Jones live blog

No more questions

2nd Monckton - asking about 60yr perioddicity and PDO. Can we detect AGW. PJ says yes.

1st question Tony Kelly of GWPF. Sea temps fundamentally different to land. Should measure marine air temp. Says sst doesn't agree to land. PJ says SST correlated to marine air temp

Questions next

And that's it

Satellites - same results as surface

SST - bucket adjustment is the most important. Also shift to buoys

Discussion of CRU subsampling studies - still get same result

BEST 2% series shown. We are getting the message that all the series show roughly the same thing.

homgeneity/urbanisation  adjusts have small effect

CRUTEM4 will include a full release of station data

adjustments to stations makes little difference at large scales

"Population growth is not a great metric for urbanisation trends"

Urbanisation. Little difference in urban/rural

Biases in order of importance form Jones and Wigley. 1 SST bucket 2. Thermometer ecposure, 3. Urbanisation

Some discussion of Menne et al 2009 - bimodal distribution of adjusts

PJ discussing homogenity issues

We've been told we will be evicted if we mention climategate

Lovely Georgian theatre here.

Comments

[BBD]

phi

Thanks for the graph.

What do we know?

- NH land surface T is rising faster than SH LST

- This is not affecting global TLT measurements by satellite

- Enhanced NH land surface warming rate (trend) is in line with modelled projections

- Upper troposphere (not TLT) temperatures at the equator and low latitudes are expected to show a higher trend than equatorial surface T

- The divergence in trend is expected to be less as you approach the poles

- Your example demonstrates this well

Does this affect comparisons of global land-ocean surface temperature anomalies with global TLT temperature anomalies? No.

[Shub]

BBD,
Firstly, when I said 'quality', I put it in quotes, to indicate the difference between what I percieve as quality and what is conventionally understood. Perhaps I did not convey this meaning clearly.

[1] My starting assumption is: the so-called uncorrected land record will be a reasonable proxy of climatology over large regions.
[2] Because I believe the uncorrected land record to 'alright', I don't care much for adjustments, quality control etc. This things are important, but not for the purpose of present argument
[3] Because of [1], I expect the satellite TLT data to match the uncorrected data. This is because, after all, the near-surface thermometers and the satellites TLT probes are measuring the same atmospheric unit, subject to similar influences
[4] Contrarily, we find, that there is poor correlation between the two. This is markedly true over the tropics.
[5] Since I believe the satellites are an even better measure of temperature estimation, than a field reconstruction using thermometers, I am surprised by the lack of correlation
[6] This discrepancy is what I referred to as poor 'quality'. The question for me here is: why aren't the near-surface thermometers not picking up what the satellites are measuring, and instead, measuring something else?

This is all I am saying and nothing more.

Moreover, I cannot understand why you find it difficult to understand my relatively simple questions and points. You seem to want to anticipate where I am headed at, and answer those questions. There is no need, believe me.

For instance, the apples and pears point by phi is very straightforward. how can one take the 'apple' of US land temperatures, mix it with the rest to produce a 'global average' and then compare with the satellites? Just take US land and compare with US satellite. If that can't be done, do it with the NH. Which is what phi's point was.

[phi]

BBD,

I don't see things like you, not at all. Not matter. Have you noticed that the divergence CRUTEM - UAH was almost linear? I think that before using GISS or CRUTEM for anything, this singularity should be explained as should be explained the homogenization bias affecting CRUTEM (0.5 ° C per century).

[BBD]

Shub

I'm listening carefully.

For instance, the apples and pears point by phi is very straightforward. how can one take the 'apple' of US land temperatures, mix it with the rest to produce a 'global average' and then compare with the satellites? Just take US land and compare with US satellite. If that can't be done, do it with the NH. Which is what phi's point was.

Let's consider what Roy Spencer says about global average temperature here (scroll down to May 22; emphasis added).

[...] the globally averaged temperature is not just a meaningless average of a bunch of unrelated numbers. This is because the temperature of any specific location on the Earth does not exist in isolation of the rest of the climate system. If you warm the temperature locally, you then will change the horizontal air pressure gradient, and therefore the wind which transports heat from that location to other locations. Those locations are in turn connected to others.

In fact, the entire global atmosphere is continually overturning, primarily in response to the temperature of the surface as it is heated by the sun. Sinking air in some regions is warmed in response to rising air in other regions, and that rising air is the result of latent heat release in cloud and precipitation systems as water vapor is converted to liquid water. The latent heat was originally picked up by the air at the surface, where the temperature helped govern the rate of evaporation.

In this way, clouds and precipitation in rising regions can transport heat thousands of kilometers away by causing warming of the sinking air in other regions. Surprisingly, atmospheric heat is continually transported into the Sahara Desert in this way, in order to compensate for the fact that the Sahara would actually be a COOL place since it loses more IR energy to space than it gains solar energy from the sun. (This is because the bright sand reflects much of the sunlight back to space).

