I’ve reposted this here in entirety with permission from Pierre Gosselin of “No Tricks Zone“, and it is well worth the read. Much of this work was inspired by posts that have appeared on WUWT. Ed Caryl has done a great job pulling various threads of info together. One generally doesn’t think of any Arctic circle outposts as being “urban” but the fact is that islands of humanity, essentially small towns, surround many of these stations. And in the Arctic, you produce a lot of energy (which has to go somewhere) or you die. What I find most interesting is the plot of “isolated” stations versus the Atlantic Meridonial Oscillation (AMO); a clear correlation of the driver for those temperatures.. – Anthony
By guest writer Ed Caryl
Arctic stations near heat sources show warming over the last century. Arctic stations that are isolated from manmade heat sources show no warming. The plots of “isolated stations” and “urban stations” below clearly illustrate the differences.
Stevenson Screen, Verhojansk, Russia
All the GISS temperature anomaly maps show the Arctic warming faster than the rest of the globe, especially northern Alaska and Siberia, but the satellite data shows a different pattern. See the 2 charts for 2009 that follow. The GISS surface map:
Satellite chart:
The baseline period selected for the GISS surface temperature chart is the 1933 to 1963 Atlantic Multi-decadal Oscillation (AMO) warm period. This period more closely matches today’s temperatures than the default 1951 to 1980 cool period that GISS uses. The satellite data uses the average over the satellite period since 1979, the modern warm period.
The satellite data show cooling in central Siberia, similar to the surface anomaly map, and very little warming for most of Alaska. It also shows cooling for the Antarctic Peninsula, where the surface map shows warming. But there is a scattering of hot grid squares across the HISS surface station map for the Arctic. So what is going on?
I selected the stations that correspond to those warm grid squares, as well as other stations in the same latitudes. In this age of everyone carrying a camera posting all photos on the Internet, there is a lot of information available on these stations. For some I could locate the Stevenson screens, for most I’ve found pictures of the surroundings, while others have investigated many of these sites already, and so links to that research are included. I downloaded the raw temperature data from GISS for 24 stations closest to the North Pole, which are all classified as “rural”.
“Urban” Arctic Stations
Contrary to GISS claims, many of these stations are actually not “rural” with respect to their siting quality. Many are at airports associated with sizable towns or research stations with sizable staff and infrastructure. In the Arctic, any town of more than a few families can be a large heat source. In the case of many towns in Russian Siberia, “central heating” takes on a whole new meaning. These towns have a central power plant that provides electricity and steam heat to the whole town. Large pipes, both insulated and un-insulated, carry steam, water, and sewage, up and down the streets to and from each dwelling. These pipes cannot be buried because of the permafrost, so they are elevated, and at street crossings are elevated 4 or 5 meters. The temperature differential between these pipes and the surrounding air can be 140° C in winter, and even more for a pressurized system.
But GISS applies the same Urban Heat Island (UHI) criteria to all stations globally, regardless of the latitude or average temperature. They look at the satellite night brightness and population to judge whether urban or rural. By GISS criteria, all the stations in the high Arctic are rural; there are no corrections for UHI.
But let’s look at each of these “urban” locations. Each name is also a link to the GISS surface temperature raw data.
List of Urban Arctic Stations (see the annex at the end of this post for details on each station)
1. Kotzebue, Ral (66.9 N,162.6 W), Alaska
2. Barrow/W. Pos (71.3 N,156.8 W) Alaska
3. Inuvik (68.3N, 133.5W) Inuvik, Canada
4. Cambridge Bay (69.1 N,105.1 W) Nunavut, Canada
5. Eureka, N.W.T. (80.0 N,85.9 W), Canada
6. Nord Ads (81.6 N,16.7 W Northeast Greenland
7. Svalbard Luft (78.2 N,15.5 E), Norway
8. Isfjord Radio (78.1 N,13.6 E), Norway
9. Gmo Im.E.T.(80.6 N,58.0 E), Russia
10. Olenek (68.5 N,112.4 E), Russia
11. Verhojansk (67.5 N,133.4 E), Russia
12. Cokurdah (70.6 N,147.9 E), Russia
13. Zyrjanka (65.7 N, 150.9 E), Russia
14. Mys Smidta (68.9 N,179.4 W), Russia
15. Mys Uelen (66.2 N,169.8 W), Russia
The following graphic is a temperature chart for 10 of the above stations (5 of the shorter ones were left out to avoid over-crowding). All are warming, some faster than others. Barrow, for which we have the UHI study, is not the fastest warming.
Temperature trends of the “urban” stations.
Isolated Stations
Now let us look at the isolated stations, which are located at similar latitudes like the above “urban” stations. One important thing to note about these isolated stations – there is limited electrical power, and so incandescent light bulbs in the Stevenson screens is unlikely. Detailed descriptions of these stations are listed in the annex at the end of this report.
16. Alert,N.W.T.(82.5 N,62.3 W), Canada
17. Resolute,N.W. (74.7 N,95.0 W), Canada
18. Jan Mayen (70.9 N,8.7 W), Canada
19. Gmo Im.E. K. F (77.7 N, 104.3 E), Tamyr Peninsula, Russia
20. Ostrov Dikson (73.5 N,80.4 E, Russia
21. Ostrov Kotel’ (76.0 N,137.9 E), Russia
22. Mys Salaurova (73.2 N,143.2 E), Russia
23. Ostrov Chetyr (70.6 N,162.5 E), Russia
24. Ostrov Vrange (71.0 N,178.5 W) , Russia
Now here is the chart of the temperatures of these isolated stations, not subjected to manmade structures or heat sources.
Isolated Stations
Note that most of the trends are flat or decreasing. Only Resolute and Ostrov Vrange are increasing slightly. Both of those might be slightly influenced by UHI. The longest records clearly show warming in the late 1930’s and 40’s, and cooling in the 1960’s, and none show a hockey stick. The GISS data for Alert ends in 1991, though the weather station is still there, and still reporting. Data for Mys Salaurova and Ostrov Chetyr also ends at about that time, probably due to the fall of the Soviet Union.
