Climate Audit

I noted yesterday that Allen [2006] did not contain any new data supporting the idea of a "cool medieval Pacific", but that it was based primarily on a paper by Kim Cobb et al [2003] based, as it turns out, on two individual corals at Palmyra Island (6N) – one dated to the 10th century and one to the 12th century – both dated done by U/Th measurement, which show relatively elevated dO18 anomalies, while modern corals show relatively depressed dO!8 anomalies in the late 20th century. On a quick read, I’ve got several caveats about dating and replication. However, more fundamentally, Cobb’s recent sampling and interpretation of dO18 values at Christmas Island at 2N makes me wonder whether these results can’t be readily interpreted as evidence of a more northerly ITCZ in the 10th century, which would be a very interesting result and easily consistent with medieval NH warmth.

I met Cobb a couple of years ago at the AGU conference. In addition to working on corals, she was sampling speleothems in Borneo. I’ve gone into some abandoned adits in Chile for short periods and never much liked the bats and snakes, so I chatted briefly with her about such problems. The coral islands that she samples are spectacular looking.

Here’s Cobb’s argument. dO18 in tropical precipitation is governed by an amount effect – lower dO18 anomalies occur in summer, not winter; more negative dO18 anomalies are associated with El Nino. Note that she is associating more negative dO18 with higher temperatures (the scale is reversed in the graph shown later) while Thompson at tropical ice cores associates more positive dO18 anomalies with higher temperatures.

Climate at Palmyra is dominated by ENSO variability, manifest as warmer, wetter conditions that persist during El NinËÅ“o events and cooler, drier conditions during La NinËÅ“a events (Fig. 1a). Positive sea surface temperature (SST) and rainfall anomalies that occur at Palmyra during an El NinËÅ“o event both result in lowered (more negative) coral d 18O, while the converse is true during a La NinËÅ“a event. Thus, Palmyra corals are sensitive recorders of regional-scale ENSO activity. For example, a d 18O record from a modern coral (Porites lutea) at Palmyra shares 72% of its interannual variance with the NINËÅ“ O3.4 index (Fig. 2), demonstrating that coral records from this site provide reliable proxies of ENSO activity, at least over the twentieth century23. Our ability to reconstruct ENSO from fossil corals at this site rests in part on the assumption that the spatial patterns of ENSO have not changed significantly over the last millennium. (NAture)

Her main graphic from Cobb et al (Nature 2003) – the one that’ relied on for the "cool medieval PAcific" – is shown below. The top panel undoubtedly looks familiar – it’s Mann’s hockey stick which supposedly has not influenced other studies. The second panel shows dO18 levels from Palmyra island corals. Two things: the corals are from continuous. Modern corals only go back to the 19th century – fossil corals are dated to 4 other intervals: 17th century; 14-15th century, 13th century and 10th century. The theory of a "cool medieval period" arises from the less negative dO18 anomaly for the coral dated to the 10th century (and the more negative dO18 anomaly for the living coral – again take care with the reversed y-axis scale.)

Original Caption: Figure 5 Comparison of proxy climate records and external forcing during the last millennium. a, The MBH Northern Hemisphere temperature reconstruction34 (green) plotted with the Northern Hemisphere instrumental temperature record of ref. 48 (red). The green horizontal line denotes the mean of the MBH record for the period AD 1886″€œ1975. b, The monthly resolved Palmyra corald 18O records (thin black line), shown with a 10-yr running average (thick yellow line). The black horizontal line represents the average of the Palmyra modern coral d 18O for the period AD 1886″€œ1975. The black vertical bar represents the^1j error in mean corald 18O for single fossil corals (this error applies only to tenth- and twelfth-century sequences). The dating error is ^10 yr for the tenth- and twelfth-century sequences, ^5 yr for the fourteenth”€œfifteenth- and seventeenth-century sequences, and^,0.5 yr for the twentieth-century sequence. c, Reconstruction of solar irradiance anomalies based on historical sunspot records (anomalies calculated with respect to the AD 1886″€œ1975 mean)49 (purple) plotted with 10Be anomalies (a proxy for solar activity)50 (blue), plotted as a 3-point running mean and scaled to the solar irradiance anomalies. d, Radiative forcing associated with volcanic eruptions recorded in ice cores (black)35. The approximate timing and duration of the “‘œLittle Ice Age’ (LIA), the “‘œMedieval Warm Period’ (MWP), and solar activity minima”€?the Maunder minimum (MM), the Spo¨ rer minimum (SP), and the Wolfe minimum (WM)”€?are marked by horizontal bars.

Dating
Dating is done through U/Th dating as outlined in Cobb et al (EPSL 2003a), relying ultimately on the ratio of 230Th to 238U to date the corals. The table shown below shows her sample results – you will see that the ratios here fall into two main groups – and, in fact, cover only the modern corals and the 13-14th century corals. The "cool medieval" corals are not shown in this table, but presumably have even higher 230Th/238U ratios.

