Hansen's Second Law of Climate Dynamics: If the present fails to get warmer, the past MUST become colder. 'Conservation of trends' is a fundamental principle.
http://www.climateaudit.org/?p=2049#more-2049

Given that Texas, and many other places in the south and east, have cooled considerably over the last 100 years - this principle must be invoked to maintain panic, and keep the door open for Hansen as the official civil service global warming czar.

Similarly, Gore and his buddies getting rich off carbon trading and global warming panic, are dependent on Hansen keeping the trends moving in the right direction.

This problem is one that almost all weather stations have, whether or not they have been moved. In most cases, it is not the weather station that has moved, but the civilization surrounding it, as the city encroaches into what was once the countryside. The solution to this dilemma is to adjust the temperature readings to account for the encroaching civilization.

The result then is that the temperature data are not date at all, but peoples estimates of what the data would have been had conditions not changed in the course of 125 years or so.

An interesting twist to this is a weather station here in northern lower Michigan, in Wilderness State Park. Not much encroaching civilization to cause odd data, but some odd data nonetheless. Over the last century, the readings show higher highs and lower lows. Global Warming? Probably not. Generally believed to have been caused by logging. Large tree growth removed, replaced with brush and grass. Once again, though, a theory, guess, or estimate replaces data.

I wonder how many weather stations there are around the globe supplying information to several sources? It must run into hundreds of thousands. The chances of all of them being as accurate as we aould wish, scientifically, are pretty remote. One swallow doesn't make a summer, and one (there will be more) batch of faulty data doesn't have much effect on the global picture.

Reply: mostl likely true, but as a daily weather forecaster seeing stuff like that is bothersome. Also, I've been at this for nearly 20 years and nothing beats the human touch of weather observation. The automated ASOS stuff doesn't compare, period.

That's a little like finding one roach in a luxury apartment that you're being asked to shell out big bucks for, and having the agent tell you there's not a chance in the world that there could be a problem.

All the best,

Aaron

Reply: I, just like McIntyre was showing one example. It's not the only one, believe me.

Brett, you need to learn to not take material at those sites at face value. The key point is that the amateur surveyors have looked at over a third of the total network (on the order of 400 out of 1200) and have found very few *apparent* problems of this sort. That isn't enough to have a noticable effect on national temperature figures. (I say apparent because we need to hear from GISS.)

Reply: Steve, I alluded to that to some extent. I fully realize that, but it still does not make it right. Why not have proper sighting at all locations in the GISS?

BTW, the imperfections in the existing stations are why NOAA decided years ago to build a brand new one, the Climate Reference Network (CRN), which is just now coming into full operation. Maybe you should do a post on it.

Reply: I will look at it.

How often does Hansen err on the low side?

His errors are consistently biased towards an upward trend in temperature, which is an unmistakable characteristic of someone trying to prove that the temperature is rising.

Hansen has a clear conflict of interest - responsible for both the theory and the data needed to prove the theory. If the data shows cooling, then his reputation is threatened.

Another one has been identified
http://www.climateaudit.org/?p=2741#comments
and seesms to confirm a mistake in GISS algorythm forcing past temps down... again the hockey stick.....This may start to break the camels back

It looks as though Dr. Hansen needs to "adjust" the Llano, Texas station readings. Something is definitely amiss here.

I do think this requires immediate attention.All stations should be in rural settings at least a couple miles from any asphalt or communities for accurate readings.Too much money involved for it to happen though.

It's data manipulation like this that keeps skeptics hammering away at the coffin nails of James Hansen's GISS travelling sideshow circus. If the US government can waste billions on this charade it can find a few million to fix it. Otherwise all the data should be discarded to the archive bin and the whole temperature measuring project put on hold along with any legislative measures that would hurt the economy. To do anything less is even more fraud.
You can't reach accurate conclusions with bad data!

It is sad (in fact, outrageous), that the tasks of model-checking and data quality control are left to unfunded amateurs. Can we please call a halt to the alarmism and embark on a massive, transparent data audit? Probably too much to ask...

Hi Steve Bloom,

Interesting that the same person who quotes temperatures to four digit precision (you) is so unconcerned with a 2-3 degree error. It would appear that you are happy with any reported temperature measurement that creates the appearance of warming.

