s captures the full range of late 20th century warming observed in the instrumental record.”8
Because tree ring temperatures disagree with a sharply rising instrumental average, climate scientists officially dubbed this the “divergence problem.”9 However when studies compared tree ring temperatures with only maximum temperatures (instead of the average temperatures that are typically inflated by urbanized minimum temperatures) they found no disagreement and no divergence.10 Similarly a collaboration of German, Swiss, and Finnish scientists found that where average instrumental temperatures were minimally affected by population growth in remote rural stations of northern Scandinavia, tree ring temperatures agreed with instrumental average temperatures.11 As illustrated in Figure B, the 20th century temperature trend in the wilds of northern Scandinavia is strikingly similar to maximum temperature trends of the Sierra Nevada and the contiguous 48 states. All those regions experienced peak temperatures in the 1940s and the recent rise since the 1990s has never exceed that peak.
How Homogenizing Urbanized Warming Has Obliterated Natural Oscillations
It soon became obvious that the homogenization process was unwittingly blending rising minimum temperatures caused by population growth with temperatures from more natural landscapes. Climate scientists cloistered in their offices have no way of knowing to what degree urbanization or other landscape factors have distorted each weather station’s data. So they developed an armchair statistical method that blended trends amongst several neighboring stations,17 using what I term the “blind majority rules” method. The most commonly shared trend among neighboring stations became the computer’s reference, and temperatures from “deviant stations” were adjusted to create a chimeric climate smoothie. Wherever there was a growth in population, this unintentionally allows urbanization warming effects to alter the adjusted trend.
Climate computers had been programmed to seek unusual “change-points” as a sign of “undocumented” station modifications. Any natural change‑points caused by cycles like the Pacific Decadal Oscillation looked like deviations relative to steadily rising trends of an increasingly populated region like Columbia, Maryland or Tahoe City. And the widespread adjustments to minimum temperatures reveal this erroneous process.
I first stumbled onto Anthony Watts’ surface station efforts when investigating climate factors that controlled the upslope migration of birds in the Sierra Nevada. To understand the population declines in high-elevation meadows on the Tahoe National Forest, I surveyed birds at several low-elevation breeding sites and examined the climate data from foothill weather stations.
Marysville, CA was one of those stations, but its warming trend sparked my curiosity because it was one of the few stations where the minimum was not adjusted markedly. I later found a picture of the Marysville’s weather station at SurfaceStations.org website. The Marysville weather station was Watts’ poster child for a bad site; he compared it to the less-disturbed surface conditions at a neighboring weather station in Orland, CA. The Marysville station was located on an asphalt parking lot just a few feet from air conditioning exhaust fans.
The proximity to buildings also altered the winds, and added heat radiating from the walls. These urbanization effects at Marysville created a rising trend that CO2 advocate scientists expect. In contrast, the minimum temperatures at nearby Orland showed the cyclic behavior we would expect the Pacific Decadal Oscillation (PDO) to cause. Orland’s data was not overwhelmed by urbanization and thus more sensitive to cyclical temperature changes brought by the PDO. Yet it was Orland’s data that was markedly adjusted- not Marysville! (Figure C)
Several scientists have warned against homogenization for just this reason. Dr. Xiaolan Wang of Meteorological Service of Canada wrote, “a trend-type change in climate data series should only be adjusted if there is sufficient evidence showing that it is related to a change at the observing station, such as a change in the exposure or location of the station, or in its instrumentation or observing procedures.” 14
That waning went unheeded. In the good old days, weather stations such as the one in Orland, CA (pictured above) would have been a perfect candidate to serve as a reference station. It was well sited, away from pavement and buildings, and its location and thermometers had not changed throughout its history. Clearly Orland did not warrant an adjustment but the data revealed several “change points.” Although those change points were naturally caused by the Pacific Decadal Oscillation (PDO), it attracted the computer’s attention that an “undocumented change” had occurred.
