Diagnosing the seasonal land–atmosphere correspondence over northern Australia: dependence on soil moisture state and correspondence strength definition

Decker, M.; Pitman, A.; Evans, J.

doi:https://doi.org/10.5194/hess-19-3433-2015

Articles | Volume 19, issue 8

https://doi.org/10.5194/hess-19-3433-2015

© Author(s) 2015. This work is distributed under
the Creative Commons Attribution 3.0 License.

https://doi.org/10.5194/hess-19-3433-2015

© Author(s) 2015. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 19, issue 8

Research article

|

06 Aug 2015

Research article |

| 06 Aug 2015

Diagnosing the seasonal land–atmosphere correspondence over northern Australia: dependence on soil moisture state and correspondence strength definition

M. Decker, A. Pitman, and J. Evans

Download

Final revised paper (published on 06 Aug 2015)
Preprint (discussion started on 19 Sep 2014)

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC C4384: 'Solicited review - major or minor revisions, depending on defense of the "coupling" theme', Paul Dirmeyer, 08 Oct 2014
- AC C6266: 'Author reply to Paul Dirmeyer', Mark Decker, 01 Feb 2015
  - AC C6292: 'Revised Manuscript', Mark Decker, 01 Feb 2015
RC C4501: 'Review of Decker et al', Anonymous Referee #2, 16 Oct 2014
- AC C6275: 'Author reply to reviewer #2', Mark Decker, 01 Feb 2015
  - AC C6293: 'Revised Manuscript', Mark Decker, 01 Feb 2015

Peer-review completion

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

ED: Reconsider after major revisions (24 Feb 2015) by Pierre Gentine

AR by Mark Decker on behalf of the Authors (26 Feb 2015) Author's response Manuscript

ED: Referee Nomination & Report Request started (26 Feb 2015) by Pierre Gentine

RR by Paul Dirmeyer (03 Mar 2015)

RR by Anonymous Referee #2 (18 Mar 2015)

Suggestions for revision or reasons for rejection

Review of “Diagnosing the seasonal land-atmosphere correspondence strength over Northern Australia: dependence on soil moisture state and correspondence strength definition” by Decker et al. submitted to HESS – revised version.

Summary:
The authors have revised their paper mostly by reframing their analysis and playing down the “coupling” aspect of their study in the presentation and discussion of their results (talking of “correspondence” instead – cf changed title). They also added some more observation data from flux towers (I believe as a response to other reviewers’ comments). I essentially still stand by my first review: mostly, the experimental set-up is not well suited for the intended analysis. Part of the analysis should be dropped, and part, while okay, could be attained more simply. I also question some of the claimed corrections. Please see below for details.

In my first review I essentially pointed out that because the authors are using an offline land-model, they could not really investigate land-atmosphere coupling in their study: the land surface states and fluxes simulated by their land model (CLM4) over Northern Australia did not feed back on the atmosphere, so they were not really investigating “coupling” when looking at correlations between simulated soil moisture, or evaporative fraction, and LCL derived from the forcing dataset(s).
More specifically, there were two aspects in the study:
i) Whether including root-zone soil moisture (SMrz), versus surface layer SM (SM1), in the statistical metric the authors use to evaluate soil moisture-atmosphere coupling, matters for the diagnosed coupling;
ii) Whether the mean background soil moisture content in the root-zone (varied between two simulations using different configurations of CLM4) matters for the diagnosed coupling.
I pointed out that, in particular, they could not investigate question ii), since the atmosphere does not “see” the changes in land model configuration.

The authors have tried to address this general problem (lack of L-A coupling in an offline land model simulation) by reframing their study as an analysis of the statistical “correspondence”, or “association”, between the land (SM) and the atmosphere (the LCL) – a correspondence that they indicate is necessary but not sufficient for actual land-atmosphere coupling to occur.
This may be fine for question i) – although there are still a few places in the manuscript where the authors speak of “coupling” that should be edited (e.g., p.20 l.561). Of course, this makes for a weaker main point of the study. Essentially, the story boils down to ET in SON being mostly transpiration and being driven by SMrz instead of SM1, in contrast to DJF where ET is mostly evaporation and therefore SM1 can be used to diagnose land-atmosphere “correspondence”. I have two comments here: first, I am not sure this point warrants an entire study on its own. I will leave that for the editor to decide. Specifically, note that the exact same point could have been addressed using the forcing datasets only, by looking at soil moisture and surface fluxes and their connection to the boundary layer in GLDAS or MERRA – except for sampling slightly different (corrected) versions of MERRA, I don’t really see the need to use an offline land simulation here. Second, even if it did (warrant a study), this point is still poorly demonstrated in the study, in my view. The authors could easily show the seasonal cycle of the different components of ET (evaporation, transpiration, interception), for instance – I recommended they did in my first review, but it is still not the case. I note that p.14 lines 373-384, they indicate that transpiration is only 10-30% of ET in DJF, but they do not indicate how much transpiration contributes to ET in SON. They should also show explicitly that transpiration is governed by SMrz, not SM1 (e.g., with time series and correlations), and that the variability of SM1 and SMrz is decoupled during SON.

