Originally Posted by Bob Butler 54

Could you provide a text list of those 133 stations? I might want to hit a few of them to make sure there is no cherry picking going on.

They're easy to find by the 'north of' and 'east of' and 'age of' criteria I listed in my previous post. But since you asked nicely, I've cut-and-pasted my list at the bottom. Station number, name, lat, long format (for some reason you get the 'cent' character instead of 'minutes', although Excel reads either one just fine).

I might even suggest we use a stronger word than 'tedious.'

I did on another thread and you got all 'Supreme Court' on me

The 1998 spike was a combination of near peak 11 year solar cycle with the mother of all El Nino seasons. I'm guessing that Siberia isn't effected very much by the Pacific Oscilation. Thus, I'm not surprised by the lack of El Nino effects. The prevailing winds in Siberia don't blow out of the South Pacific.

The AGWarmist line was that 1998 was "The Hottest Year Of The Century". It shows up as such in pretty much none of the Siberian data. It's not really important, either way -- just something I observed.

I too see the long term gradual warming trend. I'll add that I care most about 1960 and newer. That's when the hockey stick starts curving up.

Except that in the Siberian ones I've gone through so far... it doesn't. It pokes up from 60-80, and then level or down from mid-80s through now. The AGWarmist line is currently working to hand-wave away the 2000's cooling trend, but the Siberian data seems to challenge even the 1990s AGW trend. IIRC, all their graphs show the real crazy near-vertical spiking happening mid-80s and 90s.

I do note that the increasing solar heat in the first half of the 20th Century doesn't seem to have affected Siberia as much as the rest of the world. As a wild guess, the snow bounces a lot of the solar energy into space, so there is less effect?

I don't know. Relatively warmer winter weather in the way-the-hell-north does tend to mean increased snowfall. Most of the time if it's crazy cold, its too cold to snow (and what's there, if it is thin, tends to sublimate away pretty fast). You'd have questions of cloudcover patterns, too.
It's an interesting question that probably would justify a chunk of study. Of course, it also highlights another fatal flaw in GCMs.

I'm not seeing any "multi-decade level/cooling" since 1960. From my side of the values divide, the trend is broadly warming in the four charts I've seen thus far : Kirensk, Taishet, Ika and Perevoz. I really don't see how you are reading those charts as level or cooling.

See the trend lines. That's what they are there for. Those four stations have a quick spike of warming in the early-to-mid 80s, and then 2+ decades of cooling up to today. That is, even notwithstanding the current decade which AGWarmers accept is a cooler one, they also cooled during the one that supposedly saw warming worldwide.

Plus, the 'melting permafrost' claim certainly pertains to the last twenty years.

Unless.... Those are years at the top of the charts? 2008 on the left for Perevoz, 1941 on the right? Thus, your charts should be read from right to left??? Most charts (including mine) show time progressing from left to right.

Yeah. Sorry about that. I flipped all the data upside-down so I could have the same physical starting point (cell C4) for January 2008 on every record. It makes data analysis a lot easier, especially when we don't really care in a lot of detail about what happened in 1846. I found out how to turn the x-axis around... but not until after I had already posted the pictures.

Future graphs will no longer be oriented ass-backwards.

I guess if we want to further reduce the number of stations from 131, you may want to deal only with stations near or just north of the permafrost line. You might find the traditional 1950s permafrost line. If there are a bunch of stations around the same average yearly temperature near the traditional permafrost line, we might then have an average temperature where the permafrost is apt to survive. One might then look at where temperatures are at various stations north of that point and project a crude trend towards how much melting is apt to occur and how soon.

That's a good idea. If you don't mind, I'm going to get the bulk processing done on all the stations first before I go into that kind of detail.

