A multi-year short-range hindcast experiment with CESM1 for evaluating climate model moist processes from diurnal to interannual timescales

Ma, Hsi-Yen; Zhou, Chen; Zhang, Yunyan; Klein, Stephen A.; Zelinka, Mark D.; Zheng, Xue; Xie, Shaocheng; Chen, Wei-Ting; Wu, Chien-Ming

doi:https://doi.org/10.5194/gmd-14-73-2021

Articles | Volume 14, issue 1

https://doi.org/10.5194/gmd-14-73-2021

© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/gmd-14-73-2021

© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 14, issue 1

Model experiment description paper

|

07 Jan 2021

Model experiment description paper |

| 07 Jan 2021

A multi-year short-range hindcast experiment with CESM1 for evaluating climate model moist processes from diurnal to interannual timescales

Hsi-Yen Ma, Chen Zhou, Yunyan Zhang, Stephen A. Klein, Mark D. Zelinka, Xue Zheng, Shaocheng Xie, Wei-Ting Chen, and Chien-Ming Wu

Download

Final revised paper (published on 07 Jan 2021)
Preprint (discussion started on 15 Apr 2020)

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC1: 'An approach using multi-year short-range hindcasts to identify model parameterization deficiencies for climate simulations', Anonymous Referee #1, 04 May 2020
- AC1: 'Response to Reviewer 1's comments', Hsi-Yen Ma, 09 Jul 2020
RC2: 'anonymous review', Anonymous Referee #2, 05 May 2020
- AC2: 'Response to Reviewer 2's comments', Hsi-Yen Ma, 09 Jul 2020
RC3: 'Review of GMD-2020-39', Anonymous Referee #3, 09 Jun 2020
- AC3: 'Response to Reviewer 3's comments', Hsi-Yen Ma, 09 Jul 2020

Peer-review completion

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

AR by Hsi-Yen Ma on behalf of the Authors (09 Jul 2020) Manuscript

ED: Referee Nomination & Report Request started (22 Jul 2020) by Christina McCluskey

RR by Anonymous Referee #1 (06 Aug 2020)

RR by Anonymous Referee #2 (11 Aug 2020)

Suggestions for revision or reasons for rejection

I have to say that I am torn about how to approach reviewing the authors'
revised manuscript and associated comments.

On one hand, I do think there is value in publishing a paper that describes
the numerous details involved in carrying out such a hindcast experiment; and
it is a worthwhile contribution that the authors have released the source code
for this purpose. I would ultimately like to see this paper published.

On the other hand, I cannot immediately see a path toward revising this paper
such that it is suitable for publication: it appears to be fundamentally too
duplicative of work already in the literature. Both Reviewer #3 and I raised a
major concern about the uniquness of this manscript. The authors argue that
this experiment is unqiue relative to the previous literature--and in
particular relative to Ma et al. (2015)--because it describes a multi-year
hindcast. That is a fair argument. However, Reviewer #3 and I both also
raised the concern that the authors did not provide a variety specific details
that would be necessary for another modeling group to replicate their
experimental design. In this case, the authors' response was to argue that "We
did not want to duplicate the information because the initialization technique
is not the main focus of the manuscript and modeling groups who have the
capability to conduct hindcast studies already have their own initialization
strategy." (The authors did provide some--though not all--of these details in
the revised manuscript.)

In short, it sounds like the experimental design described in this manuscript
is simply the Ma et al. (2015) method, except that it is run for multiple
years. Is the addition of running for multiple years sufficently unique to
warrant publication of a new experimental design? Perhaps, though that
argument is weakend by the existence of other multi-year hindcast methods
described in the literature.

I would also like to note that the authors appear to have misunderstood my
intention in the section of my previous review where I stated "Specifically, a
number of questions come to mind that could impact the results from other
modeling groups implementing this experiment." I raised these questions with
the intent of suggesting that the manuscript should be revised to address the
questions. There were a number of instances where the authors instead directly
responded to me without actually modifying the manuscript. These were
collectively part of my major concern that the manuscript did not provide
sufficient technical detail for a "model experiment description paper," and I
would suggest that the authors revise the manuscript accordingly.

In summary, I cannot recommend this manuscript for publication at this time. I
am going to recommend that it again be returned for major revisions, in hopes
that the authors can provide a revised manuscript that makes a much more compelling and
impactful model experiment description.

Hide

RR by Anonymous Referee #3 (27 Aug 2020)

Suggestions for revision or reasons for rejection

The authors have provided clear responses to most of my comments and made appropriate revisions to the manuscript.

However, I am still concerned about the interpretation of the MJO development and propagation in these hindcasts. In their response, the authors discussed how they formed the 16-year timeseries from hindcasts at a constant lead time (e.g., a 16-year timeseries of day-3 hindcasts). The construction of these timeseries was clear to me in the original manuscript. My concern is how the authors interpret errors in physical processes simulated in the model, based on these timeseries. The authors discuss the pre-conditioning of the active MJO phase by shallow convection, the transition from shallow to deep convection and the propagation of the MJO across the Maritime Continent. These are processes that require more than three days to develop, either in a model or in observations. Thus, I do not believe that the authors can conclude (from these timeseries) that the model fails to simulate these processes correctly, given that the timeseries is discontinuous (i.e., it is made up of one day from each of many hindcasts, stitched together). Any propagation in this timeseries, either for individual events or for the composite, is an artefact of concatenating parts of many different simulations together. The hindcasts do not have sufficient time to simulate the propagation across the Maritime Continent. These processes could be simulated in the nudged simulation, because that is a continuous simulation, which is why I suggested comparing the hindcasts to the nudged run.

The authors need to discuss these caveats so that the reader does not treat the hindcast simulation as a continuous simulation, in which physical processes like pre-conditioning and moistening are simulated within a single integration of the model.

Hide

ED: Reconsider after major revisions (29 Aug 2020) by Christina McCluskey

AR by Hsi-Yen Ma on behalf of the Authors (10 Oct 2020) Author's response Manuscript

ED: Referee Nomination & Report Request started (16 Oct 2020) by Christina McCluskey

RR by Anonymous Referee #1 (30 Oct 2020)

RR by Anonymous Referee #2 (31 Oct 2020)

ED: Publish as is (06 Nov 2020) by Christina McCluskey

AR by Hsi-Yen Ma on behalf of the Authors (09 Nov 2020)

Short summary

We propose an experimental design of a suite of multi-year, short-term hindcasts and compare them with corresponding observations or measurements for periods based on different weather and climate phenomena. This atypical way of evaluating model performance is particularly useful and beneficial, as these hindcasts can give scientists a robust picture of modeled precipitation, and cloud and radiation processes from their diurnal variation to year-to-year variability.