I met Ken while on my Australian tour, he’s been doing some fine work.
Via Andrew Bolt
Retired school principal Kenskingdom was alarmed by this Bureau of Meteorology graph, showing a strong warming trend for Victoria, Australia:
He checked the data from which the trend, and found it had first been adjusted and turned into “high quality” data. As a BOM spokesman assured him:
On the issue of adjustments you find that these have a near zero impact on the all Australian temperature because these tend to be equally positive and negative across the network (as would be expected given they are adjustments for random station changes).
Actually, no, though. You see, Kenskingdom discovered that the adjustments served to exaggerate Victoria’s warming remarkably:
Kenskingdom goes through the individual stations for you and concludes:
There is a distinct warming trend in Victoria since the 1960s, which has been especially marked in the last 15 years.
The first half of the record shows a cooling trend. BOM’s adjustments have attempted to remove this.
2007, not 2009, was the warmest year in the past 100 years.
Three stations identified as urban in 1996 have been included.
Many stations’ data have been arbitrarily adjusted to cool earlier years
Only one station has had its trend reduced. Two are essentially unchanged.
Ten of Victoria’s 13 stations have been adjusted to increase the warming trend, to the extent that there is a warming bias of at least 133%, more likely 143%.
These adjustments, and the Australian temperature record to which they contribute, are plainly not to be trusted.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
carrot eater says: July 1, 2010 at 4:09 pm
The unadjusted data give you the same result, globally.
Then why is it adjusted?
carrot eater says:
July 1, 2010 at 12:54 pm
rbateman
“If the adjustments cancel out globally, then what is the justification for adjusting?”
You didn’t necessarily know this would be the case, before you started going about doing the adjustments…..
______________________________________________________________________
“adjustments” should not be done period.
On the one hand we are told that “station moves, instrument changes…” require “adjustments” to the data for individual stations. On the other hand we are told that “homogenizing” the data by making up “adjustments” from stations up to 1200 km away or infilling data from surrounding areas to manufacture data for areas with no data is perfectly alright.
And yes I have done an analysis for my home state.
This is what I found:
I live in North Carolina and here is a quick look at my state. It is VERY interesting.
There is nothing closer to the mountains than Chapel Hill which is just west of Raleigh. All the areas with longitudes further west are also further south and that puts them on the seacoast. Chapel Hill is on the plains. Seems the mountain areas are no longer part of the record, imagine that.
I also found that at Wunderground the Moncure NC station is no longer available, it flips you to the new Sanford NC airport. It was available the last time I looked. Asheville NC is the big city in the mountains, home of the Biltmore Estate (1895) Its weather station now only goes back to 2005 at the AIRPORT of course. The site also directs you to the “nearby” (80 miles) city of Greenville (downtown) which only goes back to 1970. That site has been declared “unofficial” The “official” station is now the Greenville/Spartanburg, South Carolina (Airport)
Here is a quick look at the only city & close by airport listed for North Carolina. The city is on the North Carolina/Virgina border and right on the ocean. Take a look at the city vs the airport! Norfolk City and
Norfolk International Airport
The Raleigh North Carolina area is in the piedmont area of North Carolina. It is far from both the mountains and the sea coast. Here is an Elevation map and North Carolina map of cities
North to south thru the middle of the state
North – Raleigh NC
Large city in the middle of NC – Fayetteville NC
South – Lumberton NC
Coastal Cities:
North – Elisabeth City
South – Wilmington NC
Rural
North – Louisburg
North – Louisburg
South – Southport
South – Southport
Here is the raw 1856 to current Atlantic Multidecadal Oscillation Amazing how the temperatures follow the Atlantic ocean oscillation as long as the weather station is not sitting at an airport isn’t it?
Oh and one other thing. As a farmer I pay close attention to the weather. At Wunderground the temperatures of the day are always adjusted up when I look at the previous day’s information. When you get up at dawn each day and check the weather it becomes pretty darn obvious the data is “adjusted” when the freezing morning weather from yesterday, that left ice on the livestock tanks, is reported the next day as a “minimum temperature” that was above freezing.
And some prominent commenters are playing up the alleged increase in global temperature this spring and sweeping under the rug last northern hemisphere’s colder than normal winter.
This post is evidence that confirms my suspicions that one can’t rely on reports that temperatures are “warmer” this spring world-wide.
The global carbon taxers are desperate and more willing to manipulate the data than ever.
