Where’s the Global Warming?

This year’s snowy winter has spawned a flurry of comments along the lines of, “where’s the global warming?”

It’s right where it was expected to be: on the globe. Global average temperature has been pretty much what was expected, given la Nina conditions and, yes, global warming. But that doesn’t stop people from confusing local weather with global climate.

One of the possible (some would even say likely) reasons for the snowy winter is the drastic loss of arctic sea ice. This has altered air circulation patterns, enabling cold arctic air to penetrate further south while warmer mid-latitude air climbs further north. Therefore some regions far to the north, which usually suffer bitter cold winters, have experience extreme heat waves (relatively speaking) this winter (and in fact during recent years). One such region is northeast Canada/western Greenland.

I located seven temperature stations within that area which have reasonably good time coverage which extends up to the present day: Coral Harbour, Egesminde, Fort Chimo, Frobisher Bay, Hall Beach, Inukjuak, and Schefferville. It’s straightforward to combine their data into a regional temperature estimate:

It’s also straightforward to translate temperature into temperature anomaly, to remove the annual cycle pattern in order to isolate the changes:

If it looks to you like this region’s temperature has risen dramatically over the last few years — well, it has.

We can clarify the recent warming by smoothing the anomalies, which are monthly data. Here are three ways: by computing annual averages, by a fast-response wavelet (WWZ) analysis, and by a slightly slower-responding lowess smooth:

Yes, the regional warming in this area over the last few years is quite dramatic. Some (not tamino to be sure) might even say the pattern resembles a hockey stick.

Most of the very recent warming has been due to hotter winters rather than summers. This reduces the size of the annual cycle in temperature, as shown by the wavelet analysis:

I’ve said it before, I (and many others) will say it again: just because the climate changes doesn’t mean we won’t still have weather. And there will still be regional differences in temperature patterns, sometimes quite dramatic ones.

But this much is constant: whatever happens to global climate, those who deny the reality of global warming will consistently cherry-pick local weather to suggest that it’s not real.


44 responses to “Where’s the Global Warming?

  1. I’ve been working on a paper on temperature changes in the NE Canada region since 1880. Your graph has a lot of similar characteristics to mine. The station combination method you used is your own one correct? Would it be advantageous for me to use RomanM’s rather than yours or is that just splitting hairs?

    [Response: Yes I used my station combination method. As for using RomanM’s instead of mine, I don’t think it’ll make a whole lot of difference but I think his is slightly better.]

  2. I’m trying to understand the processing from graph one to two. How does the last point in the first graph become higher than the ones before it in graph two? If the temp is lower, then how is the anomaly (much) higher? Is there an additional step?

    [Response: Anomaly is the difference between temperature at a given moment (in this case, for a given month), and its average for the same time of year (in this case, for the same month) during a “baseline,” or “reference” period. So: if January is a hotter-than average January but July is a colder-than-average July, the anomaly will be positive for January and negative for July. Therefore the anomaly is higher for January than July, although even an unusually cool July is still quite a bit hotter than an unusually warm January. (for the northern hemisphere of course)

    In this case, the winter was so much warmer than usual that the anomalies are “through the roof.” But winter is still the coldest season of the year.]

  3. Ahh ok – makes sense – thanks

  4. Snowy winter eh?
    I live in south western Ontario in the midst of the Great Lakes.
    When I heard last fall of Lake Superior’s record high water temperature combined with the fact Hudson Bay took much longer than normal to freeze over this winter I knew we were in for it.
    Open water just simply breeds snow and in my neck of the woods we are only too well aware of the term: “Lake effects snow”.

  5. TomG, yes, those of us who live in the Great Lakes region get it.
    Those who live in the southern, and central and western US, not so much.

  6. TomG wrote:

    Open water just simply breeds snow and in my neck of the woods we are only too well aware of the term: “Lake effects snow”.

    Lake effect is exactly what we should be referring people to — with the example of the Great Lakes. Keep the temperatures high enough that the lakes don’t get the chance to freeze over but a little below freezing and you will get plenty of snow. But of course this works for ocean even without the possibility of it freezing over — just not as dramatic. With respect to the oceans the big thing is that water has greater thermal inertia than land.

