|
Post by socold on Oct 15, 2009 19:30:34 GMT
There's probably a good reason but why is the comparison land based and not global?
|
|
|
Post by glc on Oct 15, 2009 19:59:53 GMT
Seems to me you are saying its OK to spatially adjust temperatures if GISS does it and not OK if Pielke does it. . . .or at least you offered absolutely no other explanation why you think one was wrong and the other was right to do such stuff.
It's not the land only issue it's the assumption that the amplification factor exists - which is odd because the same people are generally telling us that the models are wrong. so let's have it
Are the models right or wrong?
|
|
|
Post by sigurdur on Oct 15, 2009 20:04:47 GMT
Seems to me you are saying its OK to spatially adjust temperatures if GISS does it and not OK if Pielke does it. . . .or at least you offered absolutely no other explanation why you think one was wrong and the other was right to do such stuff. It's not the land only issue it's the assumption that the amplification factor exists - which is odd because the same people are generally telling us that the models are wrong. so let's have it Are the models right or wrong? The models of GCM are not being written well enough with enough information to be correct. That was an easy question to answer. And maybe you know where there is a paper to support the 3.2W temperature idea. I can't for the life of me find one......but that figure is quoted over and over. Where is the validity in it? ??
|
|
|
Post by magellan on Oct 15, 2009 22:45:17 GMT
Seems to me you are saying its OK to spatially adjust temperatures if GISS does it and not OK if Pielke does it. . . .or at least you offered absolutely no other explanation why you think one was wrong and the other was right to do such stuff. It's not the land only issue it's the assumption that the amplification factor exists - which is odd because the same people are generally telling us that the models are wrong. so let's have it Are the models right or wrong? Put the red herring back in your fridge and use for another day.
|
|
|
Post by glc on Oct 16, 2009 0:08:54 GMT
Put the red herring back in your fridge and use for another day.
Sorry no further discussion on this until we've cleared up the anomaly issue. I’ve got a few spare minutes, so I’ve decided to have one more go at explaining anomalies and baselines (or base periods).
Imagine a weather station (Station A) which has been operating since ~1960. Station A reports monthly temperature anomalies. It uses average temperatures from 1961-1990 as the base period. Let’s say that the 1961-90 September average is 10.0 deg. If the September 2009 average temperature is 10.5 deg then the Sept anomaly for Station A will be +0.5 deg.
One final point about Station A is the 1979-1998 September average temperature which is 10.2 deg.
In 1978 another station (Station B) is set up in a location near to Station A. Station B also reports monthly anomalies but because B has only been running since 1978 it uses 1979-1998 average temperatures as the base period. Just like A, B’s September average temperature for 1979-98 is also 10.2 deg. The September 2009 average temperature fro Station B is 10.5 deg – just like Station A. But the September anomaly for Station B is +0.3 deg. So the anomalies are
Station A: +0.5 Station B: +0.3
But both stations measured exactly the same average Sept temperature. The reason for the difference is the use of different base periods. If we want to make a valid comparison between the 2 stations we need make an adjustment to take account of the different base periods. One way is for Station A to use the same base period as Station B but sometimes we only have the anomalies (not the original raw temperatures). In this case there is a simple technique which can be used. First calculate the average anomaly for the desired base period (in this case 1979-98) compared to the original base period - and then subtract that figure from the original anomaly. So for Station A
The 1979-98 anomaly (compared to 1961-1990) is +0.2 (i.e.10.2 – 10.0) The adjusted anomaly = original anomaly – 1979/98 anomaly = 0.5 – 0.2 = +0.3. We now have the Sept 2009 anomaly for Station A adjusted to the 1979/98 base period.
