For those worried about tundra melt and methane outgassing, this study might dampen those worries a bit. A new peer-reviewed study by Clegg et al. demonstrates that modern global warming is significantly less than the global warming experienced in the higher latitudes, specifically Alaska, during the summers of the last 3,000 years. It demonstrates that the Current Warm Period (CWP) is not unprecedented, at least for Alaska. The authors suggest a tie in to solar variability.
From CO2 science:
What was done
The authors conducted a high-resolution analysis of midge assemblages found in the sediments of Moose Lake (61°22.45′N, 143°35.93′W) in the Wrangell-St. Elias National Park and Preserve of south-central Alaska (USA), based on data obtained from cores removed from the lake bottom in the summer of AD 2000 and a midge-to-temperature transfer function that yielded mean July temperatures (TJuly) for the past six thousand years.
What was learned
The results of the study are portrayed in the accompanying figure, where it can be seen, in the words of Clegg et al., that “a piecewise linear regression analysis identifies a significant change point at ca 4000 years before present (cal BP),” with “a decreasing trend after this point.” And from 2500 cal BP to the present, there is a clear multi-centennial oscillation about the declining trend line, with its peaks and valleys defining the temporal locations of the Roman Warm Period, the Dark Ages Cold Period, the Medieval Warm Period, the Little Ice Age — during which the coldest temperatures of the entire interglacial or Holocene were reached — and, finally, the start of the Current Warm Period, which is still not expressed to any significant degree compared to the Medieval and Roman Warm Periods.
C3 Headlines provided an annotated and reversed graph which you can see below:
The paper title is published in Quaternary Science Reviews
Here’s the abstract:
Six millennia of summer temperature variation based on midge analysis of lake sediments from Alaska
Benjamin F. Clegg, Gina H. Clarke, Melissa L. Chipman, Michael Chou, Ian R. Walker, Willy Tinnere and Feng Sheng Hu
Despite their importance for evaluating anthropogenic climatic change, quantitative temperature reconstructions of the Holocene remain scarce from northern high-latitude regions. We conducted high-resolution midge analysis on the sediments of the past 6000 years from a lake in south-central Alaska. Results were used to estimate mean July air temperature (TJuly) variations on the basis of a midge temperature transfer function. The TJuly estimates from the near-surface samples are broadly consistent with instrumental and treering-based temperature data. Together with previous studies, these results suggest that midge assemblages are more sensitive to small shifts in summer temperature (0.5 °C) than indicated by the typical error range of midge temperature transfer functions (1.5 °C). A piecewise linear regression analysis identifies a significant change point at ca 4000 years before present (cal BP) in our TJuly record, with a decreasing trend after this point. Episodic TJuly peaks (14.5 °C) between 5500 and 4200 cal BP and the subsequent climatic cooling may have resulted from decreasing summer insolation associated with the precessional cycle. Centennial-scale climatic cooling of up to 1 °C occurred around 4000, 3300, 1800–1300, 600, and 250 cal BP. These cooling events were more pronounced and lasted longer during the last two millennia than between 2000 and 4000 cal BP. Some of these events have counterparts in climatic records from elsewhere in Alaska and other regions of the Northern Hemisphere, including several roughly synchronous with known grand minima in solar irradiance. Over the past 2000 years, our TJuly record displays patterns similar to those inferred from a wide variety of temperature proxy indicators at other sites in Alaska, including fluctuations coeval with the Little Ice Age, the Medieval Climate Anomaly, and the First Millennial Cooling (centered around 1400 cal BP). To our knowledge, this study offers the first high-resolution, quantitative record of summer temperature variation that spans longer than the past 2000 years from the high-latitude regions around the North Pacific.
Here’s an excerpt from the conclusion:
Within the limit of chronological uncertainties, some (but not all) of these cooling events at Moose Lake coincide with periods of reduced solar irradiance, such as the solar minima centered on the middle and late LIA (250 and 100 cal BP), 1400 cal BP, and 3400 cal BP (Steinhilber et al., 2009).
Although the co-occurrence of solar minima with cooling during the LIA is well appreciated, the role of solar output in modulating surface temperature remains controversial, partially because the effect of solar activity changes on the surface energy budget is orders of magnitude lower than those of the drivers operating over shorter timescales (e.g., clouds or volcanism) (Damon and Peristykh, 2005). Nonetheless, a number of recent paleoclimate
studies have attributed decadal- to millennial-scale variation to fluctuating solar irradiance in Alaska (Hu et al., 2003; Wiles et al., 2004; Tinner et al., 2008) and elsewhere (Hegerl et al., 2003; Damon and Peristykh, 2005; Eichler et al., 2009). Thus the potential role of solar irradiance in high-latitude climate change remains an issue that warrants further research (MacDonald, 2010). Analysis of midge assemblages in lake-sediment cores
from other sites is necessary to verify our results from Moose Lake and assess the potential linkages of summer temperature variation to fluctuating solar output.
The full paper is available at the Willie Soon’s website at Harvard here (PDF)
Some commenters point out that I did not include this caveat from the paper:
The Moose Lake TJuly record is of limited value for assessing anthropogenic warming in the context of the long-term natural variability because of the relatively coarse temporal resolution and potential impacts of human activity on the lake chemistry. The youngest sample of the record spans the period of AD 1968-1972, falling within the cooler interval of the 20th-century in Alaska (Chapin et al., 2005).
