Climate Science Glossary

Term Lookup

Enter a term in the search box to find its definition.

Settings

Use the controls in the far right panel to increase or decrease the number of terms automatically displayed (or to completely turn that feature off).

Term Lookup

Settings


All IPCC definitions taken from Climate Change 2007: The Physical Science Basis. Working Group I Contribution to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Annex I, Glossary, pp. 941-954. Cambridge University Press.

Home Arguments Software Resources Comments The Consensus Project Translations About Support

Twitter Facebook YouTube Mastodon MeWe

RSS Posts RSS Comments Email Subscribe


Climate's changed before
It's the sun
It's not bad
There is no consensus
It's cooling
Models are unreliable
Temp record is unreliable
Animals and plants can adapt
It hasn't warmed since 1998
Antarctica is gaining ice
View All Arguments...



Username
Password
New? Register here
Forgot your password?

Latest Posts

Archives

How climate skeptics mislead

Posted on 13 June 2010 by John Cook

In science, the only thing better than measurements made in the real world are multiple sets of measurements – all pointing to the same answer. That’s what we find with climate change. The case for human caused global warming is based on many independent lines of evidence. Our understanding of climate comes from considering all this evidence. In contrast, global warming skepticism focuses on narrow pieces of the puzzle while neglecting the full picture.

What is the full picture? Humans are emitting around 30 billion tonnes of carbon dioxide into the air every year. This is leaving a distinct human fingerprint:

Signs of warming are found all over the globe (here are just a few):

On the question of human caused global warming, there’s not just a consensus of scientists – there’s a consensus of evidence. In the face of an overwhelming body of evidence, the most common approach of climate skepticism is to focus on narrow pieces of data while neglecting the full picture.

Let's look at an example. One popular skeptic argument has been to cast doubt on the surface temperature record. Skeptics claim thermometers are unreliable because surroundings can influence the reading. They reinforce this by showing photo after photo of weather stations positioned near warming influences like air conditioners, barbeques and carparks. The Skeptics Handbook goes so far as to say "the main 'cause' of global warming is air conditioners".

This myopic approach fails to recognise that air conditioners aren't melting the ice sheets. Carparks aren't causing the sea levels to rise and glaciers to retreat. The thousands of biological changes being observed all over the world aren't happening because someone placed a weather station near an air conditioner. When you step back and survey the full array of evidence, you see inescapable evidence of warming happening throughout our planet.

Our understanding of climate doesn't come from a single line of evidence. We use multiple sets of measurements, using independent methods, to further our understanding. Satellites find similar temperature trends to thermometer measurements. This is despite the fact that no carpark or barbeque has ever been found in space. Prominent skeptic Roy Spencer (head of the team that collects the satellite data) concluded about the HadCRUT surface record:

“Frankly our data set agrees with his, so unless we are all making the same mistake we’re not likely to find out anything new from the data anyway"

Our climate is changing and we are a major cause through our emissions of greenhouse gases. Considering all the facts about climate change is essential for us to understand the world around us, and to make informed decisions about the future.

0 0

Printable Version  |  Link to this page

Comments

Prev  1  2  3  

Comments 101 to 124 out of 124:

