Buy

Books
Click images for more details

Support

 

Twitter
Recent comments
Recent posts
Currently discussing
  • Jun 24 - Mark Hodgson on
    COP 23
  • Jun 24 - Mark Hodgson on
    COP 23
Links

A few sites I've stumbled across recently....

Powered by Squarespace

Discussion > Satellite vs surface temperatures

Atomic Hairdryer

I agree entirely that human energy use = waste heat. Like you, I wondered whether this was sufficient to have a measurable effect on TLT temperatures.

It doesn’t seem to be quite there, at least according to Flanner (2009):

Nearly all energy used for human purposes is dissipated as heat within Earth's land–atmosphere system. Thermal energy released from non-renewable sources is therefore a climate forcing term. Averaged globally, this forcing is only +0.028 W m−2, but over the continental United States and western Europe, it is +0.39 and +0.68 W m−2, respectively. Here, present and future global inventories of anthropogenic heat flux (AHF) are developed, and parameterizations derived for seasonal and diurnal flux cycles. Equilibrium climate experiments show statistically-significant continental-scale surface warming (0.4–0.9°C) produced by one 2100 AHF scenario, but not by current or 2040 estimates. However, significant increases in annual-mean temperature and planetary boundary layer (PBL) height occur over gridcells where present-day AHF exceeds 3.0 W m−2. PBL expansion leads to a slight, but significant increase in atmospheric residence time of aerosols emitted from large-AHF regions. Hence, AHF may influence regional climate projections and contemporary chemistry-climate studies.

http://www.agu.org/pubs/crossref/2009/2008GL036465.shtml

http://aoss-research.engin.umich.edu/faculty/flanner/content/ppr/Flannr09.pdf

AR4 includes the current thinking (possibly flawed as Pielke Sr reminds) that land use change might actually be holding temperatures down. I don’t pretend to know the truth.

http://www.ipcc.ch/publications_and_data/ar4/wg1/en/faq-2-1-figure-2.html

However, it does seem likely that waste heat through human energy use is far less than required to account for the upward trend in TLT temperatures.

We agree that surface and TLT temperatures are rising and that the trends are very similar. What’s missing is any evidence that waste heat (including UHI) and land use change drive and explain that relation.

Apr 21, 2011 at 11:14 PM | Unregistered CommenterBBD

Apologies. Missed link to data above:

http://www.cgd.ucar.edu/tss/ahf/

Apr 21, 2011 at 11:30 PM | Unregistered CommenterBBD

BBD
Sorry for delay in reply.

Let me elaborate on what I was driving at. Imagine a thermometer in the most pristine (ideal) of settings- the outskirts of a small village. Then let us say, its surroundings start to urbanize. A bed of asphalt appears...the temperatures make a jump. Then an AC vent appears, but at a slight distance. The temperature makes a small jump. And then nothing happens, let us say, for two decades. The thermometer tracks the UHI+climatology, instead of just the climate, for twenty years. In the meanwhile however, the village has expanded and become a small town - numerous constructions, parking lots and heat sources have appeared; but none of these affect our thermometer.

In other words, a lot more urbanization than can actually affect temperature readings appeared in our example. I think it would be reasonable to say this is the situation with the real world. We are expecting, in our null hypothesis, that this urbanization would affect the record, but it should not.

Now imagine a network of say, five similar thermometers as above covering one unit square of the global temperature grid. As and when temperature-affecting urbanization would appear, the thermometers would record point jumps but there would be no further change. The end result: eventually, the network would come to record a slightly higher temperature than the 'actual' temperature. My question at this juncture: why, and how, would the grid show an increasing trend, rather than just a jump in magnitude? The anomaly graph would run a bit higher but parallel (eventually) to the 'real' anomaly.

In order for such a grid to record an increasing trend with urbanization, one would have to add temperature-affecting heat sources serially, at every recording site. Asphalt this year, an A/C vent three years later, a shiny metal plate another two years later, a regularly parked truck four years later etc. Even then, the saltatory jumps each addition would produce, would create only an apparent trend and not a real one, because given enough time, it would become evident that the such a station is recording a consistently higher temperature which will bear a constant relation to the actual temperatures.

The other way one could end up with a trend, is if temperatures themselves had an independent effect on the urban heat island effect. That is, a pristine thermometer would record 10 and 13 degrees as 10 and 13, but a UHI-afflicted thermometer would measure 10 and 12 degrees as 11 and 16.