Here is an average of all the isolated stations:
Isolated Stations – Average
Note that the peak-to-peak trend is nearly zero. The linear trend is about 0.4°C/century, but the R2 value (the statistical significance for the trend) is very low, 0.023.
Here is a plot of the AMO versus the average temperature of the isolated stations.
The temperature as measured at stations isolated from any UHI is simply tracking the AMO.
Looks like an awfully good fit. There is very little, if any, global warming. We need to wait until the bottom of the next AMO cycle to get a decent reading of global temperature change. That will be in about 2050 if the AMO cycles as it has since 1850.
———————————————————————————————-
Annex – station descriptions
The “urban” stations, nos. 1-15
1. Kotzebue, Ral (66.9 N,162.6 W),
2. Barrow/W. Pos (71.3 N,156.8 W)
These towns are of similar size, and are growing at the same rate. In 1940, both towns had a population of 400. In 1980 both had just over 2000 population, and now they both have over 3000 people. Both have airports of sufficient size to handle multi-engine turboprop and small jet aircraft, and both are served daily by regional airlines. Kotzebue is on a peninsula and the airport is across the middle of the peninsula, somewhat restricting the growth of the town. Barrow has somewhat the same problem due to a series of small ponds around the town and the airport. Barrow was studied for UHI effects in 2003. That paper was in the International Journal of Climatology here. That paper describes the UHI average temperature increase in winter as 2.2°C compared to the surrounding hinterland. GISS data indicates that Barrow average temperature has increased over the years as population has increased. (See below, or click on link above.)
Barrow, AK
Source: http://en.wikipedia.org/wiki/File:BRW-g.jpg
Source: http://en.wikipedia.org/wiki/File:OTZ-g.jpg
The Barrow NWS station (Stevenson Screen) is here. On the airport picture, it is at the base of the rotating beacon tower. Kotzebue NWS station is not visible in published pictures.
Inuvik is a relatively new town, begun in 1954. The population as of 2006 has grown to about 3500 people. Because it is a “planned” community in the arctic, built on permafrost, the water and sewage infrastructure is above ground in heated and insulated “utilidors”, like the heating systems in Siberia. The weather station, from weather reports, Google Earth and Google Street View, appears to be at the airport, in a compound just north of the entrance.
4. Cambridge Bay (69.1 N,105.1 W), Cambridge Bay, Nunavut, Canada
There is a Wikipedia picture of Cambridge Bay here. The population has grown from just a few people in the 1940’s to about 1500 today. It also has an airport with daily regional airline service.
5. Eureka, N.W.T. (80.0 N,85.9 W), Eureka, N. W. T., Canada
There are the only four stations at or north of 80° latitude, Eureka, Alert, Nord and Krenkle (Gmo. I.M.ET). Only Eureka has an unbroken temperature record to the present date, and it begins in 1947. The population at Eureka has
Eureka station
never been high. In winter it has always been 4 or 5 men. In summer, the population increases to as high as 20. The station infrastructure though, has expanded through the years. Each year, some of those 20 workers add or expand buildings. In the beginning, it was one or two buildings, with water and sewage handled in tanks and barrels internal to the buildings. The Stevenson screen was originally placed where the blue New Main Complex building is now. When that was built, the Meteorological instruments were moved to the current location. Over time, the water supply, plumbing, and sewage treatment was upgraded and the outfall pipe installed. It, of course, must be heated to facilitate flow to the sewage lagoon. All the water pipes exposed to the outdoors must be heated to prevent freezing.
Image from a recent article by Anthony Watts on WUWT here.
6. Nord Ads (81.6 N,16.7 W, Northeast Greenland
Nord Station
Nord is the furthest north inhabited place on earth, on the Northeast coast of Greenland. It was built in the period from 1952 to 1956 as an emergency airfield for aircraft operating out of Thule. Access is impossible by sea because the sea ice never moves away from the coast there. Legend has it that “Blowtorch” Murphy, a mythic arctic construction worker, scraped the first runway, using a parachute dropped caterpillar tractor after he himself parachuted onto the site. His nickname came from his habit of wearing a lit blowtorch hanging from his waistband on a wire; a lit blowtorch being somewhat useful when working outside when it’s 40° below zero.
There are about 40 buildings at Nord. Not all of them are continuously heated, but those near the Stevenson Screen are. The winter population is 5 or 6 men. More pictures here.
7. Svalbard Luft (78.2 N,15.5 E)
8. Isfjord Radio (78.1 N,13.6 E)
These two stations are only 47 kilometers apart. But data for both is fragmentary for 1976 and 1977, and there is no overlap. Svalbard Luft (airport) has been discussed on WUWT here and here, so I won’t cover it in detail here. Warwick Hughes has an article on Isfjord Radio here that makes the case for warming of Isfjord Radio due to moving of sea ice away from the islands in summer since 1912. Neither station in Svalbard shows on the anomaly map because there was no common station in both the base period and the anomaly period. Here’s a map with 1998 to 2008 as the base period where Svalbard appears.
This is the Krenkel meteorological station on Hayes Island, or Ostrov Kheysa in Russian, in the Franz Josef Land Archipelago, Russia. Link The station has been moved or re-built twice since it was established. It was moved from Hooker Island (article in German) in 1957/58. A fire destroyed the power station in 2000, and it was rebuilt in 2004 closer to the shoreline. The GISS record is from 1958 with a gap from 2001 to 2009. The population was as high as 200 during Soviet times, but is down to 4 or 5 now. The population and the temperature seem to track roughly during Soviet days, and the move in 2004 was to a warmer location. In the picture you can see the old buildings on the ridge in the distance. The red grid-square on the anomaly map above corresponds to this station.
