Forging Arctic Heat

Arctic warming is still very much in the news and there is on-going concern that this may cause accelerated melting of permafrost, release of even more CO2 and methane and a form of runaway warming. A little known “fact” is that many parts of the Arctic were just as warm around 1940 as they are today. This is a theme I will return to shortly with a few more comprehensive data sets. In this short post I simply want to take another look at the two records close to Yamal – Ostrov Dikson and Salehard – that I mentioned in my recent post on the Yamal “vent”.

Figure 1 GHCN V2 “raw” in blue and GHCN v3 “adjusted” records in red for Ostrov Dikson, near Yamal, Arctic Russia. Scales offset by 3˚C to facilitate comparison. Can you spot the forgery from the Da Vinci?

For the uninitiated, GHCN stands for Global Historic Climatology Network a part of NOAA in the USA and charged with curating climatic records. GHCN V2 are often referred to as the “raw” records but have in fact been subjected to some adjustments. GHCN V3 homogenised records have been subjected locally to very significant adjustments aimed supposedly at removing non-climatic artefacts from the data, such as moving the recording site or the time of day recordings were made. The V1 actual raw records are evidently not available on line but are rumoured to be circulating on a CD. The story so far is that within any region, some records are warmed and some are cooled with a close to zero net effect.

Figure 2 Running regressions through the raw and adjusted records we see that a gradual cooling trend is turned into a significant warming trend. It needs to be noted that 1943 and 1945 were significantly warmer than the present day, a “fact” removed by the adjustments made to the records. Temperature scales are offset by 3˚C to enable comparison.

Figure 3 So how is this magic achieved? Quite simply by adding 1.6˚C to the raw record from 1957. This is one of the more blatant and significant versions of data adjustment I’ve seen.

Figure 4 In Salehard its even more difficult to spot the forgery from the Da Vinci. A close inspection shows that in the raw record, 1942 ties for warmest year. In the adjusted record 1995 is the clear winner.

Figure 5 A more subtle approach to homogenisation is applied in Salehard with gradual but progressively more cooling added to the warm leg of the raw data in exact decimal fraction increments. 

Figure 6 The average of 21 GHCN V2 raw records for E Siberia has the same form as Ostrov Dikson and Salehard with characteristic warming leg, cooling leg and a recent warming leg.

Are these adjustments justified? As far as I have seen so far, this form of record is common in the Arctic. In fact it seems to be characteristic of the Arctic dominating in E Siberia (See Figure 6) and in Northern Scandinavia. So what we are left with is a methane vent that probably isn’t on Yamal that formed in response to warming manufactured by a GHCN bot.

This entry was posted in Climate change and tagged , , , , , , , . Bookmark the permalink.

22 Responses to Forging Arctic Heat

  1. Dave Rutledge says:

    Hi Euan,

    Great post. The arctic is also a place where I would be on the lookout for a warming trend bias because of increased residential heating since the 1800s.


  2. Syndroma says:

    If you’re interested to see how people live and work in that part of the world, check out this trip on a nuclear icebreaker from Dudinka port to Dikson island: (lots of photos)

  3. Hugh Sharman says:

    By all acounts, the short-lived Eemian was a warmer interglacial than our own, with sea water levels 5 meters higher than now.

    Forests spread right up the east Greenland coast. The conditions for a run-away “catastrophic” GW were even more “ideal” than now, yet the ice returned and sea levels fell 125? meters in about 100,000 years, until about 18,000 years before the present.

    And by the way, that makes the Holocene look especially long lived! Yikes!

  4. William says:

    A silly question perhaps, but do you not find the mid-50s step rather unlikely? Just about all the data in the Stalin-era is one or two degrees warmer than just about all of the data in the decades after Stalin’s death. Only after 2000 did temperatures return to those of the Stalin era – so if you do take the data at face value, there has clearly been a rise of 1 or 2 degrees in the last decade. It seems more likely that the pre-55 data is too hot – maybe a Gulag Heat Island effect, all those millions of ‘zeks’ distorting the measurements?

  5. manicbeancounter says:


    You are now on the trickiest region of the world to understand. It has much bigger temperature fluctuations than elsewhere in the world and the differences between temperature stations are also huge. There are possibly two influences at work here, that might be working together.
    First, the further North you go, the bigger the temperature fluctuations both year on year and in the recent warming cycle. this might be a combination of cold near to seas that freeze for part of the winter, such as is found in the Antarctic Peninsula.
    Second is the warmer sea currents from the Atlantic, changes to which might have some influence over longer periods.
    The data is very sparse, particularly for the period prior to 1950. Coupled with this natural variation it makes it extremely difficult to develop a temperature contour map for the whole region. Any simple homogenization bot will make quite massive mistakes. Even rigorous evaluation using the same data by the same methods by different people could obtain quite different results. One example is to compare Svalbard at 78N with Reykjavik at 64N. For Svalbard there are two separate station records about 20 miles apart that need to be compared together. When this is done, the warming of 1910-1940 is very similar in magnitude to the warming of 1976-2006. The temperature rise in each case of 5C compares to a 2C rise in each period at Reykjavik.