Similarly, the frigid surface temperature of the Arctic or Antarctic in wintertime is prevented from getting even colder by heat transport from lower latitudes.

In this way, the temperature of one location on the Earth is ultimately connected to all other locations on the Earth. As such, the globally averaged surface temperature — and its intimate connection to most of the atmosphere through convective overturning — is probably the single most important index of the state of the climate system we have the ability to measure.

Granted, it is insufficient to diagnose other things we need to know, but I believe it is the single most important component of any “big picture” snapshot of climate system at any point in time.

Okay, now we look at uncorrected HADCRUT vs UAH.

You say:

[4] Contrarily, we find, that there is poor correlation between the two. This is markedly true over the tropics.

(What reference are you using for the tropical divergence?)

[5] Since I believe the satellites are an even better measure of temperature estimation, than a field reconstruction using thermometers, I am surprised by the lack of correlation

[6] This discrepancy is what I referred to as poor 'quality'. The question for me here is: why aren't the near-surface thermometers not picking up what the satellites are measuring, and instead, measuring something else?

All I have tried to do is show that global average anomalies compensate effectively for spatial variance.

In other words, that the big picture is essentially correct.

When you and phi point to regional variance it raises interesting questions but they do not impinge on the broad accuracy of the land-ocean GATA as validated by the TLT GATA.

Now, are you arguing that the NH land surface trend is artificially elevated vs UAH NH land because of bad station siting (or UHI if you prefer)? Or do you think the NH land surface trend is elevated vs UAH NH land because of a general surface warming correctly captured by the thermometers?

[BBD]

phi

I don't see things like you, not at all. Not matter.

I'm not surprised. My comment at 7:54pm is a complete mess. My apologies.

Please disregard and move on the next.

Your illustration of the divergence between CRUTEM and GISS NH and UAH NH is of course interesting. I ask you the same question as I ask Shub:

Are you arguing that the NH land surface trend is artificially elevated vs UAH NH land because of bad station siting (or UHI if you prefer)? Or do you think the NH land surface trend is elevated vs UAH NH land because of a general surface warming correctly captured by the thermometers?

[BBD]

Just a reminder of what this argument/debate/whatever is all about, at least for me.

The decadal trend for global average TLT is 0.14C 1979 - present.

This is very likely ;-) the best measure of a warming trend we have.

The multi-model mean referenced by AR4 is 0.2C/decade.

It seems to be too high. If so, the median estimate for climate sensitivity to CO2 doubling of ~3C at equilibrium is too high.

[Maurizio Morabito]

if the format is speech plus two questions, this secret trip is going to be quite a waste of time for the Bishop

[phi]

BBD,

I agree with your analysis (12:36 AM).

In my opinion, the divergence CRUTEM - UAH is essentially an estimator of site disturbance (not exactly poor siting nor UHI). This estimate is similar to that which can be drawn from the divergence of MXD (the bias of 1 ° C would appear around 1920).

[Mac]

The missing tropospheric Hot-Spot in the tropics is fundamental to the AGW hypothesis. No Hot-Spot - no AGW - end of.

We know that scientists have been trying very hard to find the Hot-Spot, but the best excuses, and I mean excuses, that these scientists have come up with is that the uncertainties and problems in recording and determing tropospheric temps are so great as to be nearly insurmountable that the Hot-Spot could be there but to-date remains undetected. But the problem here is that they are in effect discounting all the satellite and radiosonde data. That is a very weak scientific arguement, and is now a source of acute embarrassment.

Lets move on, if the cause remains undetected can we measure its effect?

Again the AGW hypothesis comes to hand. We know the predictions for both surface temps and oceanic heat content.

We have the data at hand, the results are in.

1. Phil Jones was forced to concede that there has been no significant surface warming for the last 15 years.

2. There has been no surface warming at all for the last 10 years.

3. There has been no significant increase in oceanic heat content.

All this runs counter to predictions.

The conclusion is stark: increasing levels of atmospheric CO2 has not resulted in additional heat being trapped in the troposphere. All the rest is simply hand waving.

When a doctrine is subject to criticism and fails it's continued acceptance means it has become dogma.

[Shub]

I did not say 'divergence', I said correlation. A warm bias in the tropics is referred to Karl et al 2006 in the Klotzbach et al 2009 paper.

I am unable to find any reference that clearly quantifies what the amount of the 'tropical bias' is. Incidentally, this is the same global region that shows poor correlation between uncorrected land and satellite TLT

The passage you posted from Roy Spencer: I do agree with it. But, that is useful from a perspective of justifying the meaningfulness of a global average temperature. But our question relates to the inter-comparability of different series. Things which behave differently as series (for eg, large land masses, US land mass, Europe) should be kept differently and then compared. So if you want to show a 'global temperature', show a handful of climatologic regions, each with their own separate trend, error bars etc. For support for this way of thinking, see Le Mouël et al 2008.