To understand the PDO’s effect, it is useful to see the PDO as a period of more frequent El Niños that ventilate heat and raise the global average temperature, alternating with a period of more frequent La Niñas that absorb heat and lower global temperatures. For example heat ventilated during the 1997 El Nino raised global temperatures by ~1.6°F. During the following La Niña, temperatures dropped by ~1.6°F. California’s climate is extremely sensitive to El Niño and the PDO. Reversal in thr Pacific Decadal Oscillation caused natural temperature change-points around the 1940s and 1970s. The rural station of Orland was minimally affected by urbanization, and thus more sensitive to the rise and fall of the PDO. Similarly, the raw data for other well-sited rural stations like the Cuyamaca in southern California also exhibited the cyclical temperatures predicted by the PDO (see Figure D, lower panel). But in each case those cyclical temperature trends were homogenized to look like the linear urbanized trend at Marysville.
Marysville however was overwhelmed by California’s growing urbanization and less sensitive to the PDO. Thus it exhibited a steady rising trend. Ironically, a computer program seeking any and all change-points dramatically adjusted the natural variations of rural stations to make them conform to the steady trend of more urbanized stations. Around the country, very similar adjustments lowered the peak warming of the 1930s and 1940s in the original data. Those homogenization adjustments now distort our perceptions, and affect our interpretations of climate change. Cyclical temperature trends were unwittingly transformed into rapidly rising warming trends, suggesting a climate on “CO2 steroids”. However the unadjusted average for the United States suggests the natural climate is much more sensitive to cycles such as the PDO. Climate fears have been exaggerated due to urbanization and homogenization adjustments on steroids.
Skeptics have highlighted the climate effects of the PDO for over a decade but CO2 advocates dismissed this alternative climate viewpoint. As recently as 2009, Kevin Trenberth emailed Michael Mannand other advocates regards the PDO’s effect on natural climate variability writing “there is a LOT of nonsense about the PDO. People like CPC are tracking PDO on a monthly basis but it is highly correlated with ENSO. Most of what they are seeing is the change in ENSO not real PDO. It surely isn’t decadal. The PDO is already reversing with the switch to El Nino. The PDO index became positive in September for first time since Sept 2007.”
However contrary to Trenberth’s email rant, the PDO continued trending to its cool phase and global warming continued its “hiatus.” Now forced to explain the warming hiatus, Trenberth has flipped flopped about the PDO’s importance writing “One of the things emerging from several lines is that the IPCC has not paid enough attention to natural variability, on several time scales,” “especially El Niños and La Niñas, the Pacific Ocean phenomena that are not yet captured by climate models, and the longer term Pacific Decadal Oscillation (PDO) and Atlantic Multidecadal Oscillation (AMO) which have cycle lengths of about 60 years.”18 No longer is CO2 overwelming natural systems, they must argue natural systems are overwhelming CO2 warming. Will they also rethink their unwarranted homogenization adjustments?
Skeptics highlighting natural cycles were ahead of the climate science curve and provided a much needed alternative viewpoint. Still to keep the focus on CO2, Al Gore is stepping up his attacks against all skeptical thinking. In a recent speech, rightfully took pride that we no longer accept intolerance and abuse against people of different races or with different sexual preferences. Then totally contradicting his examples of tolerance and open mindedness, he asked his audience to make people “pay a price for denial”.
Instead of promoting more respectful public debate, he in essence suggests Americans should hate “deniers” for thinking differently than Gore and his fellow CO2 advocates. He and his ilk are fomenting a new intellectual tyranny. Yet his “hockey stick beliefs” are based on adjusted data that are not supported by the raw temperature data and unsupported by natural tree ring data. So who is in denial? Whether or not Gore’s orchestrated call to squash all skeptical thought is based solely on ignorance of natural cycles, his rant against skeptics is far more frightening than the climate change evidenced by the unadjusted data and the trees.