  1. @97, Berenyi, I doubt anyone here is suggesting that teh UHI effect does not exist. What they are doubting is that it has changed in the way you suggest. But take that article you refer to: The UHI effect is only significant in winter, and the annual average UHI is much smaller (it's possibly even negative in summer, though the authors attribute this to other effects). But you still have the problem of assessing how, in each case (seeing as how you like the details), the UHI is able to change substantially with a rising population. The supposed relationship does not take into account any other factors, as Gneiss pointed out, and can be subjected to a more rigorous analysis. But as the UHI is clearly a local effect (your reference shows this nicely), quite how does it drive the global temperature rise as measured by satellites, or the loss of sea ice, or the retreat of glaciers, or the myriad other observed warming effects in areas far from urbanisation? That's the whole point of John Cook's post. Far more likely, given the multiple independent lines of evidence, is that your UHI hypothesis is incorrect, and you are failing to apply a 'skeptical' mind to the single blog post you're using as evidence... c'est la vie...
    0 0
  2. Berényi Péter, in line with what skywatcher wrote, I have to clarify two misunderstandings. First, I have no doubt that at high population density the UHI effect tend to saturate. Second, I do not claim that UHI is non existent nor negligibly small; indeed, all the surface temperature datasets correct for it. Go back to the figure you showed in comment #10. The problem with Specer findings (and you) is the trend in the low population density regime. It is claimed that already at 5/Km2, i.e. one family in one Km2, already produce a warming of 0.35 °C. Add three more families in the same Km2 and it jumps to 0.8 °C. This is an extraordinary claim, i'd say. You seem to suggest that the metadata define as rural sites that should not be considered as such and quoted Barrow as an example. At first I did not want to comment on this extreme cherry picking, one particular site in a particular environment and in a particular season (note that they detect a negative UHI effect in summer). But let me spend a few words just to show how far one may mislead the readers. I guess you did not read the paper nor check where the met station is in Barrow. The infamous 2.2 °C were measured by a set of ad hoc thermometers placed inside the village; they wanted to assess the problem of building stability due to permafrost melting. The metereological stations, on the contrary, are outside the village, at the airport for a while and now even further away as part of the Climate Reference Network. Should we presume that they give the same reading as the thermometers inside the village in the season of largest energy use and soil-air temperature difference? One last thing. I don't care if you may or may not do the calculations, which I consider wrong, from the data I show. What that graph shows is that when calculations are done properly and with population densities wildly different, we get an overall difference in the trends relatively small. Your try to assign the measured global warming to the UHI effect ("Not much warming is left" back in your comment #10) is unsupported. It is really surprising that you try to revive the now largely abandoned mantra of the UHI effect.
    0 0
  3. ProfMandia at 13:17 PM on 13 June, 2010, more relevant to any comparison is that between the northern and southern hemispheres. There is a relatively significant difference between temperatures and understanding and explaining that difference would be of more value, yet it's been completely ignored in this thread. Why is that?
    0 0
  4. Answering #61: Scientists may not have collected the data themselves, but scientists ARE responsible for a certain sort of "due diligence" to make sure the data they used are sound. That is where they did not do well here. Even if it is true that UHI has had little effect, the 'proofs' of this really do sound too much like circular reasoning. Or, if, as some seem to be doing in this forum, we insist on referring to satellite measurements to justify the accuracy of the surface measurements, then all we have really done is replaced our confidence in one measurement with total confidence in another. We may as well have skipped the one to begin with. The epistemology behind such an approach is obviously disappointing.
    0 0
  5. #100 Ned at 01:58 AM on 15 June, 2010 I stand by the comment that your proposed UHI effect (0.29C/century over land) works out to approximately 5 or 6 percent of the current global trend It is much worse than that. As it stands, both SST history and satellite temperatures are consistent with GHCN derived series. If trend in the latter one is decreased by 45%, they become inconsistent. In that case something has to be done. Computational climate models also have to be recalibrated using revised data. So if I would be right, it could get pretty inconvenient for folks involved in diverse branches of climate science. It is better to debunk it ASAP. However, not by rhetoric, but valid arguments.
    0 0
  6. #102 Riccardo at 03:34 AM on 15 June, 2010 The metereological stations, on the contrary, are outside the village, at the airport for a while Come on. That's how far away is the airport from Barrow.
    0 0
  7. Berényi Péter, do you want me to believe that being in an open space at least hundreds of meters away from the village or in between walls much warmer than air is the same thing?
    0 0
  8. Berényi Péter... I just looked at a more recent google map satellite image of Barrow and there is virtually no change in urbanization in Barrow since this 1997 image you're posting here. I'm not the expert here but I don't quite understand where the UHI effect is coming from.
    0 0
  9. @johnd 103 I believe the difference in temp increase between the Northern and Southern hemispheres is because a much larger proportion of the surface area in the southern hemisphere is ocean. Of course, the ocean has a much higher heat capacity, so the same amount of heat results in a smaller temperature change.
    0 0
  10. FYI... You can actually see the ASOS station in the google image. It's just to the left of the little club shaped thing on the opposite side of the airport from town. It's at the end of a small road.
    0 0
  11. #105 "both SST history and satellite temperatures are consistent with GHCN derived series. If trend in the latter one is decreased by 45%, they become inconsistent." Extraordinary. Three independent datasets agree, yet one may have a theoretical inconsistency, so they all must be wrong? That's bass-ackwards. Especially when you add in the other datasets that tell the same story. For example, here's a snippet from a report of the Meteorological Service of Canada in 2000, which used 210 stations scattered across Canada: "the area with significant upward trend has expanded from the Prairies to include northern B.C. and Manitoba. The greatest warming during spring is well over 2 deg C for the 1900–1998 period in the Prairies." From the prairies to northern BC and Manitoba? Let's look at Manitoba, as an example, where the 20th century population growth looks nothing like the GISS temperature record. BTW, Manitoba has an area of approx. 650000 sq km with a 2006 population of 1.15 million (same reference as 'population growth', above). How can urban heating be responsible for 2 deg C in 99 years in a place as un-urban as Manitoba?
    0 0
  12. BP, given that Barrow quite clearly does not show any increase in urbanisation round the weather station, as your own image helpfully shows (this report also has an image of the ASOS station, which is clearly not built-up around it), have you any examples where this theoretical effect of yours can be shown to happen. I would think, for example, one like where the weather station is not surrounded by buildings, say 30 years ago, and is clearly surrounded by buildings now? In a small settlement, with such tiny population densities as you and Spencer suggest? Barrow's weatehr station is visible on Google Earth at 71deg 17' 00.30"N and 156deg 46' 53.00"W, in the location described by robhon. It's in utterly un-urbanised open ground. Once again, how would building some houses the far side of Barrow affect that? Riccardo's point on the required densities is worth repeating - the largest change in Spencer's graph is between 5 and 20 people per square km, speculatively relating to 0.5C temperature change. That is moving from one house, to at the very most 10 houses in a whole square km. They would have to be tightly clustered round the weather station to have any effect at all as the referenced report nicely shows. Barrow's 'urban' area, measured on Google Earth, is (very generously) 6sq km, whose population of 4600 (~766/sq km) has declined slightly between 2000 and 2008. Barrow's population density, according to Spencer's model, is already well off the scale and in the region of negligible warming. So this goes to show that you and Spencer ar attributing most of global warming to the addition of very small numbers of houses to virtually uninhabited sites at enough weather stations around the world to bias the temperature record. Truly remarkable. I ask again, show me a site where this theoretical effect is observably real. And as David Horton said earlier (#11) "How do I explain this to the glaciers?"
    0 0
  13. #107 Riccardo at 07:13 AM on 15 June, 2010 do you want me to believe that being in an open space at least hundreds of meters away from the village or in between walls much warmer than air is the same thing? It is obviously not the same thing, but weather station at the Will Rogers Memorial Airport (red dot in figure) is located well within the heat island (for 09 Feb 2002) which spans some 2 km around the village.