And they have a point, I should have included this. So I’m rectifying that now. They also say:
The inferred TJuly from this sample (13.76 +/- 1.43 °C) compares favorably with the mean of instrumental
July temperatures of the same period (13.77 +/- 1.13 °C,
corrected for a dry adiabatic lapse rate of 9.80 °C per km) as
recorded at a nearby weather station (Gulkana Airport). The relatively coarse resolution of the Moose Lake midge record, along with the brevity of weather-station records from our study region (w50 years), precludes a further assessment of our midge-based TJuly estimates through comparison with instrumental climate data.
However, the midge TJuly estimates of the past 350 years (Fig. 4A) can be compared with treering temperature estimates of the same period from tree line sites in the Wrangell Mountains (Davi et al., 2003; Fig. 4B). In general, the midge temperature inferences parallel the treering temperature patterns. For example, the two records exhibit similar magnitudes of climatic warming after the Little Ice Age (LIA) and both capture low temperatures corresponding to the middle and late phases of the LIA, which also coincide with local advances in valley glaciers within the Wrangell Mountains (Wiles et al., 2004). However, the specific peaks and troughs do not always match between the two records, which is expected given the chronological uncertainties associated with our 210Pb ages and with the age-depth model for samples older than 64 cal BP.
These results contribute to a growing body of evidence
demonstrating the utility of midge assemblages for reconstruction of relatively small TJuly variation on both historic and Holocene timescales (e.g., Heiri et al., 2003; Larocque and Hall, 2003). Together, these studies indicate greater sensitivity of midge assemblages to TJuly variation than implied by reported error envelopes of midge temperature transfer functions.
Clegg et al thinks that the TJuly agrees with a local instrument record, some tree-ring study, and suggest TJuly signal is greater than the error bands. However, this brings up an interesting point.
…changes seen in the sediment cores since about 1950 indicate expected climate cooling is being overridden by human activity like greenhouse gas emissions.
So we have one study, Clegg et al saying that this midge-paleo is too coarse to use for AGW signal determinations, and another similar study Yarrow et al saying midge-paleo (with others) does have enough resolution and it shows a modern impact of humans emitted GHG. Quite the contradiction.
In the Yarrow et al Baffin Island study, they do in fact look at more recent core data than the Clegg et al Alaska lake study. In reading the Clegg et al study, they say:
The youngest sample of the Moose Lake midge record (from 3.0 to 3.5 cm core depth; we did not have adequate amount of sediment from 0 to 3 cm for midge analysis) encompasses sediment deposition of AD 1968-1972.
Yet, in the Yarrow study they apparently did have enough sediment to make a determination and then claim that it shows unprecedented warming and human influence. Interestingly though, they cite a “statistical uncertainty of 2.2 °C”
As with any transfer function, chironomid-inferred temperatures contain some statistical uncertainty (14, 34). Although absolute temperature values have a statistical uncertainty of 2.2 °C, reconstructed trends in past temperature at this site are likely robust because the amplitude of these trends exceeds the statistical uncertainty of the model; furthermore, these trends are supported by many other proxies from the region (36).
So they also compared to other proxies. I find it odd though that Yarrow says this in the CU-Boulder press release here, emphasis mine:
But the cold-adapted midge species abruptly began declining in about 1950, matching their lowest abundances of the last 200,000 years. Two of the midge species adapted to the coldest temperatures have completely disappeared from the lake region, said Axford.
This seems to point to a sample problem for recent layers such as Clegg et al lament. I wonder what chironomid data Clegg et al had from 1972 forward and why they deemed it insufficient.
Apparently though, the lack of certain species wasn’t a problem for Yarrow et al, and they used that to bolster the claim that human caused warming was reflected by that species loss. I pointed to the fact that in Alaska and Canada, post World War II DDT use for mosquito control was the norm, so perhaps the lack of modern midges was a consequence of that DDT use in both cases. It is an uncertainty.
I’m reminded though of the Mann-Briffa Yamal tree ring debacle, where if that data didn’t fit near the present, you throw it out post 1960 and splice on the instrumental temperature record. Yarrow’s insistence that the cold species midge disappearance implies human caused warming is on par with the leap of “Mikes’ Nature trick”. Both ignore other potential influences.
While some commenters complain about the lack of Clegg et al data since 1972, the same posters IIRC did not complain about the proxy data truncation at 1960 and substitution of post 1960 instrumental data in Mann-Briffa’s studies.
While the lack of a useful sample post 1972 may simply be the lake biology, I think I’ll ask Clegg why they decided the post 1972 sample was insufficient and why Yarrow et al 2009 wasn’t referenced in the context of the modern midge data sample, and if they reply, and post a follow up note here.
I’ll close by pointing out Clegg et al’s closing sentence:
Analysis of midge assemblages in lake-sediment cores from other sites is necessary to verify our results from Moose Lake and assess the potential linkages of summer temperature variation to fluctuating solar output.
Replication is the basis of science, it is good to see them calling for that.