Animating GHCN Global Temperature Anomalies from NCDC

With the recent announcement from NCDC that June 2009 – second warmest on record globally I thought it might be interesting to go back and look at some of the older NCDC announcements.

Many commenters have questioned how NCDC arrives at some of the temperature anomalies on this NCDC graph:

June's Blended Land and Sea Surface Temperature Anomalies in degrees Celsius

June's Blended Land and Sea Surface Temperature Anomalies in degrees Celsius - click for larger

Here is what NCDC says in their official announcement.

As an aid to investigation and understanding, I have compiled all such NCDC global temperature anomaly maps that I could find and made them into a flipbook animation. NCDC only made this map style back to May 2007, and I’ve captured every month up to June 2009.

You’ll be able to watch it after clicking through, please be patient, it is a 1.4 MB file and will take bit to load.

click for larger image with faster animation

click for larger image with slower animation

For those that like a slower animation, click the image and a larger one at the original resolution (not scaled for blog width) will pop up with 4 seconds between frames.

I’m beginning to think that tracking anomalies may not be providing a complete picture.

Advertisements

62 thoughts on “Animating GHCN Global Temperature Anomalies from NCDC

  1. This HAS to be the “second warmest on record” … there is important legislation pending in Congress that has billions of dollars for these people in the balance.
    The anomaly will go away after the fate of the bill is known.

  2. Anthony do you have any info whether NCDC take into account the map projection exaggeration effects at higher latitudes when constructing these anomaly maps.
    Eurasia seems very dominant in terms of +ve anomalies, yet the grid spacing of the “anomaly dots” is regular right across the globe. This would mean that a “dot” in Siberia is covering an area about 3 to 4 times the same size “dot” near the equator.
    Is NCDC just counting the “dots” and weighting them or are they taking into account the exaggeration of latitude? Couldn’t find from their site.

  3. I am struck by how many of these alarming anomaly presentations use nonrepresentative base periods, which would seem to be justifiable only if the intent is to maximize a chosen trend.

  4. The map has the total sum value of “Eat at Joe’s” at 2am in the Black Rock Desert in Nevada.
    Mostly continent sized splotches of molten lava thrown at the canvas from 3 paces. Siberia has long been reduced to a rock-strewn pile of ash.
    There are stations in places strung across Greenland where no foot has been placed in 20,000 years.

  5. Dave Wendt,
    The choice of base period doesn’t affect the trend. It’s just a constant offset. On a time graph, it determines where you put the zero axis, but the shape of the plot remains the same.
    For these color dot plots, they have chosen to change from red to blue at the zero anomaly. That is a real visual effect, and you could argue for a different color scheme.

  6. Why do we have temperature anomalies in 2007 in northern Greenland but not in 2009? I see other areas of the globe like this as well from the maps presented. You’d think that since Greenland is an area of specific interest for the the study of climate we would be increasing observations there, not decreasing them. Are that many stations dropping out just in the last couple of years?

  7. And what about that horrible winter in Alaska that started in September?
    The temps in the Fairbanks AK area plunged to -19 F as soon as the Arctic Sun dipped on the horizon in late Sept. Never looked back for months on end. There surely was no heat wave anomaly in the Yukon, either, as wretched cold caught many unprepared, conditions so bad rescue was impossible.
    So, where did those red dots come from?
    Super Gorio Bros. dropped bombies on the ice releasing the lava monster.
    Oh, I almost forgot, Godzilla breathes fire, and he got loose from the movie set. They haven’t caught him yet.
    Anything is possible on a computer game.

  8. It is just many weather spots changing. If they were climate then they would keep the same along time…
    Or this is just a free screensaver from noaa

  9. With the downtick on UAH only lasting one day and temperatures going higher AGAIN, I’ll take a wild guess we’ll see the media exploding with stories of how we’re all doomed and why we should lose all hope and give up on our dreams.
    Strangely enough it doesn’t feel like the warmest July ever here in Wichita, we’d need like an average of over 100 degrees or so to have that.

