[alancalverd]

It's also true that good records of temperature only exist for populated areas, but don't forget that, those are the only areas where we have a direct interest in knowing what the climate is doing.

Very, very wrong!

Professional scaremongers are interested in the melting of unpopulated Greenland and Arctic ice. Hurricanes begin their lives as depressions over unpopulated oceans. It's these temperatures that determine our lives!

And the current rate of change means that over the last 30 years the temperature has risen by about half a degree.

The temperature of what? The 1 sq km average over the entire Pacific and Atlantic oceans? Or just Heathrow Airport?

[MoreCarbonOK]

AlanCalvard says
The temperature of what? The 1 sq km average over the entire Pacific and Atlantic oceans? Or just Heathrow Airport?

Hi Allan,

Thx, you had a few good comments there on the accuracy of thermometers in the past which echoes what I have been suspecting. From what you say it seems we cannot rely much on what we have from before the war, really.

As to your question: "the temp. of what?",
in statistics it is of course possible to take a random sample that is representative of a population and to make an estimate. The key is in "representative"
In my sample of 47 weather stations I balanced the sample by latitude and by 70%/30% @sea/on land. Longitude does not matter as I was looking at the average yearly data at the specific station.

in addition I looked at the average change from the average over a reported time period, which excludes the influences of differences between temp. recording devices at various places, mostly

http://blogs.24.com/henryp/2013/02/21/henrys-pool-tables-on-global-warmingcooling/

if you go to the 2nd table, on means (the blue figures),  you will see that we warmed at an average rate of 0.013 degree C per annum globally over the past 32 years. That is ca. 0.42 degree C over the past 32 years.
This result is in fact confirmed by the satellite data, like dr. Spencer's  who reported exactly the same result.
So in this respect, BC is correct.
However, by not looking at the rest of the results, particularly the current cooling trend, many are acting like ostriches.

[alancalverd]

In my sample of 47 weather stations I balanced the sample by latitude and by 70%/30% @sea/on land. Longitude does not matter as I was looking at the average yearly data at the specific station.

Longitude makes an enormous difference! Moscow is pretty much the same latitude as Glasgow, but utterly different, and most of Canada that lies north of London is inhospitable or uninhabitable.

The sampling frequency is also very important. It's colder today than yesterday, but warmer than this time last week. Does this indicate a negative or a positive trend?

Nice table of data, but it's all from inhabited areas and airports, so it simply shows the obvious: concrete has less evaporative cooling than forest, and people like making heat. 

What interests me is the "single-point" curves of temperature and carbon dioxide concentration. Geologically, from the Vostok and other ice cores, and recently , from Mauna Loa observations, the CO2 curve lags behind the temperature (or its proxy) curve, not the other way around. Now I don't know what planet the professional scaremongers of IPCC live on, but around here, if A follows B, it cannot be the cause of B. Furthermore the geological temperature curve is a sawtooth, with very fast rises followed by slow declines, and this behaviour can be modelled by the superposition of sinusoids, as you would expect if temperature was driven entirely by atmospheric water content.

[MoreCarbonOK]

Alan says
Longitude makes an enormous difference!
Henry says
in my thoughts about this,  I considered that earth turns once every 24 hours and that during one whole year (which are the average temps. I took),  I cancelled out the seasonal shift. Therefore longitude does not matter.
The differences you refer to, are in fact also visible in the tables. Some stations are running exactly opposite the wave.
\Believe it or not, but this has to do with the Greenhouse effect.
Namely, in a period of warming the differential between zero and [90] latitude causes more clouds at higher latitude and somewhat less at lower latitudes. In a period of cooling, such as now, the differential temp. increases, causing more clouds and rain at lower latitudes and less clouds and rain at higher latitudes.
This amplifies the cooling effect (since insolation at the equator is 2x the average)
At some stage, I expect in about 6 years from now, there is a bit of a standstill in pressure difference, causing droughts in many parts, similar to the dust bowl drought 1932-1939
http://www.ldeo.columbia.edu/res/div/ocp/drought/dust_storms.shtml

[Bored chemist]

It's also true that good records of temperature only exist for populated areas, but don't forget that, those are the only areas where we have a direct interest in knowing what the climate is doing.

Very, very wrong!

Professional scaremongers are interested in the melting of unpopulated Greenland and Arctic ice. Hurricanes begin their lives as depressions over unpopulated oceans. It's these temperatures that determine our lives!

