Summarizing the state of the terrestrial biosphere in few dimensions

Kraemer, Guido; Camps-Valls, Gustau; Reichstein, Markus; Mahecha, Miguel D.

doi:https://doi.org/10.5194/bg-17-2397-2020

Articles | Volume 17, issue 9

https://doi.org/10.5194/bg-17-2397-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/bg-17-2397-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 17, issue 9

Research article

| Highlight paper

|

05 May 2020

Research article | Highlight paper |

| 05 May 2020

Summarizing the state of the terrestrial biosphere in few dimensions

Guido Kraemer, Gustau Camps-Valls, Markus Reichstein, and Miguel D. Mahecha

Download

Final revised paper (published on 05 May 2020)
Preprint (discussion started on 21 Aug 2019)

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC1: 'Review of bg-2019-307', Anonymous Referee #1, 16 Sep 2019
- AC1: 'Response to Reviewers', Guido Kraemer, 11 Nov 2019
RC2: 'Review of bg-2019-307', Anonymous Referee #2, 19 Sep 2019
- AC1: 'Response to Reviewers', Guido Kraemer, 11 Nov 2019
RC3: 'Review for Manuscript bg-2019-307', Anonymous Referee #3, 20 Sep 2019
- AC1: 'Response to Reviewers', Guido Kraemer, 11 Nov 2019
RC4: 'Review for "Summarizing the state of the terrestrial biosphere in few dimensions"', Anonymous Referee #4, 20 Sep 2019
- AC1: 'Response to Reviewers', Guido Kraemer, 11 Nov 2019

Peer-review completion

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

ED: Publish subject to minor revisions (review by editor) (25 Nov 2019) by Kirsten Thonicke

ED: Reconsider after major revisions (26 Nov 2019) by Kirsten Thonicke

AR by Guido Kraemer on behalf of the Authors (20 Jan 2020) Author's response Manuscript

ED: Referee Nomination & Report Request started (28 Jan 2020) by Kirsten Thonicke

RR by Anonymous Referee #4 (10 Feb 2020)

Suggestions for revision or reasons for rejection

The authors addressed all my concerns and improved the manuscript substantially. Therefore, the manuscript could be published after minor revisions (mainly spell checking, and a thorough proof reading is necessary).

L9-10: I still think that those variables do not fully describe the state of the biosphere, so it seems more accurate to write: account for 82% of the variance of the selected biosphere variables in space and time
L13: Rephrase ‘they also allow us’. What does they refer to (the anomalies, or the indicators/pcas)?
L19: add: Díaz, S., Settele, J., Brondizio, E. S., Ngo, H. T., Agard, J., Arneth, A., … Zayas, C. N. (2019). Pervasive human-driven decline of life on Earth points to the need for transformative change. Science, 1327(December). https://doi.org/10.1126/science.aaw3100
L41: change
L41: what do you mean by qualitative state?
L49: decide whether to use ‘high-dimensional’ or ‘high dimensional’
L51: known
L78: rephrase, maybe just simplify to ‘and analyses such as mean seasonal cycles …’
L99: for this study,
L102: available variables
L106: of analysis
L109: rephrase: Because the area of the pixel changes with latitude in the equirectangular coordinate system used by the ESDL
L142: remove ‘n’
L145: hols?
L149: using
L197: something is missing in this sentence
L202: Rephrase and clarify the first sentence. What type of dynamics are you referring to?
L216: Rephrase to be more readable (and add the section numbers accordingly): In the following, we first briefly present and discuss the quality of the global dimensionality reduction and interpret the individual components from an ecological point of view. We summarize the global dynamics that we uncovered in the low-dimensional space. We characterize the contained seasonal dynamics, including spatial patterns of hysteresis. … finally, we present global trends and their breakpoints.
Figure 2b) I agree that it makes sense to visualize the variable loadings to the PCs in this way, but it is still pretty hard to read the contributions for the first three components, because the values all lie close together at maybe around -0.5 to 0.5. Is there a way to rescale the colorbar/colormap (are the extremes of -1 and 1 even reached by the data range)? Or choose a colormap, which is more distinguishable at the lower values.
L241-246: Check spelling of that whole paragraph!
Figure 3: This figure could be improved: add titles for top, middle and bottom rows, don’t cut the top MSE distribution at 1.2, or maybe use log-scale if needed, and I think it would be ok to use different value ranges for each subfigure (b-d-f) to illustrate relative error distributions, since the color already indicates the absolute differences.
L263: Be more precise about the definition of albedo! See the following paper describing the differences between black-sky and white-sky albedo: Schaepman-Strub, G., Schaepman, M. E., Painter, T. H., Dangel, S., & Martonchik, J. V. (2006). Reflectance quantities in optical remote sensing-definitions and case studies. Remote Sensing of Environment, 103(1), 27–42. https://doi.org/10.1016/j.rse.2006.03.002
L265: can be caused by many reasons
L279/280: Check spelling!
L325: check spelling
L338: third component
L340-341: It is indeed hard to say what exactly PC3 represents, especially since the northern boreal forests should show a very low albedo, especially during summer. There are a few other aspects though, potentially related to bare soil albedo (using cover-crops vs bare soil in agricultural practices) and changing bare soil albedo due to soil wetness. So if there is a large soil fraction, PC2 and 3 could be related due to lower surface albedo of wet soils.
Figure 6: nice figure, but I would also optimize the color range of column two and three. There is only light colors, which indicates that a smaller range could show the patterns more clearly (e.g. just disregard extreme values when defining the color-range).
L355: It seems to track surface reflectance changes, which are not caused by vegetation phenology (since that is included in PC1). So maybe soil albedo plays a role too.
L395: instead of
L402 onwards: check spelling
L488 onwards: split into several sentances
L532: formed by
L553: three indicators
L553/554: the second … while the third
L566: Future research should consider non-linearities, adding data streams describing other important biosphere variables (e.g. related to biodiversity and habitat quality), and including different subsystems…