  10. 3 I’m baffled by the very lurid and large red dots in Siberia during the winter months. Is this a measurement of released “heat” or an actual and significant delta?
    I think is fairly disingenuous using an artificiality by manufacturing the 1961-1990 temperature record as a basis to measure against.

  11. There is something that doesn’t add up. I live near Montreal, Qc. The graph show that the temp for June were higher than normal.
    Yet the reality is that June had only one day that temp reach 30 C, and 3 or 4 days that were between 25-30, the rest was cold, low high teens, low twenties at daytime and about 10 at night. Only this week did night were hot enough for me to remove my jacket while distributing the newspaper. This was similar for the entire province where people complained of the cold (and rain).
    Maybe NCDC knows something that we in Québec don’t know.

  12. Ok, look at things this way … NCDC’s Contiguous US data agrees with the satellite data in direction … “Most Recent 12 month period” here shows that the most recent 12-month period is the 89th warmest since 1985 (0.66 degrees above average) with cooling the past 3 years. The month of June was 67th warmest since 1885 (0.25 degrees above average), also cooling the past three years.
    So whatever is causing the NOAA data to heat up isn’t coming from the contiguous US. US data “agrees” with satellite data in direction. NCDC’s “base period” for average in this case is 1901 to 2000.

  13. Base Period-
    Eliminate the notoriously globally cold decade of the 1960`s…
    with a more recent(and representaive) 1971-2000 base period.
    The big red anomalies simply disappear.
    Its a “hat” trick.

  14. Nick Stokes (16:49:05)
    Trend was indeed a careless word choice. As you indicated “visual effect” would more accurately have conveyed the point I was trying to make. Selecting a lower zero point exaggerates the positivity of the anomaly and the lurid color choices only serve to compound the exaggeration. In these trying financial times I suspect, if it were possible to find a commodities future play that covered the constituent components of hooker red inks and pigments, there would be a significant killing to be made, given the alarmists’ well established proclivity for laying it on with a trowel in virtually all their recent offerings.
    There may be a silver lining in this black cloud since, if they are able to convince the world that we are indeed in the midst of a profound warming, the large segments of the population of the developed world that are experiencing extended periods of below normal to normal temps right now might finally begin to question what has for me always been the weakest part of their argument i.e. that any warming that occurs will be necessarily catastrophic.

  15. Siberia, unlike Alaska, has 50% ice up against the Arctic Ocean for most of it’s length. Makes no sense.

  16. sorry for repeating me, but I think my message hasn’t arrived yet.
    Besides the known warming bias of NOAA, there is another problem with these maps. They do not match with the NOAA average values !
    Look at June 2009
    http://www.ncdc.noaa.gov/img/climate/research/2009/jun/map-blended-mntp-200906-pg.gif
    and try to average the anomaly, do a bit of counting, trade off the few blue points with the red ones, put higher weight on the low lattitude points and the average will be somewhere between 1° and 2°.
    NOAA however gives a global mean of 0.7°.
    I wouldn’t be surprised, if the map average is exactly at 1.7°, and NOAA just made an error with the assignment of the dots to the temperatures. Maybe, sombody could do the counting,
    A difference of approx. 1° everywhere is also supported by comparing the NOAA map with this australian map:
    http://www.longpaddock.qld.gov.au/SeasonalClimateOutlook/SeaSurfaceTemperature/index.html

  17. Nick Stokes (16:49:05) :
    Dave Wendt,
    The choice of base period doesn’t affect the trend. It’s just a constant offset. On a time graph, it determines where you put the zero axis, but the shape of the plot remains the same.
    For these color dot plots, they have chosen to change from red to blue at the zero anomaly. That is a real visual effect, and you could argue for a different color scheme.

    It is not only arguing about the color, though there could have been two neutral ones.
    In color maps the 0 is important for the visual effect.
    Since there has been some warming since 1961, about 0.1 or so per decade , taking the 0 anomaly from 1961 to 1990, allows for reddening by 0.25 or so degrees on average, so the plots will always be red since that is the third in size dot in their scale.