My "long-term Siberian" station list:
20069 Визе 79° 30¢ 76° 59¢ 20087 Голомянный 79° 33¢ 90° 37¢ 20289 Русский 77° 10¢ 96° 26¢ 20292 Им.Е.К.Федорова,ГМО 77° 43¢ 104° 18¢ 20476 Стерлегова 75° 25¢ 88° 54¢ 20667 Им. М.В.Попова 73° 20¢ 70° 03¢ 20674 Диксон 73° 30¢ 80° 24¢ 20891 Хатанга 71° 59¢ 102° 28¢ 20982 Волочанка 70° 58¢ 94° 30¢ 21432 Котельный 76° 00¢ 137° 52¢ 21647 Шалаурова 73° 11¢ 143° 14¢ 21802 Саскылах 71° 58¢ 114° 05¢ 21824 Тикси 71° 35¢ 128° 55¢ 21908 Джалинда 70° 08¢ 113° 58¢ 21921 Кюсюр 70° 41¢ 127° 24¢ 21946 Чокурдах 70° 37¢ 147° 53¢ 21982 Остров Врангеля 70° 59¢ 181° 31¢ 23074 Дудинка 69° 24¢ 86° 10¢ 23256 Тазовск 67° 28¢ 78° 44¢ 23274 Игарка 67° 28¢ 86° 34¢ 23345 Ныда 66° 38¢ 72° 56¢ 23365 Сидоровск 66° 36¢ 82° 18¢ 23383 Агата 66° 53¢ 93° 28¢ 23472 Туруханск 65° 47¢ 87° 56¢ 23552 Тарко-Сале 64° 55¢ 77° 49¢ 23662 Толька 63° 59¢ 82° 05¢ 23678 Верхнеимбатск 63° 09¢ 87° 57¢ 23867 Ларьяк 61° 06¢ 80° 15¢ 23884 Бор 61° 36¢ 90° 01¢ 23891 Байкит 61° 40¢ 96° 22¢ 23946 Угут 60° 30¢ 74° 01¢ 23955 Александровское 60° 26¢ 77° 52¢ 23986 Северо-Енисейский 60° 23¢ 93° 02¢ 24105 Ессей 68° 28¢ 102° 22¢ 24125 Оленек 68° 30¢ 112° 26¢ 24136 Сухана 68° 37¢ 118° 20¢ 24143 Джарджан 68° 44¢ 124° 00¢ 24266 Верхоянск 67° 34¢ 133° 24¢ 24329 Шелагонцы 66° 15¢ 114° 17¢ 24343 Жиганск 66° 46¢ 123° 24¢ 24371 Усть-Чаркы 66° 48¢ 136° 41¢ 24382 Усть-Мома 66° 27¢ 143° 14¢ 24507 Тура 64° 16¢ 100° 14¢ 24641 Вилюйск 63° 47¢ 121° 37¢ 24661 Сеген-Кюель 64° 00¢ 130° 18¢ 24671 Томпо 63° 57¢ 135° 52¢ 24679 Восточная 63° 13¢ 139° 36¢ 24688 Оймякон 63° 15¢ 143° 09¢ 24738 Сунтар 62° 09¢ 117° 39¢ 24790 Сусуман 62° 47¢ 148° 10¢ 24802 Стрелка Чуня 61° 45¢ 102° 48¢ 24817 Ербогачен 61° 16¢ 108° 01¢ 24908 Ванавара 60° 20¢ 102° 16¢ 24923 Ленск 60° 43¢ 114° 53¢ 24951 Исить 60° 49¢ 125° 19¢ 24959 Якутск 62° 01¢ 129° 43¢ 24966 Усть-Мая 60° 23¢ 134° 27¢ 24967 Тегюльтя 60° 28¢ 130° 00¢ 25034 Амбарчик бухта 69° 37¢ 162° 18¢ 25042 Остров .Айон 69° 56¢ 167° 59¢ 25138 Островное 68° 07¢ 164° 10¢ 25173 Мыс Шмидта 68° 54¢ 180° 38¢ 25206 Среднеколымск 67° 27¢ 153° 43¢ 25248 Илирней 67° 15¢ 167° 58¢ 25325 Усть-Олой 66° 33¢ 159° 25¢ 25356 Эньмувеем 66° 23¢ 173° 20¢ 25399 Уэлен 66° 10¢ 190° 10¢ 25400 Зырянка 65° 44¢ 150° 54¢ 25428
Омолон 65° 14¢ 160° 32¢ 25503 Коркодон 64° 45¢ 153° 58¢ 25538 Верхне-Пенжино 64° 13¢ 164° 14¢ 25551 Марково 64° 41¢ 170° 25¢ 25563 Анадырь 64° 47¢ 177° 34¢ 25594 Бухта Провидения 64° 25¢ 186° 46¢ 25621 Кедон 64° 00¢ 158° 55¢ 25656 Березово 63° 27¢ 172° 42¢ 25703 Сеймчан 62° 55¢ 152° 25¢ 25705 Среднекан 62° 27¢ 152° 19¢ 25744 Каменское 62° 29¢ 166° 13¢ 25777 Гавриила бухта 62° 25¢ 179° 08¢ 25913 Магадан 59° 33¢ 150° 47¢ 25927 Брохово 59° 39¢ 154° 16¢ 25932 Тайгонос 60° 41¢ 160° 24¢ 25954 Корф 60° 21¢ 166° 00¢ 25956 Апука 60° 26¢ 169° 40¢ 28493 Тара 56° 54¢ 74° 23¢ 28698 Омск 55° 01¢ 73° 23¢ 29023 Напас 59° 51¢ 81° 57¢ 29111 Средний Васюган 59° 13¢ 78° 14¢ 29154 Усть-Озерное 58° 53¢ 87° 45¢ 29231
Колпашево 58° 18¢ 82° 53¢ 29263
Енисейск 58° 27¢ 92° 09¢ 29282 Богучаны 58° 23¢ 97° 27¢ 29313 Пудино 57° 34¢ 79° 26¢ 29328 Бакчар 57° 00¢ 82° 04¢ 29348 Первомайское 57° 04¢ 86° 13¢ 29418 Северное 56° 20¢ 78° 22¢ 29467 Ачинск, ж.д.ст. 56° 17¢ 90° 31¢ 29539
Болотное 55° 40¢ 84° 24¢ 29541 Тайга 56° 04¢ 85° 37¢ 29557
Тисуль 55° 45¢ 88° 18¢ 29570
Красноярск, оп.п. 56° 02¢ 92° 45¢ 29580 Солянка 56° 10¢ 95° 16¢ 29594 Тайшет 55° 57¢ 98° 00¢ 29605 Татарск 55° 13¢ 75° 58¢ 29612 Барабинск 55° 20¢ 78° 22¢ 30028
Ика 59° 17¢ 106° 10¢ 30054 Витим 59° 27¢ 112° 35¢ 30069 Перевоз 59° 00¢ 116° 55¢ 30089 Джикимда 59° 01¢ 121° 46¢ 30219 Максимово 57° 06¢ 104° 58¢ 30230 Киренск 57° 46¢ 108° 04¢ 30309 Братск 56° 17¢ 101° 45¢ 30328 Орлинга 56° 03¢ 105° 50¢ 30372 Чара 56° 54¢ 118° 16¢ 30385 Усть-Нюкжа 56° 35¢ 121° 29¢ 30433 Нижнеангарск 55° 47¢ 109° 33¢ 30437 Карам 55° 09¢ 107° 37¢ 30471 Средний Калар 55° 52¢ 117° 22¢ 30493 Нагорный 55° 58¢ 124° 53¢ 31004 Алдан 58° 37¢ 125° 02¢ 31026 Учур 58° 44¢ 130° 37¢ 31062 Югаренок 59° 46¢ 137° 40¢ 31088 Охотск 59° 22¢ 143° 12¢ 31137 Токо 56° 17¢ 131° 08¢ 31152
Нелькан 57° 39¢ 136° 09¢ 31168 Аян 56° 27¢ 138° 09¢ 32252 Усть-Воямполка 58° 30¢ 159° 10¢ 32287 Усть-Хайрюзово 57° 05¢ 156° 42¢ 32363 Эссо 55° 55¢ 158° 43¢ 32389 Ключи 56° 19¢ 160° 50¢ 32411 Ича 55° 35¢ 155° 35¢ 32618 Остров Беринга 55° 12¢ 165° 59¢