Therefore (in my words), in order to say that there is no trend divergence between the surface record at the satellites and therefore the urban heat island effect on the global grid is immaterial, we have to first demonstrate that the urban heat island will produce a trend divergence and explain why. We don't know what exactly a heat island does to a temperature record, do we? (At least, that is how it seems to my ignorant brain)

I checked the IPCC report and, oh the shame, Wikipedia, on this. To my mind, there are a handful of problems. Firstly, the global anomaly clearly underestimates the actual heat felt by people living in cities and cooking in their heat islands. There is far more heat than the thermometers are recording and these guys are busy decontaminating whatever little the thermometers are picking up. Secondly, the official consensus is to ignore UHI, it seems, and therefore it is evident that a fraction of the official record rise in temperature is actually due to UHI. Thirdly, folks like McKitrick and Michaels have shown that UHI actually does affect trends - they show that the UHI effect on trends is 'robust', i.e., its significance remains unaffected even if corrected for numerous confounders. This is a serious issue because MM go to the root of the problem - they compare uncorrected, unhomogenized CRU data with the official record, which I believe is a better way of addressing the UHI question than a 'corrected surface record vs satellite' comparison. Even a Jones Chinese network-style study would be suited.

Sorry for the length.

Apr 23, 2011 at 1:20 PM | Unregistered CommenterShub

Shub

Thanks for this. Really interesting and deserving of proper contemplation.

Let me have a think ;-)

Apr 23, 2011 at 7:27 PM | Unregistered CommenterBBD

Shub

You say:

Now imagine a network of say, five similar thermometers as above covering one unit square of the global temperature grid. As and when temperature-affecting urbanization would appear, the thermometers would record point jumps but there would be no further change. The end result: eventually, the network would come to record a slightly higher temperature than the 'actual' temperature. My question at this juncture: why, and how, would the grid show an increasing trend, rather than just a jump in magnitude? The anomaly graph would run a bit higher but parallel (eventually) to the 'real' anomaly.

I agree that for individual thermometers, the effect of encroaching urbanisation is likely to be a step increase in T followed by relatively minor further warming from UHI. No significant trend from UHI.

But gridcells where urbanisation is taking place would yield higher average anomalies over time as urbanisation progressively influences more thermometers within the cell. As these are averaged in, the cell average (and anomaly) will rise. Without correction, this would creat or amplify a warming trend.

Every location affected by UHI would contribute to an amplification of any climatological warming trend (positive interference) and a reduction in sensitivity to climatological cooling (masked by UHI - negative interference).

So even gridcells already dominated by urbanisation will 'pick up' higher anomalies and under-report lower anomalies. Without correction, this would amplify a warming trend.

Only gridcells unaffected by urbanisation would provide unbiased regional anomalies.

Presumably the substantial global urbanisation over the last three decades has affected enough thermometers in series to elevate trends in a large number of gridcells.

That's why I'm not sure your example holds true:

In other words, a lot more urbanization than can actually affect temperature readings appeared in our example. I think it would be reasonable to say this is the situation with the real world. We are expecting, in our null hypothesis, that this urbanization would affect the record, but it should not.

I'm not disputing this:

Thirdly, folks like McKitrick and Michaels have shown that UHI actually does affect trends - they show that the UHI effect on trends is 'robust', i.e., its significance remains unaffected even if corrected for numerous confounders.

Since it doesn't support your argument that UHI does not influence trend for surface T ;-)

Perhaps, as Atomic noted earlier, UHI contamination has simply elevated the surface temperature calculations enough to bring the various records into line with TLT, so masking the 'missing' tropospheric amplification.

This is a serious issue because MM go to the root of the problem - they compare uncorrected, unhomogenized CRU data with the official record, which I believe is a better way of addressing the UHI question than a 'corrected surface record vs satellite' comparison.

I would tend to agree. Does this, too, suggest that satellite TLT temperatures (and trend?) may be higher than surface T and trend?

Happily, this may mean that:

1/. Surface temperatures are lower than currently estimated, further reducing the possible effect of CO2 forcing on T to date.

Or

2/. Surface temperatures are about right as currently estimated, and the tropospheric amplification expected from CO2 forcing is absent.