    From Hinkel 2003 Fig. 9

    But this Barrow thing is getting overdiscussed. The whole point of it was to demonstrate UHI is present even at GHCN stations flagged "rural". Mission accomplished, move on folks.
    0 0
  14. Berényi Péter, i agree that this discussion has been overstated. More, as already said before its extreme cherry picking is totally irrelevant and based on faulty arguments. This discussion clearly demonstrates two things. One is the attitude of skeptics to mislead people by focusing on the tiniest part of the picture. The other is how easy it is, while it's not so easy doing real science. Overall it has been very clear confirmation of the topic of this post which makes me more and more think that in general no real scientific curiosity guide the self-proclaimed skeptics.
    0 0
  15. Responding to an astonishingly unskeptical comment by Berényi Péter, muoncounter writes: Extraordinary. Three independent datasets agree, yet one may have a theoretical inconsistency, so they all must be wrong? That's bass-ackwards. It's quite a bit worse than that, muoncounter. The first two data sets (lower troposphere temperatures and sea-surface temperatures) are both measured by multiple separate groups using data from satellites processed according to objective, well-documented methods. Satellite retrievals of SST in particular are very well validated. The "theoretical inconsistency" is with the land surface temperature record. Actually, there's no inconsistency with the land surface temperature record either; there's an alleged inconsistency with one commenter's assumption about the effect of UHI on the land surface temperature record. That assumption in turn is apparently based on one speculative blog post by Roy Spencer, which the commenter here then extrapolated to the whole world. Obviously, nobody is going to throw out two completely independent and mutually consistent satellite-based global temperature data sets because they allegedly disagree with some random blog commenter's off-the-cuff estimate of UHI. But this does make for a fascinating illustration of the theme of this website (and this thread). When it comes to climate change, many so-called "skeptics" are completely asymmetrical in their application of skepticism. There's an extreme unwillingness to accept even the most straightforward evidence in support of AGW, combined with a peculiar eagerness to promote any argument against AGW no matter how tenuous and unjustifiable. I like John Cook's subtitle for this site: Getting Skeptical About Global Warming Skepticism
    0 0
  16. A couple of valid points.
    • UHI is strongest in winter and at night. Just like warming attributed to CO2.
    • Local population density is only a proxy to the actual level of changes in land use, road network, architecture, industrial activity which are the direct causes of UHI.
    • Some changes to the environment are to be expected even at very low population densities. If there are people at all, they are there for a reason, therefore they do something. Like cutting down trees, growing crops, mining, conducting scientific research programs on climate change, etc. These activities need supply, transportation, housing, infrastructure, energy, recreation facilities, sewage & garbage disposal, things like that.
    • Weather stations are not located randomly relative to local population density clumps. They have a much higher chance to occupy a place not too far from human habitation, otherwise data collection, repair work and performance monitoring gets expensive.
    • It can get tricky to define "local population density" properly. Population distribution has a fractal-like structure over spatial scales of several orders of magnitude. Defining averages over fractals is not an easy business, but can be done.
    • The logarithmic relation, though valid over a wide range, has to break down somewhere, otherwise for zero population density (the case before the first human was born) UHI anomaly would be °C, which is smaller than absolute zero. The empirical question is: where does it break down exactly and in what manner?
    • GHCN temperature trends for urban and rural sites do not show dramatic differences, not even the ones located in very sparsely populated areas. However, if logarithmic relation would break down too early, one would expect significant divergence.
    • One does not need historic temperature time series to check the limits of logarithmic dependence of UHI on local population density. It can be done here and now.
    • It would require a major research program to do it properly and would obviously cost money. One reason being satellite data are useless for this particular job, because they do not have sufficient vertical resolution. We would like to have temperature readings at 2 m above ground while as a result of calculations based on satellite brightness temperature measurements the entire lower troposphere is included.
    • Present day reanalysis UHI adjustment practice of using nearby "rural" stations as reference is inadequate for removing long term UHI bias due to general population increase.
    0 0
  17. #115 Ned at 21:22 PM on 15 June, 2010 this does make for a fascinating illustration of the theme of this website (and this thread) Yes, it was my intention to serve you and make this thread an illustration. Unfortunately I have not seen yet any well reasoned refutation of a sizable UHI effect on temperature anomaly trend as measured by surface stations, just fallacies ("big picture"), empty rhetoric ("it is really surprising that you try to revive the now largely abandoned mantra of the UHI effect"), refusal to take a closer look ("focusing on some tiny detail, and blowing it all out of proportion"), attributing vile intentions ("the attitude of skeptics to mislead people by focusing on the tiniest part of the picture"), labeling ("random blog commenter's off-the-cuff estimate"), declaration ("the urban heat effect is a red herring"), utterly false accusation ("you did not read the paper nor check where the met station is in Barrow") and the like. I am frightfully sorry for you. And this passion to debunk. Instead of searching for truth. Guys, the first thing to do in cases like this is to actually understand what is said. No amount of peer reviewed literature can make up for the lack of effort. Then you'll be able to spot errors easily if there is any. Or simply help. This is also the only legitimate way to use published papers. I mean by understanding them, learning what is there to be learned and then - then simply use your own head. These publications are not to be used as infallible authorities, this attitude belongs to the worst scholastic tradition (not even followed by outstanding medieval scholastic thinkers).