“On the issue of adjustments you find that these have a near zero impact on the all Australian temperature because these tend to be equally positive and negative across the network (as would be expected given they are adjustments for random station changes).”
Do they send their data to JH for homogenization?
carrot eater says: “If the changes due to these were truly random, you could not bother making the adjustments, if all you cared about was the global mean. But it made sense for somebody to start doing adjustments, to check.”
True. But if that is the case, assessments of aggregate temperature trends should just use the unadjusted data.
The only sensible use of an aggregate of adjusted data would be to demonstrate no statistical bias in the adjustments (when taken in aggregate).
On that logic, the approach of GISS and others still appears to be incorrect. They are using the wrong data. And my guess is that it would probably be quite simple for them to substitute unadjusted data into their methods (or to turn-off the adjustment steps if it is embedded in their code).
It would make sense for somebody to step into the “gap in the market” to publish an aggregate of the unadjusted data. It would also be interesting to see whether the other producers agree with the new series …. or more to the point, their reasoning if they have any issues with it.
Shows the importance of open source raw data with transparent reporting on weather station issues both now and in the past.
Jordan:
You perhaps don’t realise it, but for everywhere except the US, GISS takes as its input the unadjusted data. It then applies only one adjustment*, and that is to impose on ‘urban’ stations the trend of its ‘rural’ neighbors. That is the only adjustment that GISS itself does. ‘Urban’ and ‘rural’ are currently identified by how bright the areas are at night in satellite images; you can argue whether that’s the best way.
Here you can see the effect of not doing that adjustment, vs doing that adjustment:
http://clearclimatecode.org/gistemp-urban-adjustment/
Keep in mind that for this graph, for a small section of the global land mass (the US 48), the ‘no adjustment’ series does include adjustments previously made by NOAA. For the entire rest of the world, there is no adjustment at all at any point, in the ‘no adjustment’ series. *Well, OK, with the exception of 2 oddball stations, St. Helena and Lihue.
“True. But if that is the case, assessments of aggregate temperature trends should just use the unadjusted data.”
Why? The adjustments are done for a reason – to remove spurious issues from the raw data that are not related to climate. For GISS, the attempt is to remove UHI. For NOAA, the attempt is to remove the effect of station moves, etc, and the new method they’re currently testing may (or may not) also be effective for UHI (we shall see when it’s released). So if you think your adjustments are done well, then it makes perfect sense to publish the overall results with them in there.
It just so happens that the adjustments have little effect on the overall result (as copiously documented by Zeke at Lucia’s, among others), looking globally, and that’s partially due to some of the spurious issues being random in their effect (for NOAA), and in the case of GISS, it’s because the original raw dataset simply doesn’t have a ton of urban stations that warmed much more quickly than their rural neighbors.
Note the obfuscation in carrot eater’s definition of what “raw” data is.
Unadjusted is not raw.
DR:
‘unadjusted’ is simply the raw daily or hourly data, converted into monthly means.
RACook
“False. And false. Hansen (GISS) fought for years to avoid releasing ANY data even under repeated FOIA lawsuits. They still hide raw data. They still hide their conversions and their code, but have released “some” code. Under protest. We don’t know how they refuse to release, nor whether what is released is actually used.”
None of your response is based in any reality.
People have taken GISS’s published code, run it for themselves, and recovered the same results as GISS’s published results.
People (clear climate code) have taken GISS’s published code, read through it line by line, found a couple minor errors, translated into a different platform and made it easier to read, and.. recovered the same results as GISS’s published results.
People have taken the unadjusted data, written their own code from scratch that uses slightly different methods, and gotten about the same results (Zeke at the Blackboard, Tamino, Chad at http://treesfortheforest.wordpress.com/2010/05/19/better-late-than-never/, Nick Stokes).
People have even taken an entirely different source of unadjusted data (daily temperature means, instead of monthly temperature means), and gotten about the same result (http://rhinohide.wordpress.com/2010/06/29/gistemp-with-gsod-round-2/)
I’m sorry, but there’s simply no conspiracy here.
Oh, by the way: As for the claims of Watts, d’Aleo and EM Smith that global warming is at least in some part an artifact of an intentional station drop-off at 1990 – that claim has already been comprehensively refuted by several people, but Ron’s work at http://rhinohide.wordpress.com/2010/06/29/gistemp-with-gsod-round-2/ takes it one step further, by using a different source data set that has no station dropoff at 1990. His source set has a high number of stations since 1973. His analysis is still preliminary, but the station drop-off conspiracy meme is as dead as can be.
