Man with one graph declares war on mosses (more on Miller et al 2013)

Miller et al’s (2013) study on very old dead moss emerging from beneath ice caps on Baffin Island, indicating unusual warming, must have touched a nerve with the so-called climate skeptics. I’ve previously shown that attempts by Anthony Watts and Steve McIntyre to denounce this paper were flawed. Today it is the turn of ex-geologist Don Easterbrook to bring out his one graph and embarrass himself. He does it with aplomb.

a. Miller et al. assume that the ice caps are cold-based (i.e., basal ice is frozen to the ground below) and that there is no basal sliding of the ice and no basal erosion. However, deep fiords and ice-scoured scoured bedrock in the area attest to active subglacial erosion (i.e., basal sliding rather than frozen to the ground), although most of the obvious erosion is probably related to Pleistocene glaciation. The Greenland ice sheet just across the Davis Strait at the same latitude is not frozen to its base, and the average summer temperature at Clyde (north of the sample sites) is 3°C above freezing during June, July, August, and September (Fig. 5). Summer temperatures of all of the more than half dozen weather stations along the east coast of Baffin, where the sample sites are located, are above freezing during June, July, August, and September. Thus, the Miller et al. conclusion that the small ice caps in this study are frozen to their base is highly questionable and most like not true.

I thought Easterbrook’s expertise was in glaciology, but obviously not, as this paragraph is packed with errors.

Miller et al find still-rooted dead moss and lichen with a radiocarbon date >44,000 years melting out from under the icecaps. Moss and lichens are fragile. Basal movement of glaciers grinds rocks to clay.  Had there been any basal ice movement during the last <44,000 years the plants would have been ripped off their roots.  Therefore the places where Miller et al find their mosses cannot have experienced basal movement of glacial ice in at least 44,000 years. Either the ice was frozen to its bed – a cold based glacier – or the ice was so thin  – a snowfield – that it was not capable of movement even though its bed was not frozen. Preservation of the moss is possible under either situation, it matters not which (but the ice must have survived or the plants would have been destroyed). There is plenty of evidence to support the existence of cold-based ice on Baffin Island, for example at  the Penny Ice Cap.

But back to Easterbrook’s “evidence”. There are deep fjords on Baffin Inland that were obviously formed by warm-based, erosive ice. These fjords would have been filled with thick  glaciers, thick enough to insulate the bed from the winter’s cold, so geothermal heat can warm the base to melting point.

Across Baffin Bay, the Greenland Ice Sheet, is only partially frozen to bedrock. Despite the great thickness of insulating ice in the interior of the ice sheet, only where the geothermal heat flux is high is there basal melting.

Clyde has an elevation of 26.5m. The  icecaps yielding old moss are all at elevations exceeding 900m. Please can someone inform Don Easterbrook about lapse rates (~5°C/km).

b. Miller et al. contend that the Laurentide Ice Sheet did not cover the area of the ice caps and that there has been no erosion since the Eemian Interglacial 120,000 years ago. However, the Laurentide Ice Sheet (LIS) extended eastward beyond this area to the coast (Fig. 1) and reconstructed ice surface elevations show the area to be close to the 1000 m and 2000 m contours, i.e., close to or above the present ice caps. The scale of the ice surface reconstructions is not detailed enough to show exactly how high the LIS surface was at the sites, but at least suggest a good possibility that the area was overridden by the LIS. The importance of this is their conclusion that the older sites have not been disturbed for 120,000 years, but to make this assertion they need to provide adequate evidence.

Easterbrook is simply unfamiliar with the literature, and does not care. It does not take much effort to find Miller et al (2002), with the delightful title “The Goldilocks dilemma: big ice, little ice, or “just-right” ice in the Eastern Canadian Arctic”. Miller et al discuss three hypotheses for the Laurentide Ice Sheet: that it covered Baffin Island completely; that it terminated inland of the mountains; and that it filled the fjords but otherwise stayed behind the coastal mountains. The evidence against the first two possibilities is strong, whereas cosmogenically dated glacial trim-lines in the fjords  and other evidence support the filled-fjord hypothesis. The mountains were not over-run by the Laurentide Ice Sheets. The adequate evidence has already been published.

c. The Miller et al. assertion that the ice caps were not more than 70m thick is highly questionable. The ice caps expanded noticeably during the Little Ice Age and even if the LIS didn’t overrun the ice cap sites, the ice caps must surely have thickened, especially since the surrounding lower areas were filled with LIS ice. Thus, their contention that the ice caps could not have been more than 70 m thick is most likely not valid.