    0 0
  18. Depends what you call valid, BP... your point (1): different processes, different causes, same apparent temporal effect. I might as well correlate the number of stars visible from Edinburgh to global temperature (most in winter and most at night!), but of course that would be silly. (2), so what? (3) Do you think one of those effects might be to release greenhouse gases, must have some effect! You've provided no valid evidence, bar a single blog post to show that the largest effect should be in *very* sparsely populated places. (4) How close is subjective - as shown in the Barrow example, it's not in a place affected by urbanisation. And BTW, you brought up Barrow, so we're quite happy to debunk you on it. (5) Precisely, hence why using kilometre-scale population averages as Spencer has done is an invalid exercise. You've still provided no concrete examples. (all remaining points) based on a spurious logarithmic dependence for which there is no evidence. The figure in #113 misleading for your point, as it is not showing the local variations - ie that the heat island will be present/stronger in the built-up areas of Barrow and marginal on open ground beyond the airport. The contours are averages, and won't capture the local variations. And w've already established that there's no evidence of further urbanisation here near the weather station anyway (required, surely, for an increase in temperature by your spurious hypothesis)...
    0 0
  19. Just to add, the map that Spencer uses has a density of 6-25people/sq km evenly spread around Barrow over an area that is approximately 10km x 5km. Clearly there's a fair bit of smoothing and blurring going on to construct the map, as the density in the core of Barrow is much higher, and in neighbouring areas of tundra is clearly zero. Large areas of Highland Scotland, which I know are uninhabited, suffer from the same blurring that is placing 2-5 people/sq km. The people live in small communities of higher density, but the averaging covers many empty regions. Ditto Iceland and Greenland. This observation alone puts Spencer's hypothesis on incredibly shaky ground, as one of his two fundamental datasets is not good enough for the task he asks of it (though it may be useful to its designers for other studies). And as has already been described by Ned and muoncounter and others, the independent validation of the temperature record means that we already know the UHI does not operate as Spencer suggested.
    0 0
  20. #118 skywatcher at 22:46 PM on 15 June, 2010 a spurious logarithmic dependence for which there is no evidence I wonder if you understood what you have said. IPCC TAR WG1 2.2.2.1 claims there was no significant difference between rural and urban trends. There is no doubt the logarithmic relation holds for high population density (AKA urban) areas (absolutely independent of Spencer). If the logarithmic dependence breaks down early as settlement size is decreasing and turns into linear, quadratic or whatever around urban -> rural transition, the similar trends can only be explained by much shorter population doubling times at rural sites than at urban ones. It is not likely demographics supports this claim in an age of urbanization. An alternative solution would be to suppose an almost zero UHI even at urban sites, but it is plainly contradicted by a plethora of facts. Therefore you can only negate logarithmic dependence (with a coefficient similar to the ones found in urban heat island studies) by refusing to accept IPCC TAR WG1 2.2.2.1 and all the papers referenced there. Is that what you are trying to tell us?
    0 0
  21. Chris #94 it might be small when averaged GLOBALLY Chris, but human released waste heat would be significant when the total is divided by the urbanized and arable land areas. Your 0.028W/sq.m globally if applied to an arable land area of about 10.5% of the Earth's surface would equate to a heat flux of 0.26W/sq.m. Applied to an urban area of about 1.5% of the Earth's surface if would be a large 1.86W/sq.m, and applied to a number somewhere in between (about 6%)it which would represent a heat flux of about 0.46W/sq.m. So the number of 0.3-0.5 W/sq.m across the continental USA is not far out of the ballpark. If this does not have a significant impact on the surface temperature measurements of the USA, then perhaps you could explain why. BP's point about the geothermal heat flux warming the deep oceans is relevant. A very small temperature diffential from the bottom up must be present to drive the heat upward through the water column - and I wonder if Willis' 0.1W/sq.m equivalent deep ocean warming can be separated from the geothermal heating effect.
    0 0
  22. Ned #115 and BP #117 BP's forensic approach is to be admired. Even the title of this thread "How skeptics mislead" does imply some intention to mischief rather that getting at the truth of what is going on. And the point is well made - resorts to peer review as some absolute authority without understanding the technicality of the arguments contained in said papers is superficial. Note that BP has made some very telling points in the OHC thread showing that the warming was not 'robust' at all (with which my own small contributions agree) - and no serious counter argument was made by the well furnished regulars on this blog including the owner. Peer reviewed scientists should have seen that the large jumps in OHC were not 'messy data' but impossible leaps due to instrumental transition offsets.
    0 0
  23. Ken, I disagree with your interpretation of the OHC thread, but that's neither here nor there. In this thread, BP has assumed a very specific relationship between local population density and local temperature (0.16C per doubling of population density). I've showed that even if we accept this assumption, UHI would constitute about 5% to 6% of the current warming trend. I don't think there's much more that needs to be said.
    0 0
  24. #123 Ned at 00:19 AM on 16 June, 2010 if we accept this assumption If, and I mean if you accept this assumption, your least concern is how large the UHI contribution is compared to land+ocean temperature combined. Before venturing to such calculations you have to explain why sea surface temperature increases significantly faster than atmospheric temperature 2 m above land. That's a tough question. Such behavior is neither predicted by computational climate models nor do we have a simple physical mechanism to transfer the better part of heat trapped by atmospheric CO2 immediately to the sea with an ever higher efficiency as time goes by.
    0 0