I can tell you one thing, raw or not, it’s COLD in Sydney’s inner west! No question! A 61 year cold low record for June. There is no conspiracy there. Shame I am gassing myself on CO emissions because, in Australia, it’s always warm (No need for flued gas heatring, just vent into the living space. So 1950’s).
Until it was so clearly demonstrated in the commentary chain for this post, I had not realized eating carrots could lead to such severe myopia…….. and cognitive suspension of reality.
It’s Just Weather: 55F and raining, here in western Washington state. 12F (7C) below ‘normal’, as has been most of the last month and much of the spring. Our Global Warming/Carbon Tax Governor, Christine Gregoire, is petitioning the O’blame-a administration for ‘federal disaster relief aid’ because the winter and spring have been so ‘unusually cold and wet’, causing about 40% of the fruit crops to fail. Late frosts combined with continual cold rains caused poor pollination and fruit development. Note that both ‘correlation’ and ‘causation’ are established firmly.
Do you suppose Governor Gregoire is a carrot eater also?
Mac the Knife says:
July 2, 2010 at 8:51 am
Until it was so clearly demonstrated in the commentary chain for this post, I had not realized eating carrots could lead to such severe myopia…….. and cognitive suspension of reality.
Do you suppose Governor Gregoire is a carrot eater also?
_______________________________________________________________
I do not know but I think Rabett is.
It’s been comprehensively shown by several independent workers that the results of GISS and NCDC can be replicated, both using GISS code as well as written-from-scratch code, using both GISS algorithms as well as independent algorithms, using completely unadjusted data, and even using a completely different data set.
There is no conspiracy in any allegedly hidden code, there is no conspiracy in the adjustments, there is no conspiracy to drop stations to artificially bring about a rise in temperatures. No, no and no.
Your response to that is to note that it happens to be cold right now wherever you happen to live, and then accuse me of myopia?
Interesting.
I think carroteater has the best of this lot. As a skeptic, I question, challenge, prod and poke at, but ultimately go with the evidence.
So it appears that the several conspiracy theories just don’t hold up. And it’s pretty low to reply to reasoned discussion with the ‘myopia’ line!
Does the raw or ‘High Quality” data from this data set go to GHCN?
What is the case for other records sent to GHCN outside the U.S.?
DR:
It should always be the raw that is sent for the GHCN. I always have trouble finding my way around the Aussie BoM website, but you should be able to confirm that for yourself if you can manage to find the raw data on the Aussie page, and then compare it to v2.mean.
The GHCN specifies that it only picks up data if it can find the unadjusted version. NOAA then does its own adjustments. I think CRU is a bit different in this regard, and has accepted data that has already been adjusted by the providing country.
Somewhere on the BoM page, you should also be able to find station histories, which would give the reasoning behind the various adjustments for the stations in Victoria. Somewhere.. I know I’ve seen these for Australia, but never know what to click to find them.
OK. All of you fools. On both sides. Listen up! You can chit your chat for whatever you “think” is happening re: global climate. If you think humans are causing it, I pity you. If you think you can change it, I pity you even more. If the Earth is warming, its because of the Sun. If the Earth is cooling, its because of the Sun. God help us if any of you think there is a political solution (oxymoron) to this. This might be too religious for some of you!
[REPLY – We have some Sun Worshipers around here. But we also have our share of Sea Witches. And land use or “dirty snow” (etc.) may well account for some of it. We know the heat comes from the sun, but there is quite some dispute over whether the changes in TSI (or UV or whatever) are enough to account for the temperature change. It’s all in the delta, or lack thereof. (And we don’t even really know how great that change has been, considering the pitiful state of the historical records and the climate stations, themselves.) ~ Evan]
carrot eater says:
Oh, by the way: As for the claims of Watts, d’Aleo and EM Smith that global warming is at least in some part an artifact of an intentional station drop-off at 1990 – that claim has already been comprehensively refuted by several people,
No, it hasn’t. Folks have done a rather crude comparison of bulk station averages at one span of time and asserted that means they stay the same in other periods of time. They don’t. Take stations in a cold excursion in cold places (such as a cold PDO) and you get cold results. That’s the “baseline”. Then change to warmer stations in warmer places during the warm phase of the PDO, that’s the “warming”, now repeat endlessly that it will stay warmer, even as the snows fall…
The “problem” is that the range of a station at, for example, Reno will have a greater downward run during a cold time than during a warm time. And SFO will never have that kind of excursion to the low side. Have Reno “in” during the cold, then out during the next cold with SFO instead, and you can never recapture that low excursion.