  18. Hmm this is wierd, for the last few days I have been checking the historic data for 68137 on the weather channel to see if they would include the new record loes we just set. Well tonight there was a change in the data, but not the new record low. Instead all of the records seem to be now within the last 50 some years and the avereage high for mid July miraculusly went from 88 to 86. WUWT??

  19. rickM (18:27:54) : “…I’m baffled by the very lurid and large red dots in Siberia during the winter months. Is this a measurement of released “heat” or an actual and significant delta?…”
    Well, there are at least two things about Siberia you have to keep in mind: (1) during the Communist regime in the USSR, stations may have had financial incentives to report low temperatures–reporting -40° below may have gotten increased fuel budgets, etc. (2) the base temperature for most of Siberia is very low, anyway, making, say, -10°C into a red dot based on average temp of, say, -15° for that time of year.
    More important, if a place has greater snowfall than usual, the heat released by the freezing of all that water can show up in the atmosphere as a higher temperature for that month. Overall, the place may be cold as Hillary’s tent*, with snow six feet deep and net heat content much lower than average, but the reported anomaly could be positive, since air has less heat capacity than water.
    * Sir Edmund Hillary, of course.

  20. What I find more interesting is that the graphic doesn’t seem to correspond to the local [Michigan] anomalies that I track.
    Monthly variance °F
    Feb-08 -4.3
    Mar-08 -4.9
    Apr-08 1.3
    May-08 -4.3
    Jun-08 0.2
    Jul-08 -1.7
    Aug-08 -1.8
    Sep-08 0.1
    Oct-08 -3.5
    Nov-08 -3.1
    Dec-08 -3.9
    Jan-09 -10.1
    Feb-09 -1
    Mar-09 0.3
    Apr-09 -0.1
    May-09 -2.3
    Jun-09 -2.9

  21. I’m sorry to be a nuisance, but what do the actual temperature measurements show?
    Never mind averaging and adjusting them against this and that, never mind anomalies. In terms of real measured heat at ground level, was June the second warmest ever across the whole planet?

  22. Russia’s red spot.
    Apparently Russia is redefining the cold war. How do we know these stations aren’t being over-reported?

  23. Somewatt Off Topic
    You may now already, but Mr. Watts, you have been mentioned in Scientific American August 2009 issue, page 9 “Stumbling Over Data”. It states “. . . bloggers such as meteorologist Anthony Watts. His blog, Watt’s Up with That? . . ”
    Good Job

  24. I too am baffled by the red dots over Scotland, where apart from a few short periods in the last couple of years (April to mid May 2008, mid March to mid April 2009, May 28 to June 4 2009, and a week in late June 2009) we have had noticeably colder then average temperatures. This is weather, but yesterday at 2pm it was only 12C in Tyndrum, (credit to the Met Office for the forecasts over the last few days, they have been very accurate and geographically precise – Aberfeldy had rain but Rannoch stayed mostly dry). But even the Met Office forecaster commented that these temperatures are not warm for the time of year. And apart from the week of blue skies from May 28 and the warm/humid week in late June it certainly has not been the warm summer we were promised by the Met Office’s long range team.

  25. AndrewP – they’re probably including within our landmass the SSTs all around the UK by some artificial smoothing.
    I’ve noticed that before, and that is the stock excuse, despite the numerous problems with it.
    Another dreadful July for us once again. Lovely BBQ Summer isn’t it?

  26. Andrew P (21:47:32) :
    I too am baffled by the red dots over Scotland, where apart from a few short periods in the last couple of years (April to mid May 2008, mid March to mid April 2009, May 28 to June 4 2009, and a week in late June 2009) we have had noticeably colder then average temperatures.
    The red dots are inevitable by the choice of the 0 in this plot. It is a visual statement that there has been a bit of warming since 1990. The 0 has been calculated from 1961 to 1990. Assuming it will be sitting in the middle of this interval , 1975, this gives two and a half decades extra warming at your choice, 0.1 per decade or 0.2 per decade. This moves all points by two or three settings towards the red. I am surprised any blue has been valiant enought to show up with this PR trick.