Source: http://www.sevmeteo.ru/foto/15/88.shtml
This is the town of Ust’-Olenek, Russia.
The town doesn’t look like much, but notice the Tundra Buggies parked next to the Stevenson Screens. It is on the Laptev Sea, on the northern Siberia Coast, but on a peninsula on a south-facing beach. The buildings are right on the shore. The wide view above was taken from out on the ice. This is one of the few places in Russia that the Google Earth satellite view actually has enough resolution to see the Stevenson Screens. They are much too close to the heated building.
11. Verhojansk (67.5 N,133.4 E)
This is one of the “centrally heated” towns in Russian Siberia. The picture at the top of this article is of the Stevenson Screen. Verhojansk is called the “cold pole” of the earth, but the measurements are too warm by far. Look closely at the picture. Any photographer will note that the warm glow inside the Stevenson Screen is just the color temperature of an incandescent light bulb. If the steam heat in the town isn’t enough, or the cattle in the pole-barns in the distance, the heat from the light bulb will warm up the measurements. This site was covered on WUWT here and here. Anthony Watts notes that warm anomalies would appear and disappear in this part of Russia “as if a switch were thrown”. Could it be as simple as the switch on that light bulb?
Also spelled Chokurdakh. The population has been dropping in recent years, but was still over 2500 people in 2002. The town is sandwiched between the Indigirka River and the airport. There is no way to tell where the Stevenson Screen is located, but the infrastructure at the airport blends right into the town. See an aerial photo here.
13. Zyrjanka (65.7 N, 150.9 E) Also spelled Zyryanka, another steam-heated town in eastern Siberia, well inland. The airport is in this picture on the north edge of town, along the Kolyma riverbank. This airfield was built during WWII as a stop for aircraft being ferried to the Soviet Union from Alaska. A second airport 7 miles west of town was probably built during the cold war for the military. The town was established in 1931. The population is currently about 3500. During the Soviet Union it was up to 15,000.
14. Mys Smidta (68.9 N,179.4 W)
Or Cape Schmidt. John Daly wrote a bit about this location in 2000 (scroll way down in the article). The population was nearly 5000 in 1989, but has dropped since the fall of the Soviet Union. The population now is probably less than 1000. It is on the north coast of eastern Siberia, nearly at 180° longitude. The airbase there was built in 1954 as a staging base for any bombers headed for the U. S. It is still used by a regional airline.
15. Mys Uelen (66.2 N,169.8 W)
Or Cape Uelen. This is on the easternmost tip of Siberia, across Bering Strait from Kotzebue, Alaska. The current population is about 700 people. It is also centrally steam heated. The town is restricted by the geography, on a narrow spit sticking out into the sea, backed by a cliff on the landward side. The airport is a helipad. Cargo and fuel arrives by barge in the summer.
Below is a temperature chart for many of the above stations. All are warming, some faster than others. Barrow, for which we have the UHI study, is not the fastest warming.
16. Alert,N.W.T.(82.5 N,62.3 W), Alert, Canada
Alert, Canada has had a weather station since 1951. The population has never been more than 4 or 5 in the winter, with a higher population in the summer. I could not definitively locate the Stevenson screen, but there are two possibilities in this photo, both well away from the buildings.
THE ISOLATED STATIONS, NOS. 16-24
17. Resolute,N.W. (74.7 N,95.0 W)
The population of this Canadian station rose from zero prior to 1947, to 229 in 2006. There is an airport here, and the Stevenson Screen can be seen across the aircraft parking area from the airport terminal at the left edge of the photo.
Pictures of the station are here, and a web site is here. The 18 people on the island live at Olonkinbyen, or Olonkin “City”. The meteorological station is 2.6 km away. The 4 people that work there live in Olonkin City. The Stevenson Screen appears to be well away from the station building, and the surroundings have probably not changed since the station was built.
19. Gmo Im.E. K. F (77.7 N, 104.3 E)
This is a Russian station on the Tamyr Peninsula at Cape Chelyuskin (Mys Chelyuskin). Nothing is visible at that location on the Google satellite view, but the resolution is very low. I found an article by Warwick Hughes dated September 2000 that speaks of cooling of the Tamyr Peninsula here. He also talks about “non-climate” warming of Verhojansk and Olenek.
20. Ostrov Dikson (73.5 N,80.4 E
Dikson, Russia airfield
This is Dickson Island in English. There is a town of Dikson 10 kilometers away on the mainland. The airport is on Dikson Island at the point called Ostrov Dikson on the map below. Pictures of the airport can be seen here. The town is pictured on this 1965 stamp.
Wikipedia link
21. Ostrov Kotel’ (76.0 N,137.9 E)
The full name is Ostrov Kotel’nyy. In English this is Kettle Island. The first documented explorer found a copper kettle, so obviously he was not the first person to find the island. A single building is barely visible on Google 3D mapsat the “settlement” known as Kalinina. This may be the meteorological station. No other signs of civilization can be seen on the whole island.
22. Mys Salaurova (73.2 N,143.2 E)
This also spelled Mys Shalaurova. The station is on the south-facing shore of an island and is visible on Google Earth here. There is a tide gauge, and the tide data is on that same page.
23. Ostrov Chetyr (70.6 N,162.5 E)
The full name is Ostrov Chetyrekhstolbovoy. This is a small island in the East Siberian Sea in the Medvezhy Island (Bear Island) group.
Map source here.
A description of the place is found: here. “A polar meteorological station and a radio station are situated on the shore of a small bay which indents the S side of the island.”