Again, quoting Le Mouël et al 2008 does not mean I am interesting in a 'solar theory' as well.

[BBD]

No Mac, let's not 'move on'.

I am waiting for references supporting various of your unsupported claims (I first asked yesterday at 1:17pm):

The radiosonde analysis is factual - no additional heat is being trapped in the tropical troposhere, indeed this part of the atmosphere has experienced cooling - the opposite to what the AGW hypothesis predicts.

References? Especially for your assertion that 'the tropical troposphere' (at what height? All of it?) has cooled.

Measurements of Oceanic Heat Content show that the rise in heat is a lot less than predicted.

Reference?

There has been no significant increase in oceanic heat content.

Over what time scale? Reference?

Come on Mac. Back it all up.

As, for example, I did yesterday by referencing Titchner et al.2009 and Randel & Wu 2006 which demonstrate that the radiosonde data are biased cool. Contra your claim.

Before going on with this exchange, I want to see your sources. This is because I suspect you are making things up as you go along.

[BBD]

No rush - meetings for a couple of hours now.

[sherlock]

When I was a kid I was sure I was going to be a scientist someday, but my path in life took me in a different direction, something I have always looked back upon with some regret.

Until now.

[Shub]

BBD,
About the radiosonde data being cooler, please see Thorne et al 2011, page 12. There is a nice graph showing nice inter-dataset agreement, and consistently lower values for C/decade than the modelled expected values.

[BBD]

Shub

A link would be helpful - busy as hell this afternoon and cannot find this paper (google).

Thanks

Dominic

[BBD]

Shub

Your comment is extremely confusing.

I did not say 'divergence', I said correlation. A warm bias in the tropics is referred to Karl et al 2006 in the Klotzbach et al 2009 paper.

Where are you addressing the remark about 'divergence'? Lost here.

I am unable to find any reference that clearly quantifies what the amount of the 'tropical bias' is. Incidentally, this is the same global region that shows poor correlation between uncorrected land and satellite TLT

So the 'tropical bias' claim is unsupported?

The passage you posted from Roy Spencer: I do agree with it. But, that is useful from a perspective of justifying the meaningfulness of a global average temperature. But our question relates to the inter-comparability of different series. Things which behave differently as series (for eg, large land masses, US land mass, Europe) should be kept differently and then compared. So if you want to show a 'global temperature', show a handful of climatologic regions, each with their own separate trend, error bars etc. For support for this way of thinking, see Le Mouël et al 2008.

I have dealt with this more than adequately above. GATA to GATA is not problematic at all. It achieves exactly what you are talking about. It is a non-issue except in your mind. Re-read. Getting fed up with this now.

Will you please stop waffling and confusing the issue and address the following:

All I have tried to do is show that global average anomalies compensate effectively for spatial variance.

In other words, that the big picture is essentially correct.

When you and phi point to regional variance it raises interesting questions but they do not impinge on the broad accuracy of the land-ocean GATA as validated by the TLT GATA.

Now, are you arguing that the NH land surface trend is artificially elevated vs UAH NH land because of bad station siting (or UHI if you prefer)? Or do you think the NH land surface trend is elevated vs UAH NH land because of a general surface warming correctly captured by the thermometers?

And quote your reference points, with links to the original papers. Thanks.

[BBD]

Mac

References please.

[Shub]

BBD,
This discussion will not go anywhere because (a) you wish to pigeonhole every point into a category of argument (b) you wish to cordon off and funnel all arguments and points, as though they were all offered to answer some imaginary single question which you have unilaterally decided, is the only point of contention. We have largely talked past each other because our concerns are not shared.

I thought Klotzbach et al 2009 is a fairly well known paper, so I did not want to rehash it. This is what they write about the tropical bias, quoting Karl et al 2006:

(Please note: Karl et al 2006 = CCSP report)

‘‘Systematic local biases in surface temperature trends may exist due to changes in station exposure and instrumentation over land, or changes in measurement techniques by ships and buoys in the ocean. It is likely that these biases are largely random and therefore cancel out over large regions such as the globe or tropics, the regions that are of primary interest to this report.’’

and in the immediately following passage, responding to the above paragraph the authors write:

However, it is unclear whether the assumption of ‘randomness’ has any scientific ground, as there exists recent research documenting spatially nonrepresentative warming biases in the surface temperature data that were not considered in the CCSP report [see Hale et al., 2006; Pielke et al., 2007a]. Indeed, for the latitudes 20N to 20S, the CCSP acknowledges that an unexplained difference
between the surface and tropospheric trends still exits (Executive Summary of Karl et al. [2006, p.2])

Look at Table 1:

NCDC Surface (land trend): 0.32C
UAH Lower Troposphere: 0.16C

This is what I referred to earlier. The uncorrected tropical area land stations are poorly correlated with the satellite data - this is a fact. The tropical and the northern hemisphere land mass shows a 'warm bias'/discrepancy/warming - call it what you will, compared to the satellite data.