  27. Andrew P (21:47:32):
    I agree. England seems to be mainly a red dot, with often a big red dot. Yet the temperatures for the last 2 years seem to have been pretty much below par from my experience of the last 40 years of weather here in the countryside. Ten to one it’s the UHI they’re measuring in the CET (central England is one big and growing urban heat island).

  28. Jeff Id (21:34:02) :
    “Russia’s red spot.
    Apparently Russia is redefining the cold war. How do we know these stations aren’t being over-reported?”
    There also seems to be a hot spot over the Middle East, China is trending redder and Bangladesh has a positive anomaly for essentially the entire series. Western governments are making gigantic financial decisions based on temperature data provided by countries that are seeking economic advantage and may desire the demise of our system of governance. I’d be interested to see “global warming”/ NCDC monthly anomolies broken out by country. I wonder if there is a correlation between temperature increase and a countries’ likelihood to benefit from the limitation of developed nations’ anthropogenic carbon dioxide emissions.

  29. “Nick Stokes (16:49:05) :
    Dave Wendt,
    The choice of base period doesn’t affect the trend. It’s just a constant offset. On a time graph, it determines where you put the zero axis, but the shape of the plot remains the same.”
    I’d agree with you about the “trend”, but these charts are not designed to show a “trend”.
    They are taking the current June, comparing it to all of the Junes in the record (IIRC, about 125 years or so), against a “standard” base period.
    From that, they say that this June was the second warmest, which means there were about 123 years in which June was cooler than the current month.
    By their choice of an averaging period of 1961-1990, they ARE selecting the 0 degree anomaly. They still can’t explain why another period can’t be used, especially since the WMO suggests using a period that ends in the latest decade (1971-2000).

  30. It didn’t creep higher than 17C here yesterday and the temperature seems fit to not creep much higher today. Most of the previous week has shown little signs of summer with temperatures more fitting to autumn. Apart from a brief, anomalous heatwave back in 2003 the weather in the UK has consistently failed to exceed warmist and Met Office expectations.
    It seems that the NCDC chart is being rather over-optimistic about global temperatures too, especially the most recent ones in my part of the world.
    To be taken with a large pinch of salt perhaps?

  31. If it wasn’t for Russia manipulating the data, there would not have been so much warming, I think.
    Russia wins both ways by inducing a warming trend. Firstly, frightening the USA into joining Kyoto will bankrupt much of US industry. Secondly, Russia was issued with millions of Carbon Credits (CCs) to cover its old industries, many of which are gone already. Without the USA in the Carbon Credit scam, those CCs would be nearly worthless. But now the USA needs to purchase CCs on the open market, to cover any new industries, the price of Russian CCs will rise enormously.

  32. Check to see what countries have an economic interest in the US curtailing production and energy consumption compared to the temperatures they report to NOAA. If a nation has an interest in seeing the US enact regulations that curtails production, they might be tempted to manipulate the reported data.

  33. Henry,
    There’s a very good reason why organisations like NOAA don’t change their basis. They have a big accumulation of published records, in print, graphs and electronic, using that basis. As soon as they change, people have to make conversions whenever using older data. There will be lots of confusion and errors. There are many copies out there which the NOAA doesn’t control. And in terms of the numbers, there’s no reason to change.
    For plots like this, they could change their color scheme.

  34. Six issues here, perhaps, all together making unlikely red spots in the Heat War.
    (1) poor baseline – too far away from now – this really gives us no feeling for the way the temperature is trending right now
    (2) the wretched UHI advance
    (3) poor use of adjustment factors
    (4) station siting problems
    (5) the ongoing effect of loss of many rural stations around 1990
    (6) big red spots From Russia With Love – their district heating
    How can we, as a team, do an “engineering quality” report on this? Isn’t this what the Surface Stations project is meant to help start to crack?
    Just my two cents pennies worth.

  35. Manfred (20:35:33) :
    I just realized, that NOAA data points in the map are not rounded to the closest integer. The map average could then be lower than my estimate.