I find the lack of continuity in the dates to be disquieting – I realize that sampling was catch-as-catch-can, but why would there be a concentration of 14th century corals and a seeming absence of 18th century corals? Note that there is only one coral for the 12th century and one for the 10th century, so these results are not well-replicated.

Cobb also notes that natural thorium is present and that this has the effect of making the samples seem older than they really are. She considers a max and a min-adjustment, with the min-adjustment using a 230Th/232Th ratio of 2.4*10^-6, which is the value at secular equilibrium with earth’s crustal ratio of 232Th/238U. Her max correction is 2.0*10^-5, which is said to be the "highest value required to correct samples in the present study". I can’t see where this value is calculated. In her text, she discusses some 230Th/232Th measurements at the island: seawater values 1000 km from Palmyra were 5-10*10^-6 in surface water and 2*10^-4 in deep water, noting that upwelling could affect Palmyra’s seawater. She also notes that carbonate sands produced by ongoing erosion of the coral reef may contribute Th with 230Th/232Th ratios as high as 1*10^-2, in the case of 5000 year old corals. This value is obviously much higher than the values used in Cobb’s maximum correction. I don’t know enough about dating issues to comment further – these are merely issues that caught my eye.

ITCZ Movement
In discussions of the Cariaco (Venezuela) time series, Haug et al 2003 convincingly associate changing values with north-south movements of the ITCZ. While I’m far from having fixed positions on such matters, it seems like a plausible proximate cause for many centennial and millennial effects, although no consensus is reached on explaining centennial changes in ITCZ location (precession undoubtedly affects ITCZ location on Milankowitch scales). Cobb refers to the ITCZ in her most recent presentation.

If you re-examine the above information in light of her most recent results at nearby islands, I wonder if what we’re seeing here might be more resaonably explained by north-south ITCZ movements (noted elsewhere.) First, here’s the location of the three islands sampled by Cobb – although only the most northerly Palmyra is reported in the 2003 articles. The color coding shows El Nino effects.

Caption: During El NiàƒⰯ events, positive SST and precipitation anomalies both contribute to negative coral àŽⲱ8O anomalies in the CTP Interpretation of coral àŽⲱ8O on lower frequencies relies on assumption that warm SST drives higher precipitation in the CTP, and vice versa SST and rainfall anomalies during the 1982 El Nino

Next here is a very interesting figure from Cobb’s recent presentation at Florida State. Note that the dO18 anomalies at Palmyra (6N) are more depleted than at Christmas Island (2N). I’m not sure what she expected – but intuitively if the relationship was that more negative dO18 was associated with El Nino, wouldn’t you expect to see more negative dO18 values at Christmas Island? In fact, her results are the opposite! It’s too bad that a modern sample wasn’t taken yet at Fanning Island, but these are not easy places to get to, and one can hardly quibble.

If you re-examine the values in the top figure, you will see that the 10th century values at Palmyra – supposedly showing a "cool medieval Pacific" are in the -4.1 to -4.2 range – i.e. in the range of modern corals at Christmas Island about 200 miles to the south. Assuming that the corals have been dated correctly, isn’t it quite possible that the ITCZ in the 10th century was about 150-200 miles further north, thus yielding dO18 values at Palmyra corresponding to Christmas Island today? As I read it, the still unpublished dO18 coral values from Christmas Island completely overturn the cool medieval Pacific theory insofar as it was based on less negative dO18 anomalies at two Palmyra Island corals dated to the medieval period.

References:
Cobb, K.M., D.E. Hunter, C.D. Charles. A central tropical Pacific coral demonstrates Pacific, Indian, and Atlantic decadal climate connections. Geophys. Res. Lett. 28(11): 2209-2212, 2001.

Cobb, K.M., Coral records of the El Nino-Southern Oscillation and tropical Pacific climate over the last millennium. Doctoral dissertation, Univ. of California, San Diego, 165pp., 2002.

Charles, C.D., K.M. Cobb, M.D. Moore, R.G. Fairbanks. Monsoon-tropical ocean interaction in a network of coral records spanning the 20th century. Marine Geology 201(1-3): 207-222, 2003.

Cobb, K.M., C.D. Charles, R.L. Edwards, H. Cheng, M. Kastner. U/Th-dating living and young fossil corals from the central tropical Pacific. Earth and Planetary Science Letters 210(1-2): 91-103, 2003.

Cobb, K.M., C.D. Charles, R.L Edwards, H. Cheng, M. Kastner. El Niï〢½o-Southern Oscillation and tropical Pacific climate during the last millennium. Nature 424: 271-276, 2003. http://shadow.eas.gatech.edu/~kcobb/cobb03b.pdf
Cobb, K.M, 2006, Presentation at Florida State. http://shadow.eas.gatech.edu/~kcobb/cobb_fsu.pdf
Kim Cobb Website: http://shadow.eas.gatech.edu/~kcobb/