Ross McKitrick writes:
I give a demonstration of why the Parker and BEST analyses don’t disprove the evidence of contamination of temperature data, and outline what it would likely take to settle the issue properly.
Cheers, Ross
ENCOMPASSING TESTS OF SOCIOECONOMIC SIGNALS IN SURFACE CLIMATE DATA: I have a new paper out in Climatic Change on the question of whether surface climate data are biased by non-climatic factors relating to socioeconomic development:
Rather than try to settle the debate once and for all, I focus on why the various attempts to show the data are not contaminated do not disprove the results showing that they are. The problem has been that authors use non-overlapping data sets and different methods, and end up talking past each other. The way to settle matters, I argue, is to adopt an encompassing framework in which both types of results can be demonstrated on the same data set, where one arises in a restricted subset of another model, and the restrictions can formally be tested.
I give two examples, one replicating a Parker-style equivalence between nighttime minimum trends in calm and windy conditions, then showing that this persists in a temperature data set that can be shown to be correlated with population growth. I also replicate the BEST-type results that rural trends are slightly greater than those of urban areas, and show that this result appears in a restricted subset of a larger model in which socioeconomic growth is significantly correlated with temperature trends. In both cases the restrictions necessary to yield the model that supposedly shows no data contamination are rejected. Data/code archive here.
Posted at http://www.rossmckitrick.com/.
McKitrick, Ross R. (2013) Encompassing Tests of Socioeconomic Signals in Surface Climate Data
Climatic Change doi 10.1007/s10584-013-0793-5
http://link.springer.com/article/10.1007%2Fs10584-013-0793-5
Abstract
The debate over whether urbanization and related socioeconomic developments affect large-scale surface climate trends is stalemated with incommensurable arguments. Each side can appeal to supporting evidence based on statistical models that do not overlap, yielding inferences that merely conflict but do not refute one another. I argue that such debates are only be resolved in an encompassing framework, in which both types of results can be demonstrated as restricted forms of the same statistical model, and the restrictions can be tested. The issues under debate make such data sets challenging to construct, but I give two illustrative examples. First, insignificant differences in warming trends in urban temperature data during windy and calm conditions are shown in a restricted model whose general form shows temperature data to be strongly affected by local population growth. Second, an apparent equivalence between trends in a data set stratified by a static measure of urbanization is shown to be a restricted finding in a model whose general form indicates significant influence of local socioeconomic development on temperatures.
I have just read an interesting book that bears on this subject. I read a book “With Brass and Gas” by Munson Baldwin. This is a history of ballooning in the 19th century. In this book toward the end he relates from period articles some of the things that make a balloon go up and down. Over water the balloons lose altitude (These were all hydrogen balloons) due to the cooler air over water. The balloons gain altitude over towns, fields, and even individual farms due to the warmer air over these locations.
This phenomenon has never properly been documented temperature wise much less integrated into regional or general circulation models…
Ross – I think it would have been worthwhile pointing out that Watts 2012 did demonstrate that surface climate data are biased by non-climatic factors relating to station siting, and that factors relating to socioeconomic development are only another possible source of bias from non-climatic factors, not the only one.
Now don’t you suppose the Chinese knew what was going on in their cities?
============
No one disputes urban areas are warmer, but the AGWers argue that there is no evidence for an increasing trend in urban temperatures. This is partly true.
They miss that the urban influence in the warming in the global surface data is mainly due to increasing urbanization, both horizontally and vertically (taller buildings). Population changes are a good proxy for urbanization changes.
Is part of the point here that on windy days, the UHI effect will be dispersed over a wide area and therefore the upward impact on temperature will be severely diluted within the heat source urban area? I guess that makes sense.
The point then is you need to use a dataset on non-windy days or nights, which immediately opens you up to charges of cherry picking. Worse still, it would probably give the alarmist data keepers the opportunity to somehow manipulate the past into being cooler yet again.
Data-diddling leaves fingerprints? Who knew?
The other issue is that raise urban temperatures are then given a equal weighting with rural temps that would generally cover a much larger area per station.
Urban areas occupy but a small percentage of the land surface and should ONLY be applied to the area they represent. There is a large urban bias in the global surface temperature calculation because of this.
Slightly OT But Nuticelli is still ascribing the Moscow heatwave in 2010 to AGW.
http://discussion.guardian.co.uk/comment-permalink/24402058
The idea that it is measurement errors, UHI or natural variability is verboten.
I tried to post a link to NOAA pointing out that it was probably natural variability but am currently banned.
http://www.noaanews.noaa.gov/stories2011/20110309_russianheatwave.html
However, another post has been deleted as well. They weren’t on ‘pre–moderation’ so they can be seen to be deleted.