  6. William says:

    Why not print my comment Euan? It must be obvious to all but the blindest of readers that there really is a step in the mid 50s that needs explaining. Of course the chances are that many people have pointed that out to you, as I have, but only uncritical comments are published. Your comment thread is as much a work of fiction as your articles.

    • Euan Mearns says:

      That step in the 1950s is seen throughout E Siberia as detailed in the average of 21 stations shown in Figure 6. It is also present in N Scandinavia, spread among stations from 3 countries where the records are among the most strongly congruous I’ve seen.

      In E Siberia, the temperature variance is not spread evenly among the seasons. The Summers are pretty flat, most of the variance is in the winter record. This from memory since I’m not at my desk right now.

      But your comment hits an important point. Does automated homogenisation remove genuine climate signal?

      • Roger Andrews says:

        I don’t know what distances GISS, NCDC and CRU homogenize records over, but BEST homogenizes them over an area of 12.6 million sq km. This will certainly smooth out if not remove entirely a lot of local or even regional climate signal.

        A specific regional signal that gets smoothed out by homogenization in the Arctic is the 1940-1970 cooling period (Ostrov Dikson is an example). This cooling period was confined largely to higher northern latitudes, and the smoothing makes it appear as a flat spot rather than a cooling period in the global record. This makes it somewhat easier for CO2-driven climate models to “hindcast” the global temperature record over this period, although they still can’t hindcast the cooling in the Arctic.

      • William says:

        Not allowing people to defend what they said is dishonest Euan. Such selective posting of comments from me and others just manipulates what is said and misleads your readers. Please either prevent me from posting entirely or allow entirely unless I’m outside blog rules. Here’s what I said:

        Of course there could be a genuine climate signal there. There also might be a non-climate effects from the gulag camp complex that apparently existed at Dikson, the large influx of prisoners in the post war Stalin era, a probable drop in prisoners after Krushchev took over, the emptying of the area after the fall of comunism, instrument changes, location changes, etc. But you see only “forgery” because that is what you wan to see.

    • Euan Mearns says:

      PS – in Russia, this type of record is typical E of the Urals. Go west of the Urals and the three legs are not there. This is what identifying congruous climate zones is all about.

  7. Maybe this comment is too late, but anyway.

    Just imagine we had no thermometers, no adjusted, homogenized, twisted, screwed, “intentionally manipulated in order to get more grants/money for research” etc. data. Imagine, for a while, we had only sea level rise (SLR) data (where +93 % percent of heat is being absorbed, compared to 2-3 % of heat going into the atmosphere). Would we have a clue that Earth is warming, and rapidly so?

    Well, I would say yes. Here are some links. Global satellite based altimetry data are showing us the sea level is rising at a rate of 3.28 mm/year since 1993. Except that SLR is spiking right now as a result of El Nino and its associated effect on global ocean temperatures, how fast is the 3.28 mm/year of the SLR?

    According to this PNAS study…:

    Of particular note is that during the ∼6,000 y up to the start of the recent rise ∼100−150 y ago, there is no evidence for global oscillations in sea level on time scales exceeding ∼200 y duration or 15−20 cm amplitude.

    so that even 2 mm/year is fastest in at least 6000 years. Now we have almost 3.3 mm/year, and it will get even faster by the end of the century, probably closer to 4-5 mm per year. And then for centuries, no matter what.

    I think this information per se should be enough to convince most people we are undergoing a rapid global climate change.


  8. JerryC says:

    Let’s see, 3mm a year. So if I go out and stand exactly at sea level now, in 100 years I’ll be standing in water to about halfway up my calf.

  9. mbe11 says:

    If you look at the sea level rise, global satellites have not been around all that lone. so lets look at something that has, tide gauges in the middle of the ocean far from continent edges which tend to move up and down a lot as any look at the records shows. try 1619910 Midway Atoll, PO, 1.19 +-0.02 mm a year which is not that far off from similar islands in the middle of nowhere that are not sitting on active volcanoes, some others are even less, a lot less. The ocean is not actually rising at the fastest in at least the last 6000 years only the paperwork.

Comments are closed.