Literature cited
1. Mildrexler,D.J. et al., (2011) Satellite Finds Highest Land Skin Temperatures on Earth. Bulletin of the American Meteorological Society
2. Lim,Y-K, et al., (2012) Observational evidence of sensitivity of surface climate changes to land types and urbanization,
3. Karl, T.R. et al., (1993) Asymmetric Trends of Daily Maximum and Minimum Temperature. Bulletin of the American Meteorological Society, vol. 74
4. Karl, T., et al., (1988), Urbanization: Its Detection and Effect in the United States Climate Record. Journal of Climate, vol. 1, 1099-1123.
5. Erella, E., and Williamson, T, (2007) Intra-urban differences in canopy layer air temperature at a mid-latitude city. Int. J. Climatol. 27: 1243–1255
6. Goodridge, J., (1996) Comments on Regional Simulations of Greenhouse Warming Including Natural Variability. Bulletin of the American Meteorological Society. Vol.77, p.188.
7. Fall, S., et al., (2011) Analysis of the impacts of station exposure on the U.S. Historical Climatology Network temperatures and temperature trends. Journal Of Geophysical Research, Vol. 116
8. Wilson R., et al., (2007) Matter of divergence: tracking recent warming at hemispheric scales using tree-ring data. Journal of Geophysical Research–A, 112, D17103, doi: 10.1029/2006JD008318.
9. D’Arrigo, R., et al., (2008) On the ‘Divergence Problem’ in Northern Forests: A review of the tree-ring evidence and possible causes. Global and Planetary Change, vol. 60, p. 289–305
10. Youngblut, D., and Luckman, B., (2008) Maximum June–July temperatures in the southwest Yukon region over the last three hundred years reconstructed from tree-rings. Dendrochronologia, vol. 25, p.153–166.
11. Esper, J. et al. (2012) Variability and extremes of northern Scandinavian summer temperatures over the past two millennia. Global and Planetary Change 88–89 (2012) 1–9.
12. Shen, S., et al., (2011) The twentieth century contiguous US temperature changes indicated by daily data and higher statistical moments. Climatic Change Volume 109, Issue 3-4, pp 287-317.
13. Steirou, E., and Koutsoyiannis, D. (2012) Investigation of methods for hydroclimatic data homogenization. Geophysical Research Abstracts, vol. 14, EGU2012-956-1
14. Wang, X., (2003) Comments on ‘‘Detection of Undocumented Changepoints: A Revision of the Two-Phase Regression Model’’. Journal of Climate; Oct2003, Vol. 16 Issue 20, p. 3383-3385.
15. Nelson, T., (2011) Email conversations between climate scientists. ClimateGate 2.0: This is a very large pile of “smoking guns.” http://tomnelson.blogspot.com/
16. Lundquist, J. and Cayan, D. (2007) Surface temperature patterns in complex terrain: Daily variations and long-term change in the central Sierra Nevada, California. Journal of Geophysical Research, vol. 112, D11124, doi:10.1029/2006JD007561.
17. Menne. M., (2009) The U.S. Historical Climatology Network Monthly Temperature Data, version 2. The Bulletin for the American Meterological Society. p. 993-1007
18. Appell, D. (2013) Whither Global Warming? Has It Slowed Down? The Yale Forum on Climate Change and the Media. http://www.yaleclimatemediaforum.org/2013/05/wither-global-warming-has-it-slowed-down/
Adapted from the chapter Why Average Isn’t Good Enough in Landscapes & Cycles: An Environmentalist’s Journey to Climate Skepticism
Read previous essays at landscapesandcycles.net
@Nick Stokes
The TOBS may actually have a decent justification for the adjustments made. It is odd, however that the adjustments results are colder past and warmer present with a 0.3 degree slope.
