Coherence of global hydroclimate classification systems

McCurley Pisarello, Kathryn L.; Jawitz, James W.

doi:https://doi.org/10.5194/hess-25-6173-2021

Articles | Volume 25, issue 12

https://doi.org/10.5194/hess-25-6173-2021

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

https://doi.org/10.5194/hess-25-6173-2021

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

Articles | Volume 25, issue 12

Research article

|

06 Dec 2021

Research article |

| 06 Dec 2021

Coherence of global hydroclimate classification systems

Kathryn L. McCurley Pisarello and James W. Jawitz

Download

Final revised paper (published on 06 Dec 2021)
Supplement to the final revised paper
Preprint (discussion started on 26 Oct 2020)
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: 'Comments', Anonymous Referee #1, 27 Oct 2020
- AC1: 'Reply on RC1', James Jawitz, 31 Dec 2020
RC2: 'Review of “Coherence of Global Hydroclimate Classification Systems” by Pisarello and Jawitz', Anonymous Referee #2, 29 Oct 2020
- AC2: 'Reply on RC2', James Jawitz, 31 Dec 2020
RC3: 'Review McCurley and Jawitz', Anonymous Referee #3, 08 Nov 2020
- AC3: 'Reply on RC3', James Jawitz, 31 Dec 2020

Peer-review completion

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

ED: Reconsider after major revisions (further review by editor and referees) (13 Jan 2021) by Roger Moussa

AR by James Jawitz on behalf of the Authors (13 Jan 2021) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (14 Jan 2021) by Roger Moussa

RR by Anonymous Referee #3 (27 Jan 2021)

RR by Anonymous Referee #2 (28 Jan 2021)

Suggestions for revision or reasons for rejection

I find the provided responses to review comments insufficient in several places. I had outlined six major comments in my previous review. Below are brief descriptions of which ones I think deserve further consideration.

---- The lack of independent evaluation data is a critical flaw of this study.

The discussion now includes several sentences that mention that the proposed new classification systems are evaluated on data that is either the same or dependent on the calibration data, which in my opinion is insufficient to address this concern. I consider the lack of independent evaluation data not a limitation but a methodological flaw that needs to be addressed before this paper can be published. Without independent evaluation data, no honest comparison between the established and new schemes is possible.

---- The complexity criteria are still insufficiently supported and clear:

1a. Hydroclimatic zones do not have equal sizes in reality. The justification that zones must have equal sizes for visibility of the small zones in my opinion simply means that zones get redefined to be less hydroclimatically distinct. In other words, regions that could be distinct hydroclimates get merged to conform to the minimum zone size criterion, meaning we lose hydroclimatic insight as a consequence of the zone size requirement. As far as I can tell, this argument is not addressed in either the response or the updated manuscript.

1b. In addition, pixels are an inappropriate unit of measurement for area, because the area a single pixel represents is not constant within the regular lat/lon projection used in this manuscript. This also means that any pixel-based criterion is conditional on the geographical projection used for the source data and thus not on the data itself. This is avoidable by using actual area values.

2. The authors rely on Meybeck et al. (2013) for the statement that zones should be delineated in one piece. To my knowledge the authors address neither in their replies, nor in the main manuscript my concern that this criterion rewards schemes that are not very good at what they are supposed to do, namely define regions with similar hydroclimates regardless of spatial proximity. Meybeck et al. (2013) have clear reasons for wanting to do so (mostly to not separate a river’s headwaters and its lowlands) and an experimental design that supports their reasoning (i.e. using basin shapes in addition to climatic data) which do not seem to apply in this paper. Given that the streamflow data used in this manuscript is simply local P – EP and does not consider catchment aggregation at all, I would say that Meybeck cannot be used to support the use of this criterion.

3. I now understand that the authors use the number of zones as part of a trade-off (given equal hydroclimatic coherence, the system with fewer zones is preferable) but I think this is insufficiently clear in the main manuscript. Most importantly, this tradeoff is neither mentioned in the description of the results (for example Table 1 simply lists ETA as having the best – lowest – number of zones despite it not having very high coherence), nor in the discussion.