Apr 23, 2011 at 11:17 PM | Unregistered CommenterBBD

The simple point is if you take the surface records as being sound, the gold standard in the AGW debate, then the expected trend in the lower troposphere (satellite data) would be around 0.2 °C/decade on a global basis. It isnt.

The satellite data does not validate the surface data

Given that the satellite data gives far better coverage than surface measurements, that there is likely far less known biases in the satellite data, and that the satellite algorithms have gone thru in large part a process of peer-reviewed validation, then the satellite data is more reliable.

The data, all the data, is telling us something else.

For arguement sake accepting the AGW hypothesis and accepting that the satellite data is sound, the new gold standard, then the expected trend in the surface temperature should be of the order 0.1°C/decade.

Why is the surface warming at a faster rate than the troposphere, when it should be the other way round?

Here is another point of interest. Whilst the surface and the troposphere appear to gain heat at similar rates and magnitudes the surface appears to lose less heat than the troposphere. It would appear that the surface data is telling us heat is being retained, but the oceans the great repository of heat are not storing any of it.

The data, all the data, is telling us that whatever is going on CAGW, or AGW isn't it.

Apr 27, 2011 at 3:38 PM | Unregistered CommenterMac

Mac

The data, all the data, is telling us something else.

Yup. As above:

Happily, this may mean that:

1/. Surface temperatures are lower than currently estimated, further reducing the possible effect of CO2 forcing on T to date.

Or

2/. Surface temperatures are about right as currently estimated, and the tropospheric amplification expected from CO2 forcing is absent.

Apr 27, 2011 at 10:10 PM | Unregistered CommenterBBD

NOTE

Wood for Trees has now updated its satellite temperatures.

UAH has moved from v5.3 to v5.4

RSS has moved from v3.2 to v3.3

The v5.4 UAH data product has a new, extended baseline from 20 years (1979 – 1998) to a standard 30 year climatology (1981 – 2010).


UAH's Roy Spencer comments:

This change from a 20 to a 30 year base period has 2 main impacts:

1) because the most recent decade averaged somewhat warmer than the previous two decades, the anomaly values will be about 0.1 deg. C lower than they used to be. This does NOT affect the long-term trend of the data…it only reflects a change in the zero-level, which is somewhat arbitrary.

2) the 30-year average annual cycle shape will be somewhat different, and more representative of “normal” of the satellite record than with 20 years; as a result, the month-to-month changes in the anomalies might be slightly less “erratic” in appearance. (Some enterprising person should check into that with the old versus new anomaly datasets).

RSS adjustments (v3.3) yield a very slightly lower trend. There's a good analysis of this by Bob Tisdale here.

All this makes no difference to what I say above, but the full re-baselined comparisons between HADCRUT, GISTEMP, UAH and RSS show a vertical separation between UAH and the other records.

This is an artefact of the changed UAH baseline, and does not affect trend comparisons.

Apr 28, 2011 at 1:04 PM | Unregistered CommenterBBD

Mac

Sorry, this is disjointed but there is rather a lot happening at the moment.

I do not think the evidence points to significant influence on surface temperatures from UHI (see above at Apr 23, 2011 at 11:17 PM). If this were happening to a significant degree, there would be a clear trend divergence, with surface temperatures trending higher than satellite TLT.

Although there is a slight difference, it is not significant.

Here are the trends for GISTEMP, HADCRUT3, UAH and RSS from 1979 – present (UAH v5.4; RSS v3.3)

Decadal trend (degrees C):

GISTEMP 0.16

HADCRUT 0.15

UAH 0.14

RSS 0.14


So the likelihood is that it's (2):

2/. Surface temperatures are about right as currently estimated, and the tropospheric amplification expected from CO2 forcing is absent.

Good agreement for interannual variability and trend between surface and satellite records does validate the surface temperature record.

Scientific scepticism should be directed at the apparent absence of tropospheric amplification.

You say:

The satellite data does not validate the surface data

Please have another look at the whole thread. Jumping in with this assertion at this point is, to put it gently, unhelpful.

Apr 28, 2011 at 1:24 PM | Unregistered CommenterBBD

Fully updated and re-baselined comparitive graph

I have now had a chance to recalculate the offsets and re-baseline the four series for the same accurate comparison attempted in the original post.

I have used the new (v5.4) UAH baseline of 1981 - 2010 as the common baseline for the other three series.

HADCRUT, GISTEMP, RSS re-baselined to UAH 1981 - 2010. 12 month smoothing.