What the author is referring to, however is the homogenization process which is very poorly explained in the USHCN article. It also is not represented in their handy little graphs. I’d be interested to know the net results of the changes made actually are…
The idea that vetted observers in Reading, Massachusetts back in the 1930s and 1940s were misreading peak temperatures on the order of several degrees Centigrade is so untenable as to require either the entire data set and all such similar sets be discarded or the “adjustments” be rescinded.
“Janice Moore says:
September 25, 2013 at 9:20 pm”
Here in Sydney, Aus, we are experiencing a record breaking spring. We already had the warmest winter on record apparently. We also have a new, nationwide, record average temperature…all determined by the use of 112 thermometers, 1 for every ~68,500 square kilometres, or less. no weather reporter mentions wind direction with these records but always compares a measured temperature with a, calculated, average, stressing any measurement that is ABOVE that average.
If a station has changed its time of observation during the historical observing period, then an adjustment is necessary. Unfortunately it is difficult to know the proper size of the adjustment, since this is site and season dependent. The adjustment can often be as large as the signal you are trying to resolve, which in this case is the change in temperature over many years. For instance, if a site changed its TOBS from 5 PM to 8 AM, that site will report cooler temperatures for no other reason than they changed their time of observation. A 5 PM observer will often sample a warm daily maximum temperature twice, once on the day it occurred and again the next day. This leads to a warm bias in reported temperature which is removed when the TOBS changes to 8 AM. For this reason I prefer to work with data from sites that report calendar day max/min temperatures so I don’t have to deal with a rather large and uncertain TOBS correction.
Hi Janice
Spring has sprung here in Canberra, and the parrots have all gone bush, where there are rich pickings at this time of the year. It seems that there is wisdom in the old rhyme:
Spring is sprung
Da grass is riz
I wonder where them boidies is?
“Da liddle boid is on da wing.”
“How absoid!
Da liddle wing is on da boid!”
This was taught to us in 4th class as an example of a New York accent.
Anyway, there are not a lot of birds about just now; they will return as the weather heats up and they want a drink and a bath at my place. What’s happening at yours?
I’m sure that Jim won’t object to this interlude.
Best – J
Johanna, If you are going to add an interlude, when you mention “lots of “boids” around, then you should provide a few names. It helps the visualization.
Jim Steel @ur momisugly 9.36: Well said.
Nick Stokes says:
September 25, 2013 at 6:27 pm
I’m saying is that the adjustments were made, by stated algorithm, for reasons that have been extensively described, in numerous papers, and those reasons should be looked at.
============
The excellent article did just that and explained how the computer algorithms behind the adjustments were faulty. They saw urbanization as a natural process not needing adjustment, and the PDO as an unnatural process needing adjustment.
The author has gone on to argue that this was not a result of malicious intent, it was a result of ignorance. At the time climate scientists did not realize that there were cyclical natural processes, so they did not allow for them in the computer code.
I would add to that the experimenter expectation effect. The adjustments coincided with what the researchers believed was happening – temperatures were increasing due to human activity – so they didn’t question the accuracy of the adjustments.
Had the adjustments shown that temperatures were not increasing, that instead they were cyclical, this would have gone against what was expected, and the assumption would have been that the adjustments were faulty and in need of revision.
So, over time the result meet expectations, whether they are correct or not. And since the effect is subconscious, researchers that are involved are powerless to detect it without extremely careful experimental design.
Conspiracy to falsify the temperature record?
Or, no conspiracy — just incompetence?
You decide.
But notice that every USHCN temperature ‘adjustment’ goes in the most alarming direction…
“Nick Stokes says:
September 25, 2013 at 4:12 pm
[…] And the adjustments aren’t all homogeneity. At least as large a component is TOBS. This is simply an adjustment that arises where the time of observation has undergone a recorded change, which has a predictable effect on measured minmax temp. […]”
I read you linked site. Apparently the time-of-observation adjustments aren’t based on actual recorded data, but are just estimated adjustments based on assumptions regarding time-of-observation variations, which has essentially the same problems as the homogeneity assumptions.
