In this case they can find no instrumental data that correlates with the proxy data, but they proceed with their paper anyway? I wonder if the Mann hockey stick was their “instrumental” record.

Regarding coastal upwelling, it’s a function of both windspeed and wind direction. Changes in upwelling at a location may reflect shifts in mean wind direction as well as changes in windspeed.

We reconstructed the monsoon winds for the past 1000 years using fossil Globigerina bulloides abundance in box cores from the Arabian Sea and found that monsoon wind strength increased during the past four centuries as the Northern Hemisphere warmed.

That seems like a quite clear conclusion, monsoon winds are on the rise, and pretty sharply if their figure is any indication. Figure 2 of the Anderson paper (see original post above) says that monsoon wind speeds have increased fourfold from 1840 to 1975, and by 2-1/2 times since 1890!

But then they go on to say:

Comparing the G. bulloides record directly to instrumental climate records is difficult because of the short length and large variance of the instrumental ship data. The trend since 1890 in the average Arabian Sea southwest monsoon wind speed (29) (16° to 22°N, 56° to 62°E, June-July-August average) is negligible, although the variance has decreased by 30% (perhaps as anemometer use increased). The trend in all-India rainfall, although positive, is small and statistically insignificant (30, 31); however, the regional (32) and year-to-year variability is large (10%). Using historical data, Mooley and Pant (33) found no change in the last 200 years in the frequency of extreme drought events; however, it is not certain that extreme drought events should correlate directly with the upwelling index. We found no instrumental evidence suitable to verify or reject the increasing trend observed in 50-year averages for 1875, 1925, and 1975.

Translation: instrumental records show no change in the last century in monsoon wind strength, drought events, or rainfall, but by gosh, that doesn’t mean that our conclusions are wrong.

Say what?

w.

The two records move about in parallel from ~1150 to ~1300. From there, in Hole RC2735, the coarse fraction is steadily decreasing for about 400 years, up to ~ 1700. In RC2730, on the other hand, it is steadily increasing for about 500 years, to about 1800. From there, both of them increase rapidly to the present.

Part of the reason may lie in the undersea geography of the region:

As you can see, RC2735 is at an escarpment the edge of an upper plateau, which falls down to a lower plateau where RC2730 is located. Because of this, the coarse fraction will be very sensitive to the exact direction and depth of the currents in the area. If the current is running across the escarpment, there will be a large speed difference over the two areas, which will change the coarse fraction.

Alternatively, the presence of the escarpment may mean that the deeper plateau is swept by an entirely different current than the upper plateau.

Next, the coarse fraction is related to the dissolution rate of the shells. There’s a discussion of this in the paper Coarse fraction fluctuations in pelagic carbonate sediments from the tropical Indian Ocean: A 1500-kyr record of carbonate dissolution. In that paper, they say the changes in coarse fraction:

… probably result from changes in surface productivity associated with monsoon variability. Dissolution at this site may be ultimately controlled by the oxidation of organic matter which appears to be incorporated into the sediments in greater quantity during periods of weak SW monsoon and/or increased dry NE monsoon.

However, if monsoon strength were the only reason, we’d expect the two records to move in parallel since the winds will be quite similar at two oceanic sites that are only 10 km apart … but they don’t move in parallel.

Finally, the fact that the increase is all in the top couple of centimetres at least allows for the possibility that the fines have been partially washed out of the top of the sampled core.

w.

In the absence of any details on how the correlation was effected, I can’t see that the proposed correlation is conclusive. The MWP match between RC2735 and RC2730 doesn’t look very convincing to me.

This is likely because Anderson’s archived data don’t match the description in Anderson’s Science paper. The Science paper says:

We correlated 40 mm and 93 mm in RC2730 with 60 mm and 100 mm in RC2735 (Fig. 1).

The archived data, on the other hand, correlates 34 mm and 93 mm in RC2730 with 50 mm and 100 mm in RC2735. The correlation in the Science paper provides a much better fit, at least visually, to the MWP period.

w.

First, Anderson et al. 2002 discusses the connection between cores RC2735 and RC2730 – not 723A. Here is a plot from the archive for Andeson et al 2002 which is related to the statement:

“second core was stratigraphically correlated to the dated core by a combination of the G. bulloides, calcium carbonate, and coarse fraction stratigraphies.”

In the absence of any details on how the correlation was effected, I can’t see that the proposed correlation is conclusive. The MWP match between RC2735 and RC2730 doesn’t look very convincing to me.

Another point that I hadn’t noticed before: look at the tremendous increase in coarse percentage in the top 1 cm of RC2730. Is there any climatological reason to expect such a dramatic change in coarse percentage? Or could it be an artefact in which coarser fractions are washed leaving the fines. Here’s the top of the core:

depth_cm bp coarse carb bull
1 0.1 -36 27.1 NA 38
2 0.3 -17 23.9 55.6 34
3 0.5 2 20.6 54.6 27
4 0.7 21 16.0 52.8 24
5 0.9 40 17.0 53.1 25
6 1.1 59 13.5 52.6 25
7 1.3 78 10.0 50.1 28
8 1.5 97 8.8 48.3 17
9 1.7 116 8.4 49.1 15

There’s a strong association between pct G Bulloides and pct coarse, so if there is any artefact in the pct coarse, there will be a knock-on effect for pct G Bulloides. Maybe, by the time that the top 1.5 cm settles into core mode, it will be a lot lower.

Also note that there is no matching coarse and carbonate information for 723A. It’s possible that it was done and then not archived. But it also might not have been done. In any event, they don’t say that they matched 723A through use of all 3 variables and in the absence of a direct statement, I don’t think that you can assume that this is what they did for 723A.

A secondary point: note that using RC2730 without RC2735 enhances the modern-medieval differential.