    Therefore the temperature of the water drops less than the temperature of land. Moisture over water doubles for every 10°C — so there is going to be more moisture in the air. But when the air passes over colder land the humidity of saturation drops and you are going to have a whole lot of precipitation. The land may be warmer than previous years, but if the temperature is still below freezing chances are you are going to have a whole lot of snow.

  7. John Mashey

    “TomG, yes, those of us who live in the Great Lakes region get it.
    Those who live in the southern, and central and western US, not so much.”

    Well, actually, those of us who live in California, especially skiers, are rather aware of what happens when moisture comes in from the Pacific and hits the Sierras. See this for some examples.

    Note: over time, warming tends to reduce the total snowpack (bad, since we use it as a huge natural reservoir), but as long as it is cold enough to snow, we still get massive dumps. Skiers cheer the dumps … but of course, that means you sometimes cannot *get* there. 3-6″/night is much nicer.+

    “Top Three Snowstorms USFS Central Sierra Snow Laboratory (CSSL) 6

    * March 27 to April 8, 1982 — 15.5 feet
    * January 20 to 31, 1969 — 13.7 feet
    * January 10 to 17, 1952 — 12.8 feet

    (Twelve other snowfall events at the CSSL have each dumped nearly ten feet, with a mean storm-duration of 11.4 days. Snowstorm totals greater than six feet are not uncommon.) “

    • Andrew Dodds

      We got 6 inches recently, stopped the country for a week…

      But the question is, could you handle 4 weeks of continuous overcast and drizzle..

  8. Yes, but in the South few get this. Here, snow happens when it’s colder than usual, ergo snow = cold in the minds of most. Unfortunate, but there it is.

  9. John Brookes

    Robert | March 4, 2011 at 12:05 am |


    “I’ve been working on a paper on temperature changes in the NE Canada region since 1880. ”

    That is amazing Robert, you don’t look a day over 100!

  10. One of personal concerns is whether or not a warmer world will bring more or less snowfall. This is important because water that falls as snow can easily be twice as likely to find it’s way into rivers and lakes than water that falls as rain.

    In my part of the world this is much more important for our water resources than shrinking glaciers, but I have never seen a thorough analysis of trends in observed snowfall.

    My relatively limited work on this subject suggests that warmer winters leads to a shorter snowfall season, more sublimation, and therefore smaller snowpacks available for spring melt. This in turn leads to drier conditions in early summer and therefore less runoff from summer rains. In the end, this means a lot less surface water.

    However, this is based on regional precipitation predictions from GCMs, which I know are not exactly the strong suit of the models.

    If anyone knows of any high quality, large-scale snow data I’d love to know about it.

  11. Perth, Western Australia has had a very hot Summer… Hottest on record I believe. That following on from our driest Winter on record. Methinks people in Western Australia are getting a good understanding of what global warming means…

  12. Gavin's Pussycat

    > Some (not tamino to be sure)

    Some people might say (not the Pussycat to be sure) that tamino should get a myxomatosis vaccination

  13. The temperature anomaly “going through the roof” is really impressing (even if the accelaration to +4 and even +6° on an annual basis is still a short term effect). I know the IPCC projects a quicker rise in temperature in this region, but if I’m right, by far not as quick as you show here.

    To interpret this acceleration in recent years: how does this acceleration compare to what is “normal” in this region, e.g. what tell historical records about temperature changes (eg. the Little Ice Age, the MWP or others) in this region?

  14. “The drastic loss of arctic sea ice … has altered air circulation patterns, enabling cold arctic air to penetrate further south.”

    The last two winters were indeed very cold in Europe. But I thought that 2007 was the year with minimum artic sea ice. Shouldn’t that have led to a European winter even colder than the last two?