---

Edit: crap. I lost formatting on the list. Trust me, it was a beautiful set of rows and columns before I hit 'submit'. I can re-try if you want, although it is technically all there... and this way takes up less space

Originally Posted by Odin

Justin's sophistry is getting tedious.

Got one of those 'word-a-day' calendars for your birthday?

Get a dictionary Read a book.

Originally Posted by Odin

Justin's sophistry is getting tedious.

Again, I might well agree if you chose to use stronger words.

Originally Posted by Justin '77

They're easy to find by the 'north of' and 'east of' and 'age of' criteria I listed in my previous post. But since you asked nicely, I've cut-and-pasted my list at the bottom. Station number, name, lat, long format (for some reason you get the 'cent' character instead of 'minutes', although Excel reads either one just fine).

The list you sent includes the station number, which might be all I need to verify I'm using one of the stations in the area in question. Roman names might be nice if easily possible. (Not important). Once you get a finalized list of stations, I might request you e-mail the result with an attached Excel or .csv file. If you are further refining the list, might as well wait. Let me know if you need my e-mail address.

Originally Posted by Justin '77

See the trend lines. That's what they are there for. Those four stations have a quick spike of warming in the early-to-mid 80s, and then 2+ decades of cooling up to today. That is, even notwithstanding the current decade which AGWarmers accept is a cooler one, they also cooled during the one that supposedly saw warming worldwide.

Plus, the 'melting permafrost' claim certainly pertains to the last twenty years...

Yeah. Sorry about that. I flipped all the data upside-down so I could have the same physical starting point (cell C4) for January 2008 on every record. It makes data analysis a lot easier, especially when we don't really care in a lot of detail about what happened in 1846. I found out how to turn the x-axis around... but not until after I had already posted the pictures.

Future graphs will no longer be oriented ass-backwards.

I see the trend lines as showing warming. After you orient the graphs ass-forwards, do you still perceive the trend lines as showing cooling?

Originally Posted by Bob Butler 54

I see the trend lines as showing warming. After you orient the graphs ass-forwards, do you still perceive the trend lines as showing cooling?

Um.. you are looking at the long-term (that is, the uncontroversial natural-recovery-from-the-Little-Ice-Age trend). I'm talking about the later (you know, corresponding to the 'blade' of the hockey stick) two-decade trend. I put two lines on each graph...

At most you have something like Perevoz, which shows practically a null trend from since the middle of the 80s. Alternately, there are places like Ika and Taishet, where the cooling trend since '88 is much more pronounced. Note: the twenty-plus-year-cooling-trend.

That flies in the face of every warming map I've ever seen (hint: they all show an up-slope, steady or increasing since 1960).

Just to clarify, I've flipped Taishet around forward.



The black, steeply-sloping line is the traditional fit-to-all-data. However, there's no real legitimate reason to do what it implies and set trends based on the arbitrary (for the climate at least) date at which we started measuring. Much more sensible is to identify trends in the data themself and look at those. I suspect there are a handful of statistical algorithms to pick term-periods, but I don't know any of them (yet; add to list of things to check in to). So I used the old Mk-I eyeball.

In the above graph, there appear to be three possible trendlines. One takes us from the beginning of data to around the early 80s. That's the orange line and the round dots (I haven't been able to figure out how to make it not shoot past the second mark yet; ignore the line past '86). That trend (I'm being told by the computer, a half-degree-per-century warming) is the uncontroversial LIA-recovery.
Next, there is a steep jump that my eyeball wants to put somewhere between 1968 and 1994. It's not clean, and the bulk of the 80s (call it 20-25% of the range of that 'trend') fall way off it, so maybe it's not a trend at all. I don't know; that's the Mk-I for you
Then there's the clear trend from 1988 going forward. That's the line between the pink squares. Excel says that works out to a .6deg/century decline. It's that one that is particularly damning for the AGWarmist case, since it falls right over -- according to canon -- the period of the greatest temperature increase. Station data says: for the last twenty years, that just isn't so!

....

I've managed another few stations, and that pattern keeps cropping up. I'm hoping soon to stumble upon one of the AGWarmist-cherry-picked ones so I can see where they got their story from. I'll post a picure as soon as I do.

----

Edit2
Looking at the graph a bit more, I found another trend that I kind of like. It gets the rise and the drop.

Also pretty well falsifies the AGWarmist models. Funny thing about data, you know...

Originally Posted by Justin '77

Um.. you are looking at the long-term (that is, the uncontroversial natural-recovery-from-the-Little-Ice-Age trend). I'm talking about the later (you know, corresponding to the 'blade' of the hockey stick) two-decade trend. I put two lines on each graph...