It might be funny if a whole string of deleted comments turned up. It would be obvious that he was on weak ground… he he he.
Here’s an example of urban bias… a simple little calc that even AGW catastrophists may be able to manage.
Suppose in an area of 1000km^2 there are 3 temp stations.
Station A shows an increase of 1 deg C in 10 years
Station B shows an increase of 1.2 deg C in 10 years
Station C shows a decrease of 0.2 C in 10 years
What is the average temperature change for the region in those 10 years?
Now suppose I tell you that A is in an urban area of 50 km^2
and B is in another urban area of 100km^2
C is a rural temp unaffected by UHI rises.
What is the area weighted average temperature rise ?
An of course, under GISS or similar, they would probably “homogenise” the rural reading up to something like 0.8C increase in 10 year (totally changing the real rural temperature trend) to sort of match the 2 urban temps, then calculate on unweighted areas.. which gives a change of .. ?
Then they would apply an “adjustment” that lowers 10 year old temps by 0.3 degrees
And suddenly …. !!
Assuming that satellite data are insensitive to urban heat island effects. I wonder if it can be used to settle the whole thing. Maybe it’s possible to select a grid where the strongest urbanisation has taken place after 1979, the start of the satellite data, and compare the local temp trend with the satellite trends.
Look at slide 52 here:
https://docs.google.com/viewer?url=http://wattsupwiththat.files.wordpress.com/2012/07/watts-et-al-station-siting-7-29-12.ppt
That’s an order of magnitude difference from the official. 10 (ten) times difference.
The very definition of urban vs. rural is silly, that is, e.g. all settlements with less than ten thousand inhabitants are classified “rural”. At the same time it was amply demonstrated that even pretty small settlements may have considerable UHI, roughly proportional to the logarithm of local population density (~0.25°C/doubling). This relation only breaks down at a very low threshold, happily exceeded by most rural sites.
Therefore what counts is not immediate population density, but its trend over the last century. As global population has doubled almost twice during that timespan and the entire world has got urbanized, temporal UHI must be present almost everywhere around people.
It is a well known fact global population distribution is fractal-like and measurement points (weather stations) are not randomly distributed over land surface relative to this fractal. This is so, because it is much cheaper to operate / maintain stations close to spots of habitation and/or economic activity (like airports) than in the middle of nowhere.
Ross, there’s a small typo here.
‘I argue that such debates are only be resolved in an encompassing framework’
Land surface warming is highly variable given the chaotic weather patterns that overlay it and the onfluence of UHI effects amonst many other things.
But there are no socio-economic factors in ocean heat content, it represents at least 90% of the energy accumulated by the effect of rising CO2 and shows no sign of urban effects of any oth anthropogenic influence except the warming from the rising anthropogenic CO2.
http://onlinelibrary.wiley.com/doi/10.1002/grl.50382/abstract
Volcanic eruptions and El Niño events are identified as sharp cooling events punctuating a long-term ocean warming trend, while heating continues during the recent upper-ocean-warming hiatus, but the heat is absorbed in the deeper ocean. In the last decade, about 30% of the warming has occurred below 700 m, contributing significantly to an acceleration of the warming trend. The warming below 700 m remains even when the Argo observing system is withdrawn although the trends are reduced.
I am monitoring temperatures in my garden, one sensor in a shady corner the other next to the house. They are within 50ft of each other. temperature differences vary with conditions and the maximum difference between the max. temperatures is 5C. This is far more than I expected but an indication that UHI is alive and well and that all temperature data must be accurately taken and recorded on temperature sensors that conform to basic positional standards to try to remove UHI errors.
@ johnmarshall
Trouble is, how do you accurately identify all UHI factors and remove them from the temperature station. They may be very close to the device, (eg evap units), there may be a UHI effect over a larger area. Could also be prevailing wind issues at a particular time of day in certain weather patterns.. You just don’t know !! UHI exists, but is obviously VERY badly accounted for.
This is what makes the land surface readings so meaningless, especially once you let GISS or HadCrud at them !!
On but sort of off-topic to this: Skeptics have been talking about UHI for years… and yet, True Believers are now shrieking that we deny the effect of UHI? I’m surprised there has been nothing on that farce.
@- AndyG55
“This is what makes the land surface readings so meaningless, especially once you let GISS or HadCrud at them !!”
Until recently land surface weather stations were intended to measure the macroscopic changes in weather from day to day and season to season. They were certainly NOt designed to detect small trends over decades and its a minor miricale that ANY significant data can be derived from them.