And the current rate of change means that over the last 30 years the temperature has risen by about half a degree.

The temperature of what? The 1 sq km average over the entire Pacific and Atlantic oceans? Or just Heathrow Airport?

They may well, but what we were discussing at the time was the change in temperature in a place or fixed array of places (as it happens, one chosen by Henry)
We weren't looking at sea levels or hurricane frequencies.

The data I was talking about for the past 30 years are here
http://www.skepticalscience.com/pics/RFC12_Fig1.jpg
Can you explain why you need to measure that change of about half a degree to the nearest 0.01 degree?

[alancalverd]

Namely, in a period of warming the differential between zero and [90] latitude causes more clouds at higher latitude and somewhat less at lower latitudes. In a period of cooling, such as now, the differential temp. increases, causing more clouds and rain at lower latitudes and less clouds and rain at higher latitudes.

Possibly the first time I've seen a sensible assessment of cloud effects from anyone else,  in any discussion on climate change!

The effect of water on atmospheric temperature is essentially nonlinear (self-amplifying) which can account for the rapid rise  during warming periods, and is bounded and damped by cloud formation - hence the slow downward drift from a fairly consistent maximum over previous geological cycles. Then since the balance  between carbon dioxide uptake by plants and its emission by cold-blooded animals depends on temperature, it is not surprising that the CO2 graph lags behind the temperature graph where the two are derived from independent proxies, as in ice cores.

The effect of longitude is more subtle than Henry's presumption. Temperature in the British Isles, for instance, is determined principally by the vagaries of the Gulf Stream and the jet stream. Most of the time these islands are covered in cloud and our weather is whatever the advected Atlantic depression gives us, with very little influence from the local greenhouse effect or insolation (it's quite often warmer at night than during the day). But around 30% of the time we have clear, dry  arctic skies, under which the temperature is governed by radiative transfer (hot days, cold nights) as there is very little wind to transfer heat laterally on clear days. Eastern Europe at the same latitude does not get warm, wet, Atlantic air, and the climate is principally driven by radiative transfer and local convection of water.  In short, maritime climate on the west coast of anywhere is determined by events further west, whilst continental climate is dominated by local physics.   

[Bored chemist]

And, for the third time now.
The data I was talking about for the past 30 years are here
http://www.skepticalscience.com/pics/RFC12_Fig1.jpg
Can you explain why you need to measure that change of about half a degree to the nearest 0.01 degree?


I'm asking because, if there isn't a good reason, then it looks like you are seeking to set an unnecessarily difficult target of measurement before you accept that there's any warming.


Seriously, do you really think you need a hundred years of measurements that are good to 0.01C to measure a change that seems to be occurring at about 1.5 degrees over that time?

[MoreCarbonOK]

alan
it is not surprising that the CO2 graph lags behind
henry@alan
forget about CO2. It is not a factor at all. Not in warming and not in the weather.
Look again carefully at the results in my 3 tables.
1)
Note the ratio maxima-means-minima is 6:2:1
(if you take it over the longest period)
If increased CO2 or H2O were a factor, we should see minima rising, pushing up the average.
That is not happening.
2)
If you know how to do curve fitting in excel, you should try and set out the speed of warming/cooling
in K/year,  against time, on each of the 4 final results (on the bottom) for maxima, means and minima.
That gives you K/year ²
Tell me what correlation you get, especially if you try binomials?
3)
note the differences between NH and SH
Does it not seem that most of the (maxima) heat ends up in the SH oceans and is taken by weather and currents up to the NH (means)?

Once you figured out why we see these results happening, you are on your way....!!
You cannot miss it.

[alancalverd]

BC: Nice graph, but no title! The temperature of what? Measured how? Calculated how?

If I measure outside air temperature at my local airport over 24 hours, a 15C range is within normal bounds. The automatic system reports every 20 minutes, but if I use 4-hourly manual measurements I'm likely to miss the max (usually around 1300) and min (about 0300). If I take a true 24 hour mean, it will vary between +5 and -5 over four typical consecutive winter days, and anything less than a 2 week average (two Atlantic depressions or one winter high, say) is pretty meaningless.

The reason for asking these nitpicking questions is due to the apparent finesse of the graph. It shows about 18 - 20 ripples per quinquennium, suggesting 3-monthly means. OK, so let's look at "1992.5", halfway between 1990 and 1995. The slope suggests that midsummer was colder than the winter either side, which does not correspond with my recollection.