According to the World Wildlife Fund, amongst 14 of the 20 polar bear populations worldwide it surveyed, ten populations are stable, two are increasing, and two are decreasing. One of the two diminishing populations, in Baffin Bay, is declining because the air is getting colder - not warmer. Polar bears do face threats from pollution (which the carbon cultists have shoved off the environmental agenda), but they seem to take warming in their loping stride.

http://www.theregister.co.uk/2008/01/21/philip_pullman_wartime_misery/

Brett, most of this network depends on volunteer observers, so apparently the stations get shifted around depending on where an observer is available. The only alternative is automated transmission or paid observers, both of which would be major budget items as these things go. But rather than try to fix this network, NOAA (not GISS, BTW) decided about ten years ago to build the CRN (which features automated data transmission among other things). As it turns out the last of the CRN stations (100+ total) are just now being completed, which I think gives a sense of what sort of budget priority climate monitoring has been (i.e., not much). Remember that the satellite network was allowed to degrade over the same period.

That said, the network is so dense that it's still possible to get good results out of it. (Reply: Agreed, when you combine it all for the U.S.)That this is so is a well-known property of large data sets called "the power of large numbers." See here for a discussion. Take-home point:

"However, it turns out to be a fundamental property of statistics that the average of a large number of estimates is more precise than any single estimate. The more data go into the average, the more precise is the average � even though the source data are all imprecise."

Of course there are lots of other issues; see here for a detailed discussion.

Finally, I would note that Anthony Watts has occasionally let his enthusiasms carry him away. Steve McIntyre has as well, but he makes a habit of deleting such mistakes from his blog.

Hi Steve Bloom,

"Very few 'apparent' errors of this sort"?? And you call sceptics "deniers"?

Have a quick look at their site:

http://www.surfacestations.org/

70% of the 482 stations surveyed (out of 1221) have a CRN rating of either 4 or 5 - with 56% showing an error of >= 2C and 14% showing an error of >= 5C. Just in case you need a basic primer - a higher CRN rating and/or error is bad, lower is good.

Its no wonder you are such an ardent proponent of AGW despite associated IPCC uncertainties given your apparent belief that an error of the magnitude noted above is both acceptable and constitues "very few apparent errors".

Just so we are clear I'll state it again - 70% of the surveyed stations show an error of between ~ 3 to 7 times the "apparent" total warming this century and unless at least half of the remaining stations are under water, in the middle of an indoor skating rink, or situated inside a meat packing plan, its unlikely that the errors upward will be averaged out by errors downward.

Brett is probably correct in that it will be claimed that there is not enough money in the budget to correct/fix corrupted stations but if AGW proponents such as yourself spent as much time and energy lobbying for immediate action to resolve such "apparent" errors as you do lobbying for a full range of punitive CO2 emissions standards likely to cost hundreds of billions if not trillions, it should be easy to shake a few million loose and correct the problem.

Perhaps someone could suggest to Leonardo di Caprio that he organize a benefit concert - or the next round of climate talks could be held by video conference and the money saved on Hawaii beach resorts could be diverted to fix corrupted climate stations.

But I suspect neither will happen - "Fix the Climate Stations" is not quite as sexy on a concert t-shirt as "Save the Planet".

Climate Audit has two additional new threads on this topic, including accusations that the GISS algorithm is flawed.
http://www.climateaudit.org/?p=2740#comments
http://www.climateaudit.org/?p=2741#comments

GISS does not have very many stations any more, and the staff should be personally familiar with and accountable for each and every one.
http://www.climateaudit.org/wp-content/uploads/2008/02/sc2006.gif
http://www.climateaudit.org/?p=2711#more-2711

Perhaps Dr. Hansen should spend less time making public speaking appearances, and more time actually doing his job? I am amazed that he allows obviously flawed data into the GISS database.

I wish I had the kind of memory many of you guys and gals posting seem to have. BUT, it seems to me that the complaints of data collection stations being skewed by urban heat island effects were delt with in a recent artical by several researchers who actually visited data collection sites and found them to be indeed surrounded by asphalt and concrete artificial environments. Not only that, the artical claimed that these particular researchers were dealing with great secrecy as to where many data collection sites were located, hampering thier investigations relative to propper data collection site location. I understand many airports are data collection sites, obviously these locations are not good. I need to say that the spraying of cirrus clouds over NYC area is still happening. It is offensive and alarming few people seem to either notice or care.

Brett,

Thanks for posting this. My chief fear is that now, to be fair to the other side, you must post something from Realclimate, and I will want to scream.

This is one of my favorite sites, though often studying their posts gives me a migrane. They do work I could never do, meticulously going through data with a fine-toothed comb, and finding all sorts of bugs in Hansen's data.

