One of my long-standing interests is the location of ocean sediment series that enable apples-to-apples comparison of the 20th century to the mid-Holocene. These are not nearly as common as one would think. Ocean sediment series covering the Holocene typically stop prior to the 20th century due to core recovery problems and, on the other hand, high-resolution series (especially from box cores) that provide detailed 20th century information are not necessarily accompanied by corresponding Holocene information (even on “intermediate” resolution.)
Last week, Sicre and coauthors archived two very high resolution alkenone series from Placentia Bay and Bonavista Bay, offshore Newfoundland (5 and 9 years respectively), covering the last two millennium, with their most recent portion dated through the 20th century. While the proximity is not ideal, the cores do appear to be close enough to three Holocene alkenone SST series from Sachs et al 2007 to compare 20th century and mid-Holocene SSTs on the North American East Coast, an exercise that I will carry out in today’s post. The exercise has some added interest because the three Sachs 2007 series were used in Marcott et al 2013 (though their recent portions were foolishly re-dated by Marcott.) As high resolution ocean SST data over the past two millennia, the new Sicre data is also relevant to the popular two-millennium reconstruction period, but the new data is about as opposite to a Hockey Stick as one can imagine. Unsurprisingly, the new data was not press released and has thus far attracted no attention.
Location Map of Sicre 2014 and Sachs 2007 Locations
The two new alkenone series from offshore Newfoundland (Bonavista Bay and Placentia Bay) are from the same authors as the high resolution Iceland MD99-2275 series. Both new series are high resolution (5 and 9 years respectively). The two new cores are located to the north three alkenone series from Sachs (2007), that were incorporated (after re-dating) into Marcott et al 2013. The three Sachs 2007 series were in the Laurentian Fan, offshore Nova Scotia and offshore Virginia, with GGC26 (Laurentian Fan) at a similar longitude to Placentia Bay but several hundred miles to the south.
Somewhat surprisingly, the alkenone SST of the more northerly series (Bonavista Bay AI07-03G) is slightly warmer than the alkenone SST of the more southerly series (Placentia Bay AI07-12G).
Figure 1. Location map showing location of three Sachs 2007 cores and the two new Sicre 2014 cores.
High Resolution Sicre 2014 Data
Here is a plot of the two high-resolution new series. Neither series shows any recognizable uptick in the 20th century. The Placentia Bay series shows a noticeable decline over the two millennia. The most recent portion of the data is from box cores the tops of which have been convincingly dated to the end of the 20th century by Pb-210 dating.
Figure 2. Alkenone SST offshore Newfoundland, from Sicre 2014 archive. Both series are splices of gravity (early) and box core (late) with a short discontinuity in each series. Red – Bonavista Bay; blue – Placentia Bay. Both samples were taken through combining gravity and box cores, with the box core results coming to the present (2006). Note the slight gap between box and gravity cores.
In a Holocene Context: Comparing to Marcott GLB
In the next graphic, I’ve compared the East Coast offshore alkenone SSTs from the Laurentian Fan and Emerald Basin cores to the Marcott GLB (red) and Northern extratropic (NXT) reconstructions (green). I’ve centered each series to have the same mean over the 0-1870 AD period (left axis has Laurentian Fan GGC26 absolute scale, while right axis is anomaly basis 0-1870 AD). The most striking aspect of the East Coast alkenone SSTs is the remarkable Holocene warmth: 4 to 8 deg C warmer than the present. Even in the alkenone data prior to Sicre 2014, the variation over the past two millennia was very small relative to the Holocene variation. Even though the Sachs 2007 was much lower resolution than the Sicre 2014 data, it also showed a non-HS decline over the past two millennia – a pattern shown more clearly in the higher resolution newer data.
Figure 2. Alkenone SST. Left panel – Holocene; right panel – most recent two millennia. Laurentian Fan (GGC26) and Emerald Basin (GGC30) compared to Marcott GLB (red) and NXT (green) reconstructions. Sicre 2014 Placentia Bay shown in blue. Note that the Placentia Bay absolute temperatures are several degrees lower than Laurentian Fan (see Appendix.)
During the Holocene, mid-latitude and high-latitude summer insolation has decreased dramatically – by over 40 wm-2 in high latitudes. Since many, if not most, usual “proxies” respond to summer temperatures (with the invocation of annual temperatures mostly armwaving), it has long been recognized that one would expect a very pronounced signature for summer proxies on a Holocene scale. Esper et al 2012 pointed out that, even since Roman times, summer insolation at 60N has declined by ~6 wm-2 (about 1 wm-2 since medieval times), and acutely observed that one could hardly rely on tree ring “proxies” to record the smaller forcing arising from additional CO2 if they did not also record a Holocene signal – a point unfortunately not mentioned in AR5.
While summer high-latitude NH insolation has decreased sharply over the Holocene, winter high-latitude NH insolation has increased. Alkenones, as Richard Telford has recently emphasized, are responsive to summer temperature and thus the decline in East Coast alkenone SSTs since the early Holocene is to be expected. However, the size of the decline in East Coast alkenone SSTs since the Early Holocene is considerably greater than the corresponding decline in the European sector of the North Atlantic.