Boy do I ever agree with you Jim Steele. If you have the capabilities keep posting such articles, they needs to be held up to their eyes weekly. This is why I keep saying temperature CHARTS, not the actual temperatures, are never going to return to 70’s level even if they actually do in reality.
Manufactured artificial warming.
Janice, what we have here in Banora Point (just south of the Queensland border) is “WEATHER”. When the wind blows from the south, it is cool. When it blows from the north, it is warm. We are officially in spring – for some reason Australia works on 1 September, etc, instead of the equinoxes and solstices to determine the season.
These prevailing winds can last for a fortnight or more. Sometimes it rains, sometimes it doesn’t. And sometimes it absolutely buckets down, like when in June a few years ago we had 530 mm in 36 hours. And we had a few years of drought, so several of the states built desalination plants, because the “Climate Commission” said the dams would never fill again. When the desal plants were ready, it rained (they haven’t been used since!), and the operators of a dam west of Brisbane had to let water out of the dam as they were worried that the water level would rise so high as to overtop the dam – which might have been too much for the strength of the dam. Unfortunately, when the water reached the valley below, it met a bit of even heavier rain – disaster and then floods in Brisbane. See http://en.wikipedia.org/wiki/2010%E2%80%932011_Queensland_floods
for comment. Luckily, unlike New York and New Orleans. Queensland Rail parked its electric trains on high ground to wait it out!
We have seen the photos of water flooding Colorado – a well known “desert”! Hope that the missing 500 people have been found safe by now.
I just checked Orland in the Historical Observing Metadata Repository (HOMR) and it shows no change in ObsTime since 1931. The ObsTime seems to have remained at 0800 LT throughout the period from 1931-2013:
http://www.ncdc.noaa.gov/homr/#ncdcstnid=10100110&tab=PHR
The metadata shows and SurfaceStations.org documentation confirms Orland has not changed over to the electronic Maximum/Minimum Temperature Sensor, MMTS, so no adjustment should be necessary for change of instrumentation. It looks like the ObsTime changed from 1700 LT to 0700 LT in 1934, meaning an ObsTime adjustment should have been applied to the portion of the temperature record preceding 1934. Here is a raw observer form from Orland which shows the ObsTime in Sept 1929 was 1700 LT:
http://www1.ncdc.noaa.gov/pub/orders/IPS-8F4F6DD6-5EFE-4FA5-BDC7-B38262D85EEA.pdf
As is often the case, the metadata from HOMR seems to conflict with the ObsTime found on the raw observer forms. Yes, working with climate data is messy given the sparse and conflicting metadata available.
BTW, re accuracy of data. Many years ago, the Australian Met Office got worried about the data received from one of the outback weather stations. It bore little or no resemblance to the data from other stations, which were probably a few hundred miles away, but it should have been consistent. Eventually a man was sent out to check the instruments. When he arrived, he found that the lady’s husband had died during the previous year, and she did not know how to read the instruments. As she needed the honorarium her husband had received for telegraphing in the data, she just telegraphed in the data for the previous year. I believe that the gentleman decided the best thing to do was to teach her how to use the instruments correctly. All well from then on.
I’m not sure that I want to go there, Jim. 🙂
However, for the more literal minded, Canberra is awash with parrots during autumn and winter, principally because it is an artificially created national capital in the middle of bare sheep paddocks. It has millions of European trees which produce nuts and seeds, plus lots of well watered grass, in an otherwise bleak landscape (from a parrot’s point of view).
There are also some artificial lakes, and thousands of acres of heavily planted suburban backyards, which many other kinds of birds thrive in. Because summers are very hot and dry, just having a daily refreshed birdbath and some sympathetic plantings in my modest backyard brings me a daily nature documentary of birdlife in summer.
wayne says: “If you have the capabilities keep posting such articles, they needs to be held up to their eyes weekly.”