Secondly, two time series data can be poorly correlated and yet show roughly the same trend. These things may not be important to someone who wants to only say: "see....warming", but they are interesting, and may hide important physical phenomena nonetheless.

Therefore if the land thermometers are recording a higher temperature than UAH, my approach is: leave the data be. Don't touch both. Try to think of a physical explanation which could explain the phenomenon. Instead of cooking up a grand global warming greenhouse theory, coming up with one single expected answer, and then deciding one or the other data is "artificially elevated", "cold-biased" etc. Believe your instruments - that is the first rule. (not blindly of course, but that is understood).

[BBD]

Shub

You say:

This discussion will not go anywhere because (a) you wish to pigeonhole every point into a category of argument (b) you wish to cordon off and funnel all arguments and points, as though they were all offered to answer some imaginary single question which you have unilaterally decided, is the only point of contention. We have largely talked past each other because our concerns are not shared.

And I say this exchange is suffering because when I ask, twice:

Now, are you arguing that the NH land surface trend is artificially elevated vs UAH NH land because of bad station siting (or UHI if you prefer)? Or do you think the NH land surface trend is elevated vs UAH NH land because of a general surface warming correctly captured by the thermometers?

Your eventual response is:

Therefore if the land thermometers are recording a higher temperature than UAH, my approach is: leave the data be. Don't touch both. Try to think of a physical explanation which could explain the phenomenon. Instead of cooking up a grand global warming greenhouse theory, coming up with one single expected answer, and then deciding one or the other data is "artificially elevated", "cold-biased" etc. Believe your instruments - that is the first rule. (not blindly of course, but that is understood).

Which I find evasive.

William of Ockham comments:

Numquam ponenda est pluralitas sine necessitate.

You... move about a lot.

For example, you have repeatedly urged that I use uncorrected surface temperature records.

I insisted, and rightly in my opinion, that one ought to look first at an uncorrected temperature record in order to do any kind of meaningful analysis vis a vis its comparision with the satellite data.

Yesterday, in response to that request, I posted this:

Here's HADCRUT unadjusted global mean vs UAH on a common 1981 - 2010 baseline.

http://woodfortrees.org/plot/hadcrut3gl/from:1979/offset:-0.26/plot/uah/plot/hadcrut3gl/from:1979/offset:-0.26/trend/plot/uah/trend

Guess what? The decadal trends 1979 - present are exactly the same as per the adjusted comparison:

HADCRUT 0.15C

UAH 0.14C

Wow. Who'd have thought?

And you simply ignored it (and I should have cut the snark).

I’m interested in the very close agreement in trend between surface GATA and TLT GATA because these variously flawed measurements still mutually confirm an increase of energy in the climate system. Surprisingly well.

I'm pretty sure you know that UHI is energetically insignificant. It may influence surface temperature reconstructions through methodological weakness, but it is localised and trivial in terms of energy flow within the climate system. It can no more warm the global TLT that I can jump over the moon.

So it isn't responsible for the nearly matching warming trends for surface and TLT.

You say:

We have largely talked past each other because our concerns are not shared.

We have certainly talked past each other but I think our concerns are shared. We both want to know more about the real-world effects of adding CO2 to the atmosphere. (To be clear, I am a lukewarmer sceptical about the actual climate sensitivity to CO2 forcing).

What is interesting about the surface GATA vs the TLT GATA is that even if the surface trend is elevated by ‘weirdness’ it is below the 0.2C/decade multi-model mean referenced by AR4.

The awkward question of tropospheric amplification at the equator rears its head… but it’s a distraction.

What matters is that the most reliable trend we have – TLT GATA – is 0.14C/decade 1979 – present. There has been mention elsewhere of ‘missing energy’. The hypothesised median estimate for climate sensitivity to a doubling of CO2 is looking wobbly.

As sceptics both, I’m certain that we can agree on this.

[Shub]

BBD,
Not a reply to your post above, but I just found this:

http://www.weatherbell.com/jb/?p=1749

[phi]

BBD,

Here's HADCRUT unadjusted global mean vs UAH on a common 1981 - 2010 baseline.

This serie does not include adjustment of the variance, but it has nothing to do with the raw temperatures (unadjusted or not homogenized). HadCRUT does not provide the raw data, it seems that the CRU does not simply have these data for some countries.