Or if the abscissal marks indicate mid-year, look at 1995, where once again it apparently got hotter towards the end of the year!   

There's a significant discrepancy between the pink and the red line. Which indicates the truth, and which the opinion of the author?

For what it's worth, the last calibrated meterorological thermometer I owned was scaled in 0.05 degree increments but had mid-range (10 - 30 C) corrections of as much as 0.2 degrees. You simply cannot use these instruments "out of the box" to compare temperatures at different times or places. Problem is that as far as as aviators are concerned, absolute accuracy of +/- 1 degree is of no consequence: we just want to know if we can take off with a reasonable margin of safety, or land without encountering fog or ice, so the kit is generally used without reference to the cal chart and most met reports are actually pretty crude compared with the rate of climate change.

A good rule of thumb in metrology, if not meteorology, is to use an instrument at least one order of magnitude more accurate than the effect you are trying to measure, hence my suggestion of +/- 0.01 deg as the acceptable specification for examining climate change, and for the figures to be meaningful you need to average each location over a year.

There certainly is climate change, always has been and always will be, but I object to a temporary correlation being used as an excuse to raise taxes and screw my life up by politicians jumping on a fatuous bandwaggon.

[alancalverd]

MCOK: I'll believe your arithmetic. What are the answers? I don't have time to fit the curves myself!

[Bored chemist]

Let's face it, the important thing about graphs of global temperature is that they look more like this
/ than like this\
And you don't need a thermometer which reads to 0.01C to spot a change of 0.5 do you.
It doesn't matter where it's measured or by whom as long as it's consistent so ,yes if you measure the temperature every 4 hours you will miss the max and min- but, if you are consistent, you will always miss it by the same extent. And you would have missed it just as much in 1980 as today.

As I have said a number of times you don't need brilliant measurement precision to show a trend.

"A good rule of thumb in metrology, if not meteorology, is to use an instrument at least one order of magnitude more accurate than the effect you are trying to measure, hence my suggestion of +/- 0.01 deg as the acceptable specification for examining climate change, and for the figures to be meaningful you need to average each location over a year. "
Yes, but what said was "you need to be able to measure absolute temperature to better than 0.01 degree over a range of -30 to +40 degrees, for about 100 years."

Over that time the predicted change is about 1.5 degrees so 0.1C would be good enough,  and that's still not allowing for the benefits of averaging many thermometers but you wanted 0.01 degrees.

[JP]

A good rule of thumb in metrology, if not meteorology, is to use an instrument at least one order of magnitude more accurate than the effect you are trying to measure, hence my suggestion of +/- 0.01 deg as the acceptable specification for examining climate change, and for the figures to be meaningful you need to average each location over a year.

True, and that's why day-to-day measurements of local temperature don't tell us much about climate.  What matters is the uncertainty around the measurements over time, accounting for the law of large numbers, which tells us that combining many measurements actually decreases the uncertainty in the mean relative to the uncertainty in any individual measurement.  If many independent measurements are taken at roughly the same time, then the uncertainty decreased by a factor of ~ the square root of the number of measurements taken. 

I'd go so far as to say that any claims about climate change without an analysis of uncertainties are suspect, because you don't know if they show a trend or just measurement error.  Fortunately, most scientific publications which track temperatures do have error bars, e.g. http://data.giss.nasa.gov/gistemp/2008/  The important part is not that the trendline increase, but that it increases so much relative to measurement uncertainty that we can be nearly certain that the rise is real, and not due to uncertainties in the measurement.

Edit: BC beat me to most of this, but the TL;DR version is that uncertainties can be reduced by averaging and a proper analysis shows global temps increasing.

[MoreCarbonOK]

henry@alan
r²> 0.95
in the case of maxima,
r²=0.996

ergo:
must be natural, such a trend cannot be man made.

[alancalverd]

Let's face it, the important thing about graphs of global temperature.....It doesn't matter where it's measured or by whom

It makes an enormous difference! Just staying with airfields for a moment, in the period of your graph Stansted grew from a small strip of tarmac surrounded by grass and a few huts, to a hundred acres of concrete surrounded by steel hangars, whilst the runway at Wrexham was broken up and removed. Every two minutes, someone dissipates several megajoules of exhaust heat or kinetic energy on the Stansted runway, but you can't land at Wrexham any more. So the local temperature at Stansted went / and that at Wrexham went \  Of course nobody measures the temperature at Wrexham now the runway has been sold as hardcore, so the apparent mean of all British airfields is // !