|
|
|
Post by magellan on Oct 16, 2009 0:18:01 GMT
Put the red herring back in your fridge and use for another day.Sorry no further discussion on this until we've cleared up the anomaly issue. I’ve got a few spare minutes, so I’ve decided to have one more go at explaining anomalies and baselines (or base periods). Imagine a weather station (Station A) which has been operating since ~1960. Station A reports monthly temperature anomalies. It uses average temperatures from 1961-1990 as the base period. Let’s say that the 1961-90 September average is 10.0 deg. If the September 2009 average temperature is 10.5 deg then the Sept anomaly for Station A will be +0.5 deg. One final point about Station A is the 1979-1998 September average temperature which is 10.2 deg. In 1978 another station (Station B) is set up in a location near to Station A. Station B also reports monthly anomalies but because B has only been running since 1978 it uses 1979-1998 average temperatures as the base period. Just like A, B’s September average temperature for 1979-98 is also 10.2 deg. The September 2009 average temperature fro Station B is 10.5 deg – just like Station A. But the September anomaly for Station B is +0.3 deg. So the anomalies are Station A: +0.5 Station B: +0.3 But both stations measured exactly the same average Sept temperature. The reason for the difference is the use of different base periods. If we want to make a valid comparison between the 2 stations we need make an adjustment to take account of the different base periods. One way is for Station A to use the same base period as Station B but sometimes we only have the anomalies (not the original raw temperatures). In this case there is a simple technique which can be used. First calculate the average anomaly for the desired base period (in this case 1979-98) compared to the original base period - and then subtract that figure from the original anomaly. So for Station A The 1979-98 anomaly (compared to 1961-1990) is +0.2 (i.e.10.2 – 10.0) The adjusted anomaly = original anomaly – 1979/98 anomaly = 0.5 – 0.2 = +0.3. We now have the Sept 2009 anomaly for Station A adjusted to the 1979/98 base period. So what is the baseline difference between UAH and HadCRUT? Oh wait, I already told you....
|
|
|
Post by glc on Oct 16, 2009 9:26:04 GMT
So what is the baseline difference between UAH and HadCRUT?
Depends on the month but the average is ~0.17
Oh wait, I already told you
You might have told me but it didn't show up in your plot. In your plot all the HadCrut-UAH values were positive. In other words even allowing for the baseline adjustment HadCrut was always warmer than UAH. That can't be right.
You quoted a baseline difference of 0.146 - Ok fair enough. That means that the HadCrut anomalies need told to be adjusted by that amount before HadCrut-UAH calculation is performed.
This table shows the figures and calculations for 2009
2009 Raw Had Had adj UAH Had-UAH
Jan 0.384 0.238 0.3 -0.062 Feb 0.364 0.218 0.35 -0.132 Mar 0.371 0.225 0.21 0.015 Apr 0.415 0.269 0.09 0.179 May 0.407 0.261 0.05 0.211 Jun 0.499 0.353 0 0.353 Jul 0.499 0.353 0.41 -0.057 Aug 0.532 0.386 0.23 0.156
Raw Had: Hadcrut anomaly relative to 1961-1990 Had Adj: Hadcrut anomaly relative to 1979-98 (i.e 0.146) UAH: UAH anomaly relative to 1979-98 Had-UAH: Difference
Note there are 3 negative differences (i.e. Hadcrut cooler) in 2009. There are none in your plot. The largest difference is +0.353 (Jun) while you have a difference of +0.481. Hang on a minute .......
Are you sure you haven't adjusted by 0.0146 rather than 0.146 or done something similar.
|
|
|
Post by magellan on Oct 16, 2009 18:15:40 GMT
So what is the baseline difference between UAH and HadCRUT? Depends on the month but the average is ~0.17 Oh wait, I already told youYou might have told me but it didn't show up in your plot. In your plot all the HadCrut-UAH values were positive. In other words even allowing for the baseline adjustment HadCrut was always warmer than UAH. That can't be right. You quoted a baseline difference of 0.146 - Ok fair enough. That means that the HadCrut anomalies need told to be adjusted by that amount before HadCrut-UAH calculation is performed. This table shows the figures and calculations for 2009 2009 Raw Had Had adj UAH Had-UAH Jan 0.384 0.238 0.3 -0.062Feb 0.364 0.218 0.35 -0.132Mar 0.371 0.225 0.21 0.015 Apr 0.415 0.269 0.09 0.179 May 0.407 0.261 0.05 0.211 Jun 0.499 0.353 0 0.353 Jul 0.499 0.353 0.41 -0.057Aug 0.532 0.386 0.23 0.156 Raw Had: Hadcrut anomaly relative to 1961-1990 Had Adj: Hadcrut anomaly relative to 1979-98 (i.e 0.146) UAH: UAH anomaly relative to 1979-98 Had-UAH: Difference Note there are 3 negative differences (i.e. Hadcrut cooler) in 2009. There are none in your plot. The largest difference is +0.353 (Jun) while you have a difference of +0.481. Hang on a minute ....... Are you sure you haven't adjusted by 0.0146 rather than 0.146 or done something similar. In other words even allowing for the baseline adjustment HadCrut was always warmer than UAH. That can't be right. Ok, I'll try one last time. I used absolute values, and was NOT intended to compare trends (although in this case the end result is similar) or how much warmer or colder any particular period is. Example: UAH .1 HadCRUT .3 HadCRUT - UAH = .2 Absolute value = .2 Example: UAH .1 HadCRUT -.1 HadCRUT - UAH = -.2 Absolute value = .2 The largest difference is +0.353 (Jun) while you have a difference of +0.481.