    • Not necessarily. If you read carefuly, there is more cold air penetrating to the south, but also, more warm air penetrates to the north and penetrations are deeper and deeper (think of it like instead having “border” between cold and warmer air fixed at given latitude, this border is actually wavy and now amplitude of this waves is bigger). So at given longitude, you can have more warm air or more colder air. Now, it can happen, that at given location, for example Europe, “border wave” can be more north, meaning warmer weather or south meaning colder weather.
      But on average, at mid latitudes (all around the globe at same latitude), winters would be colder due to this effect.

  15. “Well, actually, those of us who live in California, especially skiers, are rather aware of what happens when moisture comes in from the Pacific and hits the Sierras.”

    Different situation entirely, John. There’s nothing even remotely similar to the Sierras around the Great Lakes, except the Adirondacks. And the Pacific doesn’t freeze over along the coast of California, and LA and the Bay don’t get lake effect snow.

    • Daniel Bailey

      Those of us who live in the shadow of the Big Lake (Superior) are well-acquainted w/ lake effect precipitation of all forms. That being said, here in Marquette winter is continuing it’s recent trend of average temps w/ occasional thaws w/ below-average precipitation:
      2011 graph
      2010 graph
      Great Lakes surface water temps
      Great Lakes ice cover
      Marquette climate data and graphs
      While conditions would seem to favor big snows, wind directions have been not conducive for lake effect snows, while most system snows have stayed south, resulting in below-normal precipitation.

      Back to the studio,

      The Yooper

  16. “The last two winters were indeed very cold in Europe. But I thought that 2007 was the year with minimum artic sea ice. Shouldn’t that have led to a European winter even colder than the last two?”

    Martin, 2007 had the lowest summer minimum–though your larger point is (perhaps) valid; this year is challenging for the lowest winter maximum, but 2006 (per IJIS data) was a whisker lower for this date. Some folks are expecting little or no further growth this year, and thus a new “minimum maximum.” (Not sure what grounds there are for such expectations.)

    But it’s not just low extent, according to those who assert a link between sea ice and WACCy weather; it’s low extents in certain specific areas. It relates to a circulation pattern called the “Arctic Dipole.” (And “WACCy” comes from “Warm Arctic, Cold Continents.”)

    NSIDC–the National Snow and Ice Data Center–has a new update, linked below. This February tied 2005 for lowest monthly sea ice extent ever. They also discuss WACC, but remain resolutely agnostic on the connection of the low ice to the cold winter weather.

    NSIDC Arctic Sea Ice Analysis:


    [Response: If I’m not mistaken, this year saw the record lows for December and January extent. And, Hudson’s bay was remarkably late freezing over.]

    • @Kevin McKinney
      “But it’s not just low extent, according to those who assert a link between sea ice and WACCy weather; it’s low extents in certain specific areas.”

      The last two cold winters in Europe are probably too short a period to constitute proof.
      Do you know where I could find a comparison of sea ice extent in ‘certain specific areas’ (which areas?) to temperature anomalies in Europe over the past 30 years? Perhaps somebody has already done that and created a nice graph showing the correlation?

      • Martin,

        I know of no such study and suspect there is not enough data at present to establish such a link.

        Overland[1] finds annual variability is still a major factor in the pressure field, however the signature of reduced Arctic sea ice is consistently detectable in the geometric pressure field thickness. i.e. “While we saw a positive anomaly in the thickness field in every year with extensive September open water, there is considerable year-to-year variability in pressure fields.”

        In terms of oddity of the last 2 cold periods in Europe (most of last winter and last year’s Late November and December) the warm Arctic cold continents pattern (WACC), has happened only 4 times since 1850, it happened in 1969, 1963, 1936, and 1881 (Overland at AGU poster session). So rather than look at the cold in Europe specifically it’s necessary to consider the pattern over the whole northern hemisphere. Here is a link to an image of the WACC pattern:

        The original was generated from GISS Maps.

        Overland[1] Large-scale atmospheric circulation changes are associated with the recent loss of Arctic sea ice. Tellus 2009
        (You might find this isn’t paywalled, apols if it is.)