...confirm an increase of energy in the climate system.

If we include the ocean in the climate system, it becomes less than certain for most of the warming.

I'm pretty sure you know that UHI is energetically insignificant.

Disturbances by urbanization reported over surfaces completely sealed and drained and taking into account the energy consumption is in the same order of magnitude than the solar gain in terms of mean power.

Böhm 2001, relied upon by the CRU, said the disturbances by urbanizqtion affecting the stations were 0.55 °C higher in 1890 than in 2000!

[BBD]

phi

UHI is energetically insignificant, as stated. It does not, and can not heat the TLT. It can warm small areas of the boundary layer and those thermometers sited within them.

Why this point is proving so difficult to communicate is a mystery. Unless of course it is an inconvenient truth.

If the increase in energy in the climate system is an illusion made from UHI or other land-based phenomena, please account for the increase in GLOBAL TLT temperature.

Since the sea is 70% of the surface area, it is obviously having a larger effect on GATA than the 30% land surface area. This is all so obvious I am amazed that I seem to have to repeat it.

The choice of variance adjusted/unadjusted surface records is, frankly an irrelevance. The agreement in slope of the trends for surface and TLT is what matters.

Where's the energy coming from that explains this?

Does the TLT simply warm itself?

Is it magic? Fairy dust? Instrument error? Fraud?

[BBD]

Shub

D'Aleo has used Ch4 and Ch6 as well as Ch5. See Spencer's techical note below:

THE DISCOVER WEBSITE: NOAA-15 PROBLEMS STARTING IN MID-DECEMBER
For those tracking our daily updates of global temperatures at the Discover website, remember that only 2 “channels” can be trusted for comparing different years to each other, both being the only ones posted there from NASA’s Aqua satellite:

1) only ch. 5 data should be used for tracking tropospheric temperatures,
2) the global-average “sea surface” temperatures are from AMSR-E on Aqua, and should be accurate.

The rest of the channels come from the AMSU on the 12 year old NOAA-15 satellite, WHICH IS NOW EXPERIENCING LARGE AMOUNTS OF MISSING DATA AS OF AROUND DECEMBER 20, 2010.

http://www.drroyspencer.com/2011/01/

(Scroll down to Jan 03 2011)

Furthermore, we are currently still cool from the tail-end of La Nina. ENSO shows up more strongly in the satellite measurements - especially Ch5 (TLT).

So I'm not buying. But thanks for the link.

[BBD]

Shub

Here's a quick example of the more extreme TLT response to ENSO. It affects both peaks and troughs and is clearly visible in this comparison:

HADCRUT vs UAH. 1995 - present; common 1981 - 2010 baseline.

[phi]

BBD,

UHI is energetically insignificant, as stated. It does not, and can not heat the TLT. It can warm small areas of the boundary layer and those thermometers sited within them.

I totally agree with you on this point. If I insist, is that disturbances from urbanization have enormous importance for our understanding of continental temperatures of the twentieth century. The TLT are only known for too little time (1979). Rising global temperatures as measured by TLT may well be dominated by natural causes such ocean.

Since the sea is 70% of the surface area, it is obviously having a larger effect on GATA than the 30% land surface area. This is all so obvious I am amazed that I seem to have to repeat it.

Yes but the continents of the Northern Hemisphere are expected to warm more strongly than the oceans. What can this mean if we discover that there is no warming trend of the land since 1850 ? This is what seems to tell us the MXD, the TRW and glaciers (among other things). 30 years is too short to draw meaningful conclusions, especially when it starts in a hollow and includes 15 years of rise and 15 years flat.

The choice of variance adjusted/unadjusted surface records is, frankly an irrelevance. The agreement in slope of the trends for surface and TLT is what matters.

Indeed, this is exactly what I said. However it would be of great interest comparing raw data to homogenized data.

[BBD]

phi

Okay; we are in agreement that:

- UHI (or disturbances by urbanisation, if you prefer) may account for some or even most of the NH land surface warming trend of the last century.

- The global ocean may be warming, and in turn warming the TLT

- This may have a natural explanation

You say:

Yes but the continents of the Northern Hemisphere are expected to warm more strongly than the oceans. What can this mean if we discover that there is no warming trend of the land since 1850 ? This is what seems to tell us the MXD, the TRW and glaciers (among other things). 30 years is too short to draw meaningful conclusions, especially when it starts in a hollow and includes 15 years of rise and 15 years flat.

It is interesting that you mention glaciers. I'm not a huge fan of pointing to glacier retreat and shouting 'global warming!' because this is simplistic.

As I am sure you know, the advance and retreat of glaciers is determined by mountain precipitation vs annual mean T at the snout. Physical obstruction of the snout (eg ice sheets inhibiting the flow of glaciers of the WAIS) and debris atop the glacier are also significant.