Easterbrook has either not read or not understood the Supplementary Online Materials where the estimate that the icecaps cannot exceed 70m in thickness is explained:

Maximum size of a small ice cap

The relationship between ice-cap thickness and ice-cap area for an idealized circular ice cap on a horizontal plane is given by Nye, J. F. (1952). “The mechanics of glacier flow.”J. Glaciology 2: 82-93. Assuming a basal shear stress of 1 bar, the equation collapses to h = (23x)0.5, where h is the ice thickness at distance x from the ice edge, both expressed in meters. For a circular mountain summit of 200 m radius, the maximum ice thickness is calculated to be 68 m.

If Easterbrook want to contest the maximum thickness of the icecap, he needs to engage with  the methods the authors used rather than arguing from personal incredulity.

[2] Miller et al. claim that recent exposure of moss by melting ice proves that modern temperatures at the site were as high or higher than at any time since the moss was covered by ice and that therefore present temperatures have not been exceeded in 120,000 years. But is this necessarily true? If a block of ice is placed on the floor of a room and the thermostat is turned to 90°F, the ice will begin to melt. If the thermostat is then turned down to 40°F before all of the ice has melted, ice will continue to melt until the floor is uncovered, but to conclude that the temperature had never risen above 40°F since the floor was first covered with ice would be totally false. The same is true of the Baffin ice caps—if moss is uncovered at today’s temperatures, that doesn’t mean that higher temperatures haven’t occurred previously. Thus, the Miller et al. conclusions that“temperatures of the past century must have exceeded those of any century in more than 44 ka” and “there has been no intervening century during which warmth exceeded that of the last 100 years” are illogical and badly flawed. One wonders how this bad logic got past peer review. …

Easterbrook’s analogy is painfully flawed and divorced from any understanding of how icecaps respond to climate. In the analogy, the early 90°F conditions are far from equilibrium, the ice would have all melted had this state continued. In contrast the early-mid Holocene was long enough for the icecap to reach an equilibrium with climate. It does not need long, as it would take less than a century to go from the maximum possible thickness to the present dimensions. In the analogy, the cooler 40°F conditions immediately follow the 90°F. In reality, there were several thousand years where ice accumulation exceeded ablation and the icecaps expanded. If the last century was cool compared with the Holocene thermal maximum, the icecaps would have reached an equilibrium size, where ablation equalled accumulation that was larger than the Holocene thermal maximum icecap and Eemian mosses would not be being exposed.

[3] Among the 145 14C dates on exposed moss in this study are10 dates ranging in age from 23,900 to 50,700 years, leading to their conclusion that temperatures today are the hottest in >50 ka and most likely in the past 120 ka. They explain the disparity between these old dates and the multitude of young Holocene dates as due to higher elevations of the older samples so the younger sites could be exposed by melting of ice while the higher, older sites remained ice covered. But as shown by their data, this really isn’t true. Figures 1 and 3 show site M10-231v as an ‘Eemian’ site with dates ranging from 23,900 to 44,300 years. But ages at two nearby sites, M10-B226v and M10-223v, whose ages are shown as 2-3,000 and 4-5,000 years old, are higher than the site with old dates (Figure 4).

Elevation is not the only factor that determines whether an icecap melts or not. The aspect of the site will affect how much insolation and hence ablation occurs. Sites on the lee-side of icecaps or other topographic features will accumulate more snow. These other factors could easily be important enough to overcome the influence of elevation, especially when the icecaps are not in equilibrium with climate.

Summer temperature records at Clyde, north of the sample sites, show no warming from 1940 to 2009 (Fig. 5). How is it that “temperatures of the past century must have exceeded those of any century in more than 44 ka” when temperature records clearly show no warming over the past 70 years? This makes no sense at all!

If Easterbrook could show that there was no warming over 700 years, he would have a point, but as 70 years is less than a century, he doesn’t.

Finally Easterbrook reverts to his role as the man with one graph. He goes on at length about the isotope-inferred temperature reconstruction from the GISP2 ice core on Greenland. However, as Easterbrook has been told before [see comments by Nick Stokes and wheelsoc], this record stops at 95BP i.e. 1855. There is no 20th Century in this record, consequently it is absurd for Easterbrook to use it to claim that “virtually all of the period from 1500 years ago to 5000 years ago was warmer than modern temperatures” when all that can be demonstrated is that “virtually all of the period from 1500 years ago to 5000 years ago was warmer than 1855”.

Easterbrook’s post is a mix of nonsense and ignorance, garnished with his one graph.