At most you have something like Perevoz, which shows practically a null trend from since the middle of the 80s. Alternately, there are places like Ika and Taishet, where the cooling trend since '88 is much more pronounced. Note: the twenty-plus-year-cooling-trend.

That flies in the face of every warming map I've ever seen (hint: they all show an up-slope, steady or increasing since 1960).

There is a clump of warm years in Ika in particular and Siberia in general from 1988 to 1998. Yes, if you start your trend from the start of a warm period, you can generate a small down slope.

There is a clump of cool years just preceeding it, from 1984 to 1987. If one starts the trend from 1984, just four years earlier, one can generate an entirely different trend line showing a sharp increase in temperature building towards the present.

The process of drawing trend lines to get the results that favor one's values would be called 'cherry picking.' I can take the long term trend going all the way back to 1944 farily seriously. With a graph as noisy as Ika, I wouldn't trust shorter term trend lines. If you have strong values, you can cherry pick start and end points that can confirm whatever political ideas you happen to have.

How would one generate a politically neutral start point for the last 20 years or so? It looks like 1994 was the nearest hot year, 1987 the nearest low year. If you average the temperatures for those two years, that's a reasonable place to start a trend line. If you do that, you are real close to the long term 1944 trend line, which is going up at about 2.3 degrees a century.

***

On another tangent, I'd like to talk about the hockey stick. The following chart starts from a diagram in the 2001 IPCC report, which first made the hockey stick famous. I added a bent green line to make the hockey stick overly obvious. I also added a deep red line which is to the best of my understanding a reflection of the skeptic's theory that we have had a slow gradual warming trend since the bottom of the Little Ice Age.

From my perspective, the Little Ice Age is one of many little bumps on the shaft of the hockey stick. There are many minima and maxima. The LIA is a little bit bigger than any of the others, but not by that much. The shaft of the hockey stick, however, is very straight long before the LIA. Prior to the Industrial Revolution, the Milankovitch Cycles had been pulling us slowly cooler towards the next ice age. Around 1900, there was another solar maxima. There would have been a bump upwards in temperature around 1900 without human intervention. However, around 1950 the amount of oil being burned took off. Thus, instead of another bump on the shaft, we got the hockey stick.

Several points.

1. The heel of the hockey stick where the trend curve abruptly higher took place around 1900.
2. The green bent AGW hockey stick line is a lot better fit to the data than the skeptic's red 'slow steady recovery from the Little Ice Age' hypothesis Justin is pushing.
3. Since 2001, a lot of people have been generating a lot of temperature history charts. The results aren't as pretty as the 2001 picture. The hockey stick isn't as clear and obvious in newer papers as it was in 2001.
4. Did I mention that the straight line fit for 'slow steady recovery from the LIA is really bad? The green hockey stick line drives pretty much through the center of the gray yearly noise. The red recovery from LIA line is in wide open white space over most of the length of the line. It just doesn't reflect the data.

Originally Posted by Bob Butler 54

There is a clump of warm years in Ika in particular and Siberia in general from 1988 to 1998. Yes, if you start your trend from the start of a warm period, you can generate a small down slope.

There is a clump of cool years just preceeding it, from 1984 to 1987. If one starts the trend from 1984, just four years earlier, one can generate an entirely different trend line showing a sharp increase in temperature building towards the present.

One can, and I referred to that on my above comments on Taishet. The thing is, the trendline one would draw (were it to be linear) misses close to 25% of the datapoints in its range. That makes it a pretty iffy trendline. If one wanted a trend for that range, the polynomial-fit I showed in a later post works better -- though it shows a decelerating trend over the 80s-00s, rather than the accelerating one that AGWarmist models keep spitting out.

I'm okay with that interpretation, btw. I can't say whether the linear-cooling trend or the decelerating-warming trend is the more reflective of reality. But both of them fit the data -- whereas the AGW models don't.

How would one generate a politically neutral start point for the last 20 years or so?

Why are we talking politics? This is supposed to be addressing matters of science. So far as I know, there isn't necessarily one correct way to pick starting/stopping points for trend analysis. In fact, the only semi-objective scientific criteria I'm aware of is the fits-lots-of-data one. Though since we're not expecting 100% overlap on many points, the terms 'lots' and 'fits' render even that criterion unacceptably squishy.

..

Good deal. So we've agreed that while some trendlines may be politically 'good' (or 'bad'), there's not really such a breakdown scientifically. That's progress.

On another tangent, I'd like to talk about the hockey stick. The following chart starts from a diagram in the 2001 IPCC report, which first made the hockey stick famous. I added a bent green line to make the hockey stick overly obvious. I also added a deep red line which is to the best of my understanding a reflection of the skeptic's theory that we have had a slow gradual warming trend since the bottom of the Little Ice Age.

Much as I love your use of that long-discredited visual trash, you've ended up [unintentionally, I'm sure] misrepresenting the data. Allow me to quote from Greenland ice cores:

15,000-year scale




2000 year scale





300-year scale



That's a recovery of in the neighborhood of a degree C a year, starting in the early 1800s (back when a fair number of the Siberian stations started measuring, too). It's also data that clearly displays -- one might even say, supports -- a variant of the red line you put on Mann's graph.

Originally Posted by Justin '77

They're easy to find by the 'north of' and 'east of' and 'age of' criteria I listed in my previous post.

I cnnot use that app now, I am getting the 502 errors. I got through thr other day and played with it, but the output was not useful so I must have been doing something wrong.

What I did was pick one station, picked a month and put in a an interval that corresponded to the time the station had date for (based on the other site you linked to). I got a single column of what looked like temperatures but no years or any other identifying information.

I'm guessing that the data are monthly averages for the months selected for the years in the interval I selected, starting with the most recent one first. Is this correct?