However, there is absolutely no doubt that globally, glaciers have retreated (with exceptions and reversals, of course) since the late 1800s.

This figure from the World Glacier Monitoring Service (Global Glacier Changes report p24 fig 5.1) makes this absolutely clear.

I suggest that the thermometer records are corroborated by the broad picture of predominant glacial retreat, and that there has indeed been an increase in surface T over the last century.

Of course the onset of the glacial retreat appears to be too early to be explained by CO2 forcing. This is not supposed to be climatically significant until after 1950. There is the possibility that (as now) black carbon played a significant part in triggering the process.

Whatever the case, we can still point to glacial retreat as strong (arguably incontrovertible) evidence for increasing land surface temperatures over the last century.

WGMS home: http://www.geo.uzh.ch/microsite/wgms/

Report download page: http://www.grid.unep.ch/glaciers/

See here for associated text: http://www.grid.unep.ch/glaciers/pdfs/5.pdf

[BBD]

Now, let's go back to the points of potential agreement and consider them again:

- UHI (or disturbances by urbanisation, if you prefer) may account for some or even most of the NH land surface warming trend of the last century.

- The global ocean may be warming, and in turn warming the TLT

- This may have a natural explanation

Glacial retreat suggests the land surface temperature really is increasing (it's NOT just UHI and bad methodology etc).

If land surface T is being forced by a warming global ocean (which is also warming the TLT) this could account for the glacial retreat. It may also explain the early onset of the retreat, which appears to pre-date the hypothesised detectable effects of CO2 forcing.

My personal view is that we are seeing two different things.

The first warming period initiated the glacial retreat post-1885 and ended in 1940. It was almost certainly natural.

The second began post 1950 (1975 is a good 'start year', with the Pacific Climate Shift coming in 1976) and ended (paused?) in 2000. This is more likely to be AGW than natural, although it may well contain a natural component (possibly multi-factoral).

[phi]

BBD,

Thank you for the links.

If the glaciers retreat since the nineteenth century, this means that the conditions for retreat were met at the time. If these conditions have remained unchanged to this day, it is normal that the retreat continues. There is no need to take into account a warming in the twentieth century to explain it.

Progress and retreat of glaciers indeed also depends on precipitations, which is why glaciologists use the notion of melting anomaly when they aim to highlight the correlation with temperature. That is what do Huss et al. 2009 for 3 alpine glaciers. In this case the melting anomaly shows no general trend to increase during the twentieth century (the melt was larger in the 1940s than in the 2000s).

In general, I do not have a very particular opinion about the real effect of CO2, I'm just highlight the important gaps that affect our understanding of temperature trends in the twentieth century.

[Shub]

In general, I do not have a very particular opinion about the real effect of CO2, I'm just highlight the important gaps that affect our understanding of temperature trends in the twentieth century.

Yes, indeed. For some reason however, BBD insists that I confront the greenhouse demon with Occam's razor in hand, not 'move around a lot', etc. :)

This is another approach to 'uhi'. Please point out where I am wrong.

[1] The amount of records disturbance caused by poor station exposure (concrete under thermometers, jet exhaust, a/c unit exhaust pointed at thermometer) is small, and is referred to, as the 'urban heat island effect'.
[2] But at the same time, the urbanization/land use change that causes this poor station exposure to occur, is quite substantial, and has occurred over the entire 20th century.
[3] The effect this urbanization/land use change has on the climate, in turn can be quite significant.
[4] The urbanization/land use change, the poor exposure, plus 'global warming' and other influences conspire to produce the decadal trend.

I am getting [3] from Klotzbach et al 2009, and other Pielke Sr papers and articles.

[BBD]

Shub

Yes, indeed. For some reason however, BBD insists that I confront the greenhouse demon with Occam's razor in hand, not 'move around a lot', etc. :)

It's called 'the scientific method'.

And this is where it leads:

Take the undisputed radiative physics of CO2.

Ask:

If more CO2 is added to the atmosphere will it

- do nothing

- cool

- warm

When you have considered that for a moment (with Occam's razor at your throat), perhaps you see my problem with your position.

If you can show me any widely accepted evidence that land use change is a major climate forcing, comparable with that from GHGs, do so.

I maintain my position: UHI is energetically insignificant. Land use change is significant, but not sufficient to account for observed trends. Even if you discount all records pre-1979, this remains true.

[BBD]

phi

If the glaciers retreat since the nineteenth century, this means that the conditions for retreat were met at the time. If these conditions have remained unchanged to this day, it is normal that the retreat continues. There is no need to take into account a warming in the twentieth century to explain it.

Progress and retreat of glaciers indeed also depends on precipitations, which is why glaciologists use the notion of melting anomaly when they aim to highlight the correlation with temperature. That is what do Huss et al. 2009 for 3 alpine glaciers. In this case the melting anomaly shows no general trend to increase during the twentieth century (the melt was larger in the 1940s than in the 2000s).