24. Ostrov Vrange (71.0 N,178.5 W)
This is otherwise known as Wrangle Island. It is about 125 kilometers off the Siberian coast on the 180th meridian. The weather station is at Ushakovskiy on a spit at Rogers Bay, at the right in this picture, well separated from the village. One building in the village is visible at the left. Link
The population in the village grew to as many as 180 people in the 1980’s, but when the Soviet Union dissolved, subsidies declined and the population moved to the mainland. The last villager was killed by a polar bear in 2003. The population at the weather station, when occupied, has always been 4 or 5.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Al Tekhasski says:
evanmjones wrote: “So we would need over 120,000 stations? That’s a lot of stations.”
Sure it is. But I am afraid you might need more. The example of stations 50km apart having opposite long-term trends means that we don’t know what trend is in between, and what is around in the same proximity. […]
More, we still have no idea if the 25x25km is enough to capture complexities of local micro-climates, so be prepared to another half-scale, which would quadruple the number of necessary stations. Without this uniform sampling grid of data it is not serious to discuss any mathematics of subsets or else. This is what physical science says. Sorry.
Al, you may like this article where I look at some of the mathematical issues of sampling surface temperature. As topology is fractal (mountains, coastlines, etc.) the temperatures from them ought to also be fractal. A black pebble next to a snow melt stream will have quite divergent temperatures… Measuring a fractal gives different answers based on the ‘size ruler’ you use. And we’re measuring ‘climate change’ with a ruler who’s size constantly changes over time.
http://chiefio.wordpress.com/2010/07/17/derivative-of-integral-chaos-is-agw/
Ben D. says: I will not say that the anomaly approach is incorrect, but there are issues with it as well. As far as I can tell, its the best method known right now, but it is not perfect. To argue that its the end-all is kind of ignoring the issues that it also brings up.
Very well put. Also, there are many kinds of anomaly method and they have many different modes of failure. One of the simplest to ‘get’ is that of the splice artifact.
It doesn’t much matter if you use anomalies or not, when you take a station that warms 1 C as it grows, then in a later decade add a new station that warms 1 C as it grows, then in the final decade swap to a third station that warms 1 C as it grows. Tack them all together and you a 3 C “warming trend”. It matters not if this is done via averaging, direct splicing, or “homogenizing” using each to “adjust” neighbors.
The temperature series codes like GIStemp are FULL of that. And anomalies make it easier to have happen rather than harder. (No station has to reach unheard of record highs and call attention to itself…) I saw this effect in the region near Marble Bar Australia where an all time ever record was set back near the ’30s and never exceeded, yet the ‘region’ has a ‘warming trend’.
So that’s why I periodically anchor myself back in ‘real temperatures’ and why I start by looking at the profile of ‘real temperatures’. I’ve taken a great deal of flack from folks asserting that it is sheer stupidity to do that, and they are wrong. It is only stupid to think that they show accurately the temperature trend. Just as it is stupid to think that averaged anomalies show the accurate temperature trend if for no other reason than ‘splice artifacts’. IMHO, the ‘splice artifact rich’ nature of the ‘homogenizing’ done is a target rich environment in the temperature series codes.
But wait, there is more…
To clarify, anomalies are based on the area-weighted global average,
For one kind of anomaly…
The climate codes use a “grid / box to grid / box” anomaly. They have one set of thermometers in the box at the baseline and a different set now. This is horridly broken. It would be like me saying cars have gotten faster as my home “grid / box” had an average VW fastback in 1970 and has and average Mercedes SL now and the “max speed anomaly” has risen by 55 mph.
Yes, they take steps to mitigate the problem. But mitigation is not perfection. We are basically betting the global economy on the perfection of their mitigation and coding (and their coding is fairly sloppy.)
So I started by looking at plain temperatures, and found things that were not in keeping with AGW and CO2 theories. Then moved on to anomalies. But I wanted a more controlled beast. So I do anomalies only “self to self” for a given thermometer.
This brings up the issue of ‘baseline’. But you don’t need a baseline to do anomalies. ONE kind of anomaly is based on a common period of time, the baseline. But as Steve Mosher points out, you need a common time period in your baseline. And many thermometers don’t have it. So a whole load of ‘homogenizing’ and ‘splicing’ (that GISS calls joining) and box to box and infill and… well, junk… gets done to try to make a complete enough record to use a ‘baseline’ method. And IMHO it adds too much error to be usable to 1/10 C. But you can use non-baseline anomalies, such as First Differences. And I use one like that (but fixing an issue in First Differences that makes it fail on data with lots of gaps in it.)
To put it simply, you can use this approach in the data above and it will probably change what you see simply because of the transformation of the data so to speak.
BINGO! And that is what the temperature data codes like GIStemp do. They change the data. So we end up with the past cooling by whole degrees…
But I must also interject here and say one thing: If everyone uses the same method and that is ALL they use, how would we know this method is actually “correct”. I might be playing devil’s advocate there, but at some point I will take a shot at adjusting the data myself and the first thing I would do is NOT use the anomaly system.
You have it exactly right. The three major labs all use the same basic approach, data, and methods with all the same flaws. Any attempt to look at it from a different angle gets rocks thrown at you (though it does point up their flaws…). And yes, starting from the basic temperature data gives you the context to know when something is straying from reality.
http://chiefio.wordpress.com/2010/04/03/mysterious-marble-bar/
And a whole lot more in:
http://chiefio.wordpress.com/category/dtdt/
Re: “the warm glow inside the Stevenson Screen is just the color temperature of an incandescent light bulb”
I am a skeptic because I see lots of problems with the the data used by the “climate disruption” camp, but in the interest of complete honesty I feel compelled to point out an alternative source for the illumination.
The gentleman on the step appears to be taking a picture of the inside of the enclosure, and the light could very well be coming from the focus-assist LED. They can be quite bright. I took a very charming picture of my granddaughter last Christmas in which she is squinting because it, and well illuminated by that characteristic “warm glow”.
Thanks for bringing back one of my favorite pictures ever, the Stevenson screeen at Verhojansk. I just love the light bulb inside. No spurious warming from that…
Briney
I believe I have seen a picture from a different angle and it is in fact a light bulb.