About richard telford

Ecologist with interests in quantitative methods and palaeoenvironments
This entry was posted in Fake climate sceptics, Peer reviewed literature, Silliness, WUWT and tagged , , . Bookmark the permalink.

20 Responses to Man with one graph declares war on mosses (more on Miller et al 2013)

  1. John Mashey says:

    No surprise, given commetns by WWU geology faculty about their long-retired emeritus colleague.

  2. JCH says:

    Layman here. I have a question about resolution. On CA they have a graph from the Miller paper. In that graph the direction of the mean OHC changes repeatedly, often several times within a 100-year span. In a video one of the authors says the changes in OHC are smooth. Another author speculates their resolution is around 100 years. He seemed unsure about that. If the mean OHC changes directions multiple times within 100 years, doesn’t that indicate their proxies do include evidence of changes on a time scale of significantly less than 100 years?

    Also, in the taped interview done by Revkin, Yair Rosenthal seems to be saying that the modern surface air temperature has either come close to or exceeded the surface air temperature of the MCA, so I don’t think he has concluded his paper establishes the MCA was warmer than today.

    • Hei,

      I think you are asking about Rosenthal et al (Pacific temperatures estimated from Mg/Ca) rather than Miller et al (Arctic moss). I’ve not yet looked into the details of Rosenthal et al. but you have to remember that there is some noise in the proxy measurements. Some of the high frequency variability is this noise, rather than anything real.

    • Paul S says:

      It’s not necessarily the case that relative trajectories of OHC and surface temperature have to be congruent over these long timescales. I’ll give an example using the MPI-ESM-P simulation of the past millennium. This plot shows thermosteric sea level change over that period, which would strongly correlate with OHC, and this plot shows surface temperature evolution.

      In the first plot, relating to ocean temperatures, the “MCA” period is clearly much warmer but current temperatures are clearly warmer at the surface.

      Note that this shouldn’t at all be considered an accurate simulation of what was shown in the Rosenthal paper*. Much the same trend can be seen in the control run for this model so it’s the result of “drift”. Other models produce drift of opposite sign. The point is simply that finding a warmer ocean around the medieval period shouldn’t have much weight in debate about relative surface temperatures.

      *On the other hand, I guess the processes involved in what we call “model drift” could be broadly similar to those which produce Holocene length trends like those seen in the Rosenthal et al. paper.

  3. Paul S says:

    To paraphrase a character in political satire In The Loop: “In the land of truth, my friend, the man with one graph is King.”

  4. Carl says:

    What’s your comments on this? I’m confused..

    • That’s the summer temperature record from Clyde (26.5 m asl), showing a warm period in the 1940s-1950s, followed by cooling and then warming again. Some information about decadal fluctuations in temperature, which generally corresponds well with the Arctic-wide pattern, but no information about how the 20th century compares with earlier centuries. Therefore it provides no evidence for or against Miller et al’s conclusions that the 20th century was the warmest century since at least 44,000 years ago.

  5. Carl says:

    Hi, thanks for the answer. But the lack of warming over the last 60-70 years? And since human induced warming is supposed to be small/negligible before this record starts (0.2-0.3 degC at most)? I guess the conclusion must be that the recent warm century at Baffin Island is due to natural causes if this study is correct?

    • Paul S says:

      If you roll two dice and get six on both you might think yourself lucky. If you did that ten times in a row you might think something was up.

      Consider that we’re basically talking about anomalously warm multi-decadal periods in the context of the Holocene here. I would wager that there have been some periods of anomalous warmth similar to the 1940s earlier in the Holocene. What perhaps has not been the case is that this anomalous warmth has persisted, to some degree, for several more decades, leading to prolonged perturbation of the ice cap.

      In other words, we appear to be continuing to roll sixes. Even if the temperature hasn’t increased since 1940 (more on that shortly) there’s still plenty to suggest something unusual is occurring. Also consider that a wealth of other evidence suggests Arctic temperatures have generally declined over the millennia since the mid-Holocene, so it is highly unusual to find current temperatures being the highest in the Holocene, and that’s difficult to reconcile with only natural factors being involved. Admittedly the micro-climate on Baffin might be different to the wider Arctic in this respect, I’m not sure what data is available to check that.

      On to the temperature record: The Clyde NWT plot you posted doesn’t appear to correspond to the data included in the latest GHCNv3. Presumably it’s from an earlier version? GHCNv3 includes station homogenisation procedures which attempt to find breakpoints in station data and it appears to have found at least one here, suggesting that temperatures in the early record were artificially higher. BEST use a different procedure but appear to have found substantially the same thing:

      This plot shows the GHCNv3 Clyde data alongside output from the relevant grid in the BEST gridded dataset. Linear trends suggest warming at this location over the whole record has been about 0.6 – 0.8ºC.