Sachs observed that a relatively small coastward displacement of the Gulf Stream could account for the difference and plausibly speculated that the Gulf Stream hugged the East Coast much more closely in the mid-Holocene. Even so, the differences in East Coast alkenone SST appear to be considerably higher than the SST differences in PMIP models (but this is a large topic that is outside the scope of today’s post.)
Readers need to keep in mind that SH insolation has changed more or less opposite over the Holocene – thus, at face value, the proxies ought to look a lot different in their patterns. I’ve spent a lot of time recently looking at SH Holocene proxies, but they do not appear to me to reflect summer SH insolation nearly as well as one would expect from the theory of NH proxies. But that’s a very long story.
Obviously the Sicre 2014 results provide further evidence against Marcott’s supposed early-20th century blade. At the time, I pointed out that the Marcott blade does not exist in the data and is entirely an artifact of incorrect data handling. To borrow a term from Mark Steyn, the Marcott blade was f……..flawed. It is reprehensible that Marcott and coauthors have failed to issue a corrigendum. And that specialists in the field, knowing of the error, have permitted the result to be reproduced and disseminated. While the blade in the original article may have been merely f….lawed, one could perhaps describe its promulgation as a more virulent form of the flaw. Or perhaps, to coin a word, flawed-ulent.
Defenders of Marcott have argued that, even if the uptick is flawed, this is immaterial to the Marcott estimate of the difference between Holocene and modern warmth. This is a large topic in itself – one that is made more difficult by defects in Marcott’s data analysis that are outside the scope of today’s post. In the above comparison, I’ve centered the data on the most recent two millennia (rather than 6000 years ago as in Marcott.) In my opinion, Marcott’s centering method disguises the real variation of mid-Holocene estimates and is very inappropriate for the objective of his study. But again a long topic.
But even with this limited preview, readers can readily see that the very large “residual” between the Marcott NXT estimate and the East Coast alkenone SSTs: up to 9 degrees for the Laurentian Fan SSTs. Regardless of whether this is an artifact of the proxies or a regional variation, this is the sort of thing that deserves attention. At face value, the alkenone proxies show that summer ocean temperatures along the North American East Coast between Washington and Newfoundland were 4-9 deg C warmer around 10000 BP than in the 20th century (with the larger differential to the north). In the far north, remnants of the continental ice sheet were still eroding.
In 2001, at the time of the IPCC TAR, a major report from the U.S. Global Change Program stated that the Holocene was a “relatively stable” period relative to the LGM or to the “20th century”, as, for example, the following:
Taken together [paleoclimate information from ice cores, ocean sediments etc], this information demonstrates that the Earth’s climate over the past 10,000 years has been relatively stable compared to the 10,000 years that preceded this period and compared to the 20th century.
While this statement may ultimately hold, the more that I look at Holocene data, the less obvious it seems to me. From the perspective of East Coast alkenone SST data, both the levels and rate of change in the 20th century seem unexceptional. Over the past decade, there have been many important new results from specialists in the longer Holocene period, results that, in my opinion, have been one of the most positive paleoclimate developments since AR4. For the most part, specialists in the field have attempted to understand their data, rather than throwing their data into a poorly understood black box in Mannian style, Marcott et al 2013 being an exception. Unfortunately, the many interesting Holocene results from specialists were given negligible coverage in AR5, which, instead, overfocused on diminishing returns over the past two millennia. Even worse, to the limited extent that AR5 addressed Holocene changes, it cited Marcott et al 2013, while neglecting numerous articles by more knowledgeable authors and specialists.
In addition to today’s post on East Coast alkenone series, I plan to look at other locations where high-resolution 20th century data can be compared to Holocene series used in Marcott.
The next diagram shows the original diagram of the three alkenone series from Sachs 2007. The diagram below also shows the Virginia Slope series, not shown in my re-plot above. The horizontal axis is (obviously) in opposite direction to my diagrams.
Figure A1. From Sachs 2007 Figure 2. Alkenone-derived SSTs over the Holocene in three slope water cores. Diamonds – Virginia Slope core CH07-98-GGC19 (36 52N, 74 34W, 1049 m); open circles – Scotian Margin core OCE326-GGC30 from the Emerald Basin (43 30N, 62 8W, 250 m); and solid circles – Laurentian Fan core OCE326-GGC26 (43 29N, 54 52W, 3975 m) (solid circles).
In the figure in the post, I re-centered the Placentia Bay series to the same mean over the past two millennia as the Sachs 2007 GGC26 series. The figure below shows the two series in their (original) scale. Unsurprisingly, given its more northerly location, the Placentia Bay series is about 4-5 degrees colder than GGC26, but both series have more or less similar declines over the past two millennia. Using alkenone SSTs from Placentia Bay to project GGC26 alkenone SSTs seems like a more sensible way of trying to compare 20th century alkenone SSTs to mid-Holocene SSTs than a Marcott black box. Needless to say, one has to examine many other examples from different geographies before trying to summarize conclusions.
Figure A2. GGC26 (Sachs 2007) and Placentia Bay (Sicre 2014) alkenone SSTs.
Here is KNMI HadiSST for 47N 55W (Placentia Bay) h/t Mosher for suggestion.