Pretty obviously, most of our weather data comes from populated areas because that's where the present and forecast conditions are most immediately important, but nearly all of our climate depends on conditions in unpopulated areas like the middle of the Pacific, Arctic, Gobi, northern Canada... which account for 95% of the earth's surface.The problem is that the 5% on which we live is increasingly concreted over and heated by our activities as well as the sun, so the data is both overrepresented and unrepresentative.

So the questions remain: what data is represented by your graph, and why does it show that some winters are warmer than the adjacent summers?

[Bored chemist]

Not sure why you are so hung up on airfields since we have data like this
http://data.giss.nasa.gov/gistemp/2008/
 but...
How fortunate then that, as I have said repeatedly, the average (which would include Stansted and Wrexham) would give a better result than either of them.

[MoreCarbonOK]

 JP says
Fortunately, most scientific publications which track temperatures do have error bars, e.g. http://data.giss.nasa.gov/gistemp/2008/  The important part is not that the trendline increase, but that it increases so much relative to measurement uncertainty that we can be nearly certain that the rise is real, and not due to uncertainties in the measurement.

Henry says
sorry JP, but I do not see any error bars in that graph that you quoted?

In fact, subsequent to satisfying my curiosity about error, I did a comparison of same gistemp data set with the wood for trees temperature index, which is an average of anamolies of all available data sets,
here is the result of my analysis
http://www.woodfortrees.org/plot/wti/plot/gistemp/from:1930/to:2014/plot/gistemp/from:1930/to:1980/trend/plot/gistemp/from:1980/to:2002/trend/plot/gistemp/from:2002/to:2014/trend/plot/wti/from:2002/to:2014/trend/plot/wti/from:1980/to:2002/trend

It appears that gis temp. anomalies  are considerably higher than that of other data sets,

which points to a considerable,  apparently consistent, error in the gis temp. data set.

[Bored chemist]

Interesting set of graphs.
It illustrates my point nicely.
Even though there's a lot a scatter and different biasses on the two data sets, they both show the same trend.
The data on the right hand side are higher than those on the left- i.e. they show that the world is warming (with some scatter).

[MoreCarbonOK]

bored chemist says
the two data sets, they both show the same trend.

henry says
true
both show that temps have been gradually going up from 1930 -2000
but going down from 2002
exactly as predicted by me here
http://blogs.24.com/henryp/2012/10/02/best-sine-wave-fit-for-the-drop-in-global-maximum-temperatures
(note there is a lag between energy-in /maxima and energy-out /means)

consequently, seeing that I was right, correctly predicting history, 
unfortunately we will continue cooling down until around 2040

clearly

we cannot trust the data before 1930 as we do not have a global baseline before that time,
because of inaccuracy and different methods of recording means

[alancalverd]

Not sure why you are so hung up on airfields since we have data like this
http://data.giss.nasa.gov/gistemp/2008/
 but...
How fortunate then that, as I have said repeatedly, the average (which would include Stansted and Wrexham) would give a better result than either of them.

I'm not "hung up" on airfields, it just happens that nearly all the credible historic data comes from them.

Not sure what you mean by a "better" result. If the average temperature over the entire UK had decreased during the period when Stansted airport was expanding, in what respect would today's Stansted temperature (there is no mean because Wrexham no longer has a runway so it no longer reports temperature) be "better"?

The GISS "data" is fascinating. It goes back to 1880, when one continent (Antarctica) was completely unexplored, another (Australia)  had no established meteorological service, and there were no regular reports from anywhere in the Pacific. So how did they deduce a global mean? I smell bullshit!

So I repeat my question: you showed a graph earlier (reply 126). Where did the data come from and what do the curves represent? Simple enough, surely, to deserve an answer?

[Bored chemist]

"Where did the data come from"
Here
http://www.skepticalscience.com/
you can tell from the web address.
But take your pick. There are plenty of web pages out there with data and they seem to show a pretty much consistent rise over the last 30 years or so.
Here's another (chosen pretty much at random from heaps on google).
http://metaclimate.org/2010/02/14/a-ghcn-analysis/

Perhaps you can find the ones where an estimate of the global temperature over that period (rather than cherry picking since the last el nino) actually falls.
Its fair to say that if you look at the results from a google search for images of climate change graphs
there are a lot more like  this / than like this\ and it's hard to see how they could all be wrong.