I intend to adapt an essay from the book every few weeks. I wrote the book to be scientifically robust yet readily understood by the layperson. So far comments on Amazon suggest I succeded. My hope is the book will be read widely by the public as well as used in environmental studies classes. I just got word it already made it to the suggested reading list at one such University class.
Re Mark Albright
If the OBS Time changes, you note that a break in series occurred at that date, and DO NOT MODIFY the previous data. Same if there is a change of location, or the immediate vicinity changes from open grassland to suburban car park. Why would only climate persons (cannot call them scientists) modify previous data? No financial, unemployment or other record that I know of has ‘modified data’ – always there is an indication of “break of series” when the definitions or something else has changed.
Thanks for explaining this – a worthy piece.
Mr Steele, your Fig. A (their figure 5b not 1b) is in fact GDCN not GHCN, a minor difference perhaps for the US but NOAA do say “Unlike GDCN, however, GHCN-Daily contains
numerous data streams for updated data that enhance the latency of the dataset through
rapid and frequent updates. Relative to GDCN, GHCN-Daily also contains a much more
comprehensive set of QA checks as well as a more expansive set of historical data sources. “, so there is a difference. Also the graph is the average of a graph the authors say is “only a rough and visual assessment of the change points. Rigorous detection of change points in a time series needs to go through a statistical procedure” My main point is though that you use that graph and say “The results seen in Figure A (their figure 1b) (sic) suggest recent climate change has been more cyclical” whilst they say…..”such data sets must be homogenized before they can be used for climatological analysis”. Is there a reason you can make a climatological judgement from that graph when the authors you site say specifically you can’t?
TIA Nick
From Nick Kermode on September 25, 2013 at 11:04 pm:
Reference in question, Shen et al (2011), “The twentieth century contiguous US temperature changes indicated by daily data and higher statistical moments”
Available here on Springer for $40 (has abstract, references, but no supplementary info):
http://link.springer.com/article/10.1007/s10584-011-0033-9
Free version:
http://www-rohan.sdsu.edu/~shen/pdf/shen_climc_2011.pdf
Jim Steele,
You might find our recent paper on urbanization and homogenization interesting: http://onlinelibrary.wiley.com/doi/10.1029/2012JD018509/abstract
It turns out that pairwise homogenization tends to do a good job at removing urban-correlated biases even under very stringent definitions of urbanity and when only rural stations are used in the homogenization process. Its also worth pointing out that the net effect of CONUS homogenization is actually to lower TMin relative to TOBS-only adjustments.
As far as general reasons for homogenization goes, they are covered pretty well here: http://rankexploits.com/musings/2013/a-defense-of-the-ncdc-and-of-basic-civility/
Similarly, MMTS transition biases are one of the larger contributors to the increase in CONUS max temps post-homogenization: http://rankexploits.com/musings/2010/a-cooling-bias-due-to-mmts/
In general, raw data is problematic due to the fact that weather stations were not designed to be climate stations. Over the last century most have moved two or three times, had at least two different instruments, have had time of observation changes, and other issues whose impact can dwarf the background climate signal. Some of these are true biases (e.g. instrument changes), while others are changes in true local condition that are not representative of the broader regional climate (e.g. urbanization). One of the current challenges is that homogenization cannot effectively differentiate between the two.
You might also find the Berkeley homogenization process worth a look. We provide charts showing specific breakpoints as well as difference series used to detect step changes. Here is Reading, MA, for example: http://berkeleyearth.lbl.gov/stations/163049. The station has had three document station moves and two documents TOBs changes in its history, as well as a number of other notable undocumented step changes relative to surrounding stations.
Hey, that is a perfect anagram of Stephan Lewandowsky …
kadaka (KD Knoebel),
Thanks, I should have included a link to the paper.
I stand corrected. According to tonight’s weathercast on Channel 10, Australia has had it’s warmest 12 MONTHS on record.
Excellent contribution. Just superb!