As for question ii), I recommended the authors drop this part of the analysis and focus on i) instead. They have chosen to retain this part of the analysis. I don’t think this is very fruitful. In the absence of land-atmosphere feedback, the fact that SMrz-from-CLM4/LCL-from-the-forcing correlations are similar in the CTRL and DRY CLM4 runs (with and without soil-groundwater interactions) only goes to show that the variability in SMrz and L-A fluxes is similar in both runs –despite the different mean SMrz and fluxes. This conclusion could (and actually, should) be reached separately, with time series and correlations between both runs, without computing correlations with a third variable (i.e., the LCL) from the driving dataset. I don’t think it is fair to say, as in the conclusion p.21 lines 591-592, that ”the diagnosed L-A coupling is insensitive to the mean vertical profile of soil moisture”. If the land were truly coupled to the atmosphere, the difference in mean ET in SON between CTRL and SON, for instance (up to 30 w/m2, or 100% of the CTRL value…) could very well impact land-atmosphere interactions. The similarity in SM and fluxes variability between CTRL and DRY may be interesting for its own sake with respect to the land model behavior; but it does not really inform on “the dependence of L-A correspondence on soil moisture state”, as the paper’s title implies. I note that this aspect of the analysis is actually not even addressed in the abstract, which I take it reflects the fact that no real conclusion can be derived from this part of the study.

Finally, I asked in my previous review that the authors address the issue of seasonality potentially driving the SM/LCL correlations diagnosed here (a potential reason why CTRL and DRY might have shown similar signals, too). The authors claim to have detrended the data “separately over each season” (p.9, line 248) prior to computing the correlations. First, I am not sure I understand what this means – a linear detrending each year over three months? Why not classically and simply remove a mean seasonal cycle from the whole data? If this is not the case, I strongly recommend the authors do the latter. (Here I would also encourage the authors to describe more explicitly how they compute the correlations: with daily data, monthly data? I though I understood daily, but p.9 lines 249-242 got me confused, this time).
But more importantly, I am strongly skeptical of whether the data was appropriately “de-seasonalized”, given that figures 6-8 in the revised version are exactly, pixel for pixel as far as I can tell, the same as figures 5-7 in the first version (except for the different color keys). I find it unlikely that removing the seasonality in SM, EF and LCL, and, e.g., computing correlations on anomalies, would result in exactly the same results.

Hide

ED: Reconsider after major revisions (07 Apr 2015) by Pierre Gentine

AR by Mark Decker on behalf of the Authors (28 Apr 2015) Author's response Manuscript

ED: Reconsider after major revisions (19 May 2015) by Pierre Gentine

Dear authors,
One of the reviewers still seemed dissatisfied by your responses. In particular you overlooked some of his/her comments. In particular I echo his/her comment regarding the discussion on land-atmosphere coupling.
In the manuscript there are many instances which I believe are incorrect. You mention that you are evaluating the full cycle Delta SM->Delta EF-> Delta PBL but this statement is incorrect. You do not have Delta PBL. In reality you are testing the correspondence of EF to the observed LCL (and thus PBL). I believe the strategy is fine but I would not call a coupling because you have offline observations. The term correspondence is also rather fuzzy. IN fact in a paper I was trying to do something similar: inferring EF from LCL
Gentine, P., C. R. Ferguson, and A. A. M. Holtslag (2013), Diagnosing evaporative fraction over land from boundary-layer clouds, J Geophys Res-Atmos, 118(15), 8185–8196, doi:10.1002/jgrd.50416.

I would suggest re-framing the parts mentioning the coupling to PBL and LCL. You should rather mention an inference of an accurate EF and SM for CLM. You should also modify the title accordingly: What about: Inferring evaporative fraction and the role of root zone moisture though maximization of land-atmosphere correlation?

Specific comments:
In the abstract I would remove correspondence it is rather fuzzy.
l122 is incorrect you are not evaluating Delta PBL, I would rather directly reframe this in terms of trying to maximize the EF and observed PBL/LCL information (see Gentine et al. 2012).
l169 is incorrect: Delta SM to Delta EF instead of Delta PBL
l295 which Tdew is used is unclear
l452-455: the SM-LCL association should eb the way to go but it has to be clarified. You are really trying to maximize the information content between your simulated EF and the observed LCL.
l463: DJF coupling: you are not really looking at coupling here same line 539-540
IN the conclusions the land-atmosphere association is unclear. I would really suggest re-framing those discussions as a maximization of the EF correlation and a way to infer the soil moisture dynamics. This will make your paper stronger and clearer.

Hide

AR by Mark Decker on behalf of the Authors (30 Jun 2015) Author's response

ED: Publish as is (04 Jul 2015) by Pierre Gentine