Trends for same.


Decadal trend, 1979 – present:

HADCRUT 0.16

GISTEMP 0.16

RSS 0.15

UAH 0.14

NOTE:

Means for Jan 1981 - Dec 2010 were derived from this.


Means for the period Jan 1981 – Dec 2010:

HADCRUT 0.26 (0.256722)

GISTEMP 0.35 (0.349472)

RSS 0.10 (0.098775)

UAH 0.00 (0.0009777)

Apr 28, 2011 at 7:59 PM | Unregistered CommenterBBD

A lot of good stuff in this thread - it has taken a while to read through as I have been absent for a while. One has to do something other than worry about climate and energy once in a while.

UHI - there is no doubt in my mind that the US temperature reporting network is heavy contaminated. Many think that is true in other parts of the world, including China, Australia and New Zealand. And given that we are only looking at small trends - it must be significant. Whether is a game changer - who knows, but why is the temperature inaccuracy always in the same3 direction, except before 1950 that is.

Having had a fair amount to do with the exposure of meteorological instrumentation over many decades, to me the sites shown in Anthony Watts survey in the US are just a joke.

I was once involved (40 years ago) in moving an instrument enclosure about 400 yards, because the thermometer screen, was in our opinion, too close to a building nearby - but it was better sited than many of the current US sites. Still we were after perfection (a laughable concept now in many ways). Within 5 to 10 years a step change was obvious in the temperature data, especially in night minima (as one might expect).

I repeat for emphasis that this site originally was an absolute king compared to some in the US today.

The criteria for the siting of meteorological instruments is quite clear.

http://joannenova.com.au/2010/10/is-the-western-climate-establishment-corrupt-part-i/

The ones in my back garden are better exposed!!!!!!

Apr 30, 2011 at 7:48 PM | Unregistered CommenterRetired Dave

Retired Dave

Yes, this is getting a bit long-winded. Also the recent update at Wood for Trees to the latest satellite data products broke a number of the graphs above (see Apr 28, 2011 at 1:04 PM).

Perhaps I should have started a new thread when WfT updated. Anyway, all graphs below are corrected. Their equivalents above pre Apr 28, 2011 at 1:04 PM are obsolete.

The v5.4 UAH and v 3.3 RSS satellite data are in very good agreement and show a decadal warming trend of 0.14C since 1979.

The satellite TLT trend is very similar to surface temperature trends.

Bear in mind that:

- Global tropospheric temperatures are not influenced by UHI.

- Even the total ‘anthropogenic heat flux’ has no significant effect on TLT temperatures (see Apr 21, 2011 at 11:14 PM).

If UHI has artificially influenced surface temperatures its effect on trend is small: 0.01C per decade for HADCRUT and 0.02C per decade for GISTEMP.

My understanding is that it is GISTEMP’s Arctic interpolation methodology that makes it trend higher than HADCRUT (as opposed to a more pronounced contamination by UHI).

GISTEMP and HADCRUT are land-ocean surface temperature anomalies. The global ocean accounts for ~70% of the Earth’s surface and is not subject to UHI.

The assumption that most or all of AGW actually consists of UHI and adjustments to the surface temperature records is demonstrably false over the 32 year period of satellite observations.

Whatever is causing the recent warming is also warming the lower troposphere at essentially the same rate.

UHI cannot warm the global troposphere, therefore UHI is not a significant factor in recent warming.


TRENDS UAH v5.4 and RSS v3.3 1979 – present.

TRENDS RSS and UAH on 1981 – 2010 baseline, 1979 – present.

TRENDS RSS and UAH on 1981 – 2010 baseline, 12 month running mean.


When HADCRUT and GISTEMP are re-baselined for accurate comparison with the satellite TLT measurements, agreement remains good for interannual variability and for overall trend.

TREND AND TIME SERIES, HADCRUT, GISTEMP, UAH, RSS (common 1981 – 2010 baseline; 12 month running mean).


Decadal trend 1979 – present:

HADCRUT 0.15

GISTEMP 0.16

RSS 0.14

UAH 0.14

May 1, 2011 at 7:20 PM | Unregistered CommenterBBD

But gridcells where urbanisation is taking place would yield higher average anomalies over time as urbanisation progressively influences more thermometers within the cell. As these are averaged in, the cell average (and anomaly) will rise. Without correction, this would creat or amplify a warming trend.