Hide

RR by Anonymous Referee #3 (14 Feb 2020)

Suggestions for revision or reasons for rejection

Review of manuscript bg-2019-307-v2

The manuscript has been considerably improved after the first round of review, including thanks to the considerable restructuring of the text. There are still the following issues that should be addressed before publication.

The paper is a lacking a discussion of some of the caveats of the input variables used in the analysis. This does not undermine the methodological contribution of the work, but rather affects the interpretation the authors make of it. This should probably include some statements on how the data from Tramontana et al for instance depend strongly on the spatial distribution of fluxtowers, which typically represent better temperate regions (due to their higher number there) than tropical ones for instance. Also, I believe that data is known to not be able to represent the CO2 fertilization effect. Could you comment on that? Finally, I would also add some words on the possibilities of now using sun-induced chlorophyll fluorescence (SIF) to better estimate GPP and photosynthesis, for instance. This could all be part of the conclusion, but should be there in my opinion.

In the conclusion suggest how these 3 principal components you detect could perhaps be used for operational monitoring of the state of the terrestrial biosphere. Could you link this to the Copernicus Climate Change Services (C3S) for instance?

I have the impression I did not find the temporal range of the data cube explicitly in the text (e.g. in the data section description). Furthermore, if I understand correctly from figures later on, it ends around 2012? How do you justify that in 2020 we would not have a time record that lags by 8 years? I know some datasets are not all open access and free in the same way, but now a days, most of the variables are effectively available beyond 2012 (even if from other sources than those used by the authors). Please add somewhere more justification explaining this, perhaps precising that you are doing a demonstration and do not need the latest years. But in the conclusion, stress that this work should be extended up to present times.

It should be specified up-front (from the abstract and throughout) that albedo refers to surface albedo, and not planetary albedo (which is much different due to cloud cover). Also, please specify somewhere that this is broadband albedo (i.e. across the shortwave spectrum), and that this combines both the albedo from the visible range and then near-infrared, two components that react quite differently with respect to vegetation (darkening the first and lightening the second).

L40: explain here what the Bowen ratio is to be more complete in your illustration

L61: Land does not need capital letter… also, why not add a reference to this phrase? And event to the next?

L100: to be clear, please confirm in the text: if there is one missing data in one variable, all are considered missing, correct?

L158: correct “deviatiosn”

L176: correct “booth”

L181: you have not yet said you consider only 3 components, so here there is an issue with specifying up to PC3 only… make it generic

L207: correct “whiche”

Why do Hovmoller plots in B1 have so many gaps? If this is due to occasional data gaps can’t more tolerance be added?

Hide

ED: Publish subject to minor revisions (review by editor) (17 Feb 2020) by Kirsten Thonicke

AR by Guido Kraemer on behalf of the Authors (21 Feb 2020) Author's response Manuscript

ED: Publish as is (25 Mar 2020) by Kirsten Thonicke

AR by Guido Kraemer on behalf of the Authors (31 Mar 2020) Author's response Manuscript

Short summary

To closely monitor the state of our planet, we require systems that can monitor the observation of many different properties at the same time. We create indicators that resemble the behavior of many different simultaneous observations. We apply the method to create indicators representing the Earth's biosphere. The indicators show a productivity gradient and a water gradient. The resulting indicators can detect a large number of changes and extremes in the Earth system.