  36. The coverage is utterly weird particularly at high latitudes.
    Antarctica was apparently abandoned at the end of 2007. In Greenland there were apparently stations right across the icecap in northern Greenland in 2007 but only on the southern icecap after that (in fact there is only one weather station on the icecap, Summit, and that is in the north).
    For some odd reason there is one dot out in the middle of the Fram strait between Greenland and Svalbard, but none on Jan Mayen, Svalbard, Björnöya, Hopen or Franz Josephs land (which all have one or more weather stations). That single dot might be Jan Mayen, but if so it is displaced about 1000 km to the north.
    Further there are apparently no weather stations along the Siberian coast or on Novaya Zemlya, Severnaya Zemlya, The New Siberian Islands or Wrangels land, nor in most of the Parry Archipelago.
    On the other hand they have 100% coverage of the Sahara Desert, which is more than anybody else has. There is no coverage in Angola, which might be true, but why no coverage in Madagascar, or northern Namibia?

  37. Looks like they never bothered corrected the reporting of sept 2008 figs again in oct 2008 for large swathes of Russia.
    The sequence is interesting for watching the lingering la nina and for the speed of progression of large warm and cool areas round the globe in the higher latitudes.

  38. rbateman (17:00:38)
    “Anything is possible on a computer game.”
    In the words of the late Prof. Auer:
    “Play Station climatology.”

  39. I like that impressing ret dot above Hungary, while June 2009 was actually below normal for most of the country.
    You can see the anomaly map for June 2009 here (from the Hungarian Weather Service): http://www.met.hu/data/mmTr/mmTr20090601_0000+1.jpg
    The base period used by our NWS is 1971-2000, but last June would be below normal with respect to the 1961-90 period as well.

  40. I haven’t been able to get the animation. Either I’m stupid or my software is not working properly. I get a few seconds of the download “circle” animating then nothing. Any suggestions? (not on correcting my stupidity please!)

  41. Look at this map, I really believe that the higher temperatures are a result of either sensors in less than ideal locations or old, failing sensors. I will give you an example. This past week, the official sensor at the Raleigh-Durham airport was twice abnormally high. Friday, “officially” the high was 97, a record. However, that was 4 degrees warmer than any other nearby sensor. This happens so much that the meteorologist I watch lamented about the sensor being up to its “old tricks again”. Those were his words, not mine. Usually the sensor is fine, but twice in one week indicates a quality control issue. Will it be replaced or serviced? Unless it was reading cold, it won’t.
    So the question is, how many of these sensors are reporting faulty data? How many times does the NWS know the data is faulty and yet do nothing about it? Why is nothing being done to get the most accurate data possible?

  42. Lets figure this out a step at a time.
    An anomaly is a difference from an average value.
    If the average value is taken from all the available data, and the anomaly plotted in red and blue dots, then there should be as many red dots as blue dots and the plot(s) would be neutral.
    If the plot is predominantly red, then it means that the “average value” is not an average value including all the data, but only some.
    Since we know that we are coming out of the little ice age, if we took as average value the values of the 19th century, we would be all deeply red, the closer to our time the redder, until 2000, and then a small shade less red from 2000 on.
    The choice of taking the average of 1961 to 1990, as the value against which temperature has become deviant, is just another way of showing that there is a trend showing heating at something like 0.1 to 0.2 per decade, except it can be used for PR purposes to claim “overheating” since the ten year global temperature hiatus has no chance to be seen against this color scale.

  43. I am confused. From my vantage point, June was a rather pleasant month. June seemed below average in terms of high temeperatures, except for the beginning of the month. I have no data to back up these perceptions, but I am nonetheless skeptical about the red dots over Egypt. I am wondering out loud about this: Where might one find historical weather station data for Egypt?

  44. We have seen that as climate datasets get revised, the adjustments tend to raise the recent temperatures and lower the older temperatures, thus raising the observed warming trend.
    I wonder if as they release the new maps, they have recalculated the 1961-1990 baseline temperature as well? This would also increase the amount of red on more recent maps.