Originally Posted by Justin '77

Bob,

I'll start with comments on your Tara graphs. The reason your numbers are nearly 10-degrees different is that the upper graph is showing (appears to be showing) something most definitely not summer temps.

When you ask for temperature data for a specific month what does the site give out? Is it an average temperature, is it the temperature at one particular time during that month?

Originally Posted by Bob Butler 54

I figured a way to get the monthly data into a spreadsheet without too much trouble, so I'm no longer looking for a single month to work with. A yearly average seems the better choice.

As no one provided a letter of the alphabet, I am going with 'M'. I started out from the top of the Russian site looking for station names that begin with M. I intend to work my way down the list through the M stations. I got the first three Ms done when I came up with a hypothesis. The stations are sorted by station number. It is possible that the older stations have lower station numbers, thus the first three stations chosen might be older, better developed, and thus more asphalt prone than those lower down the list. Thus, should I get ambitious, the next few stations will be from the bottom of the list, and the next few from the middle.



The following graphs are generated using the monthly data, asking for the semicolon divider, and cut pasting the result into a word processor. All graphs are for 1960 through 2008. I did a find and replace to change the semicolons into commas, saving the result with a .csv extension, then loading the .csv file into a spreadsheet.

For some stations, there are missing months and years. I've been filling in the holes with the same month's data from the previous year. This isn't ideal, but seems better than to average zeroes into the curve. I'm open to other suggestions.

The third M on the list was Murmansk, which is definitely not a rural wilderness station. That was what made me consider whether the list might be sorted in a non-random fashion.

Still, a trend seems visible.

How are you getting this data, and why are their no years on the x-axis?

Originally Posted by Mikebert

I cnnot use that app now, I am getting the 502 errors. I got through thr other day and played with it, but the output was not useful so I must have been doing something wrong.

What I did was pick one station, picked a month and put in a an interval that corresponded to the time the station had date for (based on the other site you linked to). I got a single column of what looked like temperatures but no years or any other identifying information.

I'm guessing that the data are monthly averages for the months selected for the years in the interval I selected, starting with the most recent one first. Is this correct?

Originally Posted by Mikebert

How are you getting this data, and why are their no years on the x-axis?

I selected all twelve months in the Meteo site, rather than one month as you apparently did, did not request year or station ID, and asked for the semicolon dividers. I then changed the semicolons to commas in a word processor, then saved the result as a .csv file that my spreadsheet could import. I then averaged the 12 months of data for each year inside the spreadsheet.

But that was yesterday. The site seems to be down today. I suspect the server wasn't designed for front line duty in the climate wars.

All the above graphs started at 1960 and run to 2008. The years actually are on the original graphics... every single year. As there are more years than room to print the years, the result is absolutely unreadable... so I cropped it off before posting the image. I should likely fix it, but the discussions at the time seemed to be about broad trends rather than specific years.

Originally Posted by Bob Butler 54

I selected all twelve months, did not request year or station ID, and asked for the semicolon dividers. I then changed the semicolons to commas in a word processor, then saved the result as a .csv file that my spreadsheet could import. I then averaged the 12 months of data for each year inside the spreadsheet.

But that was yesterday. The site seems to be down today. I suspect the server wasn't designed for front line duty in the climate wars.

So what did you get? This list shows 476 stations (I put this list into Excel format in case you want it). Did you get 476 x 12 columns of monthly temperature data?

Originally Posted by Mikebert

So what did you get? This list shows 476 stations (I put this list into Excel format in case you want it). Did you get 476 x 12 columns of monthly temperature data?

I requested only one station at a time, and am building a separate .csv file for each station. With a fancy data base program, I might have gone for the whole thing. With just a spreadsheet, it seems easier to graph things out one request per station. I could go for 5 or so stations on a single graph if I felt like it, but thus far I haven's seen cause to get that fancy.

Just for laughs, the following gives five years of data from Perevoz.csv, starting at 1941. The semicolons have already been changed to commas.

-35.3, -26.2, -20.5, -6.0, 5.4, 15.6, 20.2, 15.0, 6.3, -2.5, -18.1, -30.3
-22.0, -27.3, -16.6, -5.2, 4.4, 13.1, 16.9, 11.8, 5.3, -5.9, -19.4, -28.0
-24.3, -18.9, -13.4, -1.3, 8.3, 12.2, 16.3, 12.5, 7.9, -3.6, -20.2, -24.6
-24.2, -20.7, -12.1, -1.6, 5.2, 12.3, 17.1, 13.6, 7.2, -6.0, -25.3, -27.2
-29.0, -24.8, -16.0, -0.2, 7.4, 14.3, 17.0, 13.6, 6.9, -1.1, -17.3, -34.8

At this point, I'm not as interested in digging into this as I was a few days back. The data is noisy. If one cherry picks start and end points for straight trend lines, one could draw many and varied trend lines that could support many and varied views on the data. The most commonly accepted way to smooth out the noise, to make longer term trends visible, seems to be rolling averages. This seems a more politically neutral and mathematically simple approach towards generating smoother curves. It doesn't seem constructive for us to engage in exchanging trend lines that begin or end at various points.

I'm switching back to an older version of Appleworks. The new iWorks has fancier graphics, but the old one gives more control of axis labeling. Anyway, what follows is the data for one particular station, with and without the five year rolling average. Other stations around Siberia will of course vary somewhat, but Perevoz seems more or less representative to show the points Justin and I are discussing.

Data from aisori.meteo.ru, plotted by myself

One can read the chart as fairly flat from 1990 to present. That flatness might be found in many stations around Siberia, judging from the semi-random sample of stations I've already looked at.