  25. I agree completely that this would be a rather strange outcome. Most people, I think, would take this as an indication that your estimate of the effect of UHI is inflated. You, apparently (?) take it as an indication that the global sea surface temperature trend is incorrect, though you don't actually come out and say that in plain language so I might be misinterpreting you.
    0 0
  26. Berényi, you do realize that the data has already been corrected for the UHI effect? That several studies done (as prompted by the skeptic argument about the decline in number of stations not being properly handled, also here, which I highly recommend you re-read) on drop-out rates, sub-set selection, rural vs. urban stations, etc., all show the UHI and area corrections work, leaving no warming biases via station selection? The slight bias found indicates that the ground station data as currently processed is actually underestimating global temperatures. That directly contradicts your argument. I would point out that the demonstrated independence from sub-set selection is an excellent validation of data treatment for this temperature dataset. UHI does not lead to overestimation of warming. Add to that the two independent satellite data sets, all in agreement - that's a significant confirmation of the local station data. Your claim that the satellite data is incorrect really doesn't fly, so to speak. This has been discussed ad nauseum - I think you're making imaginary mountains out of molehills.
    0 0
  27. Ken Lambert at 23:47 PM on 15 June, 2010 "If this does not have a significant impact on the surface temperature measurements of the USA, then perhaps you could explain why." It doesn't have a significant impact on the surface temperature anomalies of the USA since this effect is taken into account in assessing the anomalies. This is done either by correcting urban sites by reference to non-urban sites, or assessing the anomalies purely from the non-urban sites. This is described extensively in the scientific literature. Let's not pretend that we don't know what we do know! "BP's point about the geothermal heat flux warming the deep oceans is relevant. A very small temperature diffential from the bottom up must be present to drive the heat upward through the water column - and I wonder if Willis' 0.1W/sq.m equivalent deep ocean warming can be separated from the geothermal heating effect." The geothermal heat flux has been warming the ocean bottom for eons. It's obvious that this heat is distributed through the oceans and dissipated entropically by movement of ocean waters. Indeed it's rather well established that the geothermal flux contributes to deep water mixing. So it's likely that "Willis 0.1 W.m^2 equivalent deep ocean warming" can be separated from the geothermal heating effect, since any recent deep ocean warming is a supplement to the background heat content that results from all natural contribution to ocean heat including the geothermal flux. Again this is extensively described in the scientific literature. We should use current knowledge as a starting point for a deeper understanding of phenomena in the natural world.
    0 0
  28. You misunderstood me BP: Spencer does not provide good evidence for such a trend (hence why I called it spurious), and as you are basing your assertions on Spencer, your correlation is spurious. The fact that one of the key datasets in Spencer's correlation is not up to the task he sets of it is the reason the correlation provided by Spencer cannot be trusted. I did not even need to verify his temperature data - if one of the two measures in a correlation is dodgy, the relationship, if one exists, that emerges will also be dodgy. That leads neatly onto what KR (and many others here) post regarding the many independent verifications of the warming trend, namely that the trend as identified in multiple independent sources is in all probability real, and there is still no reason to doubt it. Despite your best attempts at muddying the waters...
    0 0
  29. Berényi - you seem to see a problem with the ground station data (that others have not) and say that As it stands, both SST history and satellite temperatures are consistent with GHCN derived series. If trend in the latter one is decreased by 45%, they become inconsistent. In that case something has to be done. You're absolutely correct. If three independent data sets are in agreement, and an objection is raised that moves one set 45% away from the others, some thing should be done. You should drop the objection. Occam's razor and general parsimony should lead you to conclude that the objection is not valid, rather than the rather odd conclusion that two other datasets are completely wrong based on an objection that has already been examined and invalidated.
    0 0
  30. Might I suggest that this conversation continue on the more appropriate Heat Island or Temp record unreliable threads?
    0 0
    Moderator Response: Perhaps this conversation will serve as a case study appropriate to the topic of the post?
  31. Ken BP left unaddressed as far as I know the question of whether the entire record of OHC from left to right was simply an error.
    0 0
  32. So the reasoning is that a spurious correlation, postulated in a non reviewed blog post, based on a less than adequate dataset, in disagreement with previously published work, should be used to make one dataset vary more than 45% from the others, so as to invalidate all datasets. Interesting.
    0 0
  33. Ken, you might have missed the beginning of this discussion, where I pasted a link to an example of how climate skeptics mislead, from one of the most prominent "skeptic" internet sources. I should add that it does not employ any of the more subtle methods described in the OP, just really bad maths, which the "skeptics" are all too eager to accept. If this is what you consider "getting at the truth" we may not be able to communicate at all. http://www.skepticalscience.com/How-climate-skeptics-mislead.html#16040
    0 0
  34. @Arkadiusz Semczyszak #85 I have a copy of the Green-Armstrong-Soon paper you mention. It seems to me to have a flaw - it applies purely business forecasting methods to a physical phenomenon. You can (as I have) apply Autoregressive Moving Average Models to the temperature data and obtain good fits. I have used it to forecast temperature in future years, but I think the forecasts are useless. I am in a different discipline to business forecasting or climate science, but I do apply statistical models to physical phenomena (e.g. wearout of a mechanical component). There is a danger in all such models in extrapolating them outside of the conditions in which they have been fitted e.g a relationship showing thermal expansion of plastic will break down when the plastic approaches its melting point (trivial example, I know). Hence if global warming has a physical basis which is forcing the change, purely statistical models fitted to past data cannot be simply extrapolated into the future without taking the changing physical influences into account. If it was as simple as Green, Armstrong and Soon make out, we would all be doing it!
