Stenni et al (2017), Antarctic climate variability on regional and continental scales over the last 2000 years, was published pdf this week by Climate of the Past. It includes (multiple variations) of a new Antarctic temperature reconstruction, in which 112 d18O and dD isotope series are combined into regional and continental reconstructions. Its abstract warns that “projected warming of the Antarctic continent during the 21st century may soon see significant and unusual warming develop across other parts of the Antarctic continent [besides the peninsula]”, but no Steigian red spots of supposedly unprecedented warming.
Long-time CA readers will be aware of my long-standing interest in Antarctic ice core proxies, in particular, the highly resolved Law Dome d18O series. One of my first appearances in Climategate emails was a request for Law Dome data to Tas van Ommen in Australia, who immediately notified Phil Jones in Sauron’s Tower of this disturbance in the equilibrium of Middleearth. Jones promptly consulted the fiercest of his orcs, who urged that the data be withheld as follows: ” HI Phil, Personally, I wouldn’t send him [McIntyre] anything. I have no idea what he’s up to, but you can be sure it falls into the “no good” category.” I’ve discussed incidents involving Law Dome data on several occasions in the past. This is what the data looked like as of 2004: elevated values in the early first millennium, declining up to and including the 20th century.
Law Dome – Holocene Perspective
Recently, I’ve commented on many occasions on the benefits of looking at proxy data in a Holocene (10000 year context) rather than just the last 2000 years. A longer perspective permits one to see Milankovitch factors at work and this is true for Law Dome d18O as well. Although Law Dome d18O analyses were carried out nearly 20 years ago, results have been archived only for the deglacial period (~20000-9000 BP) and for the last 2000 years – shown in the graphic below. The inset shows (unarchived) Law Dome dD values over the Holocene, available only in a panel in a 2000 survey of Antarctic cores (Masson et al 2000). Though the data is frustratingly (and pointlessly) incomplete, the story is clear: d18O values were very low in the Last Glacial Maximum, then increased fairly steadily for 10000 years reaching a maximum ~9-10000 BP (in the early Holocene), then declined in the past 9000 years. Modern values are neither as high as in the early Holocene, nor as low as the Last Glacial Maximum. Variation over the past two millennia is relatively modest.
Accumulation during the Holocene is more than four times greater than in the glacial period. Elevation of Law Dome has decreased over the Holocene – an important factor which needs to be accounted for in temperature estimation – Vinther et al 2008 made a really excellent effort at disentangling elevation changes in Greenland d18O data, but no one seems to have made a corresponding effort in Antarctica (including Stenni et al 2017).
Stenni et al 2017 Reconstruction
Stenni et al 2017 calculated a variety of composites from the 112 series considered in their reconstruction, featuring reconstructions weighted by positive correlation to “target” temperature series (which had strong increases in West Antarctic and weak increases in East Antarctica), with negatively correlated isotope series screened out (weight of 0). This is disclosed in SI as follows:
The problem with this recipe is that, when the target has an upward trend (as do key target instrumental series), this methodology has the effect of enhancing the blade-ness of the resulting composite. The blade bias arises because the series are intrinsically very noisy – but series with too “big” a blade are left in, while series which go down are left out. The defective procedure is made worse when there are a lot of short series, as here. At least this methodology doesnt turn series upside down (Manng-nam style).
Stenni et al 2017 are somewhat evasive about their results and their graphics contribute to the evasion. I’ve re-plotted their Antarctic continent reconstruction (decadal version) from archived data in the figure below. Like the Law Dome series, the composite shows elevated values in the first millennium, declining through the last millennium, with the decline continuing well into the 20th century. Values in 1950 and 1960 were among the coldest in the past two millennia, with a very late uptick (1980- 2000). Stenni et al show this series as the dashed orange series in their Figure 8 which has negligible vertical resolution (see inset below). The very modest blade at the end of this series is almost certainly exaggerated by the defective screening and weighting procedures noted above. But even with their fingers on the scales (so to speak), the main message of the series is that values in the first millennium are consistently elevated above modern values.
Their main reconstruction graphic (their Figure 7) is, if anything, much worse than the panel shown in the above inset, as shown below. It too shows elevated first millennium values, though you’d barely know it from looking at the figure. Its 10:1 horizontal-to-vertical panel size disguises rather than highlights the difference between the first millennium and modern values.
By now, we’re all familiar with the fevered prose of abstracts when climate reconstructions supposedly show “unprecedented” modern results. Needless to say, Stenni et al does not contain colorful and excited descriptions of high first millennium values. The lede to their abstract is relentlessly flat:
“Climate trends in the Antarctic region remain poorly characterized, owing to the brevity and scarcity of direct climate observations and the large magnitude of interannual to decadal-scale climate variability. Here, within the framework of the PAGES Antarctica2k working group, we build an enlarged database of ice core water stable isotope records from Antarctica, consisting of 112 records.”
Continuing the abstract, they report “a significant cooling trend” to 1900 CE, followed by “significant warming trends” after 1900 CE in three regions which are “robust” to something or other and which are “significant” in the weighted reconstructions.
Our new reconstructions confirm a significant cooling trend from 0 to 1900 CE across all Antarctic regions where records extend back into the 1st millennium, with the exception of the Wilkes Land coast and Weddell Sea coast regions. Since 1900 CE, significant warming trends are identified for the West Antarctic Ice Sheet, the Dronning Maud Land coast and the Antarctic Peninsula regions, and these trends are robust across the distribution of records that contribute to the unweighted isotopic composites and also significant in the weighted temperature reconstructions.
This is a pretty outrageous spin, given that the continental Antarctic reconstruction continues the downward trend to 1950-60 – despite the use of a defective method which will enhance the most meager blade. Despite these adverse results, they close with the obligatory warning of “significant and unusual warming” – none of which is evident in their data.
However, projected warming of the Antarctic continent during the 21st century may soon see significant and unusual warming
develop across other parts of the Antarctic continent.
Discussion
As noted above, Law Dome has been a long-standing issue at Climate Audit.
It astonishes me that there is no technical journal article on Law Dome d18O data either for the Holocene or for the past 2000 years. Van Ommen planned to publish the data according to my earliest correspondence with him (2004). It’s disquieting that longer Holocene data for such an important site remains unpublished.
The characterization of Antarctic ice cores in the 2006 NAS report (discussed at CA here, especially at the press conference) was integral to their attempt to distinguish past warming from modern warming:
This [additional] evidence [of the unique nature of recent warmth in the context of the last one or two millennia] includes …the fact that ice cores from both Greenland and coastal Antarctica show evidence of 20th century warming (whereas only Greenland shows warming during medieval times).
However, this assertion in respect to Antarctica was not supported by their data or analysis. I tried unsuccessfully at the time to obtain a source. The Law Dome series, which was in circulation at the time, showed opposite results: warmth in the late first and very early second millennia and which didn’t show evidence of 20th century warming.
Drafts of IPCC AR4 showed a panel diagram of Southern Hemisphere proxies, but conspicuously omitted the Law Dome series. As an AR4 reviewer, I asked that it be included in the diagram (knowing of course that it showed a result that was opposite to what they were claiming.) The IPCC AR4 lead authors knew this as well and refused to show it in their diagram, concocting a ludicrous excuse. There was a revealing discussion in Climategate emails (discussed at CA here).
The Law Dome proxy series was important in the Gergis reconstruction as well. It met ex ante criteria for inclusion in her reconstruction. It was one of only three Gergis proxies with values in the Medieval period; if it were included in the network, medieval values would have been raised significantly. Rather than let this happen, Gergis concocted ex post screening criteria which excluded Law Dome from her network – see CA discussion here.