The World Glacier Monitoring Service report does not support this claim. Please see fig 5.8.

Mass balance measurements on entire glaciers have been available for the past six decades. Glacier mass change is a direct, undelayed reaction to atmospheric conditions. The specific mass balance can be compared directly between different glaciers. This makes it easier to establish a link to climate data, as compared to length changes. However, the limited number of long term observations – only 30 ‘reference’ glaciers have continuous data series since 1976 – renders global analysis much more complicated.

As a consequence there are three main approaches to calculating global average mass balances which are independent of climate, hydrology or climate indicator data. These are by (i) using the (arithmetic) mean value of the few continuous measurement series, (ii) averaging the moving sample of all available data series, and (iii) using regionally weighted samples (cf. Kaser et al. 2006).

However, when cumulated over the past six decades, the results of these approaches are consistent. The global averages (i, ii, iii) reveal strong ice losses in the first decade after the start of the measurements in 1946, slowing down in the second decade (1956–65), followed by a moderate mass loss between 1966 and 1985, and a subsequent acceleration of ice loss until present (Fig. 5.8 a—f). The mean of the 30 continuous ‘reference’ series yields an annual mass loss of 0.58 m water equivalent (m w.e.) for the decade 1996–2005, which is more than twice the loss rate of the previous decade (1986–95: 0.25 m w.e.), and over four times the rate for the period 1976–85 (0.14 m w.e.).

[phi]

Shub

Land use is a question for climatologists, What I'm not. In my opinion, it will anyway be properly adressed only the day we have data reasonably treated.

UHI has a definite effect on the comfort of the urban and the thermometers but probably not on the global climate. Not to be confused with the poor quality of the stations which affect only thermometers and climatologists. In both cases, in terms of overall anomlies is anyway the evolution of these characters that matters and not their absolute value.

[BBD]

Shub

[4] The urbanization/land use change, the poor exposure, plus 'global warming' and other influences conspire to produce the decadal trend.

UHI cannot affect TLT and land use change is not a major forcing and so will have only minor effects on TLT. How do we account for the +0.14C/decade trend in TLT GATA?

[phi]

BBD,

Interesting but a bit short to conclude.

On the one hand the general trend since the nineteenth century, a study dating back to 1946 at best do not say much. On the other hand 1946 is already well into the twentieth (just above the maximum detected by Huss). The methodology is also a bit, say, random. If you intend to provide anything in the climate, it is essential to have the raw data. See what's happening with the tree rings! Even the most reliable series (Schweingruber) was triked and you are forced to use the raw data to conclude anything worthwhile.

[BBD]

phi

I appreciate the properly sceptical view that the raw data are vital to verification of any study, but there is a worrying tendency here for everyone to assume that everything is unsound.

The mass balance analysis referenced above is not (to my knowledge) the subject of academic or blogospheric dispute. And it shows accelerated melt in line with the surface and latterly satellite TLT records. This corroborates the surface and now the TLT observations.

You say:

The methodology is also a bit, say, random.

Could you be more specific? What problems do you see, and what do you suggest are the likely effects on the findings as referenced above?

Regulars here know that the dendro proxies and the uses to which they have been put are questionable. This is why I have kept the focus on glaciers. I was trying to avoid another 20-comment exchange over methodology and interpretation ;-)

[Maurizio Morabito]

Careful with glacier studies. Due to mechanical stresses a lot can be lost for very small climatic changes.

[BBD]

Maurizio Morabito

Isn't this one of the things the methodology used in the mass balance study is designed to accommodate?

As a consequence there are three main approaches to calculating global average mass balances which are independent of climate, hydrology or climate indicator data. These are by (i) using the (arithmetic) mean value of the few continuous measurement series, (ii) averaging the moving sample of all available data series, and (iii) using regionally weighted samples (cf. Kaser et al. 2006).

And is this not suggestive that the approach is successful?

However, when cumulated over the past six decades, the results of these approaches are consistent. The global averages (i, ii, iii) reveal strong ice losses in the first decade after the start of the measurements in 1946, slowing down in the second decade (1956–65), followed by a moderate mass loss between 1966 and 1985, and a subsequent acceleration of ice loss until present (Fig. 5.8 a—f). The mean of the 30 continuous ‘reference’ series yields an annual mass loss of 0.58 m water equivalent (m w.e.) for the decade 1996–2005, which is more than twice the loss rate of the previous decade (1986–95: 0.25 m w.e.), and over four times the rate for the period 1976–85 (0.14 m w.e.).