  17. I’ve lived in Ohio and in N. Ca., and the experience–rate of change–is very different. In California we drive hours inland to the snow; the Lakes drive the snow to you in minutes.

    Great Lakes radar showing lake effect snow — along the lakes’ shores and inland: http://scienceblogs.com/bushwells/LakeEffectSnowRadar.jpg
    The rates of snowfall that occur in that area are astonishing: http://scienceblogs.com/bushwells/2008/02/a_lake_effect_whiteout.php
    Hat tip to Dr. Bushwell, see the article for much more detail.

    “One particularly dangerous aspect of lake effect snow is its dependence on the wind direction. A slight shift and you can go from partly sunny skies to living in the heart of a snowman within minutes. If you are unfortunate enough to have a band stall over you overnight (i.e, wind keeps blowing but does not shift direction), you may wake up to several feet of fresh snow…. On the plus side, those big, fluffy flakes are much easier to move than the dense little crystals we often see from storms which blow up from the south…. ”

    You can also in the same image see snow wrung out of wet air rising over higher land– looks different on the radar. That’s the green areas, including “the Tug Hill Plateau, one of the snowiest places in the lower 48 states…. typically receives over 300 inches of snow per year…. holds the 24 hour snowfall record for the lower 48 of some 77 inches.” and farther east, the New Hampshire mountains.

  18. Because many people intuitively but erroneously equate “more snow” with “colder,” a much better proxy is lake ice-out dates, which are much more tightly constrained by average winter temperature rather than snowfall, which is closely tied to weather. USGS has a nice dataset from 1850 to present for lakes and ponds in New England.


  19. well pissed off you only talk about the USD….how does the great nation of Ausatralia fit in…we have had a strange Summer

    [Response: ???]

  20. Response: If I’m not mistaken, this year saw the record lows for December and January extent. And, Hudson’s bay was remarkably late freezing over.

    I don’t think you are mistaken, mine host!

  21. @ wombat – AFAIK we in Australia aren’t overly affected by the warmer arctic, except insofar as it affects global temperatures :D

    Here is a short overview of how ENSO affects Australia, with a chart and links to La Nina and El Nino:

  22. Yes, thanks to La Nina, it was a little damper than usual here in Oz…
    This Bureau of Meteorology map shows the rainfall for last September, what they described as a “wet spring”.
    Yep, I’d call that wet!
    The rainfall anomaly maps are interesting too – especially stepping through month-by-month. You really start to understand why 75% of the state was flood-affected in January, if you look at the rainfall anomalies for December 2010 and earlier.
    With 10-20 inches more rain than the average, the ground was about as wet as it was going to get, the extra rain just ran straight off.
    I heard a report that peak inflows into the dam upstream of Brisbane hit a million megalitres (that’s a cubic kilometre of water) per day.

  23. “. . . a million megalitres per day.”

    Definitely non-trivial.

  24. 1 e12 liter/day=0.01 Sverdrup
    by comparison, the i think that all rivers combined is around 1 Sv, (Gulf stream around 55, ACC around 100)

    So 1/100 of all rivers was flowing toward Brisbane. Sounds damp.


  25. And it’s not just Brisbane. Lake Eyre is now filling with water for the third year in a row. Unheard of.

    3 years ago, hordes of people went to see the water because they thought they’d never get the chance again in their lifetimes. Hah!

  26. I would be interested to see the information contained in the last two graphs presented by way of a smoothed anomaly graph of some subset of the winter months. Not so interested that I’m willing to make the effort to do it myself, though.

  27. And Darwin had it’s wettest ever day in Feb, it broke many rainfall records with Tropical Cyclone Carlos…
    Been a weird year for Aussie weather…

    • Sure has been a weird year down under. Though all it has done is cemented the irrational/unthinking mindset of many who declare that if it floods or buckets rain, global warming is obviously not happening. They even say it as they comfortably wear their shorts and t-shirts out in the pouring rain.

      I’ve given up trying to unravel the logic behind it.

  28. Darwin’s 2011 wet season record is going to be hard to beat. It’s already the wettest wet season ever – and it’s got 2 months to go.