However, the warming trend is visible from 1970 to present, 1980 to present, and 2000 to present. In selecting 1990 as the start of his trend, Justin is selecting an outlier data point. When he looks at the world wide average, he chooses 1998 rather than 1990 to measure his trend, as 1998 is the greatest outlier favoring his point of view available on that chart. He also chooses the coldest point of the Little Ice Age and the few unmelting glaciers as his trend points.

There may also be a reason why 1990 was an outlier point. As the Pacific Oscillation (El Nino and La Nina) causes temperature swings in other parts of the globe, Siberia is dominated by a similar Arctic Oscillation. See washington.edu or NOAA for introductions to the phenomena.

A NOAA chart showing the warm (red) and cold (blue)
phases of the Arctic Oscillation

The Arctic Oscillation locked strongly into its warm mode in the early 1990s. This resulted in unusually warm temperatures in Siberia, generating Justin's outlier data. The warm mode of the Oscillation also generates a stronger prevailing wind pushing Arctic ice towards the Atlantic. In normal decades, when there is a good balance between the warm and cold trends of the AO, thick ice developed over many years has a better chance of staying in the north. In the early 1990s, in addition to the warmer temperatures, much thick ice was pushed into the Atlantic to melt as icebergs. This likely contributed to the Arctic ice cap melting much faster than the models predicted.

Some are concerned that the early 1990s AO locked in warm mode might be a symptom of global warming. (Note, the preceding was written in 1999, when the trend towards permanent warm mode was considerably stronger than it seems today.) Perhaps the AO might get stuck in warm mode full time. Alternatively, one can look at 1996 through the present and say that the AO has returned to a reasonably balanced mode. The early 1990s warm AO might well be just an outlier, not part of any long term trend.

Warm (left) and Cold (right) phases of the Arctic Oscillation
Twin globe illustration from J. Wallace,
University of Washington

I played about with spreadsheets a bit more. In green, above, we have the temperature trend line for Perevoz (3 year rolling average) with the effects of the Arctic Oscillation (no smoothing) subtracted out.

Nice, Bob. I like what you've done with the '-Oscillation' graph. It may not be very scientifically rigorous (although, on the other hand, it may be spot-on), but I like it, nonetheless. It has a good beat, and you can dance to it.

What makes me wonder, though, is that while removing the oscillation changes the 80's-90's-00's to more or less match the AGW orthodoxy, it does so at the expense of a BIG, FAST spike in the 60s. Followed by a pretty darn quick drop-off. I certainly hope there's a way to account for that...

Originally Posted by Justin '77

Nice, Bob. I like what you've done with the '-Oscillation' graph. It may not be very scientifically rigorous (although, on the other hand, it may be spot-on), but I like it, nonetheless. It has a good beat, and you can dance to it.

As the Arctic Oscillation is measured in terms of pressure, not temperature, I wouldn't claim my subtraction is scientifically rigorous. Subtracting millibars from degrees isn't really legitimate. It is just coincidence that the scale of the millibars on one graphs matches the scale of the degrees on the other chart to produce a good curve fit. By all rights I should have had to factor in a scaling factor in order to make it work.

Still, it does have a good beat. You can dance to it.

Originally Posted by Justin '77

What makes me wonder, though, is that while removing the oscillation changes the 80's-90's-00's to more or less match the AGW orthodoxy, it does so at the expense of a BIG, FAST spike in the 60s. Followed by a pretty darn quick drop-off. I certainly hope there's a way to account for that...

Getting rid of the spikes can be reasonably done. The temperature data was already shown as five year rolling average. On the last version of the chart, the the AO data was raw... No rolling averages. On the following chart, both the temperature and AO have five year rolling averages.

There remains a decade wide surge in temperature centered around 1961. It is clearly not caused by the Arctic Oscilation. However, there is a 1961 surge in the global temperature too.

OK. So the 1961 bulge isn't the Arctic Oscillation, and it doesn't mesh well with the global warming theory of steady addition of greenhouse gasses. If we are rounding up all the usual suspects...

Hmm... The highest peak in the sunspot record is right in the neighborhood of 1961. Solar forcing might be playing a role?

And this is just one station. I do tend to see temperature bumps at other stations. I might redo some of the other stations with rolling averages to see if they also show a similar 1961 bulge.

But this reflects my approach that there are many factors involved in climate. If you show a raw temperature chart, one shouldn't expect every single peak and valley to be caused by the same thing. I suppose the next thing I could try to do is to enter the sunspot record onto the same spreadsheet and try to find a way to subtract the solar forcing effect out of the raw temperature rise, and see if the result can be made flatter.

But I'm not feeling that ambitious just now.


Edit : I went to see if other cities corrected for the Arctic Oscillation would beat and dance as well as Perevoz. Assuming the other Siberian stations are effected similarly by the AO, they ought to. Alas, http://aisori.meteo.ru/climat is still down.

One more look at Perevoz. (http://aisori.meteo.ru/climat is up, at least for the moment, so I grabbed Ika. I may try a similar chart for Ika.)

It seems that the Siberian 1990s warming is more complex than my recent posts allege. As I'm reading the above chart, from the 1950s through the mid 1980s, the solar cycles were pulling in the opposite direction from the Arctic Oscillation. In general, the solar cycles seem to have been the stronger influence.

In the late 1980s, the two forcing factors came into phase. Instead of fighting one another, the two influences stacked on top of each other. As a result, there have been some huge swings in temperature. The warm trend in the early 1990s looks like an alignment of a reasonably typical peak in solar activity combined with a really big warm bump in the Arctic Oscillation.

