Net ecosystem carbon exchange of a dry temperate eucalypt forest

Hinko-Najera, Nina; Isaac, Peter; Beringer, Jason; van Gorsel, Eva; Ewenz, Cacilia; McHugh, Ian; Exbrayat, Jean-François; Livesley, Stephen J.; Arndt, Stefan K.

doi:https://doi.org/10.5194/bg-14-3781-2017

Articles | Volume 14, issue 16

https://doi.org/10.5194/bg-14-3781-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

Special issue:

OzFlux: a network for the study of ecosystem carbon and water...

https://doi.org/10.5194/bg-14-3781-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 14, issue 16

Research article

|

23 Aug 2017

Research article |

| 23 Aug 2017

Net ecosystem carbon exchange of a dry temperate eucalypt forest

Nina Hinko-Najera, Peter Isaac, Jason Beringer, Eva van Gorsel, Cacilia Ewenz, Ian McHugh, Jean-François Exbrayat, Stephen J. Livesley, and Stefan K. Arndt

Download

Final revised paper (published on 23 Aug 2017)
Supplement to the final revised paper
Preprint (discussion started on 13 May 2016)

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC1: 'Valuable Site but Incomplete Data Post-Processing and Flawed Methods and Interpretation', Anonymous Referee #1, 18 May 2016
RC2: 'Review of Hinko-Najera et al.', Anonymous Referee #2, 07 Jun 2016
AC1: 'Interactive comment on “Net ecosystem carbon exchange of a dry temperate eucalypt forest” by Hinko-Najera et al. (bg-2016-192)', Nina Hinko-Najera, 03 Sep 2016

Peer-review completion

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

ED: Reconsider after major revisions (08 Oct 2016) by Mirco Migliavacca

AR by Nina Hinko-Najera on behalf of the Authors (10 Feb 2017)

ED: Referee Nomination & Report Request started (14 Feb 2017) by Mirco Migliavacca

RR by Anonymous Referee #2 (23 Feb 2017)

RR by Anonymous Referee #1 (07 Mar 2017)

Suggestions for revision or reasons for rejection

The revised version has a core of methods, results, and interpretations that are acceptable for publication. The analysis of environmental drivers still needs to be improved. Below are suggestions for removal of a portion of the analysis, and replacement of this section by alternative methods that would provide more useful and powerful insights.

1) The authors have done a good job of revising their use of the eddy covariance technique, particularly the post-processing of data, to provide the best available dataset they can and in alignment with common practices. It is reassuring to see that results were so surprisingly robust to all of the changes that were made.

2) This lends needed confidence in all results up to the analyses of environmental drivers, and Figures 5, 6, and 7. Analyses of environmental drivers have also been significantly revised to remove circularity. However, the analyses still do not adequately address the second stated objective of the paper, to “identify the environmental controls of these CO2 ecosystem fluxes”. Unfortunately, this portion of the paper is still not well designed and provides disappointingly limited insight.

3) Analysis of Environmental Drivers: The random forest analysis attempts to estimate the relative importance of 4 environmental variables for determining midday-average NEE or early nighttime-average NEE for each month of the year and separately across the three years of study. Presumably there are thus at best only 30 or so observations in each bin on which the random forest is trained, which seems rather data scarce, and the reality is in fact far worse (<15) in many cases. Results in Fig 6 indicated that many of the months have fewer than 10 observations, and some have 2. How can a random forest possibly devise meaningful relationships regarding variable importance with so few observations? Apologies but this seems ludicrous. I recommend that analyses and results relating to Figures 5 and 6 be cut from the paper. (Note: For Figures 5 and 6, captions need to explain the numbers below the months on the x-axes, indicating the number of observations for each RF.) Figure 7 still has something to offer, but it could be replaced by something much better.

4) As suggested by Reviewer 2 in the prior review, the RF approach is rather indirect for gleaning insights into underlying processes. The paper does not provide a clear diagnosis of environmental controls on CO2 fluxes, but it could. Alternative, more fruitful approaches are available. As stated by R2.6, the random forest analysis could be replaced by analysis of functional relationships, conditionally sampling data to reveal light response parameters (e.g. NEE at light saturation and with low to modest VPD and for low versus high soil moisture), and similarly for VPD response, soil moisture response, and temperature response. The author’s comment in open review ignored this suggestion / critique by R2 altogether. Numerous studies have shown how this can be done with eddy covariance data but this study does not follow those leads for some reason.

5) Furthermore, the existing analysis of which variables had greatest importance for driving within-month or within-year variability leaves us wondering what the relationships look like. Is there a positive or negative relationship between sunlight and midday NEE? How strong is the relationship? Same for all of the other potential drivers.

6) The methods description does not indicate how data were binned for the random forest analysis of within-year variability.

Hide

ED: Reconsider after major revisions (08 Mar 2017) by Mirco Migliavacca

AR by Nina Hinko-Najera on behalf of the Authors (26 May 2017) Author's response Manuscript

ED: Referee Nomination & Report Request started (07 Jun 2017) by Mirco Migliavacca

RR by Anonymous Referee #1 (30 Jun 2017)

RR by Anonymous Referee #2 (05 Jul 2017)

ED: Publish as is (14 Jul 2017) by Mirco Migliavacca

AR by Nina Hinko-Najera on behalf of the Authors (21 Jul 2017)

Download

Article (1143 KB)
Full-text XML

Short summary

We undertook a 3-year study (2010–2012) of eddy covariance measurements in a dry temperate eucalypt (broadleaf evergreen) forest in southeastern Australia. The forest was a large and constant carbon sink, with the greatest uptake in early spring and summer. A strong seasonal pattern in environmental controls of daytime and night-time NEE was revealed. Our results show the potential of temperate eucalypt forests to sequester large amounts of carbon when not water limited.