Sensitivity to species selection indicates the effect of nuisance variables on marine microfossil transfer functions

Jonkers, Lukas; Kučera, Michal

doi:https://doi.org/10.5194/cp-15-881-2019

Articles | Volume 15, issue 3

https://doi.org/10.5194/cp-15-881-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

Special issue:

Paleoclimate data synthesis and analysis of associated uncertainty...

https://doi.org/10.5194/cp-15-881-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 15, issue 3

Research article

|

15 May 2019

Research article |

| 15 May 2019

Sensitivity to species selection indicates the effect of nuisance variables on marine microfossil transfer functions

Lukas Jonkers and Michal Kučera

Download

Final revised paper (published on 15 May 2019)
Supplement to the final revised paper
Preprint (discussion started on 06 Sep 2018)
Supplement to the preprint

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Review-CP-Jonkers-Kucera-2018', Anonymous Referee #1, 16 Oct 2018
- AC1: 'reply to reviewer 1', Lukas Jonkers, 15 Jan 2019
RC2: 'Review of Jonkers and Kucera', Anonymous Referee #2, 26 Nov 2018
- AC2: 'replay to reviewer 2', Lukas Jonkers, 15 Jan 2019

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

ED: Reconsider after major revisions (14 Feb 2019) by Christian Ohlwein

AR by Lukas Jonkers on behalf of the Authors (14 Feb 2019) Author's response Manuscript

ED: Referee Nomination & Report Request started (25 Feb 2019) by Christian Ohlwein

RR by Anonymous Referee #2 (27 Mar 2019)

Suggestions for revision or reasons for rejection

This authors have dealt with the problems in the first version of the manuscript and I now have only a few comments.

Line 45 "taxonomic resolution" I am not sure this is the best term to use. I see taxonomic resolution as recording the counts at the genera/species/subspecies level, whereas this ms discusses whether certain taxa can be omitted. Even higher taxonomic resolution, to cryptic species level, might solve much of the problem with e.g. G. glutinata. I don't have a better term, except perhaps "species coverage".

Line 139 "ggplot2" not "ggplot"

Line 156. The comparison of each species distribution to a Gaussian curve implicitly assumes that species are expected to have a Gaussian niche. While some theory suggests that they should, much field evidence shows that they often do not. This is important as WA assumes that the species niches are ~Gaussian. Species with strongly non-Gaussian niches may therefore appear to be unimportant or nuisance taxa. I think this is what is happening with G. glutinata. It has a strongly non-Gaussian niche with respect to temperature (which may of course be modulated by adjustment to the bloom season with latitude), being absent from cold water, most abundant in temperate water, and present at moderate abundances even in tropical waters. I strongly suspect that this non-Gaussian niche, rather than the importance of secondary variables, is what is making G. glutinata appear to be a nuisance taxon. The interpretation of the importance statistics is a little more subtle than previously realised. G. glutinata is probably the most extreme taxa in this regard.

Line 160. The estimation of the temperature sensitivity may be sub-optimal in that it does not account for the binomial nature of the data. A better (and easier) way to fit the estimated Gaussian curves would be with a binomial GLM using a quadratic model. The model pseudo-r2 can then be used. Perhaps more importantly, the reported correlation between abundance and temperature sensitivity may simply be an artefact of rare taxa being more noisy rather than any indication of their sensitivity to non-temperature gradients. It might be possible to test this with some appropriate null models.

Line 170 "for MAT inclusion of warm-water
species only appears sufficient to obtain a minimum prediction error, ..." Is this a relic from the first version? Both MAT and WA now report that the cold-water species N. pachyderma is important.

Line 190 "celsius" -> "Celsius"

Line 265 "proof" is far to strong a term here. "evidence" perhaps. I am fairly certain that many of the same patterns of importance and changing reconstructions would be observed in artificial data where all taxa are sensitive to temperature only. As it stands, I am not totally convinced that the variability in reconstructions demonstrates the importance of secondary gradients.

Fig 4b improve or delete axis labels

Table 1. "Correlation coefficients of linear regression" - is this the model R2 - the coefficient of determination?

line 275 This enjoinder to count all taxa (which I concur with) contradicts the suggestion in the introduction that a subset of taxa could be counted for speed.

Hide

ED: Publish subject to minor revisions (review by editor) (03 Apr 2019) by Christian Ohlwein

AR by Lukas Jonkers on behalf of the Authors (12 Apr 2019) Author's response Manuscript

ED: Publish as is (29 Apr 2019) by Christian Ohlwein

Short summary

Fossil plankton assemblages have been widely used to reconstruct SST. In such approaches, full taxonomic resolution is often used. We assess whether this is required for reliable reconstructions as some species may not respond to SST. We find that only a few species are needed for low reconstruction errors but that species selection has a pronounced effect on reconstructions. We suggest that the sensitivity of a reconstruction to species pruning can be used as a measure of its robustness.