[BBD]

From the report, to keep the links simple to find:

Fig 5.1 Glacier length change

And

Fig 5.8 Spatio-temporal overview of glacial mass balance change

[phi]

BBD,

I appreciate the properly sceptical view that the raw data are vital to verification of any study, but there is a worrying tendency here for everyone to assume that everything is unsound.

I also think it is a unfortunate tendency but that is the salary of climatologists and they deserve it. The global surface temperature series are absolutely useless. Outside the raw data, no palaeoclimatic study has any value. That's what the situation today.

Personally, I think that melting anomaly are excellently correlated with surface temperature. That's what I infer from the comparison of results of Huss with the raw data of the station they have chosen as reference. But to decide, you need reliable, continuous data over a relatively long period.

Very few and incompletes samples averaged over 120 degrees of longitude and 60 degrees of latitude, what do you get serious? In my opinion, not much. For 60W-60E/30N-90N, what spatial representation in 1946? and 2000? How many samples from Greenland, Alps on those dates? When is the maximum T before 1980 ? In many places it is before 1946.

For dendro, there are, in fact, few technical difficulties but a lot of smoke!

[BBD]

phi

I suggest that it is like the surface GATA and TLT GATA. It's the big picture we need to see.

We could now go on to look at Arctic ice loss, or mass balance change in the Greenland ice sheet. (And yes, I'm familiar with the possibility raised by Wu et al. 2010 that incorrect GIA has introduced over-estimates of ice loss into interpretations of the GRACE data).

But wherever you look, and however sceptically you weigh the evidence, it all says the same thing. Of course the records are variously unreliable (pre-modern, especially so), potentially biased, or short. However, my guess is (/scientific method off) that the WGMS has it roughly right.

I'm a sceptic, but a rational sceptic, so I have to follow the dictates of reason.

[BBD]

And reasonable scepticism returns us to the problem of climate sensitivity.

UAH and RSS trend at 0.14C/decade from 1979 - present, and the multi-model mean referenced by AR4 is 0.2C/decade for the same period for the boundary layer.

Even setting aside the tropospheric amplification debate, (which is actually about temperatures at ~10Km, not 14,000ft TLT), there is clear divergence.

As Lomborg said, the debate is not about whether we are warming the climate, it is about how much.

[BBD]

Here's Fig. 5.9from the mass balance study which provides a clearer view of the data.

[phi]

BBD,

This is a graph of cumulated values​​. So what is related to temperature, the slope. And the slopes of the first and last decade are quite comparable. All this is rather consistent with a hypothesis of overall stability (with oscillations) temperatures in the twentieth century.

[BBD]

phi

But this isn't:

The global averages (i, ii, iii) reveal strong ice losses in the first decade after the start of the measurements in 1946, slowing down in the second decade (1956–65), followed by a moderate mass loss between 1966 and 1985, and a subsequent acceleration of ice loss until present (Fig. 5.8 a—f). The mean of the 30 continuous ‘reference’ series yields an annual mass loss of 0.58 m water equivalent (m w.e.) for the decade 1996–2005, which is more than twice the loss rate of the previous decade (1986–95: 0.25 m w.e.), and over four times the rate for the period 1976–85 (0.14 m w.e.).

Does it?

[BBD]

Sorry - ignore 'Does it?' That should have been deleted.

[BBD]

phi

You say:

The global surface temperature series are absolutely useless.

This is far too strong. Explain the correlation between the periods of accelerating glacial mass balance change (MBC) and the surface temperature record.

We see a lagged response, with melt from the strong warming 1910 - 1940 continuing from the start of observations in 1946 until 1956. Then a lagged response to the cool period post-1940 with MBC slowing from 1956 - 1965.

Post-1976 warming shows up as moderate ice loss 1966 - 1985, which then accelerates 1985 - present.

The period of interest is of course 1946 - 1979 (start of satellite record). The correlation between MBC and surface temperature reconstructions is good, allowing for a moderate lagged response in ice loss to increased T.

[phi]

BBD,

(1:30 PM)
I commented on the graphic and not on the text. And this is what to do, the interest is to compare the two periods with higher slopes! We know very well that the 1970s were cold, what we seek to discover is whether the 1940s were warmer or colder than the 1990s and how much.

(1:47 PM)
Useless is indeed the wrong word, perhaps unusable?
In fact, high fréqunce CRUTEM is probably excellent but there is a very big problem of secular trend: linear probable overstatement of 1 ° C per century.

[BBD]

phi

You are doing a Shub. There is a clear, referenced and logically sound argument before you. And you are ignoring it.

[phi]

BBD,

There is a clear, referenced and logically sound argument before you. And you are ignoring it.

Which one ?

[BBD]

phi

I have dealt with you in absolute good faith for two days. I don't engage with nonsense like this.

You have absolutely not made your case, and are now resorting to rhetoric instead of reason. This discussion appears to have run out of road.