“”” Briney Eye says:
September 23, 2010 at 12:15 pm
Re: “the warm glow inside the Stevenson Screen is just the color temperature of an incandescent light bulb”
I am a skeptic because I see lots of problems with the the data used by the “climate disruption” camp, but in the interest of complete honesty I feel compelled to point out an alternative source for the illumination.
The gentleman on the step appears to be taking a picture of the inside of the enclosure, and the light could very well be coming from the focus-assist LED. They can be quite bright. I took a very charming picture of my granddaughter last Christmas in which she is squinting because it, and well illuminated by that characteristic “warm glow”. “””
Well I have never seen a digital camera on which the “Focus assist” LED was anything but a RED LED; and if you look more closely you will see there isn’t any light on the open door outside the Owl box. That certainly isn’t a Whie LED or even a Warm white LED, and it is definitely NOT the color of an AlINGaP yellow/Amber LED; which although extremely bright, is visually very close to a Sodium yellow (589.0/6 nm).
And hitting that exact yellow color is extremely difficult, since the human eye sees only about a 5 nm range of wavelength to be yellow. Outside of that it would be gold on the long wavelength end, and grellow on the short wavelength end; so making yellow LEDS with a standard colr is quite a chore so manufacturers aren’t going to mess with the color; unless somebody wants to pay for several tons of LEDs.
Long and the short is, it isn’t a camera LED.
“”” Maud Kipz says:
September 23, 2010 at 11:13 am
George E. Smith says:
September 22, 2010 at 2:28 pm
“”” Maud Kipz says:
September 22, 2010 at 9:08 am
the R^2 value (the statistical significance for the trend) is very low, 0.023
This betrays a fundamental misunderstanding of statistics. The R^2 value is a measure of linear dependence between two (random) variables, and nothing more. A trend with extremely high significance may still have a low R^2 if the trend is non-linear or if the observations are noisy. “””
All of which is very nice; as are the computations of any other (completely fictional) branch of Mathematics.
But NONE of it, establishes ANY Physical cause and effect relationship.
You can carry out the very same statistical analysis on the numbers in your local telephone directory; and derive the same quantities; and it still means nothing; unless the average telephone number in the book, happens to be your telephone number.
I don’t understand your joke about branches of mathematics. But unless you’re using Karl Popper as a prescription not to think, you’d realize that an event having low probability under the assumption of no causal effect is (potentially) evidence of some causal effect.
Invoking Shannon-Nyquist, I think, is distracting. We’re trying to recover the first moment from spatial data, not perfectly reconstruct a temporal signal. At least please explain why the hypotheses of the theorem are satisfied in this case. “””
Item #1 Maud. ALL of mathematics is pure fiction; we made it all up in our heads out of whole cloth. To put it another way:-
There is absolutely nothing in any branch of mathematics which exists in the real universe. There are no Points, or lines, or planes etc those things area figment of our imagination.
The formula x^2 + y^2 + z^2 = a^2 does not provide for anything like 8 km high mountains on its surface.
So I wasn’t picking on any branch of mathematics; merely pointing out that mathematics has the properties that we build into it. It is useful for describing the behavior of MODELS; which it can do with great exactitude if we choose; but that does not mean those models actually replicate reality. Of course we hope they do to some extent; but it is the models that our mathematics describes; not the real universe. So no matter how close the correlation it still doesn’t prove causality.
I have described on several occasions an incident in the late 1960 – early 1970s where a mathematical model was presented, which produced the exact value of a fundamental Physical Constant (Then Fine structure Constant) to less than 2/3 of the standard deviation of the very best experimental measurement of the fine structure constant (at that time) a value that is known to parts in 10^8.
So if the mathematical derivation agrees with experiment to a part in 10^8, it surely must be correct; most people would think.
Well it wasn’t; the model had absolutley no input data or parameters from the Physical universe; it was purely mathematics; and subsequently it was shown that a slightly different version of the same quite fictitious model gave a value that was more than twice as accurate; and every bit as phony.
But as to the invocation of Nyquist-Shannon; which you dismiss as “Distracting”; you’re darn right it is distracting. Perish the thought that we should deal with any realities.
You say you aren’t interested in a reconstruction; just a trend. I’ll grant you that; I wasn’t too interested in a recosntruction either; in fact the only thing I was interested in was the average value of the sampled function; I didn’t even care about the trend; just the average value.
And as you know from your Nyquiest Shannon, the average value of the function is simply the DC component of the signal spectrum; and the Nyquist Criterion tells us that it we indersample a band limited singal by just a factor of two, then the reconstructed spectrum folds back all the way, to produce aliassing noise at zero frequency; which of course is the average value that was being sought.
So if you do a min-max twice daily sampling of the Temperature at a point; and the diurnal
temperature cycle is non sinusoidal but contains at least a 12 hour period second harmonic component or higher; then the min-max average Daily Temperature computaion is erroneous because of Nyquist.
so even without the need for a recosntructed original continuous signal; you can’t extract even the average value of a sufficiently undersampled signal. And of course the global spatial undersampling is just a joke; orders of magnitude under what is needed simply to get a correct average; not to reconstruct the original signal.
And no the Central Limit theorem cannot buy you a reprieve from a Nyquist violation.
By the way; in ordinary Euclidean Geometry a circle is simply a special case of an ellipse. There are alternative Geometries and in one of them a circle is not an ellipse; it is a special case of a Hyperbola; and like all hyperbolas it is an infinite sized object. And every possible circle happens to pass through the same two fixed points.
But every one of the historical Geometry Theorems of Euclid can be proved in this alternative geometry.; which doesn’t even have the capability of proving that any more than seven “points” exist.
So yes mathematics is fictional as are all of our scientific models; they only approximate the real world; but we can explain their behavior very well using our equally fictional mathematics.
Only Mother Gaia has a good enough model to be able to mimic reality; and she does it all the time; she always gets the right answer.