---- The introduction and discussion are somewhat limited.

Given the focus on annual or longer time scales, the Budyko framework probably needs to be part of the introduction. Although not a climate classification system in the traditional sense, it is a well-established way of organizing locations based on long-term aridity (P / EP) and long-term ET and Q. A discussion of the Budyko framework can replace the current sentence on line 68 (“given the major gap regarding the inclusion of ET in climate classification systems, …” ) and may provide a justification for assessing the existing climate classification schemes at annual or longer scales. This could introduce a research question along the lines of “do existing hydroclimate classification schemes correspond with catchment organization as predicted by the Budyko curve?” or something similar.

I will also repeat one comment directly from the first review: What I miss in this discussion is a critical assessment of hydroclimatic understanding and how the authors’ proposed WEC adds to this. Existing classification schemes are based on hypotheses about how the world works and about which elements of the global climate are first-order controls on the resulting hydroclimate. WEC is simply a clustering method that finds regions with similar P and ET values. What does this teach us about global hydroclimatic relationships? If WEC is better than established methods that do rely on theory, then where is this theory faulty or incomplete? In other words, I think this discussion would be much stronger if the authors where to consider the question “what did we learn about the world?” I don’t see where this is addressed in section 5.

Hide

ED: Reconsider after major revisions (further review by editor and referees) (15 Mar 2021) by Roger Moussa

AR by James Jawitz on behalf of the Authors (02 Jul 2021) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (27 Sep 2021) by Roger Moussa

RR by Anonymous Referee #4 (18 Oct 2021)

Suggestions for revision or reasons for rejection

I was invited to review the article of McCurley Pisarello and Jawitz on the Coherence of Global Hydroclimate Classification Systems which is the fourth version after the second major revision. I found this paper interesting and relevant to the understanding of the hydroclimatic dynamics at the global level.

In this manuscript the authors present a comparison of within-zone hydrologic coherence and inter-zone shape complexity between four previously established climatic classifications Koppen-Geiger classification (KPG), Holdridge Life Zones scheme (HDL), Meybeck (MHR) and Knoben (KHC) and four newly suggested climatic classifications; three univariate classifications (ETA, ETV and ETC) directly based on actual evapotranspiration (ET) and a parsimonious multivariate classification, based on the mean annual precipitation (P) and potential evapotranspiration (PET) named Water-Energy Clustering (WEC).

The authors deduced from the results of the comparison that WEC has showed the highest internal coherence for different variables while also showing less complexity on boundaries delineation in comparison to KPG, results that were validated on two independent datasets (GRUN and GLEAM). They finally concluded that WEC can be used for assessing water availability and for direct management understanding. Indeed, such simple classifications are more frequently requested for understanding the hydroclimatic dynamics and constitute quick tools for assessment.

In addition to this fourth version, I also read the three previous versions of the article along with the referees’ comments and the authors’ responses. I noted that the previous reviewers have raised critical and essential comments to enhance the article, and the authors gave satisfactory answers in their last submittal. The authors have answered the main valid concerns regarding the validation of the results by using two independent datasets (GRUN and GLEAM) and the elaboration on the discussion of the results. Consequently, the quality of the article has notably improved since the first submittal therefore I found that the article merit to be published.

The authors are kindly requested to correct the following typing errors:
- Line 98: based “on” a series of …
- Line 270 Figure 4B: KPG instead of KGZ

Hide

RR by Anonymous Referee #1 (25 Oct 2021)

ED: Publish subject to technical corrections (25 Oct 2021) by Roger Moussa

AR by James Jawitz on behalf of the Authors (29 Oct 2021) Manuscript

Download

The requested paper has a corresponding corrigendum published. Please read the corrigendum first before downloading the article.

Article (1294 KB)
Full-text XML

Short summary

Climate classification systems divide the Earth into zones of similar climates. We compared the within-zone hydroclimate similarity and zone shape complexity of a suite of climate classification systems, including new ones formed in this study. The most frequently used system had high similarity but high complexity. We propose the Water-Energy Clustering framework, which also had high similarity but lower complexity. This new system is therefore proposed for future hydroclimate assessments.