That was my point as well. But, the difference between what I said and your position is that you contend that progressive temperature-influencing urbanization will 'amplify' a trend, whereas I state(d) that it would 'only create an apparent trend'.

Do we know this, know this (instead of hypothesizing what urbanization ought to do to a grid/station)? As in, do we have quantitative and qualitative data on what urbanization will do to a station or a grid?

In any case, the effect of what a particular pattern of urbanization would have on a grid, has to be established, independently. This effect must then be subtracted from that local grid. Summation must then proceed. I am not sure at all that this is being done to obtain the surface record. It therefore seems to me, that their derivation of UHI quantification from composite trends (say, like whole US surface data), to be circular, and involves mixing up of too many things.

Another way would be to compare differentially urbanized areas belonging to a single climatologic zone (and therefore could be expected to show different trends). (not nearby pairs, whole areas that are different from one another). I am sure there are such places around.

Since the above two ideas are simple and obvious lines of inquiry (for a bottom-up approach, as opposed to your top-down approach), I am sure these types of studies have been conducted already. I went looking for them and cannot find any presently.

I only saw an interesting comment about Isfjord Radio on Warwick Hughes' site that came somewhat close as an example. This example by the way leads to another interesting question: what about local effects (non-urban caused) that produce trends? Why not adjust for them too?

May 3, 2011 at 11:55 AM | Unregistered CommenterShub

Shub

That was my point as well. But, the difference between what I said and your position is that you contend that progressive temperature-influencing urbanization will 'amplify' a trend, whereas I state(d) that it would 'only create an apparent trend'.

Nope. What I said was:

Without correction, this would create or amplify a warming trend.

Which is correct.

Let's concentrate on the core of this. How can the UAH/RSS TLT temperature record be in such close agreement with the land-ocean surface temperature record if UHI is distorting the latter in any significant way?

Remember that whatever is causing the recent warming is also warming the lower troposphere at essentially the same rate.

UHI cannot warm the global troposphere, therefore UHI is not a significant factor in recent warming.

So, the assertion that most or all of AGW actually consists of UHI and adjustments to the surface temperature records is demonstrably false over the 32 year period of satellite observations.

All the obscurantist circumlocution in the world cannot explain the strong agreement between TLT and surface for trend and interannual variability.

Unless that is, you can find a study which shows UHI can elevate global tropospheric temperatures. Good luck with that.

May 3, 2011 at 7:19 PM | Unregistered CommenterBBD

Shub

I expect you've seen all this before; it's the standard stuff, but just for completeness' sake:

Talking Points related to concerns about whether the U.S. temperature record is reliable (NOAA)

Surfacestations.org has examined about 70% of the 1221 stations in NOAA’s Historical Climatology Network (USHCN) (Watts, 2009). According to their web site of early June 2009, they classified 70 USHCN version 2 stations as good or best (class 1 or 2). The criteria used to make that classification is based on NOAA’s Climate Reference Network Site Handbook so the criteria are clear. But, as many different individuals participated in the site evaluations, with varying levels of expertise, the degree of standardization and reproducibility of this process is unknown. However, at the present time this is the only large scale site evaluation information available so we conducted a preliminary analysis.

Two national time series were made using the same homogeneity adjusted data set and the same gridding and area averaging technique used by NOAA’s National Climatic Data Center for its annual climate monitoring. One analysis was for the full USHCN version 2 data set. The other used only USHCN version 2 data from the 70 stations that surfacestations.org classified as good or best. We would expect some differences simply due to the different area covered: the 70 stations only covered 43% of the country with no stations in, for example, New Mexico, Kansas, Nebraska, Iowa, Illinois, Ohio, West Virginia, Kentucky, Tennessee or North Carolina. Yet the two time series, shown below as both annual data and smooth data, are remarkably similar. Clearly there is no indication from this analysis that poor station exposure has imparted a bias in the U.S. temperature trends.

And see graph on Page 3:

http://www.ncdc.noaa.gov/oa/about/response-v2.pdf

This was preliminary to the formal publication of Menne et al. (2010), (controversially using Watts’ data without permission):

On the Reliability of the U.S. Surface Temperature Record

http://www1.ncdc.noaa.gov/pub/data/ushcn/v2/monthly/menne-etal2010.pdf

Basically, same result, same conclusion.