    0 0
  35. BP, It's really quite simple. You have a hypothesis, that hypothesis makes measurable predictions. If your hypothesis is true and UHI is not properly accounted for in the temperature record, then both the sea surface and satellite trends should vary significantly from land surface trends. This is not the case, so your hypothesis as it stands is falsified. What more needs to be said? You do not get to wave your hand and declare that all these observations must be wrong. If you think there is an error in the satellite and SST data that perfectly mimics that of UHI, you must first identify that error and prove that it exists. Only then do you get to revive your hypothesis. This is how science works. Observations drive theory, not the other way around.
    0 0
  36. #125 Ned at 00:45 AM on 16 June, 2010 Most people, I think, would take this as an indication that your estimate of the effect of UHI is inflated. I bet they would. Wrongly, of course. I hope you agree the analytic method requires breaking down problems to their constituent parts and treat them separately as long as possible. Having done to each problem what we could a synthetic step follows when we try to put together what we have got. If they don't fit, we should go back and scrutinize each problem further in itself until a satisfactory fit is achieved. It is not allowed to distort the shape of any piece at synthesis to fit it to the "big picture" without sufficient reason concerning that piece alone irrespective of the "whole". That's a basic & indispensable technique of scientific enquiry, one way for avoiding the horrendous trap of confirmation bias. And, for that matter, one of the first lessons to learn about the scientific method, and learn it by heart is not to care about how most people would take your results. That concern is for politicians and actresses. You, apparently (?) take it as an indication that the global sea surface temperature trend is incorrect No, I don't take it as an indication of anything. I simply register the logical consequences of a proposition in a conditional form. If land surface temperature trend is reduced by 45% then either SST history should be adjusted accordingly or some weird physics is going on about which we don't have a clue. The proposition above is a true one. At the moment, however, I am not concerned with SST, but land surface temperatures as they are recorded in GHCN v2 and their relation to UHI. SST can wait for its due course. #126 KR at 01:06 AM on 16 June, 2010 you do realize that the data has already been corrected for the UHI effect? I do realize an attempt was made to correct for UHI. But do you realize I have already explicated why this was insufficient? Go back in this thread, read, understand and come back. Your claim that the satellite data is incorrect really doesn't fly, so to speak. Either my English is an atrocity (sorry for that, it's my second language) or your comprehension skills are lacking. At the place you have indicated I have written "satellite data are useless for this particular job, because they do not have sufficient vertical resolution" (which is true). How could you read this as satellite data being incorrect is beyond me. #128 skywatcher at 01:21 AM on 16 June, 2010 you are basing your assertions on Spencer No, I do not need Spencer anymore. I still give him credit for discovering the fact logarithmic dependence of UHI on local population density extends well below levels usually considered "rural", but now I only have to use it as a heuristic hint, not as a premise. The figure 0.16°C/doubling derivable from his graphs is still useful, but the same (or a slightly higher, ~0.25°C/doubling) value can be obtained from independent UHI studies as well. Anyway, there is no order-of-magnitude difference in this respect. As far as I am aware of serious local UHI studies were only done to settlements with population larger than 10,000, that is, flagged "urban" in GHCN. For this range the logarithmic relation is firmly established. This is why I need another source, in this case IPCC TAR WG1 2.2.2.1 stating "neither pair of differences is statistically significant", referring to multiple peer reviewed studies on temperature trend differences for rural/urban sites. As there is no statistically significant difference in trends, for the time being let's consider them to be equal. We do know world population has doubled almost twice (1.83 times) during the 20th century. Therefore average population density over the world has also doubled twice (because land surface has not changed significantly in this interval). Local population density trends may differ according to location, but on a global level they cancel each other. Now, we have a function TUHI(d), where d is local population density and the function lends the expected value of UHI in °C. TUHI(d) is known to have a logarithmic pace for population densities considered "urban". Therefore TUHI(d) = c×log(d) for some constant c if d > d0 with a suitable value of d0. Now let d(t) be the average time dependence of local population density for rural sites, D(t) the same for urban ones. We can write D(t) = c×log(D(t)) with the same constant c and D(t) > d0 > d(t) if no site switched status during the epoch. Time derivative of UHI (trend correction) is same for rural and urban sites (IPCC): T'UHI(d(t))×d'(t) = T'UHI(D(t))×D'(t) = c×D'(t)/D(t) (1) At the same time UHI effect for rural sites is supposed to grow slower than logarithmic: c/d > T'UHI(d) if d0 > d (2) Therefore c×d'(t)/d(t) > c×D'(t)/D(t) (3) It means temporal logarithmic derivative of population density is larger for rural than for urban sites. It is equivalent to say that doubling time is smaller at rural sites, that is, relative increase of population is faster. I let you check how realistic this conclusion is in a world where urbanization is the overall trend with an ever larger proportion of population living in cities.
    0 0
  37. #135 e at 05:06 AM on 16 June, 2010 If your hypothesis is true and UHI is not properly accounted for in the temperature record, then both the sea surface and satellite trends should vary significantly from land surface trends You are absolutely right. If UHI is not properly accounted for both SST and satellite data should be re-checked. As for the latter one I have a guess. Satellites do not measure atmospheric temperature directly, but brightness temperature using various channels. To convert it to proper temperatures a pretty complicated atmospheric model is needed. If I were you I'd check that computational model in depth. But as I have already mentioned a couple of times, it is a separate job for another occasion.
