Did the floating bog quiver in the sunshine?

The current edition of The Holocene has a paper by Xu et al that reports a relationship between solar activity and millennial-scale failures of the Indian Summer Monsoon as recorded in peat deposits of Lake Xihu, southwestern China. Is this  robust evidence of the influence of solar activity on palaeoecological proxies?

The peat deposits were formed by a floating swamp. These scare me. They tremble when you walk on them and I worry about falling through. Fortunately, this swamp is now grounded.

Xu et al report bulk density, total organic carbon (TOC), magnetic susceptibility, and carbonate content. These are commonly and easily measured proxies, but normally play a supporting role rather than taking the lead, in part because they can be difficult to interpret by themselves.

Xu et al Figure 4. Comparison between Lake Xihu proxy indices and solar activities. Solar activities were indicated by detrended atmospheric 14C  concentration extracted from tree rings (a: red; 11-point moving average; Stuiver et al., 1998) and reconstructed total solar irradiance (TSI)  (b: purple; 11-point moving average; Steinhilber et al., 2009). Pink triangles show the 14C dates. Other data shown which also are derived from core Xihu-1 include bulk density (c: blue), TOC (d: green), magnetic susceptibility (e: light blue), and  carbonate content (f: orange). Grey-shaded columns highlight periods of broad-scale ISM failures during the early to mid-Holocene

Xu et al Figure 4. Comparison between Lake Xihu proxy indices and solar activities. Solar activities were indicated by detrended atmospheric 14C concentration extracted from tree rings (a: red; 11-point moving average; Stuiver et al., 1998) and reconstructed total solar irradiance (TSI) (b: purple; 11-point moving average; Steinhilber et al., 2009). Pink triangles show the 14C dates. Proxies from core Xihu-1 are bulk density (c: blue), TOC (d: green), magnetic susceptibility (e: light blue), and carbonate content (f: orange). Grey-shading denotes periods of broad-scale Indian Summer Monsoon failures during the early to mid-Holocene

Xu et al identify periods of high detrital input into the peat (high bulk density, low TOC, high magnetic susceptibility and/or high carbonate), interpret these as periods when the Indian Summer Monsoon failed, and correlate these periods with changes in solar activity. No attempt is made to test if the correlation between the purported monsoon failures and the solar proxies is better than chance. Methinks it is like a weasel.

There are two critical questions. First, is the interpretation of periods with high concentrations of in-washed detritus as monsoon failures reasonable, and second, is the chronology adequate?

I would typically associate an increase in detritus as flooding. Not the obvious response to a monsoon failure. An alternative, and opposite, explanation is that drought during the monsoon failure reduced vegetation cover, so when it rained the soil could be easily eroded and transported to the lake. A third alternative is that these periods with high detritus signify periods with enhanced oxidation of the peat, concentrating the detritus. Xu et al’s interpretation is different. They argue that the factor controlling the supply of detritus to the peat is the amount of vegetation on the bog, and that this is influenced by temperature and nutrient supply. They argue that high temperatures and high precipitation leads to a plentiful supply of nutrients to the lake give high biomass, and that monsoon failure gives low temperatures, nutrient scarcity and low biomass (water is unlikely to be a limiting factor on a floating bog). This is not an implausible scenario, but without additional proxies, it is not very convincing and it is difficult to determine if it is a better explanation than the alternatives.

It would be much better to use a site/proxy that was directly sensitive to hydroclimate variability than one where the purported relationship is mediated through nutrient supply. This makes me wonder if Xu et al had set out to study the relationship between detritus concentrations and the monsoon, or if this is a post-hoc analysis with all the risks that brings.

The paper compares the Indian summer monsoon record from Xihu with other reconstructions from the area affected by the Indian monsoon and claims that the monsoon failures are generally synchronous. Well except for the one that are not. One of the monsoon failures at Xihu is silently omitted, several of the others seem to be stretched by a hundred years or so. Methinks it is like a flock of  weasels.

Xu et al Figure 5. Comparison between Lake Xihu Indian Summer monsoon intensity indices and other precipitation indices: (a) tree ring Δ14C (red; 11-point moving average; Stuiver et al., 1998), (b) Xihu bulk density (blue), (c) Hongyuan peat Carex mulieensis δ13C (orange; Hong et al., 2003), (d) and (e) stalagmite δ18O data from Qunf Cave (green; 11-point moving average of the detrended data; Fleitmann et al., 2003) and Dongge  Cave (purple; Yuan et al., 2004), and (f) Arabian Sea G.Bull % (cyan; Gupta et al., 2005).