May 3, 2011 at 7:53 PM | Unregistered CommenterBBD

BBD,
My argument is very simple.

1) The surface temperature reconstructions you link to above, are adjusted in a myriad number of ways. How does one determine the effect or the lack of it thereof, of UHI, from such an end-product?

2) That said, I agree with you that the corrected, surface record matches the satellite measurements, with respect to trend.

3) That said, any further firmly-grounded conclusions cannot be drawn about the influence of UHI solely on the basis of the comparisons of the kind you draw attention to.

4) That said, it is very well evident that the UHI cannot be the sole origin of all the warming that has been observed during the satellite period. Hand in hand however, it is not clear to me, (a) as to who actually believed this to be true (b) whether you have not simply assumed that disproving this rather artificial point would simply negate all meaningful effects, if any, of the UHI

Secondly, it is your contention that UHI will 'amplify' a warming trend. It is only a contention, and secondly I am not clear what you mean by 'amplify'. If UHI turns a 0.2/decade trend into a 0.3/decade trend, is that amplification according to you? What I am saying is that while such a trend may well be evident in the early phase, eventually they would cease to be so (i.e., the slope will not go on increasing).

May 3, 2011 at 11:04 PM | Unregistered CommenterShub

Shub

1) The surface temperature reconstructions you link to above, are adjusted in a myriad number of ways. How does one determine the effect or the lack of it thereof, of UHI, from such an end-product?

By comparing it to the satellite measurements for TLT.

Or, did you mean that the actual graphs are in some way misleading?

The methodology used in the graphs is fully transparent so it can be examined and criticised. If I have made mistakes, please point to them and I will correct them.

2) That said, I agree with you that the corrected, surface record matches the satellite measurements, with respect to trend.

The surface records match the TLT records well both in trend and interannual variability. TLT is more responsive to ENSO, but the peaks and troughs cancel out. Hence the good agreement in trend with surface temperatures.

3) That said, any further firmly-grounded conclusions cannot be drawn about the influence of UHI solely on the basis of the comparisons of the kind you draw attention to.

Why not? How does UHI heat the global troposphere sufficiently to account for such close agreement in trend? You need to show this.

4) That said, it is very well evident that the UHI cannot be the sole origin of all the warming that has been observed during the satellite period. Hand in hand however, it is not clear to me, (a) as to who actually believed this to be true (b) whether you have not simply assumed that disproving this rather artificial point would simply negate all meaningful effects, if any, of the UHI

What artificial point? That tropospheric and surface temperatures show a (recently stalled) increase over 32 years with an average trend of 0.15C/decade? Or that surface and TLT records are in good agreement for interannual variability and trend?

Neither of which can be explained by UHI.

Nor does UHI account for the very close trend agreement between sea surface temperatures and satellite TLT temperatures.


Which really should make you pause for thought.

May 4, 2011 at 12:02 AM | Unregistered CommenterBBD

re: UHI

Apologies if this point has already been made.

UHI exists. We know it exists. It has probably been experienced by everyone at some time or other. UHI will affect temperature readings. However, that is not the same as saying that UHI will affect the temperature trend - and it is the temperature trend in which we are interested. The fact that an urban area reads, on average, 2 deg warmer than a rural area is irrelevant if this has always been so. It is no more relevant than the fact that temperatures at the equator are warmer (higher) than those in the arctic.

It does appear that, over the last 30 years at least, UHI has not had a significant influence on the global surface temperature trend.

May 4, 2011 at 10:52 AM | Unregistered CommenterJohn Finn

BBD,
I cannot understand how you expect the Hadley Center or GISS to put out a graph, that diverges from the satellite data, at all. Whatever be their adjustments, they wont do such a thing.

With that in mind however, my own understanding is very close to yours - I believe that a well-distributed and well-spaced grid would capture the essential global temperature trend, as well as any satellite can. It is the same thing they are measuring. Once again, that said, single thermometers or a handful of thermometers can pick up the climate signal (i.e., trend) as a part of a grid nicely no doubt, but have every potential to skew the regional signal when used in this way. Case in point: Eric Steig's Antarctic offerings (purely a regional example of the phenomenon I am describing), and the regional Arctic smearings seen in GISS.

In the end, I don't think I am in disagreement with you at all (to the specific point you wish to make). But I dont want to say "UHI does nothing to the surface record", simply because we don't have satellite data going well back enough for that.