    0 0
  38. "If I were you I'd check that computational model in depth." There's more than one, and they're in reasonably good agreement.
    0 0
  39. "Moderator Response: Perhaps this conversation will serve as a case study appropriate to the topic of the post?" I think the point has already been made. With emphasis.
    0 0
  40. BP, while I admire and commend your enthusiasm and dedication I think "e" at #135 shows one of the primary problems (other posters here have noted other problems) with Spencer's hypothesis and methodology. In my humble opinion, stating the current UHI corrections are way off is not borne out by the observations, and is essentially a red herring. There is always room for improvement to the UHI and homogenization adjustments (which are both a necessary evil). For example, Hansen et al. discuss some of those in the draft of his upcoming paper on the SAT record. However, if the current UHI corrections were as poor as you claim them to be, then there would simply not be such good agreement between the RATPAC, SAT, MSU, SSTs and 0-2000 m OHC data as there is. That is the reality.
    0 0
  41. Berényi, I interpreted this post as claiming that the satellite data could not be used to support the GHCN v2 ground station data, due to the satellite vertical resolution - and your repeated statements that the satellite data would need to be rechecked if there was an issue with the ground data. This would only be the case if you thought the satellite data incorrect! If you thought the satellite data accurate, and the GHCN data is in agreement, you have no basis to consider the ground data wrong, either. These three data sets all show the same trends, and are in quite close agreement. Spencers data treatment is well known to be incorrect, and the 'logarithmic' population objection you speak of doesn't appear to be an issue in properly corrected GHCN v2 data. In fact, the agreement between assorted subsets (properly corrected) of the GHCN stations is a clear indication that the UHI effect does not distort the data. (This is a point you have not addressed at all well, in my opinion - not all stations have experienced population increases) Possible issues due to UHI effects would be higher variances for the GHCN data, considerable dependence on which stations were used, and significant biases or slope changes against other data sets. None of these are visible in the properly corrected GHCN data. When compared to the satellite data there are no differences in slope or value outside what you would expect from the accuracy ranges. I'll repeat myself a bit here - if you have as Albatross points out MANY independently developed data sets, with robust internal and external checks, all in excellent agreement, and a widely debunked objection that throws one of them off by 45% - Occam's razor indicates that your objection is incorrect, not all the other data sets. You cannot ignore the additional data sets - independent investigation is _key_ to a robust conclusion. At most you might find that the single data set you're discussing is providing a false agreement, and at this point only a single investigator (Spencer) appears to have any claims to that point. This UHI thing is a red herring (false error), and a dead horse (we've beaten this topic to death long ago). I appreciate your viewpoints, Berényi, and the energy you put into them, but this UHI objection just doesn't hold up.
    0 0
  42. #138 dhogaza at 07:08 AM on 16 June, 2010 There's more than one [model], and they're in reasonably good agreement
    Of course they are. The atmospheric model behind them is fine tuned to reality, meaning to align them nicely with other datasets (including land surface temperatures). Lower troposphere is only measured at a single channel (Ch. TLT, a narrow IR band) with an awful vertical resolution. One can only recover lower tropospheric temperatures from that single channel, if, for example, water vapor distribution is known. As direct (weather balloon radiosonde humidity) measurements are dismissed by the community, it is not an easy task to get it aligned to (genuine) reality. Therefore they usually suppose a quasi-constant relative humidity profile, while many decades of balloon data indicate a decreasing trend above 700 mbar. Anyway, it is a job for another day to explore every nook and cranny of this issue.
    0 0
  43. BP, You say "The atmospheric model behind them is fine tuned to reality, meaning to align them nicely with other datasets (including land surface temperatures)." Not correct. Spencer himself states that the MSU data are not calibrated using the surface temperatures. Even so, let us ignore the MSU data for the moment. That still leaves us with the OHC, SST and RAPTPAC data. As I showed earlier on this thread, the warming trend in the mid-tropospheric radiosonde temperature data is greater than that in the SAT record shown in the Figure. There is simply no credible evidence to suggest that official SAT data are exaggerating the warming. I know you feel that you are onto something here, but surely you are smart enough to know when to call it a day?
    0 0
  44. #141 KR at 07:48 AM on 16 June, 2010 independent investigation is _key_ to a robust conclusion Dear KR, you live in another world. In the days of pre-postmodern science if a conclusion was robust, it was considered a weakness, not strength. Robust means it is resistant to attempts of falsification not because of its internal value but its external connections (networking properties). The concept of robustness was developed for software engineering purposes, where it has its proper place. It is used in a world where software gets so mind bogglingly complicated, that no one has the capacity to understand it, so one has no choice but tolerate some errors and try to compensate for them by other means. This concept was borrowed recently by various branches of science even if it is incompatible with the scientific method itself. After all a scientist is supposed to know in detail, down to the last comma what he is talking about.
    0 0
  45. BP > If land surface temperature trend is reduced by 45% then either SST history should be adjusted accordingly or some weird physics is going on about which we don't have a clue. The piece you are missing is that the consequences of this hypothesis impact the probability that your hypothesis is true. If it were true, it would mean not only that land surface temperature, sea surface temperature, and satellite temperature readings have errors, but that the source of error is different in each of these, and yet they produce essentially the same result. This conclusion is highly improbable and by extension your hypothesis is highly improbable. Possible yes, probable no. I won't disagree with you that perhaps there are interesting things to learn in investigating the nuances of UHI's. Where I - and others here - take issue is the implication that your hypothesis casts doubt on current "mainstream" climate science. I fail to see why improbable speculation should automatically cast doubt on hard science.