Ed,
I am a northern resident who has spent recent time in all of the Canadian arctic communities which you mention in this post. As such, your categorization of Inuvik, Cambridge Bay and Eureka as ‘Urban Stations’ beggars belief.
Inuvik’s weather station is located at the airport, more than 10 km from the town itself, and more than 250 m from the nearest heated building. There is No significant heat effect from the town on that weather station.
Cambridge Bay’s airport and weather station are more than 2 km from the Town itself, with the weather station located more than 80 m from the nearest heated building.
Eureka is a site established as a weather station, and primarily manned for that purpose. As you state with virtually never more than 20 people there, and normally less than 10. To argue that it is ‘urban’ is so ludicrous that it casts doubt on anything else you say, however much merit it may have.
Please don’t ‘sex up’ your point by using such poor examples. Your analysis will stand or fall on it’s own merits, and referring to these stations as urban does your argument a disservice.
George E. Smith wrote:
“And as you know from your Nyquiest Shannon, the average value of the function is simply the DC component of the signal spectrum; and the Nyquist Criterion tells us that it we indersample a band limited singal by just a factor of two, then the reconstructed spectrum folds back all the way, to produce aliassing noise at zero frequency; which of course is the average value that was being sought.”
Oh man, I am afraid you have touched the spot of knowledge that is completely foreign to climatologists 🙁 (BTW, sorry to ask, is the term “climatard” acceptable on this board?). But we need to cut some slack here. I have seen professional engineers who were designing a temperature measurement system with 10Hz sampling rate, while the input had 50% noise level with spectrum up to several GHz and was completely unfiltered. No wonder they frequently saw some “unexplained” jumps in readings in tens of degrees C (!!!) (when the system would switch to a different program).
“So if you do a min-max twice daily sampling of the Temperature at a point”
The daily sampling of data could be ok. It could be interpreted as taking a half-turn Poincare map of the quasi-periodical trajectory (of weather). (Oh, again, this gobbledygook will be clearly unforgiven!). The trouble is that (a) they do not take the data precisely at the same time, and (b) most thermometers are of min-max type, they “remember” min and max, but do not remember when these min-max happened. My sympathy to glorious climate mathematicians who embark on such a mess. 🙂
BTW, among all digicams I had, all had a yellow focus assist light. Go figure…
Cheers,
– Al Tekhasski
E.M.Smith says:
“Measuring a fractal gives different answers based on the ‘size ruler’ you use. And we’re measuring ‘climate change’ with a ruler who’s size constantly changes over time.”
Yep, exactly. Except it is not a ruler, it’s more likely a “wet finger”. Measuring fractals in practice is a non-trivial task, it is called “Hausdorff Dimension” in certain scientific circles far from climatology. Been there, tried that. Given substantial level of unrelated noise from instruments, it is nearly impossible. At least an attempt would require many “halving” of the ruler scale, a thing that is remotely acceptable in climatology.
It is well known that models with progressively finer grid scale could approach the level of details of realities, while crude sampling schemes are totally off. Precipitation patterns is one example. Here is another good example, about “disappearing glaciers”:
http://www.nichols.edu/departments/Glacier/impact_of_sampling_density_on_gl.htm
The initial impression was that the glacier’s annual balance was decreasing on the scale of decade, when they used their normal sampling density 12 sticks/km2. But when they re-calculated the balance using sampling density of 388 points/km2, the deficit in ice balance disappeared!!!. What if the sampling density would be 1200 points? Maybe the glacier was advancing, not retreating after all? This heretic thought was apparently rejected at the stage of peer review…
Also, do they monitor all 350 glaciers in Northern Cascade with 388pt/km2 sampling density? Hell no. If anything, only a handful of glaciers are monitored:
http://www.nichols.edu/departments/Glacier/quick_map_edited.jpg
Yet they arrived to unconditionally certain conclusion that global glaciosphere is shrinking like never before, all thanks to men.
@george E. Smith:
My apologies. Reading your response I see I confused Al Tekhasski’s comments about spatial sampling with yours about temporal sampling. Temporal sampling is where Shannon-Nyquist most naturally applies. But consider that reporting the minimum and maximum is not equivalent to twice-a-day sampling. Sampling is occurring at a higher frequency and it is just the lowest and highest order statistics that are reported. In a similar (toy) situation, given iid samples from a uniform distribution on interval [a,b] and trying to estimate a and b, it turns out that the lowest and highest order statistics taken together are complete sufficient statistics. In other words, they’re all you need to know if you want to know a and b.
Maud Kipz wrote:
“Temporal sampling is where Shannon-Nyquist most naturally applies.”
Duh. The Shannon-Nyquist-Kotelnikov-Whittaker Theorem, a.k.a. the Cardinal Theorem of Interpolation Theory, or simply the Sampling Theorem naturally applies to any attempted interpolation of any smooth bandwidth-limited function regardless of particular physical meaning of its domain, and regardless of its dimensionality. It is a mathematical fact. Even if no lines or differentiable manifolds exist in Nature, Mathematics is a precise language that allows us to express relationships between objects and derive conclusions with absolute certainty.
“it turns out that the lowest and highest order statistics taken together are complete sufficient statistics. In other words, they’re all you need to know if you want to know a and b”
Sufficient for what? Given that (a+b)/2 has no physical meaning and therefore no physical law could be used nor applied to “explain” its behavior, it is sufficient for nothing. That’s why here we are.
Dave in Delaware wrote:
“Temperature is a PROXY for Energy.”
Yes, the temperature is. However, “Global (average) Temperature” of a non-uniform heated globe in an outer space is a proxy for NOTHING.
