Disentangling sources of future uncertainties for water management in sub-Saharan river basins

Amaranto, Alessandro; Juizo, Dinis; Castelletti, Andrea

doi:https://doi.org/10.5194/hess-26-245-2022

Articles | Volume 26, issue 2

https://doi.org/10.5194/hess-26-245-2022

© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/hess-26-245-2022

© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 26, issue 2

Research article

|

18 Jan 2022

Research article |

| 18 Jan 2022

Disentangling sources of future uncertainties for water management in sub-Saharan river basins

Alessandro Amaranto, Dinis Juizo, and Andrea Castelletti

Download

Final revised paper (published on 18 Jan 2022)
Preprint (discussion started on 26 Jan 2021)

Interactive discussion

Status: closed

RC1:
'Comment on hess-2021-40 - Disentangling Sources of Future Uncertainties forWater Management in Sub-Saharan River Basins by Amaranto et al.', Alessio Ciullo, 22 Mar 2021

I enjoyed reading the paper, it is well written and it addresses both a methodological (i.e., integration of UA and SA into decision-robustness frameworks) and practical (i.e., the management of Umbeluzi river basin) issues. I support the effort in integrating UA and SA into established decision-support frameworks, I do however have the following issues:

- one of these frameworks proposed in the literature (perhaps the first one), i.e., Robust-Decison Making (RDM), in fact finds robust solutions heavily relying on a technique therein called Scenario Discovery, which to me can rightfully be assumed as a SA technique (i.e., factor mapping as also pointed out by Pianosi et al. (2016) http://dx.doi.org/10.1016/j.envsoft.2016.02.008). RDM has also been extensively adopted to address water planning problems https://link.springer.com/chapter/10.1007/978-3-030-05252-2_7. My suggestion would be to integrate such literature in the intro where it is stated that only few studies dealt with this problem before.

- perhaps related with the previous point: I think the UA and SA steps should more strongly be integrated into the robustness search. Afterall, robustness broadly defines the performance of the system under uncertainty and the UA and SA steps should be part of it. In the proposed framework instead (Fig. 3), the search for robustness and the UA and SA steps are reported as two different steps. How can the intial robust policies be updated/ameliorated based on SA analysis results? I think one could change the current framework either bringing the UA/SA step into robustness, or establishing some sort of feedback loop between the two. This would also affect the results, where it would be great to comment upon how, in light of the GLUE and PAWN results, one could increase the robustness of the previosuly found strategies. This could perhaps come in the form of a table where e.g. a description of how the best policies for each stakeholder change going through the three steps of the frameworks, i.e., from optimal to robust to robust+SA.

minor points:

- the model description may not be immediate to a reader not used to such models. It would be good to make clear what are the policies, the uncertainties etc. etc. - a good possible framework to follow is the XLRM (actually linked to RDM)

- better description of the GLUE method

- when introducing RID and RUD (at about line 380) it isn't immediately clear that the considered least robust alternative is not the asbolute least robust, but rather the least robust among the Pareto set (i.e., performing best for at least one stakeholder). Also, it would be ideal trying to condense the info from figures 7 to 9 into one figure - or at least plotting them together as they are "the same".

- better captions

- some remaining typos/language issues

Citation: https://doi.org/10.5194/hess-2021-40-RC1
- AC1: 'Reply on RC1', Andrea Castelletti, 05 May 2021
  
  We would like to thank the reviewer for the precious contributions. Our response is available in the attached PDF, where RC stands for 'Reviewer Comment' and AR stands for 'Author Response'
  
  Citation: https://doi.org/10.5194/hess-2021-40-AC1
RC2:
'Comment on hess-2021-40', Anonymous Referee #2, 19 Apr 2021

Please see attached PDF

Citation: https://doi.org/10.5194/hess-2021-40-RC2
- AC2: 'Reply on RC2', Andrea Castelletti, 05 May 2021
  
  We would like to thank the reviewer for the precious contributions. Our response is available in the attached PDF, where RC stands for 'Reviewer Comment' and AR stands for 'Author Response'
  
  Citation: https://doi.org/10.5194/hess-2021-40-AC2
EC1: 'Editor's comment', Dimitri Solomatine, 21 May 2021

Reviewers' comments were quite useful, and the authors provided extensive answers and clarifications (and also rebuttal to some points), making it clear how the manuscript will be revised. The paper can move to the next stage, so I invite the authors to work on the revision.

