Guest post by David Middleton
The recent return of the Warming Island AGW myth inspired me to build a climate reconstruction for the Greenland Sea region.
Temperature Reconstruction
I performed a GISS station search centered on 71.4 N latitude, 23.5 W longitude and downloaded the 12 GISS/GHCN instrumental records with at least 60 years of continuous data up to 2011.
Next I calculated a temperature anomaly relative to 1961-1990 for each of the 12 stations and then averaged them together to create a temperature reconstruction. The climate in the Warming Island area is statistically indistinguishable from that of the 1930’s.
Then I took that reconstruction back to 1000 AD with the GISP2 ice core Ar-N2 data (Kobashi et al., 2010)…
The Modern Warming is also statistically indistinguishable from the Medieval Warm Period in the Warming Island / Greenland Sea region.
Arctic Sea Ice Reconstruction
It occurred to me that there might just be a relationship between the temperature anomaly and the Arctic sea ice extent. So I went to Wood for Trees and downloaded the historical NSIDC Arctic Sea Ice Index. Then I cross plotted an annual 13-month running average of the sea ice index against the average of the station anomalies and the GISP2 reconstruction (Kobashi et al., 2010) and found a pretty good correlation (R-squared = 0.67)…
Using the equation “Sea Ice Index = (-0.5976 * Temp. Anom.)+12.374” I calculated a Model Sea Ice Index.
The “Model Sea Ice Index” (white curve) is very similar to the measured sea ice index (cyan curve)…
Using the same equation, I extrapolated the Model Sea Ice Index back to 1000 AD using the GISP2 temperature data from Kobashi et al., 2010…
The model suggests that Arctic sea ice had been steadily expanding from ca. 1150 AD up until ca. 1800 AD and has been declining since ca. 1800 AD.
Next, I carried the model back to the Early Holocene using the Alley, 2000 GISP2 reconstruction…
This suggests that the sea ice contraction during the instrumental era (1979-2011) is not particularly remarkable.
Calibrating the Model
Realizing that my model has been extrapolated about 8,000 years away from real data, I decided to compare it to some real data. McKay et al., 2008 demonstrated that the modern Arctic sea ice cover is anomalously high and the Arctic summer sea surface temperature is anomalously low relative to the rest of the Holocene…
Modern sea-ice cover in the study area, expressed here as the number of months/year with >50% coverage, averages 10.6 ±1.2 months/year… Present day SST and SSS in August are 1.1 ± 2.4 8C and 28.5 ±1.3, respectively… In the Holocene record of core HLY0501-05, sea-ice cover has ranged between 5.5 and 9 months/year, summer SSS has varied between 22 and 30, and summer SST has ranged from 3 to 7.5 8C (Fig. 7).
My GISP2 (Alley, 2000) sea ice model is generally consistent with McKay et al., 2008…
Conclusion
“Move along, there’s nothing to see here.” The Arctic sea ice has “been there and done that” many times over the last 10,000 years without any anthropogenic assistance.
References
Alley, R.B. 2000. The Younger Dryas cold interval as viewed from central Greenland. Quaternary Science Reviews 19:213-226.
Kobashi, T., J.P. Severinghaus, J.-M. Barnola, K. Kawamura, T. Carter, and T. Nakaegawa. 2010. Persistent multi-decadal Greenland temperature fluctuation through the last millennium. Climatic Change, Vol. 100, pp. 733-756.
McKay, J.L., A. de Vernal, C. Hillaire-Marcel, C. Not, L. Polyak, and D. Darby. 2008. Holocene fluctuations in Arctic sea-ice cover: dinocyst-based reconstructions for the eastern Chukchi Sea. Can. J. Earth Sci. 45: 1377–1397
Michaels, P. 2008. “Warming Island”—Another Global Warming Myth Exposed.World Climate Report.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
RE: John Marshall observed that…
“The alarmist method of choosing summer to measure ice loss and scream ‘AGW’ but ignore the winter freeze, regardless of the depth of that freeze, shows a complete one sided cherry picked system of argument which proves nothing.”
I do tend to concur that the figures pertaining to Arctic Sea Ice Extent at the end of melt season (i.e. September) may take undue pride of place in the publicity stakes. (Although a variety of Naval forces and shipping companies may beg to differ.)
If it is indeed the case that there exists somewhere a dataset which paints a completely different picture during the winter months, then I’m sure may people would love to see said data. However, if this conjecture is merely predicated upon a personal view of how the world ought to behave, then a measure of “steeling ones’s self for disappointment” may well be in order.