    • What are the odds of the orbitally-forced Holocene trend for cooler Arctic temperatures to be reversed by chance at the same time that humans emit a radiatively active gas. Minuscule.

      If natural variability was enough to melt the icecaps on Baffin Island down to their current configuration, why did it not happen millennia ago, especially as the orbital configuration was more favourable then.

      The early 20th Century Arctic warming was probably enhanced by deposits of black carbon changing the albedo.

  6. Carl says:

    Paul S, yes, the Best date shown no warming as well (of course). The other data has very bad reputation (politically motivated adjustments). I can only conclude that either the Miller study is wrong or the recent century at Baffin Island is warm due to natural causes. Actually both statements can be true the same time.

    • Paul S says:


      Not sure what you’re saying here. The BEST analysis finds the most warming, as you can see on the plot I posted. About 0.8ºC over the past 70 years.

    • Jim Bouldin says:

      “The other data has very bad reputation (politically motivated adjustments)”

      As you say something like that, you call your objectivity into serious question for some of us, especially given that the BEST results put those kind of claims to rest pretty thoroughly.

      Miller et al’s conclusions wrt Baffin Island’s temperature extremes history may or may not be correct, but not for that reason.

  7. Conclude on what grounds? That it does not fit your preconceptions?

  8. Carl says:

    Hi Richard, no because it fits the data. Regarding earlier melt at warmer times (MWP or before) we need to consider the thickness of the ice (time lag) which is very uncertain and winter deposition at those times where data is uncertain. But black carbon in 19th and 20th century could be significant also in melting the ice, I agree with that. But that is not related to the greenhouse effect. How this study is linked to global warming is beyond me. It must be because it fits someones preconceptions.

    • What data?

      The time lag is not uncertain. These are not enormous ice sheets that take millennia to melt, but thin icecaps that could melt from their maximum possible size to their current size in less than a century. Of course you could arguing that the MWP was only a few decades long, but Miller’s claim is that the last 100 years has been the warmest century, and makes no claim about warmer intervals.

      What data do you have that suggests that the 20th Century is unusually dry in Baffin Island relative to the rest of the late Holocene? What are odds that Baffin Island is unusually dry by chance in the same century that humans release enormous amounts of radiately active gases into the atmosphere?

  9. Jim Bouldin says:

    I still think there are a couple things that require further evidence and/or thought. One is the idea that the current ice melt rate in the area (0.5 m/yr) is the figure that should be applied to estimates of former reductions in ice cap mass (and hence, estimates of how long it would’ve taken them to shrink to current sizes). It’s the average melt rate over the last century that should be used, not the current rate, and this point increases in importance when one considers that black carbon (i.e. micro- not macro-temperature) might be involved for some unknown portion of that time. If the estimated melt rate is on the high side, which based on the Clyde T record it might very well be, your time-to-shrinkage estimates go up inversely, which weakens what you can say w.r.t. uniqueness of recent times.

    A second point is that they are sampling the moss/lichens as soon as they are uncovered by the melt. This means there is more material of the same age still sitting under ice, and we don’t how much, and we won’t until it all melts away. I don’t have a feel for how much ice remains at these 145 sampled caps, and thus how significant this effect might be, but it will take some more heat to melt what remains.

    The combined effects of these two points need to be considered, IMO.

    • Yes, ideally Miller et al would have melt rates across the century rather than just 2000-2005, years that are at Clyde somewhat above the mean for the last 100 years.

      I’m not sure how much this matters. Consider what the case for low melt implies. The icecaps have gone from their Little Ice Age maximum extent to their current size in a century or so. If the melt rate over most of the last century was on average much smaller than the 2000-2005 melt rate, then the integrated melt over the last century is also small, which means that the icecap cannot have been very thick at the end of the LIA. The thickness of the ice in the LIA was probably the maximum thickness reached in the Holocene (except perhaps the the early Holocene). If the icecaps have been thinner than their maximum possible thickness, which is what a low mean melt rate implies, they are even more sensitive to warmth.

      For a large part of the Holocene, the ice would have been nothing like it’s maximum thickness. During the early Holocene icecap size must have been small, it would not have taken much extra warmth to shrink them a little further, so that only Holocene moss was appearing now.

      The second point will be resolved in time. The icecaps are not in equilibrium at the current temperature. Even if the temperature remains constant, the ice caps will retreat towards an equilibrium size (which may be zero).

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s