Citation: https://doi.org/10.5194/hess-2021-40-EC1

Peer review completion

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

ED: Publish subject to revisions (further review by editor and referees) (21 May 2021) by Dimitri Solomatine

AR by Andrea Castelletti on behalf of the Authors (24 Jun 2021) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (18 Jul 2021) by Dimitri Solomatine

RR by Alessio Ciullo (04 Aug 2021)

RR by Anonymous Referee #2 (25 Sep 2021)

Suggestions for revision or reasons for rejection

Review of first revision “Disentangling Sources of Future Uncertainties for Water Management in Sub-Saharan River Basins”, Amaranto et al.

I have reviewed the revised manuscript and would like to thank the author for their careful consideration of the comments by myself and the other reviewer. The manuscript has most certainly improved, and additional details provided contribute to the understanding.
Despite the improvements the author have made, I am still struggling somewhat with the scope of the manuscript, given that it still hinges somewhat on two thoughts; the first being the optimisation of operating policies, robustness assessment and subsequent uncertainty analysis and sensitivity analysis of robust and non-robust policies; while the second is application to the Barragem Pequeños Limbobos (BPL). I think the conclusion section exemplifies this confusion, and that then is my main general comment on this revision.
The bulleted conclusions appear to be primarily related to the findings of the application of the method proposed to BPL, and are therefore contingent on the particular structure of the studied water system, as well as already noted on the relative size of demands (e.g. the size of the demand from d/s irrigation when compared to the urban demand). This includes conclusions on the trade- offs between identified demands. The second part of the conclusion, which is in fact more of a discussion despite the name of the section, focuses more on the methods applied, but conclusions on the applicability or the generality of the method itself are lacking. This still brings to question how general the conclusions that are made are. In my opinion, perhaps one of the more interesting conclusions is that where the authors conclude that robust policies are vulnerable only to hydrological perturbations, whilst non-robust policies are also vulnerable to changes to urban and agricultural demand. That is an interesting conclusion if it is indeed generic and therefore a clear contribution to the establishing of reservoir operating polices through the methods proposed. I would think that some discussion on how generic that conclusion is, or if it is indeed dependant on the configuration of the BPL water system. I would still call for the authors to add a short reflection on this. It may also be appropriate to label the section as conclusion and discussion. A separate discussion section prior to the conclusions would make the structure clearer, but I will leave this for the editor to decide on.

Detailed comments
Some detailed comments are provided below. The grammar of the revised version has indeed improved, though there are still some improvements that will need to be done. A good revision prior to submission of the revised manuscript would be recommended.
Line 82: Change reminder to remaining.
Line 86: I would change this to total discharge volume as that would be commensurate to the unit, which is a unit of volume.
Line 113: the closing bracket of the interval should be a square bracket.
Line 123: This constraint would appear (given its formulation) to be related to the capacity of the outlet through which releases are made. However, what confuses me is that the upstream release is direct from the dam itself (elsewhere it is suggested that it is pumped). So, does the constraint apply to the release to the upstream irrigation also? I am fine with it being simplified to consider releases as if these are effectuated through the same outlet, but at least make a comment to this as the optimisation space may be quite different if taking e.g. the max pump capacity upstream as the constraint.
Line 124: It is in turn and not in turns. This should also be corrected in one or two other places in the manuscript.
Line 141: I am not sure irrigation prone solution is a correct formulation. Perhaps this could be formulated as :

solutions that favour irrigation can be discovered

Line 174: yielding and not yielding to.
Line 304: The historical averages in the table for irrigation and urban demand are not, I presume, averages as determined over a time series (as was done I presume for the inflows). So, in a sense these are not averages as determined over a time series of demands I presume. Perhaps current or initial values would be a more appropriate label. Figure 2 suggest the demands were calculated based on extent the different crops and their evaporative demand as a function of climate data.
Line 323: This sentence does not make sense to me. Perhaps reformulate as below but do check the context:

In particular, GLUE allows for determining which SOW result in optimal robust policies that yield unacceptable results

Hide

ED: Publish subject to revisions (further review by editor and referees) (04 Oct 2021) by Dimitri Solomatine

AR by Andrea Castelletti on behalf of the Authors (27 Oct 2021) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (28 Oct 2021) by Dimitri Solomatine

RR by Anonymous Referee #2 (25 Nov 2021)

ED: Publish as is (10 Dec 2021) by Dimitri Solomatine

Short summary

This study aims at designing water supply strategies that are robust against climate, social, and land use changes in a sub-Saharan river basin. We found that robustness analysis supports the discovery of policies enhancing the resilience of water resources systems, benefiting the agricultural, energy, and urban sectors. We show how energy sustainability is affected by water availability, while urban and irrigation resilience also depends on infrastructural interventions and land use changes.