Deriving the global warming signal from ocean temperatures, moving plant type regions and the land ice mass balance is probably a lot more reliable than the extensive processing required to find the warming trend in data from surface temperature measurements.
However, even given the inadequacy of the surface temperature record in recording the observed warming, it is still possible to see that it is confirmed by data sources that are not contaminated by any UHI or siting issues. The surface trend without a UHI component is confirmed by measurement by satellite of global troposphere trends and sea surface trends.
Neither of which have had any significant urban build-up over the period in question.
“not contaminated by any UHI or siting issues. ”
BS !!
izen says:
June 18, 2013 at 2:55 am
////////////////////////////
The satellite data, as you suggest, does not have the same contamination problems. Apart from the step change in and around the 1998 super El Nino it is all but flat both before and after that event. There is no first order correlation seen in that record between rising CO2 levels and temperatures (at least not unless the Super El Nono of 1998 was caused by CO2, which I do not understand any one suggests). The satellit data therefore suggests that the warming seen in the land based thermometer record between say 1979 and about 1996/7 may well be the result of contamination of that record by UHI, siting issues, station drop outs and questionable adjustments/homogenisation.
Ocean temperature is the only important data since this deals with energy and because of the heat capacity of the oceans compared to that of the atmosphere. However, pre ARGO the data is riddled with uncertainties, inconsistencies and potential errors bars are so high that it is not fit for purpose (ie., it cannot shed light on whether the oceans at any layer have warmed to any significant extent). ARGO may be able to provide this info, but the duration of the data set is too short, there is still insufficient coverage given the vastness of the oceans, issues abound with its calibrationand splicing onto earlier ocean temp data sets, and of course, ARGO buoys drift such that they never make like for like measurements and the drift may in itself lead to a bias (the drifting may be influenced by currents which themselves have a distinct temperature profile different to the ocean at large).
HadCrud and Giss have clearly climbed more than RSS and UHA atmospheric temperatures.
Now since HadCrud and GISS have not been subject to any adjustments, this is an obvious sign of urban heat effects.
http://www.woodfortrees.org/plot/hadcrut4gl/from:1979/plot/rss/from:1979/plot/uah/from:1979/plot/gistemp/from:1979
QED.
John Silver says: @ June 18, 2013 at 1:03 am
Look at slide 52 here:
https://docs.google.com/viewer?url=http://wattsupwiththat.files.wordpress.com/2012/07/watts-et-al-station-siting-7-29-12.ppt….
>>>>>>>>>>>>>>>>>>>>>>>>>>
John, I get
and then a request to LOG in
Whilst discussing UHI, Tokyoboy recently posted a comment on another thread to the effect that last century Tokyo warmed by 3degC whilst an adjacent Japanese Island just 180km away showed no warming.
I suspect that Tokyo is one of the weather sites used in the world series, and the weather station on the adjacent nearby Japanese Island is not used.
izen says:
June 18, 2013 at 1:58 am
_____________________________________
This response seems to be citing the Trenbreth, et al. abstract. Isn’t this the missing heat paper that tries to reconcile a calculated energy balance discrepancy by putting it into deep ocean? I’m not sure I quite understand how you can warm the deep ocean without warming the surface. Nor do I understand how this supposed effect negates any discussion on the effects of population density on measured temperature changes.
Every day on BBC weather forecast the presenter says ” tonight it will be X degrees C in the countryside with the the towns and cities 2 or 3 degrees warmer….”.
Quelle surprise!
@- richard verney
“The satellite data, as you suggest, does not have the same contamination problems. Apart from the step change in and around the 1998 super El Nino it is all but flat both before and after that event. ”
Not correct. Do the maths and calculate the linear trend between 1979 and 1997, there is a clear rising trend of over 0.1degC/decade.
@- “The satellit data therefore suggests that the warming seen in the land based thermometer record between say 1979 and about 1996/7 may well be the result of contamination of that record by UHI, siting issues, station drop outs and questionable adjustments/homogenisation.”
Not correct, the satellite data clearly shows a warming trend with ENSO and volcanic effects overlying that. As you acknowledge that warming trend is NOT a UHI effect, it is a real increase in the energy content of the climate system.
@-“However, pre ARGO the data is riddled with uncertainties, inconsistencies and potential errors bars are so high that it is not fit for purpose (ie., it cannot shed light on whether the oceans at any layer have warmed to any significant extent). ”
But then the thermal expansion of the oceans IS unequivocal over the same time, however much you may wish to relegate the data to useless it is confirmed by the consilience of multiple paths of evidence that ocean heat content is rising.