And all the hand waving about gross averages will never ever find that fact. Nor do they prove anything.
Oh, and do remember that GIStemp uses one set of thermometer in the baseline and a completely different (AND warmer due to the dropped stations…) set in the ‘present’ for creation of the grid/box anomalies. It is NOT station to self anomaly creation. So don’t waste your time bleating about how ‘anomalies will save you’ as you are not using proper anomalies.
I normally just ignore C.eater as a waste of time, but a direct claim about me was made.
Oh, one other point, I’m working on an interesting example of the impact of The Great Dying of thermometers. I’ve found a source from some data on stations that were left out. Comparison with stations that were kept shows the left out stations NOT warming while the kept ones do. That will not show up in the analysis done by other folks when they do “kept & tossed up to 1990 vs Kept Only” after as the after 1990 data for the tossed are not there for comparison. I can’t yet say WHY they diverge, but they do. (At this point I’m suspecting either a coincident equipment change or a change of processing, but there are other possibilities). The bottom line, though, is that I have in hand example cases of “kept” rising in a 1990 hockey stick and tossed from nearby not rising so.
So you can claim to have proved a negative all you want. I’ll stick with the existence proof I have in hand. (And no, I’ve not published it yet. It takes a fair amount of work to hand transcribe the data from the forms and make a comparison, one station at a time. If only I had some Big Oil Money I could hire clerks and go a lot faster, but such massive funding is just like the rest of the warmers claims. Nowhere to be seen in the real world.)
EM Smith
You’re wrong on pretty much every count.
“No, it hasn’t. Folks have done a rather crude comparison of bulk station averages at one span of time and asserted that means they stay the same in other periods of time. They don’t. Take stations in a cold excursion in cold places (such as a cold PDO) and you get cold results. That’s the “baseline”. Then change to warmer stations in warmer places during the warm phase of the PDO, that’s the “warming”, now repeat endlessly that it will stay warmer, even as the snows fall…”
I don’t think quite grasp all the work that has been done. People have made temperature histories using only those stations that are continuous through the drop-off period you were so worried about. Using only those continuous stations, you still get the same result. What does that tell you, EM Smith? No switching of stations. Same result. As for the ones that did drop at 1990? Up to that point, also the same result. You want to tell me that those two subsets tracked each other fine until 1990, and then suddenly diverged when nobody was looking? What evidence do you have for that?
But in case you were worried about what those stations did, when nobody was looking, people have even used non-GHCN data sources (GSOD and ISH) that don’t have the 1990 station drop-off at all – data sources with a large station count through that whole period .. again, same result; nothing weird happens at 1990. What does that tell you?
And for the nth time, ‘cold’ and ‘warm’ don’t matter in the calculation you’re talking about. Only ‘cooling’ and ‘warming’ do. The offset used when combining stations completely eliminates any memory of what was ‘warm’ and ‘cold’.
“Oh, and do remember that GIStemp uses one set of thermometer in the baseline and a completely different (AND warmer due to the dropped stations…) set in the ‘present’ for creation of the grid/box anomalies. It is NOT station to self anomaly creation. So don’t waste your time bleating about how ‘anomalies will save you’ as you are not using proper anomalies.”
And you still don’t understand how GISTemp calculates anomalies, and it appears that a lot of your difficulties stem from that. On your website, you recently tried to work out a toy example of how GISS does it, and got it horrifically wrong. That was here,
http://chiefio.wordpress.com/2010/03/14/why-temperatures-matter-vs-anomalies/#comment-3881
You can see how the method actually works (and yes, the published paper and the code both do the same thing) here
http://rankexploits.com/musings/2010/not-so-spherical-cows-more-toy-problems/
at comments 38192 and 38195
When you actually understand how the RSM works, you can implement it and get results consistent with the CAM. Did you ever stop to consider that only GISS uses the RSM, but they still track CRU and NOAA anyway?
And also, by the way, I don’t think ‘baseline’ has the significance in the RSM (GISS’s method) you think it does. When combining stations, you use the entire period of overlap, not a fixed baseline.