The first plot I made used baseline adjusted numbers and absolute values, which was in the title of the chart. The difference was .335 (not .353 typo?). I made the mistake of assuming such a simple exercise would lead to this. You then went into a tirade, said I didn't understand what I was doing, and pointed out that UAH and HadCRUT were within .010 in Feb 09 by just comparing the raw anomaly data, so I created another chart without baseline adjusting and viola! Feb 09 was within .010 just as you said, but now Jun 09 is .481 instead of .335 (with baseline adjustment). .481 - .335 = .146. Can we put this to bed now?
|
|
|
Post by glc on Oct 16, 2009 19:43:15 GMT
Can we put this to bed now?
As you wish.
|
|
|
Post by icefisher on Oct 17, 2009 0:53:42 GMT
THe GISS extrapolation over the arctic offers the best and most reasonable explanation. Whether they are right or wrong to do this extrapolation (and I think they're probably right) they have, at least, been consistent. Once the arctic lost the very high temperatures GISS anomalies dropped back in line with the other records. When Hadley post their Sept figure I'll do a direct comparison for all anomalies using the 1979-98 base period. Seems to me you are saying its OK to spatially adjust temperatures if GISS does it and not OK if Pielke does it. . . .or at least you offered absolutely no other explanation why you think one was wrong and the other was right to do such stuff. Golly GLC! No response? Is this just out and out favoritism? Political shenanigans condoned as long as they align with your way of thinking?
|
|
|
Post by julianb on Oct 26, 2009 9:55:05 GMT
It seems to me that too much emphasis is put on average temperatures and trying to unravel proxies that give averages to explain historical climate.
Several accounts of extreme cold spells that caused widespread famine by crop and livestock freezing in Europe in the 17th and 18th centuries were preceded by unseasonal warm weather, and then came late winter cold that froze stock and game in the fields and split trees feet across. It is quite possible that the years average temperature was near normal.
The vagaries of the polar and mid latitude jet streams may be the key to understanding changes in climate, rather than small yearly change.
How thick does snow have to be to not melt in the summer on the Canadian prairies for example ?, a year or two ago I read that Toronto had to spread its snow dump or it would not melt in the summer.
|
|
|
Post by trbixler on Oct 27, 2009 14:17:11 GMT
|
|
|
Post by julianb on Oct 28, 2009 11:33:17 GMT
These postings by Lous Hissink on his blog on 18th and 19th Oct 2009 point out the irrelevance of temperature measurement unless it is related to the volume of the atmosphere being measured. geoplasma.spaces.live.com/blog/So unless each surface station is measuring an identical volume of atmosphere, (and cover the entire globe) conclusions drawn from their data are virtually meaningless, something I have felt but have not been able to express in scientific form. I realise that satellites approach this ideal by their nature, but it may explain why adjusted surface station data is at odds with them at times, given their limited number and spacial distribution.
|
|
|
Post by trbixler on Oct 28, 2009 14:09:34 GMT
Rather interesting manipulations by GISS. It seems to me that surface stations are needed. I will note that the compilation of the data should not be in the hands of an agenda driven, political action oriented government agency. Add, drop, position as they will, typically where no one will possible look except...... chiefio.wordpress.com/2009/10/28/ghcn-china-the-dragon-ate-my-thermometers/
|
|
|
Post by sigurdur on Oct 29, 2009 1:52:21 GMT
So far for October, we are 7.5F below normal temps. Going to be interesting to see the anomoly map for Oct. (This is NOT just a small area)
|
|