The short solar cycle is generally fairly predictable. The Arctic Oscillation is far less so. The two forcing factors could move up together again, or they might not.

Justin describes Siberian temperatures as relatively flat or slightly declining since 1990. I would suggest, if Perevoz is typical, that the Siberian temperatures have been wildly swinging since the two forcing factors came into synch in the mid 1980s. Justin can show a downtrend in temperatures if he starts a trend on a peak of both forcing factors and ends it on a trough of both forcing factors. The Perevoz record at least suggests that AO and solar working together can overcome greenhouse in the short term.

But if you measure 1986 to present, looking at the trough to trough trend rather than the peak to trough trend, there is a warming factor coming from somewhere other than solar or AO.

At the hand waving level, one can argue from the above graph that solar and AO effect temperature. I attempted to go a step beyond, creating a spreadsheet climate model. It would be nice if by subtracting off solar and AO effects from the red line, one could get a nice smooth curve showing a slow steady greenhouse warming trend, with no bumps around 1960 and 1990.

Perhaps it can be done. My first attempts were disasters. At this point, I think the pretty looking beat and dance effort shown on yesterday's charts should be treated more as luck than as rigorous and sound.

All curves on the above chart have a 5 year rolling average. The Perevoz temperature source is to the best of my knowledge raw, with no corrections for station moves, equipment changes or anything similar. The three curves on the chart are in different units of measurement, sunspot count, average pressure, and temperature. Don't treat the numbers on the Y axis seriously. The shapes of the curves ought to be good, but the absolute values don't count for much.

I got it to work, sort of. I selected about half the stations, chose month, and then selected year, station and July. I used the interval 1960-2006. I then saved it as a text file with notepad and imported into Excel as a semicolon delimited file. I had three columns of data, the station number, the year and temperature, presumably for July. I don't know if this is an average, I hope so.

I then sorted the table by temperature and removed all the blank cells. I then sorted this by year and and used boolean instructions to calculate a yearly average of the temperature from all the station.

Year Temp #stations
1960 15.5 227
1961 15.1 247
1962 14.7 250
1963 14.5 253
1964 15.2 251
1965 14.2 253
1966 14.5 252
1967 15.1 251
1968 12.9 253
1969 14.7 249
1970 15.0 252
1971 14.5 249
1972 15.3 251
1973 14.3 251
1974 15.8 248
1975 14.6 253
1976 13.7 246
1977 14.6 249
1978 13.4 251
1979 14.1 249
1980 13.5 251
1981 15.0 241
1982 14.7 246
1983 14.8 249
1984 14.9 252
1985 13.5 251
1986 13.8 251
1987 14.2 252
1988 16.3 251
1989 15.1 246
1990 15.1 251
1991 15.7 240
1992 14.0 248
1993 15.2 241
1994 14.5 238
1995 14.9 232
1996 15.1 230
1997 13.6 224
1998 16.3 226
1999 15.9 227
2000 15.7 235
2001 16.7 229
2002 16.7 220
2003 16.2 237
2004 16.0 235
2005 15.8 240
2006 14.7 240

I ran a regression on it. There is a positive slope (2.6 deg / century) significant at the 99.5% level:

Regression Statistics
Multiple R 0.406
R Square 0.165
Adjusted R Square 0.146
Standard Error 0.818
Observations 47
Significance 0.0046

The problem with this simplistic analysis is I average all the stations together. Not all of them have data for a given year, which is why the number of stations varies from 227 to 253. This will introduce error which ought to be random if the reason why data is missing is random.

If I can get in again I am going to do this again with all the stations and see if the results change at all. By changing the sample size in various ways, I can test whether the results are sensitive to missing stations.

Originally Posted by Mikebert

I got it to work, sort of...

I thought I'd plot what you got. The top two curves are two major forcing factors for Siberia. Sunspots are a proxy for solar forcing, in brown. The Arctic Oscillation is in blue.

Red is the temperature at Perevoz, with 3 degrees added just to place the curve in a pretty place on the chart. Green (Ika) is another station further north. I needed to add a larger number in order to place the curve near Perevoz. Ika and Perevoz are two stations selected by Justin, in part to emphasize the warming around 1990. These stations are in Siberia proper, while the web site Justin provided includes stations from all over Russia.

Mike's large collection of sites are broader based, but clearly they average out to be further south than Ika and Perevoz. I had to subtract 17 degrees, rather than add 3 or 4.

I note Mike's chart matches Ika and Perevoz reasonably well before 1970 and after 1997, but there is a larger gap between Mike's curve and the other two in the early 1990s. My theory is that Mike's wide selection of stations, outside of Siberia and further south, are not as strongly effected by the Arctic Oscillation.

A while ago, in post 1775 of this thread, Justin provided a list of 133 stations in Siberia proper. These stations will likely be in the area where the Oscillation is strongest. He was going to work a smaller list of station close to and just north of the Permafrost line. Mike, you might want to try to work Justin's list. Justin might want to share his shorter list.

Originally Posted by Mikebert

The problem with this simplistic analysis is I average all the stations together. Not all of them have data for a given year, which is why the number of stations varies from 227 to 253. This will introduce error which ought to be random if the reason why data is missing is random.

I vaguely recall a number of stations vanishing from the record about the time the cold war ended. I don't know that the missing data will be totally random, but is apt to have military, political and economic causes. I don't know that this would really introduce a bias, though I'm particularly interested in the early 1990s, which is just when a bunch of stations vanished.

Gulf News reports Asian glaciers melting fast For discussion purposes.

Originally Posted by Gulf News

Way above us in the Himalayan cloud are jagged, snowbound peaks Annapurna, Damodar, Gangapurna, Dhalguri.