Carrot Eater
It is quite one thing to replicate (repeating analyses to get the same result) and undertaking a forensic examination to understand where problems might lie and if they can cause a significant effect. I can dredge up the old “If you do what you’ve always done, you’ll get what you always got”
Remind me – what are we talking about in terms of warming – 0.6-0.8 deg C? So something that causes >0.1degC difference on average must be taken seriously (see here). Most people are looking at current temperatures, and they are looking at averaged temperatures (yes I know – spatially weighted). It turns out the difference is in the older temperatures, and when the spatial distribution of this is examined the differences are even greater (see here) – there is a big difference with lattitude. Don’t even get me started on adjustments.
Now that is where the warm and cold come in as different latitudes respond differently to the natural cycles in the climate (if you plot the GISS numbers here) “Cold” places (altitude as well as lattitude) show greater natural variation in temperature than “warm” ones in terms of amplitude. If we have been warming for a long time with cold stations present they will provide a large amplitude positive anomaly, then if we drop them as we start to cool this can have an effect of bias as ‘warm’ stations have a smaller amplitude variation and therefore anomaly. This is a concern.
And talking of anomalies, in the example you referenced, E.M.Smith was taking the trouble to answer the question “..what does an anomlily mean in layman terms?” [sic]. He did so simply and not ‘how GISS does it’. As for his understanding of it, do you think that with all the disucssions that have gone on that he really misunderstands it (still, if he ever did), or that perhaps he has been unclear or misinterpreted in what he has said about it and has just not been bothered to counter the ad homs tossed in his direction? He is clearly someone who has all the necessary skills and intelligence to do so and is not so concerned about what others think that he wastes time on such distractions.
Doesn’t it concern you just a little that we’ve had all these stories (with photos and eyewitness evidence) of ice melt in the arctic in the past and that we might be spending billions on something that is natural and cyclical?
“So something that causes >0.1degC difference on average must be taken seriously”
I have no idea what that graph is, seeing as it doesn’t have a descriptive title or caption. What are you trying to show?
“Don’t even get me started on adjustments.”
Go ahead, start on them. They have little effect on the global mean.
“If we have been warming for a long time with cold stations present they will provide a large amplitude positive anomaly, then if we drop them as we start to cool this can have an effect of bias as ‘warm’ stations have a smaller amplitude variation and therefore anomaly. This is a concern.”
In case you haven’t noticed, a multitude of bloggers have built up a temperature record using only continuous stations. No dropping of stations. On top of that, there are now a couple analyses that use a more complete set of stations, again with no drop-off in numbers at 1990. The results are consistent with the published ones.
The way you could get a bias through station drop is if you lost stations that had markedly different trends from their neighbors in the same grid box. But after doing analysis with data sets that don’t have station drop, and you get the same thing.. what does that tell you?
But at least your description is getting better than where it started, which was the naive idea that dropping cold stations makes it appear warmer, just because they were cold, informed by endless series of irrelevant graphs showing absolute temperatures averaged together.
“And talking of anomalies, in the example you referenced, E.M.Smith was taking the trouble to answer the question “..what does an anomlily mean in layman terms?” [sic]. He did so simply and not ‘how GISS does it’.”
Whatever he was trying to show, it does not correspond to what anybody does. What he did was nonsensical. So no, it was not a simple demonstration of what anomalies are.
“As for his understanding of it, do you think that with all the disucssions that have gone on that he really misunderstands it (still, if he ever did), or that perhaps he has been unclear or misinterpreted in what he has said about it and has just not been bothered to counter the ad homs tossed in his direction? He is clearly someone who has all the necessary skills and intelligence to do so and is not so concerned about what others think that he wastes time on such distractions.”
I have seen no evidence that he understood how GISS does it. He continually says GISS is somehow not ‘saved’ by anomalies because it doesn’t calculate anomalies at each station against itself. He’s never (that I’ve seen) explained how GISS actually does do it, or how he thinks that leaves it prone to not being ‘saved’. If he has ever clearly explained his thinking on this, please point it out. The reference station method is really not that difficult. On the other hand, he’s claimed that the GISS code doesn’t do what’s in the papers, when it very much does. And anyway, CRU and NOAA use roughly the same data source, but not the RSM, and still get consistent results… what does that tell you about the RSM vs the CAM, and whether it matters here?
And he might waste a little less time if he actually looked at the work done by others. Those aren’t distractions.
As for ad homs, you might consider the contents of this:
http://wattsupwiththat.com/2010/01/26/new-paper-on-surface-temperature-records/
I wonder if the authors still stand by the accusations made therein. They have not stood up.