You might find this graph very useful and timely.

http://diggingintheclay.files.wordpress.com/2011/04/overlay3.png

Just as expected, the rural, suburban and urban stations (R, S and U on graph) run quite parallel to each other, with a small, teensy bit of a trend divergence between the urban stations and the rural stations. And just as you point out, the inter-annual variability and overall trend are pretty well captured - in all three kinds of stations - the ups and downs are there in all curves. (just the graph is good for our purposes. The originating post has a related discussion but we can ignore that )

May 4, 2011 at 11:57 AM | Unregistered CommenterShub

Shub

I cannot understand how you expect the Hadley Center or GISS to put out a graph, that diverges from the satellite data, at all. Whatever be their adjustments, they wont do such a thing.

You appear to be implying that the land-ocean surface temperature records are ‘fixed’. In the same breath you accept that they do not diverge from the satellite TLT measurements, which we presume to be unbiased and relatively accurate.

You can’t have it both ways!

With that in mind however, my own understanding is very close to yours - I believe that a well-distributed and well-spaced grid would capture the essential global temperature trend, as well as any satellite can. It is the same thing they are measuring. Once again, that said, single thermometers or a handful of thermometers can pick up the climate signal (i.e., trend) as a part of a grid nicely no doubt, but have every potential to skew the regional signal when used in this way. Case in point: Eric Steig's Antarctic offerings (purely a regional example of the phenomenon I am describing), and the regional Arctic smearings seen in GISS.

While UHI may have regional effects, they are insufficient to affect the global picture. Land-ocean surface temperature trends and interannual variability match well with the global satellite TLT data.

In the end, I don't think I am in disagreement with you at all (to the specific point you wish to make). But I dont want to say "UHI does nothing to the surface record", simply because we don't have satellite data going well back enough for that.

This is confusing. While UHI undoubtedly has some slight regional effect on the land surface data, it is too small to show up in a comparison with satellite TLT records. So it doesn’t matter in the big picture.

UHI was presumably a smaller influence on earlier decades as there has been a significant acceleration in urbanisation over the last 30 years – happily the period validated by satellite observations.

The lack of significant divergence shows that UHI plays only a very minor role.

Everything else is basically distracting and irrelevant to this conclusion.

May 4, 2011 at 7:19 PM | Unregistered CommenterBBD

Following on, but for general interest (if any ;-), the agreement between global sea surface temperature anomalies (HadSST2) and satellite TLT measurements is very clear when the SSTs are re-baselined for correct comparison with satellite data.

Using the new UAH v5.4 1981 - 2010 baseline here.

And again here, but with a 12 month smoothing.

May 4, 2011 at 10:47 PM | Unregistered CommenterBBD

Delayed response to BBD, :)

No I don't believe that the land records are 'fixed'. They are UHI-corrected.

How can you look at a UHI-corrected record to discern the magnitude of the UHI?

I am not arguing vis a vis Watts, that even with correction, a significant proportion of the heat is/could be still accounted for by UHI. I am arguing that we should make a 'clean break' in terms of prior assumptions when approaching this issue fresh.

Look at the digging in the clay graph. The difference between the urban and the rural stations is a little over 0.2 C, unadjusted. That is a good bit, isn't it? Look also at the interannual variation. It is very similar in both. This is exactly what we are seeing with the satellite vs land WFT graph you linked.


In other words, we are saying the same thing.

May 9, 2011 at 3:56 PM | Unregistered CommenterShub

Shub

Just re-read the whole thread and pondered why, if we are in agreement, it is so long ;-)

For clarity, are you arguing that UHI has:

- Elevated the global average surface temperature anomaly 1979 - present (ie HADCRUT, GISTEMP) but not affected the decadal trend

- Had no effect on surface GATA 1979 - present

And do you agree or disagree with the concluding remark from the Digging ITC post that provided the graph you link:

Perhaps future generations will look back and laugh that we even tried to detect a warming signature in the global surface temperatures against a background of change in measurement environs that were so poorly documented and understood at the time.

(I'm not clear why the graph shows anomalies from the year 1940. Do you know why a single year was chosen as the reference? And why 1940? Sorry if I am missing something obvious.)