All without any economic impact from human activity except the physical process by which the increasing atmospheric CO2 from burning fossil fuel is increasing the energy gradient from surface to tropopause because of its radiative absorption converting outgoing LWR photons to thermal energy.
AndyG55 says:June 18, 2013 at 2:45 am
“Trouble is, how do you accurately identify all UHI factors and remove them from the temperature station.”
Extremely easy Andy. Shut down all the urban sitings all together and scrub the historical data from the record and only use accepted standardised principles for placement from rural sitings.
@David Schofield says:
Every day on BBC weather forecast the presenter says ” tonight it will be X degrees C in the countryside with the the towns and cities 2 or 3 degrees warmer….”.
I’ll know that the BBC has stopped supporting the alarmist position when they say:
” tonight it will be X degrees C in the towns and cities with the the countryside 2 or 3 degrees colder….”.
The correlation seems robust and is not surprising. However, I wonder if it has a major effect on the temperature change (trend) over time. If it did, wouldn’t the trends in the surface temperature records (GISS, HadCRUT, etc), which presumably do not correct appropriately for UHI effects be substantially different from satellite trends?
I suppose those who live in UHI’s that rarely if ever leave would be prone to disregard rural temperatures. There are those who travel from airport to airport, city to city, UHI to UHI who are basically concrete and asphalt creatures. Add the control freak mental cases who leap onto any agenda that blames man for everything; add those greedy hearts who see an opportunity to blindside non-thinkers into carbon taxes, and we end up with the momentum we have seen in the AGW agenda.
Any independent thinker can observe their surroundings and be cognitive that temperatures vary greatly just by where one chooses to stand and so forth. Attempting to measure a planets average temperature is not easy from the surface. An accurate long term thermal image record from space is not available. The long and short term influences are endlessly chaotic. At best we can generalize. We adapt.
Ross, this gent wants to fight you:
http://www.sfgate.com/science/article/Warning-from-climate-science-front-lines-4569221.php
An area where both sides of the argument should be careful about what they wish for?
If rising levels of anthropogenic CO2 cause measurably increasing temperatures on a global scale, then I might reasonably expect it to be observable in the small fraction of the Earth’s surface where most of the anthropogenic CO2 is produced, which is, errrrrr, urban areas….
I have yet to see this specific point addressed in detail.
izen says:
June 18, 2013 at 5:07 am
Not correct. Do the maths and calculate the linear trend between 1979 and 1997, there is a clear rising trend of over 0.1degC/decade.
=====
izen, this is the most rediculous thing that anyone posts…
You can’t get a trend from such a small change in temps….over such a short period of time
Look at any temp reconstruction paleo…..temps change in whole degrees…over hundreds of years..going one way
…when the overall trend is going the other
So if there has been no warming in 17 years, this means no population growth. Interesting claim.
Sceptical that is a complete non-sequitur, a logical fail: if the true global average is cooling and this is exactly offset by the UHI effect, we will see a flat reported global average. Was the fail accidental or deliberate?
Muller assumed that UHI would be greater in urban than in rural areas, but that is only true when looking at the TOTAL local human effect on temperature, not the marginal effect. It is very possible that the first marginal increments of UHI “pollution” of the temperature signal in rural areas will be larger than the marginal effects in areas where total human effects are already much larger. Thus areas that are still properly classified as rural may have on-average experienced greater increments of UHI pollution over a particular time period than more urbanized areas.
In particular, it is likely that the rural stations themselves expanded heating and air conditioning while adding paving, outbuildings, building additions, etcetera. The impact on measured temperatures of these kinds of changes in station siting would be greater in rural areas than in urban areas.
Decreasing marginal effects are so common that in economics they are referred to as “the law of diminishing marginal returns.” This is a well-known factor that should not be overlooked.
Stephen Pruett says:
June 18, 2013 at 8:00 am
The correlation seems robust and is not surprising. However, I wonder if it has a major effect on the temperature change (trend) over time. If it did, wouldn’t the trends in the surface temperature records (GISS, HadCRUT, etc), which presumably do not correct appropriately for UHI effects be substantially different from satellite trends?
1) they are different
2) the growth slowed down. In absolute numbers we see now the same population growth as for 40-60 years, however we have more then double the total population
People, Ross McKitrick is not debating UHI. His point is that we have been comparing apples and oranges, and that we should develop a methodology to bring them together so that they can be compared properly.
I agree with Mike Jonas upthread who pointed out that Watts et al indicates that the classification of weather stations is a significant issue in this regard.
@Izen
“Not correct. Do the maths and calculate the linear trend between 1979 and 1997, there is a clear rising trend of over 0.1degC/decade.”