Example:
Let a planet to have only two climate zones, 50% equatorial area with flat temperature T1, and 50% polar area, with T2. Consider the following sequence of temperatures (“climate change”):
(A) T1=295K, T2=172.8K
(B) T1=280K, T2=219.4K
(C) T1=270K, T2=236.9K
(D) T1=260K, T2=249.8K
The zonal temperatures in this sequence of “climates” are such that they give the same global OLR of 240W/m2, or no change in any energy fluxes has happened. Yet the “global average temperature” (T1+T2)/2 grew from 234K (case A) to 255K(case D), or increase of 21K , all without ANY change in radiative balance. So is your global average a proxy for energy? No.
“”” Maud Kipz says:
September 23, 2010 at 8:52 pm
@george E. Smith:
My apologies. Reading your response I see I confused Al Tekhasski’s comments about spatial sampling with yours about temporal sampling. Temporal sampling is where Shannon-Nyquist most naturally applies. But consider that reporting the minimum and maximum is not equivalent to twice-a-day sampling. “”””
Well Maud perhaps you should unconfuse yourself; because I made NO RESTRICTION of my comments to just Temporal Sampling; I’m in total agreement with with Al Tekhasski that the Nyquist sampling theorem applies to any multi-variable sytem of mathematics; whether it it is connected to any rreal physical ssytem or not. It is a property of mathematics; not of Physics.
In the case of Climate or weather science the Temperature (of the earth) can be considered to be a map diagramming in space and time what the Temperature is (everywhere) and at any time; well we typically limit it to surface or near surface temperature. That Temperature is a single valued continuous function of time and space; giving a unique value for any point on the surface for any instant of time. Now we can’t practically keep track of such a double infinity of information; so of necessity we have to sample it; and we have to sample it both in time and in space; and mathematically it doesn’t matter in what order we treat the two (in this case) variables.
In principle we can take a snapshot of the entire surface at any instant, and get a Temperature for any point. Now it is a simple integration problem to multiply each sampled Temperature by the elemental area for which it is cvalid (say one square km, or metre or even mm) and then add them all up, and divide by the total area of the surface; and Voilla !!, we have the average Temperature for the surface for that instant of time. So we repeat that one second later and get a new average, and so on, and then add those all up and divide by the total eleapsed time.
Well because of known quite cyclic variations over time; diurnally and over the course of a year; we would likely want to average a whole year of observations; to take into consideration expected variations due to orbital position and daily rotation and the like.
And of course it doesn’t matter the order. We choose to average at least daily for each and every sampling station; so we actually do the time averaging first. That doesn’t change the result if we do it properly; but we don’t because a min-max average temperature is only valid (in the Nyquist sense) for certain kinds of diurnal cyclic variations; and Nyquist itself says only for a single frequency sinusoidal cycle of Temperature since twice a day sampling is the minimum required for equidistant samples of a single frequency. Now generally a fixed regimen of 12 hourly sampling is still inadequate to reconstruct the complete cycle; that is a degenerate case; so we can’t recover the amplitude of the cycle; just the average. Min-max sampling of a sinusoid has the added benefit that it does recover the amplitude as well; even though we just want the average.
But perish the thought that it should warm faster in the morning, than it cools after sundown; which will introduce at least a second harmonic component; and don’t even talk about clouds passing by.
But I am on exactly the same page as Al here. NO ! we don’t need to be able to reconstruct the original continuous function; although in signal processing which is the most common application we would do that in effect just to get the average.
But you see in the case of the global temeprature; other than the diurnal cycle; we do in fact reconstrudct the continuous function.
That is the samples measured at each weather station; for daily average; when plotted on a map and smoothly interpolated; whcih is exactly what happens in any real communication network where sampled signals are transmitted and processed for recovery; they basically do an interpolation that conforms to rules well understood by signal processing engineers.
But the problem for climatology is that the violation of the sampling theorem with global temperature sampled data; makes even the global average unrecoverable;.
So we don’t have any idea what the global mean surface temperature is; and that is a failing of the sampling stratagem; it is not a limitation of our instrumentation or of mathematics (including statistics).
And note Al’s remonstration that the rules for sampled data processing are purely mathematical and involve no need for any connection to any real world phenomenon; it is strictly a (fictional) mathematical discipline, that has practical applicability to many real world systems.
And your Statistics likewise is a purely (fictional) mathematical discipline; that can be applied to any sets of data even totally fictitious made up sets of data; or real world numbers from some physical system (like the climjate).
And that is why I threw out the Telphone directory example.
I can make up a totally arbitrary set of data (numbers) and give it to you to do your Statistical wizardy on; and the results you get; are no less valid (ormeaningful) than a set of experimenta measured data values that might be obtained in the most well behaved well understood Physical experiment. And that is so BECAUSE your statistics is a property of a set of mathematical axioms on which statistical theory is built. Statistics is NOT a property of the Physical universe; but we can use it usefully to describe certain properties of that physical universe.
But always remember; is it Godel’s conjecture, that no formal mathematical system is complete in the sense that there will always be problems which can be meaningfully stated within the axioms of that mathematics; but which cannot be answered rigorously within those same axiomatic constraints. I believe it was his conjecture on “undecideability.”
The mathematics called “Projective Geometry; which is a plane Geometry has three very simple axioms.
#1 Two Points define a line (joining them)
#2 Two lines define a point (their intersection)
#3 There are at least 4 points.
That’s it. The four points are usually drawn at the corners of a kite, or the southern cross.
Using #1 you can construct the lines joining those four points; and because of #2, you will find that the three pairs of lines so drawn, will locate three more points. In fact that is the first theorem of Projective Geometry; that there are at least seven points.
Constrained by those axioms and the definitions of some procedures; it is impossible to prove there are any more points. There may be; probably are; but you can’t prove it within projective geometry.
Some will immediately object that parallel lines do not intersect at a point. Axiom #2 says they do; and adds that parallel lines do meet at a point on “The Line at Infinity.” Which conveniently one can always draw on the page.. Is it not true that two railway tracks merge to a single point on the horizon. The line at infinity is important in other ways, in that parabolas touch the line at infinity at two coincident points but ellipses never reach the line at infinity. Hyperbolas on the other hand intersect the line at infinity at two non co-incident points.