It pays to go through Cimateaudit's archives. People like to dismiss their findings because they only amount to .01 degree here and .005 degree there. However when their archives show you this happens over and over and over and over again in Hansen's data, you pause, and start to add up all the .01's and .005's. Before you know it you are talking about a really big number, like .1 or .2 or even .3 degrees. Considering Hansen is raising a huge alarm about .6 degrees, you start to get more and more Skeptical about his objectivity, and Mann's objectivity, and Gore's objectivity.

Then I advise people to switch over to Realclimate, and see how they absolutely loathe Steve McIntyre and Anthony Watts. Not that they can refute the data, but they can bash the two fellows amazingly.

I like to visit as many sites as I can find time for, but find this Accuweather site the best, for someone wishing a general overview, and also to hear stinging responces from both sides.

The La Nina is weakening. Time for a new post on the subject, Brett? I'm itching to debate, for I've learned a lot since the subject last came up.

Reply:In due time, I want to see a few more weeks of data from the equatorial Pacific.

Brett,

Please check out Anthony Watts blog and Surface Station sites. You will see the number of bad stations swamp the good stations. I am one of the "amateurs" doing the surveys. There are good sites and bad sites, the problems is the bad sites and good site are all mixed in the historical record. The real issue is how GISS is adjusting the existing records, good and bad. With Hansen behind the curtain making adjustments using software code that will never be revealed, we will never know the truth. Please take Blooms comments in context, he has been trashing Anthony's efforts from day one. He has an agenda.

Steve Bloom: The key point is that the amateur surveyors have looked at over a third of the total network (on the order of 400 out of 1200) and have found very few *apparent* problems of this sort.

I'm curious about Bloom's basis for the claim that the surveyors have found "found very few *apparent* problems of this sort." The surveyed sites are rated on a scale of 1 to 5 with 1 being the best and 5 being the worst. Sites rated 1 or 2 have an expected error of less than 1 deg C; sites ranked 3, between 1 and 2 deg C; sites ranked 4, between 2 and 5 deg C; and sites ranked 5, 5 of greater deg C. Of the 482 surveyed sites, the ratings are:

1 - 4% (1 < deg C)
2 - 9% (1 < deg C)
3 - 17% (1 - 2 deg C)
4 - 56% (2 - 5 deg C)
5 - 14% (> 5 deg C)

So only 13% fall into the best two categories, with an error of less than a degree, while 70% fall into the worst two categories, with errors of at least 2 degrees. As many are in the worst category (> 5 deg C) as are in the two best categories.

Steve:
You said: "....and have found very few *apparent* problems of this sort."

I assume that mean to be taken VERY literally with that comment. Specefically "OF THIS SORT".

If you look further into the info at surfacestation.org you will find that there are indeed few problems of that exact type but many that are just as bas or worse.

Here is a link to a page of the presentation Watts gave reently to CIRES/UCAR on 8/29/07
It shows clearly how big the problem is. Its BIG.

Just backs up what PH always says about GISS data fudging.

http://gallery.surfacestations.org/UCAR-slides/page92.html

Damo,

According your your stats that of 482 observed stations, 56% have an error of 2 degrees C or more and 14% have an error of 5 degrees more, this is very disconcerting since the claim is that the globe has only warmed 0.8 F in the last century.

Let's be ridiculously conservative and say that all the errors there are were in those 482 stations and the rest of the 739 stations are fine. Entirely unlikely, but let's go with that.

That means 739 + 30% * 482 = 884 stations are accurate, 270 are off by more than 2 degrees C and the other 67 are off by 5 degrees C or more.
That gives an average error of ([270 * 2) + (67 * 5)]/1221 = 0.72 degrees C. But the claimed warming of the last century is only 0.44 degrees C. That would seem to imply the average has actually dropped 0.28 degrees C here in the United States even assuming the 739 unsurveyed sites are good. And this is with what are said to be the most accurate stations in the world!

Brett,

FYI - Every Coast Guard station logs temperature, on the hour, every hour. These measurements generally come from non-digital, glass tube thermometers and are interpreted by busy watchstanders. These LESS than accurate records are not scrutinized, sent off to NOAA, which then uses them as part of their scientific database.

Hi Steve Bloom,

Re: the power or large numbers. A take home point for you which you have obviously not "taken home".

The error minimizing impact of large data series only works if you assume a random distribution of error (i.e. thermometers accurate to +/- ~0.5C with an equal chance that individual readings are + or - within this range).

This is clearly not the case here as we are not talking about thermometer accuracy but poor station siting which typically results in an error upward and seldomly if at all results in a downward error. If simple common sense won't convince you of this, and since you've referenced CRN, have a look at their rating system and you'll note that the bulk of siting error is of the type to cause a higher micro-temperature in