    0 0
  46. #145 e at 08:41 AM on 16 June, 2010 I fail to see why improbable speculation should automatically cast doubt on hard science. In that case you should train yourself to see it. Probabilistic reasoning at the meta-level has no place in science whatsoever. If you like this method, perhaps stockbroker is the proper profession for you. Science uses binary logic. A proposition is either true or false. I would really be happy to see a single attempt to discuss the problem itself in detail on its own merit instead of casting doubt from the twilight.
    0 0
  47. Berényi #146
    Science uses binary logic. A proposition is either true or false.
    This is incorrect. A proposition is either supported or unsupported. Deductive reasoning is the preserve of mathematics, science is the home of induction, which is why there will never be formal proof for anthropogenic global warming, just the balance of the evidence. At this stage the multiple independent lines for AGW are overwhelming, but there will always be wriggle room for the so called sceptics.
    0 0
  48. "you live in another world. In the days of pre-postmodern science if a conclusion was robust, it was considered a weakness, not strength. Robust means it is resistant to attempts of falsification not because of its internal value but its external connections (networking properties)." This argument is semantic red herring. "Robust" existed as a word well before software engineering and is not always used in the way you state. For instance, a theory can be "robust" in that it applies under a wide range of conditions or has survived extensive falsification. In this context "robust" means "healthy" or "strong." It's true that in some cases "robust" can mean impervious to challenge, but that was clearly not the intended meaning in this case. The message that was intended was that a hypothesis that is consistent with multiple independent lines of data will always be more convincing than a another, mutually exclusive one that is consistent with only one line of data and inconsistent with others. That sentiment is perfectly consistent with the scientific evidentiary approach (I don't know what the term "post-modern science" even means). You, on the other hand, are arguing for a new correction to land temps that generates inconsistency among independent measures of global heat which should in reality be aligned (and do with the current set of corrections). If you showed any willingness to acknowledge that this situation raises questions about the validity of your method, and that maybe it needs revision or rejection as a consequence, people would be more receptive. As it stands, its appears your idea is the one that is unfalsifiable and subject to confirmation bias.
    0 0
  49. BP: >Probabilistic reasoning at the meta-level has no place in science whatsoever. False. Absolutely and categorically false. I suggest you read up on the philosophy of science, particularly the work of Karl Popper and the problem of induction. In summary, because positive formal proof cannot exist in empirical science, the best we can ever do is claim that a current theory is more probable than any other competing theory. Popper explains why this is sufficient to place trust in the scientific method, where previous philosophers cast doubt. A good high level discussion of the topic can also be found here (the references cited on that page are also worth looking at). I quote: "A crucial related point is that modern scientific theories are probabilistic. This means that all testing of scientific predictions is carried out in a statistical framework. Probability and statistics pervade modern scientific theories, including thermodynamics (statistical mechanics), geology, quantum mechanics, genetics, and medicine." >Science uses binary logic. A proposition is either true or false. False again. Much of this is addressed in the works cited above. My favorite take on this subject though is Isaac Asimov's essay The Relativity of Wrong. I strongly suggest you take a look, it is an interesting and illuminating read. >I would really be happy to see a single attempt to discuss the problem itself in detail on its own merit instead of casting doubt from the twilight. As this is a blog about AGW skepticism, the most relevant question to ask is "does your speculative hypothesis cast doubt on the science in its current state"? You have thus far failed explain why it should.
    0 0
  50. e: ...the source of error is different in each of these, and yet they produce essentially the same result. That's exactly what bothers me, but the strange coincidence is much broader than what's being discussed here and now in this thread. In various threads on this site assertions have been made to the effect that GPS systems are insufficiently inaccurate, tide measurements mean nothing, atmospheric temperature measurements are hopelessly flawed, sediment cores are faulty, oceanic temperature measurements are unreliable, C02 measurements are not correct, ice cores are contaminated, radiometers lie to us. Various faults have been pointed out in the data produced by this diverse collection of instruments, noise has been highlighted and generally exaggerated in importance yet again and again we see a disturbing similarity and confirmations of predictable connections and relationships in longitudinal trend lines produced from these data sources. We're supposed to conclude that-- coincidentally-- all data sources connected with climate research are substantially producing bunkum. Paradoxically, many of these measurement systems are somehow still useful for other applications. How likely is that? Another conclusion--more plausible-- might be that these systems are adequately well engineered and operated and that while undoubtedly there are greater or lesser sources of noise infecting data collections they are reasonably useful, with some reinforcement of that notion coming from their mutual consistency, their meta-measurements of various related climate-related subsystems.
    0 0

Prev  1  2  3  

You need to be logged in to post a comment. Login via the left margin or if you're new, register here.



The Consensus Project Website

THE ESCALATOR

(free to republish)


© Copyright 2024 John Cook
Home | Translations | About Us | Privacy | Contact Us