A variety of sources for the relevant Sea Ice data exist, and, for example, it is straightforward to download monthly figures from the National Snow & Ice Data Centre stretching back to November 1978. Calculating the annual linear trends in Sea Ice Extent (with the equivalent Antarctic figures shown alongside for comparison) yields the following results…
(numeric values in thousands of square kilometres per annum, Arctic figures precede Antarctic)
Jan -49 8
Feb -46 5
Mar -43 12
Apr -39 16
May -33 23
Jun -44 17
Jul -68 13
Aug -72 10
Sep -81* 14
Oct -57 15
Nov -53 10
Dec -47 15
Ave -51 16
*The September 2011 figures should become available in the next couple of days, and as a consequence, the mean rate of ice loss for the month will change to approximately -84,000 sq km/annum.
Graphical representations showing trends across the entire year (as opposed to the supposedly cherry-picked summer months) are easily available from sites such as the NSIDC itself (using the average of the 1979-2000 figures as its baseline), Nansen Environmental (baseline 1979 – 2006) or the University of Bremen (baseline 1972 – 2008 for the Arctic, 1973 – 2008 for the Antarctic.)
Obviously, if the above datasets from NSIDC and the others have been falsified, then the original assertion may indeed be valid. Otherwise, it would appear to be somewhat lacking in merit.
Edim wrote…
“The most important/dominant time scale for humans is the last ~8000 years (since the interglacial maximum peak to the LIA minimum and present). The linear trend is COOLING. I don’t see the trend reversing in the future.”
Aha, good old “lies, damned lies and Statistics”.
The above assertion is, strictly speaking, correct. Unfortunately, it is also complete nonsense.
Edim, I’m quite happy to assume that you are making a genuine mistake here: however it is well past time for you to reassess your understanding of Statistics. Your starting figure of 8,000 years suggests pretty unambiguously that you are talking about the length of time from the Holocene Thermal Optimum up until today. Now, that represents a rather extended X-axis on a graph, and that is at the heart of your misunderstanding.
Can we, for the sake of simplicity, assume that the LIA petered out around 1850? The reason for choosing this figure is to allow the scale to be brought down to more manageable proportions. If we can agree that the LIA ended about 160 years ago, this gives us a rather obvious cancellation factor.
Instead of talking 8,000 years ago, we can divide this by 160, thus giving us a mere 50 or so “time units” to work with. (Each of these being a somewhat non-standard 160 years long.)
Now try out the following steps on Excel. (NOTE you can either do this on a chart, and ask for a linear trend line and formula, or, even easier, just use the SLOPE function.)
STEP 1: Along Row 1, enter the numbers 1 – 54. (This represents each of the 160-year units, and will simply provide a base for the X-axis in the chart and/or the SLOPE calculation.)
STEP 2: Starting in the same column as above, but along Row 2, enter the sequence 50, 49, 48, ….. 3,2,1 (This is a gross simplification, but is meant to be representative of declining temperatures from the Thermal Optimum. This will be the Y-axis time-series data values. The actual scale of the Y-axis is unimportant here if one is merely looking at the sign of the trend.)
STEP 3: Using either a chart and trend line, or the SLOPE function, calculate the trend. With the above figures, this will, unsurprisingly, produce a trend of -1/unit time.
STEP 4: Insert an additional entry of value 50 at the end of the Row 2 values, such that it now reads “…3,2,1,50”. Recalculate the trend, and it will now read -0.89
STEP 5: Add further values of 100,150, 200 into the next three cells of Row 2, such that it now has 54 entries and ends with “…3,2,1,50,100,150,200”. Recalculate the trend and it will STILL be negative!
Let’s try a put that into some physical perspective shall we? If you think that the difference in global average temperature between the Holocene Thermal Optimum and the Little Ice Age was (say) 3 degrees, then, if the temperatures over the next 5 centuries rose steadily to 9 degrees ABOVE the Holocene Thermal Optimum, then the trend over the 85 centuries would STILL be negative!
The problem with your reasoning, if you haven’t worked it out for yourself, is that the data from the end of the LIA till today only represents about 2% of the time series. When fitting a linear trend to the time series, this small proportion becomes utterly swamped by the other 98%. Even when we extrapolate out another 500 years, this only takes us up to about 7.5% of the whole series.
I obviously do not know what age you are, but I would be prepared to bet vast sums of money that you will not live to “see the trend reversing in the future”. For the trend to reverse sign during your lifetime, the Sun would need to go nova, and you would need to have found some method of surviving this cataclysm. (If you didn’t survive, it wouldn’t be happening within your lifetime.)
Under such extreme circumstances, I would certainly be prepared to admit that CO2 was just an unimportant trace gas.