That’s 1degree C per century( IMO noise)
This has been so since the LIA ended. No?
Surely no CO2 signal there.
Surely no need to vilify fossil fuels over 1 deg.C/Century
Surely no need to measure “Carbon footprints” of guilt.
Surely no need to alarm school children with doomsaying catastrophism.
Surely no need to limit third world prosperity.
Surely no need for any of this horsesh!t.
izen says:
June 18, 2013 at 5:07 am
///////////////////////////////
I beg to differ. I attach a plot of the UAH data set 1979 to 1996. http://www.woodfortrees.org/plot/uah/from:1979/to:1996/plot/uah/from:1979/to:1996/trend
You will note that in the 17 years the linear trend shows a temperature anomaly rise from about -0.15degC to -0.11degC, ie., a change of 0.04degC in 17 years which is less than 0.02degC per decade.
I am sceptical (for many reasons) as to the extent of thermal expansion of the oceans. But even if there has been such expansion, it does not follow that it has been caused by CO2. Infact, it is difficult to see how DWLWIR can heat the oceans (given the absorption characteristics of LWIR in water which would appear simply to promote surface evaporation thereby cooling the very top micron layers of the ocean), and the most likely explanation for a change in ocean heat content (if there has been such change) is due to changes in cloudiness and resultant solar insolation.
Further to my post at 02:35pm, the second paragraph should have read
“You will note that in the 17 years the linear trend shows a temperature anomaly rise from about -0.15degC to -0.11degC, ie., a change of 0.04degC in 17 years which is a little over 0.02degC per decade.”
Gail Combs says:
June 18, 2013 at 4:02 am
It’s a Power Point doc you can find here:
http://wattsupwiththat.files.wordpress.com/2012/07/watts-et-al-station-siting-7-29-12.ppt
Slide 52.
I have Google Docs Viewer installed in Firefox.
Anthony should have made a PDF instead.
Alec Rawls says:
June 18, 2013 at 11:34 am
///////////////////////////
It would appear likely that the point you make is true.
UHI becomes a problem as soon as urban sprawl begins to influence station data. As the influence pollutes more and more and more, it eventually becomes increasingly saturated in the station data. The resultant trend (UHI warming bias) diminishes so once a station has been fully saturated by UHI the future trend from that station may show no (no further) UHI influence (from an anomaly trend perspective).
Most urban sited station on now fully saturated so their present and future anomaly trends are no longer adversely affected. It is their past temperature anomaly rises that is the issue. Of course, each location willhave grown at different times and at different rates, but it may well be the case that the problem is most pertinent in the period 1960 to 1980s/1990s (the baby boom, post war affluence. rise in motor car use, rise in central heating, aircon use etc). But it will be a station by station factor depending upon the own special facts of the station concerned, so generalising is dangerous.
Semi-rural sited stations may presently be partly affected by UHI warming bias. As they become increasingly less rural, the UHI influence is problematic. Eventually they will become urbanised and eventually their data will be fully saturated by UHI so that at some future point their future trend anomaly will no longer show UHI warming bias.
Rural sites which remain truly rural should show no influence from UHI, but even here local prevailing winds could carry so UHI ‘heat’ from relatively nearby urbanisations, semi urban areas. So even rural sites may not be completely free from the effects of UHI.
What is needed is a station by station audit with very detailed consideration of background facts. It is a very difficult task to find out the real affects of UHI, but anyone who does not suspect the thermometer record may have a serious UHI bias is deluding themselves.
One point that is often overlooked: since on its face the effects of UHI would appear much greater than CO2 (we know that cities are some 3 to 6 degC warmer than truly rural areas), if we cannot detect the signal of UHI in the land based thermometer record what hope have we of detecting the effects of CO2 on the temperature record. If we cannot detect UHI in the record, it obviously is not sensitive enough to detect CO2; the signal of both being drowned out by wide variations of natural variability.
Next thing you know, somebody’s going to conduct a study on the effects of jet engine exhaust on temperature readings of climate stations located at major airports.
It is true that some urban stations are saturated in UHI. But air pollution can lower the daily max. So a decrease in air pollution can generate a warming signal in the daily max.
Warmists argue that UHI is not relevant as rural warming is greater than urban. I argue that rural areas also have much warming that is due to heat islands and not due to the greenhouse effect (RHI?). I cite a town I’m familar with that has a population of 4500 and is the largest town in the county– definitely rural. Yet over the last 50 years traffic has grown tenfold, houses have been converted to multiple family and offices requiring the lawns to be paved for parking, most trees have died off, etc. As you leave town now the temp drops 5-7 deg F much more than even the warmists are forecasting due to the greenhouse effect.