  45. The more the information on global temperatures unfolds, the more I’m convinced its just gibberish. The sometime scientist in me sees so many uncontrolled variables, uncalibrated technologies, malfeasance, incompetence and obscure “measured” variation at such low levels, that notwithstanding the excellence work by Steve Mc et al, no amount of number crunching can dress up this pig.
    This graphic is without meaning or even intelligent interpretation, based as it is on so many scientific fairy tales.

  46. Meanwhile, here in Nashville, TN, we are expected to hit all-time lowest high temperature, and tonight, all-time lowest low temperature.

  47. Issued by The National Weather Service
    Nashville, TN
    6:12 am CDT, Sun., Jul. 19, 2009
    … UNSEASONABLY COOL TEMPERATURES TODAY AND TONIGHT…
    CANADIAN HIGH PRESSURE WILL CONTINUE TO CONTROL THE WEATHER ACROSS MIDDLE TENNESSEE TODAY AND TONIGHT. CLOUDS WILL FORM TODAY AND KEEP TEMPERATURES IN CHECK. THE RECORD LOW HIGH TEMPERATURE AT THE NASHVILLE INTERNATIONAL AIRPORT TODAY IS 79 DEGREES SET IN 1897. A NEW RECORD WILL BE SET TODAY.
    THE RECORD LOW TEMPERATURE TONIGHT IN NASHVILLE IS 59 AND IN CROSSVILLE IS 57… SET IN 1947 AND 1970 RESPECTIVELY. NEW RECORD LOWS ARE EXPECTED.

  48. Just think, perhaps if they weren’t measuring these temperatures in the middle of an airport surrounded by asphalt and airplanes, it would likely be recorded a degree or more lower. My house, approximately 10mi. away from the airport, is typically 4F-6F cooler, you know, where people really live. If we all lived on runway zero-zero-niner, then perhaps this wouldn’t be an issue.

  49. Living in the southestern US and enjoying one of the coolest summers in a decade, I’m at a loss to explain the medium sized red dot sitting over my house. Garbage in…garbage out.

  50. My buddies and I were golfing yesterday and being the prima donna golfers that we are, we were complaining about the conditions… -which is par for the course (pun extremely intended) except that we were complaining about it being chilly and it’s (the odd part is coming) THE MIDDLE OF JULY. 😉
    Andrew

  51. Temperature of the Atlantic ocean off the new jersey shore ( long beach island) was a balmy 61f today. Kids lips turning blue. That map is complete nonsense.
    Maybe someone can get the nytimes to publish it and make a big deal of record high temps. This more than anything will kill the climate bill. People are willing to listen to scientific Pontifications that they do not understand but trust the deliverer of the news but if u put up a chart with a big red blob over the beach they couldn’t swim in because it was too friggin cold they are going to wake up and smell the capacino.

  52. “If the average value is taken from all the available data, and the anomaly plotted in red and blue dots, then there should be as many red dots as blue dots and the plot(s) would be neutral.”
    The key here is the 1961-1990 average baseline. The 1960s was the coldest decades of the 20th Century, and there the recent warming did not begin until after 1976 -some say it didn’t begin until after the 1983 El Nino event. In that case, NOAA can show significant warming because most months will warmer than one of the coldest periods of the 20th Century.
    The other thing to consider is the TOB adjustment, which coincidentally warms the 1990s, but cools the 1930s. Now NOAA can argue that almost all of the 20th Century warming has occured within the last 16 years, with 4 of the 5 warmest years occuring since 1998.
    Finally, as Anthony has constantly pointed out, the NOAA reporting stations in its data sets are comprised of 50% from North America (which make up only 8% of the total landmass), and of those reporting stations the majority are located at areodromes (plenty of concrete). Nearly 75% of the world’s reporting stations have gone offline since 1960. So, in effect the real global average temp that NOAA derives (and from which NASA gets its data) is nothing more than a reflection of the UHI trend of North America.

  53. Wade (05:25:10) : “So the question is, how many of these sensors are reporting faulty data? ”
    The other question is: How many of the dots that have no sensors are reporting valid data?

Comments are closed.