Therea re two special points on the line at infinity; and they are called the “Circular Points at Infinity”. So you can consider the line at infinity as defined to be that unique line that joins the two circular points at infinity. All circles (every one) pass through the circular points at infinity; so that means that circles are hyperbolas; not ellipses.
Within projective geometry; it doesn’t introduce any inconsistencies. But Godel still says there are undecideable problems within that confinement.
“”” Al Tekhasski says:
September 23, 2010 at 5:58 pm
George E. Smith wrote:
“And as you know from your Nyquiest Shannon, the average value of the function is simply the DC component of the signal spectrum; and the Nyquist Criterion tells us that it we indersample a band limited singal by just a factor of two, then the reconstructed spectrum folds back all the way, to produce aliassing noise at zero frequency; which of course is the average value that was being sought.”
Oh man, I am afraid you have touched the spot of knowledge that is completely foreign to climatologists 🙁 (BTW, sorry to ask, is the term “climatard” acceptable on this board?). But we need to cut some slack here. I have seen professional engineers who were designing a temperature measurement system with 10Hz sampling rate, while the input had 50% noise level with spectrum up to several GHz and was completely unfiltered. No wonder they frequently saw some “unexplained” jumps in readings in tens of degrees C (!!!) (when the system would switch to a different program).
“So if you do a min-max twice daily sampling of the Temperature at a point”
The daily sampling of data could be ok. It could be interpreted as taking a half-turn Poincare map of the quasi-periodical trajectory (of weather). (Oh, again, this gobbledygook will be clearly unforgiven!). The trouble is that (a) they do not take the data precisely at the same time, and (b) most thermometers are of min-max type, they “remember” min and max, but do not remember when these min-max happened. My sympathy to glorious climate mathematicians who embark on such a mess. 🙂
BTW, among all digicams I had, all had a yellow focus assist light. Go figure…
Cheers,
– Al Tekhasski “””
The hell you say !! a Yellow focus light ? Well you did say digicams; I must be just looking at the cheap stuff.
Well yellow LEDs are really something else if they are the newest materials Technology. The earliest Yellow LEDs were high Phosphorous composition GaAs(1-x)Px material doped with Tellurium plus the iso-electronic trap; Nitrogen. The effect of the nitrogen was to localise the carrier recombination at the site of the nitrogen atom in the lattice; and that localisation, as a result of Heisenberg, made the momentum uncertain enough to spread across the band structure diagram far enough to reach the energy gap minimum and follow an allowed transition. Since GaP is an indirect band gap Material (like silicon) but GaAs is direct gap, GaP is an inefficient light source; without the proper doping. It’s the best everyday demonstration of the Heisenberg Uncertainty principle I know of.
But today they have AlInGaP; which makes an extremely bright Yellow, (and red) lamp possible some diodes have more than 50% external quantum efficiency; so half the input energy comes out as light, and only 1/2 or less is wasted as heat; kinda spooky really.
So I’ll drop my objection to yellow focus lights but notice they were observed on Digicams. I guess I will have to look at the focus light on my Panasonic digicam; never noticed it.
Never given “Clamatard” much thought; but I have thought that “Climatism” should be a real word.
Anthony has relatively few “do it my way” rules. Keep it clean; mean ain’t clean. Try to keep within some reasonable boundaries of thread topic. Charles the Moderator (Chasmod) doesn’t like to have to wake up every now and then reach for another encyclopedia; but he’s a good chap.
We don’t do typo reconstructive surgery, since everybody has that disease.
Other than that this is THE place to be; the girls are all very pretty; lots of trolls you can make pets out of. The garden of Eden must have been like WUWT.
George
George, “So I’ll drop my objection to yellow focus lights but notice they were observed on Digicams. I guess I will have to look at the focus light on my Panasonic digicam; never noticed it.”
I just checked my Casio. The light is more like “orange”, although it is hard to conduct full spectral analysis by eye and assign a single moniker to a continuous function. I guess we can settle for “orange”. 🙂
Walter Orr
Thank you for the on-site information. That confirms the location for Inuvik. The airport there is the source of the heat, and the situation is similar to that at Barrow, where the UHI has been studied. Cambridge Bay is another airport location, again, like Barrow. Eureka has it’s own problems. Every year the infrastructure is expanded and improved.
jose says:
September 22, 2010 at 10:22 pm
“Smokey: Way to change the topic. But since you asked, the simple answer is “its complicated”. […]”
Beautiful! Pure Gavin Schmidt!
E.M.Smith says:
September 23, 2010 at 10:44 am
“Often the most ‘warming’ comes at the ‘shoulder’ months between seasons. ”
I looked at monthly trends on CET during decades where there was a yearly warming trend, and the months with the strongest warming trends are around the equinoxes. What came to mind is the increased connection with the solar wind at this time of year.
Interesting post, but you are mistaken when you write:
although GISS should share responsibility for this mistake. When the GISS station data page returns “rural” for a station, this is based only on the original GHCN metadata, and does not take the nightlight radiance into account.
Barrow/W.Pos (radiance value 40) and Zyrjanka (radiance value 16) are both above the rural limit of radiance value 10, and so will also return a data set “after cleaning/homogeneity adjustment”. Kotzebue, Ral and Mys Smidta have radiance values of 10, just failing to make it out of the rural classification. Other high latitude stations with urban radiance values which are not included in your subset are:
And here I have plotted the raw and adjusted records for Barrow:
Barrow (raw and adjusted). Trend over full record slightly reduced by adjustment. Trend over last 30 years considerably reduced by adjustment.
The image (which appeared in the preview) seems to have been lost. A link:
http://oneillp.files.wordpress.com/2010/10/barrow2.jpeg