Xu et al Figure 5. Comparison between Lake Xihu Indian Summer monsoon intensity indices and other precipitation indices: (a) tree ring Δ14C (red; 11-point moving average; Stuiver et al., 1998), (b) Xihu bulk density (blue), (c) Hongyuan peat Carex mulieensis δ13C (orange; Hong et al., 2003), (d) and (e) stalagmite δ18O data from Qunf Cave (green; 11-point moving average of the detrended data; Fleitmann et al., 2003) and Dongge Cave (purple; Yuan et al., 2004), and (f) Arabian Sea G.Bull % (cyan; Gupta et al., 2005).

What about the chronology?

Xu et al Figure 3. Calibrated 14C ages for core Xihu-1. Blue triangles show the ages of samples from the peat section, while orange and green squares denote the ages of samples from the upper and lower lake sediment sections, respectively. Snail remains are seen in the upper and lower lake sediment sections.

Xu et al Figure 3. Calibrated 14C ages for core Xihu-1. Blue triangles show the ages of samples from the peat section; orange and green squares denote the ages of samples from the upper and lower lake sediment sections, respectively.

At first sight, the chronology looks weird. The dates in the peat section are all in stratigraphic order, but those in the lake sediments above and below are all out of sequence. The young dates in the basal limnic sediments can be explained – these are the sediments that formed beneath the floating bog. The authors explain the old date in the upper limnic sediments as being the result of reworking and redeposition of old organic matter. I suspect that the hard-water effect is more important, with old carbon from the dissolution of carbonates in the catchment biasing the dates.

Xu et al radiocarbon dated some modern materials from the lake to check for hard-water effects. Dates on a water snail, a submerged macrophyte and two water samples ranged from 855 to 1637 14C yr BP. This is probably an underestimate of hard-water effect as the water will include some bomb radiocarbon, which could make the samples appear much younger.

How about the peat, which is the focus of this paper? Here the cellulose fraction is dated and it is presumed by the authors to represent reeds. One of three radiocarbon dates on modern reeds has a 103.27 percent modern carbon, indicative of the incorporation of bomb radiocarbon and is fairly close to what is expected from the calibomb calibration curve. The other two samples have dates in excess of 100 14C yr BP. The authors blame this on the Suess effect – the old carbon from burning fossil fuels. I’m not convinced. Urban areas do have less radiocarbon, but the difference at Xihu between the “clean-atmosphere” measurements in calibomb and the reed dates is as great as one would expect in a large city, and not from a rural part of Yunnan. More likely, I think, is that the reeds also have a hard-water error. Emergent macrophytes are known to incorporate carbon from the water or from the sediment at least while they are shoots. The peat is unlikely to be 100% reed, if there is a contribution from submerged aquatic macrophytes, the hard-water  error could be substantial. Conversely, the authors do not state that roots were removed from the dated samples: roots can penetrate older sediment and bias the radiocarbon date of bulk sediment. AMS dates on identified emergent/terrestrial macrofossils would have been more robust.

I think the chronological uncertainty is likely to be higher than that reported by the authors. If the uncertainty is more than a few hundred years, the phase relationship between the peat proxies and the solar variability becomes unknowable.

 The paper finishes with a spectral analysis of the proxies in two sections of the core using an unspecified spectral method. Some peaks are found, but it is not specified how (or if) they were determined to be significant (I certainly would not get excited about a 360-year cycle in a 762-year record – there are only just over two cycles). The methods do matter – the results are almost uninterpretable without knowing what was done. Even worse, the plots of the spectral analysis only show selected frequencies rather than the whole spectrum. This makes it impossible to tell if the selected frequencies are large peaks or only slightly above the continuum. Please don’t make plots like this.

Xu et al Figure 6. Spectral analysis of the proxies a) for a fast accumulation interval (9226– 8464yr BP) and (b) for a slow accumulation interval (6680–4000yr BP).

Xu et al Figure 6a. Spectral analysis of the proxies for the  interval (9226–8464 yr BP)

I don’t think this paper is robust evidence of the impact of solar variability on palaeoecological proxies.


Xu et al. 2015. Abrupt Holocene Indian Summer Monsoon failures: A primary response to solar activity? The Holocene 25: 677–685.

Advertisements

About richard telford

Ecologist with interests in quantitative methods and palaeoenvironments
This entry was posted in Age-depth modelling, Peer reviewed literature, solar variability and tagged . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s