  29. Speaking of wet weather in queensland…

    The town of Bedourie got 319.1mm on Sunday – that’s 12.5″ – in a 24-hour period.
    Their “annual average” rainfall is 193mm.
    Their previous record daily rainfall was 188mm, set about a year ago, so they well and truly blew through that record… in fact their new daily rainfall record also breaks the monthly rainfall record!

  30. Tamino (or anyone who might know or want to hazard a guess),

    Just to make sure I have this right, in the plot of size of the annual cycle in temperature (the last plot) for NE Canada/W Greenland, the size presumably is an estimate of the difference in max to min temperature for a given year or something along those lines.

    The question for me is how do we interpret the data in the graph given it is a Semi-Amplitude as indicated by the y axis label?

    [Response: The “semi-amplitude” is, roughly speaking, half (hence the “semi”) the difference between the summer mean and the winter mean. Smaller semi-amplitude means smaller amplitude, i.e., smaller difference between summer and winter. In this particular case, it’s because winter has warmed more than summer, shrinking the difference between them.

    In more precise terms, it’s half the peak-to-trough difference of a best-fit sinusoid to a brief span of the data.]

  31. This post provides a poignant indication of how significant a region NE Canada and Greenland are in terms of monitoring the effects of climate warming. At the same time this post was published Jeff Masters at Wunder Blog published a summary Greenland Update for 2010 which can be found here:


    While his report focuses specifically on Greenland it strongly supports the conclusions of the blog post here for the general region considered.

    Some of the results discussed include:

    In August a 100 square mile by about 500 ft thick portion of The Petermann Glacier in northwestern Greenland fractured off due to calving. Masters indicates that Eric Rignot of UC-Irvine (mentioned by Steve Bloom in the Mission Failure post) at the AGU meeting in SF in December implicated warmer ocean waters attacking the glaciers from underneath as a key reason cause of the calving. He has an excellent visual from NOAA of the ocean temperature anomalies in that area showing significantly high values against the 1971-2000 average in the western southern and eastern waters around Greenland. It’s the warm oceans that lead to the record temperatures and melting in the weatern Greenland area.

    Record warming at indicated at several sites including Egedesminde (used in the analysis here), Prins Christi in the south and Godthab Nuuak in the southwest .

    Maybe the most impressive is the data on loss of mass from the Greenland Ice Sheet. A plot from Rignot paper is discussed indicating two methods of mass loss estimation including mass balance and gravity via the GRACE satellite. Its seems hard to believe but Masters indicates that between 1992 and 2002 Greenland was losing 7 Gt/yr. In that 10 year period that’s 70 Gt. From 2003 to 2007 the ice sheet lost was 171 Gt.

    This post is well worth a read for anyone studying warming in that region.

  32. One for Tamino’s comment?


    Random walk lengths of about 30 years in global climate
    “We have applied the relation for the mean of the expected values of the maximum excursion in a bounded random walk to estimate the random walk length from time series of eight independent global mean quantities (temperature maximum, summer lag, temperature minimum and winter lag in land and over the ocean) derived from the NOAA-CIRES twentieth century reanalysis (V2) for 1871–2008 and the ECHAM5 IPCC AR4 twentieth century run for 1860–2100, and also the Millenium 3100 yr control run mil01, which was segmented into records of specified period. The results for NOAA-CIRES, ECHAM5, and mil01 (mean of thirty 100 yr segments) are very similar and indicate a random walk length on land of 24 yr and over the ocean of 20 yr. Using three 1000 yr segments from mil01, the random walk lengths increased to 37 yr on land and 33 yr over the ocean. This result indicates that the shorter records may not totally capture the random variability of climate relevant on the time scale of civilizations, for which the random walk length is likely to be about 30 years. For this random walk length, the observed standard deviations of maximum temperature and minimum temperature yield respective expected maximum excursions on land of 1.4 and 2.3°C and over the ocean of 0.5 and 0.7°C, which are substantial fractions of the global warming signal. “