May 9, 2011 at 10:10 PM | Unregistered CommenterBBD

And back with some fresh news, namely the eagerly anticipated surfacestations paper. I suspect that will raise more questions than answers and be assumed to answer questions it doesn't. Like UHI :)

From the WUWT summary:

Here, in brief, are the answers: The poorest sites tend to be warmer. The minimum temperatures are warming faster at poorer sites than at better sites. The maximum temperatures are warming slower at poorer sites than at better sites.

and

Q: What’s next?

A: We also plan to look specifically at the effects of instrument changes and land use issues, among other things.

Poorer sites warmer? Could be UHI, but then why would Tmax be rising slower at poor sites than good? The 10% CRN1 sites should show any global warming signal but still don't explain the cause. Usual answer I get when discussing possible USHCN data discrepencies is 'thats just the US'. So it could be a regional discrepency caused by UHI or land use changes still, but also should contain any global warming signal. If Yamal can see the world, so should the CRN1 stations. We still need to find that signal and tease out the AGW element(s). But the paper does address the station and data quality issues. Mosh summarises what I think it means nicely in the WUWT discussion:

The REAL questions are
1. what uncertainty bounds do we put around that number.
2. how MUCH of the .8C is due to changes we have made to the atmosphere?

It has gotten warmer. Some of that is of course natural variability(sun and ocean cycles) . A small part of that may be UHI bias. and some of that, physics tells us, is due to GHGs.

You want a debate IN climate science, that is where the debate is.
If you want to argue that 100% of the .8c is UHI, you’re not in the debate
If you want to argue that it is all the sun, you’re not in the debate
If you want to argue that GHG is ALL of it or NONE of it, you’re not in the debate.

Fall et al addresses data quality issues around point 2, ie is it 0.8C. Not the rest. It makes some of the detection issues more certain, but doesn't answer the attribution questions. Those are being addressed in other papers, and are a combination of relatively small numbers. If AGW is as Hansen et al 2005 says, only a 0.85 ± 0.15 W/m2 problem, then small errors in global (or even human) assumptions can quickly add up to that amount. Going back to UHI and the Flanner paper quoted earlier

but over the continental United States and western Europe, it is +0.39 and +0.68 W m−2, respectively

which is a fair amount of Hansen's global warming energy. Averaged out, it gets lost in the noise, but is averaging it out globally the correct thing to do? It's extra energy into large regional climates and will presumably have some effect on those climates. And that's just waste heat, which should also show up in TLT. Other anthropogenic effects like land use changes, aerosol emissions will also have an effect, positive or negative. I still don't think apparent agreement between surface and satellite measurements proves or disproves UHI though. Global warming certainly isn't entirely due to UHI, but some of it will be, and we still don't really know how much.

May 12, 2011 at 12:28 PM | Unregistered CommenterAtomic Hairdryer

Atomic

I despair.

Global warming certainly isn't entirely due to UHI, but some of it will be, and we still don't really know how much.

Look, UHI is an artefact. It is local, it affects thermometers. The actual amount of energy involved is so trivial as to be unmeasurably small in the context of global atmospheric temperature change.

Yes, of course it might have a disproportionate effect on surface temperature reconstructions if not corrected for carefully.

No, it cannot ever warm the TLT.

So if the global average anomaly TREND for TLT and the global average anomaly TREND for surface T are in good agreement we know that:

- warming is real

- warming is not caused by UHI (or anything else) contaminating the surface temperature records

For the last time, forget about UHI. The trend agreement between surface GATA and TLT GATA tells you all you need to know.

Look to other causes for the increasing temperature over time. Of course there will be a mix of forcings. Of course CO2 will be one of them (possibly now dominant) but still only one of them.

Averaged out, it gets lost in the noise, but is averaging it out globally the correct thing to do?

Of course it does, because the net effect is trivial. Of course you average because it is a regional atypical forcing and averaging is the correct way to smooth for this.

I only started this to show that those who dismiss the surface temperature records as corrupted, useless, dishonest etc are completely wrong and it is easy to prove it.

It is time to forget about UHI as a significant component of 'global warming'. If you prefer, just look at UAH and RSS - that's the truth, more or less: 0.14C/decade trend 1979 - present. Less than the multi-model mean of 0.2C. That's what matters. Climate sensitivity to CO2 as derived from observation not hypothesis is what matters.

And it's looking lower than the consensus median estimate of ~+3C per doubling.

May 12, 2011 at 5:06 PM | Unregistered CommenterBBD