Below us is the Thulagi glacier, a river of ancient ice snaking steeply down the Marshyangdi valley from near the top of Mount Manasulu.
Thulagi's snout is a milk-blue lake marked on few maps. It has doubled in size in just a few years and is held back only by a low wall of dead ice and earth. If Thulagi carries on melting at the present rate, nothing will stop billions of litres of water bursting through this natural dam and devastating villages, farmland and everything below.

Thulagi is one of 20 steadily growing glacial lakes in Nepal which mountain communities and scientists fear will inevitably rupture if the growth in greenhouse gas emissions is not stemmed by world leaders at the Copenhagen climate summit.

I've heard stories of long ago glacial lakes being released by huge melting ice dams. I've always wondered what such a flood would be like. It seems that we might be about to find out...

A DU thread that is amusingly relevant (The Amazing Randi's amazing global warming denial).

6. Side question: why are engineers so near-universally conservative?

Is it because they usually approach a problem thinking that there's just one tiny component at fault? Because they think political problems are the exact same thing as technical problems, and have a simple black-white answer? I have worked in I.T. for two different software companies and this even extends to software engineering: those guys are 99.9% Republican, lapping up the simplistic blame-somebody politics of today's GOP.

7. You might also ask why climate change denial is so rampant in organized skepticism

I think it's classic Dunning-Krueger. Engineers (and skeptics) understand some very complex things. Therefore they thing they're equally capable of understanding everything, or that logic and rationality alone are enough to make an informed decision on any given topic (even if it requires advance study). I actually saw someone make this exact claim in a response to Phil Plait's take on this situation on his Facebook wall.

I'm afraid I'm one of the vitriolic ones in this whole thing. Randi's non-apology doesn't really take back anything he said, other than he should have looked at The Petition Project more closely. Otherwise it's just "Hey, I was just asking questions and being provocative." Bullshit. That's what Fox News says. It's not what the most prominent leader of movement skepticism should say. Hell, Penn Jillette's apology on his AGW denialist positions was more satisfying. At least he said, "You shouldn't be listening to me."

On topics that require advanced education to reasonably understand, skeptics, unless they have that education, should not be making claims in those fields. Skeptics can pick apart logical fallacies and point out common errors in reasoning but that's the extent of their ability to dissent from scientific consensus. In other words, skeptics should take the scientific consensus as the default position, requiring dissenters to provide the extraordinary evidence that majority of mainstream scientists in any given field are wrong.

And I'm sorry if I sound vitriolic, but Randi should know better. He blew it big time here.

8. Regarding the engineers and skeptics

I've previously mentioned a friend who's doing doctoral work in soil ecology and microbiology, and he told me flat-out last year that humans have nothing to do with global warming. He justifies this claim by asserting that he's had lots of practice assessing data, and he's looked at the figures.

Well, I don't have the background to challenge him directly, so I usually nod and make noncommittal comments and then change the subject. Nothing I can possibly say will change his thinking, because he's already decided what the answer is.

I know a few engineers who are fiercely intelligent yet almost comically right-wing in their politics. It's truly baffling.

9. I wonder that too, seen it my whole career.

I've spent my whole career around aerospace/electrical engineers. A depressing number are creationists who believe that since they have to design electrical circuits, SOMEBODY had to design the Universe.

From WH's post, this is a giant red flag: They do have 10,000 engineers!!! 2,600 mechanical engineers!!

IOW, engineers trying to apply scientific principles totally outside their fields of competence. I think that is the #1 mistake engineers make, when they go wandering off into the fields of biology, climatology, etc.

And the #2 reason, which you hit on with: "Because they think political problems are the exact same thing as technical problems, and have a simple black-white answer?"

IMO, sort of. Many engineers seem to believe in a near-woo-wooish idea of Universal Engineering Principles, which operate even in non-engineering disciplines like...uh...evolutionary biology.

(Though that notion certainly doesn't operate in their own field, since we don't see civil engineers designing airplanes or electrical engineers designing bridges. Fortunately...)

Just the other day on these very internets, I read a long dissertation by a mechanical engineer on the human skeletal structure. He said that he could never design anything so complex, therefore the human skeleton was a bio-mechanical miracle, hence it had to be designed by...Guess Who.

http://www.alaskajournal.com/stories.../loc_cce.shtml

http://www.adn.com/3437/story/1048216.html

(Interesting idea that oil companies would spend money to solve a nonexistent problem.)

http://blogs.nationalgeographic.com/...northeast.html

(Of course Greenland isn't warming. This guy is imagining things. One can see this -- if one squints -- from data carefully selected from Siberian temperature measurements.)

http://www.expatica.com/de/lifestyle...er-_15221.html

(Hey, you know these illiterate reindeer-herders, they don't know Jack except what their grandparents told them, and you know how old people make crap up.)

Bob, I appreciate all your efforts to plow through Justin's data blizzard (that's the expression I was looking for earlier!), but here's a simple rule of thumb: there's something wrong with any set of data that tries to claim the sky isn't blue. Maybe it's the data itself, or maybe it's the way we're interpreting it, or the way it's being presented. The evidence that the planet is warming is not confined to temperature measurements, and so trying to use a limited set of temperature measurements to claim that it isn't has the result (if accepted) of leaving a whole slew of secondary phenomena lacking an explanation.

Myself, I'm of course unsurprised that you found a warming trend from the data Justin linked, because that was the only likely outcome, but I'm also unsurprised that Justin is finding excuses to dismiss it -- and that's why I was unwilling to go to all that work myself. But then, you've done a better job, certainly with better graphics, than I would have, so all's well.