the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
MESSAGEix-GLOBIOM Nexus Module: Integrating water sector and climate impacts
Adriano Vinca
Edward Byers
Stefan Frank
Oliver Fricko
Esther Boere
Peter Burek
Miguel Poblete Cazenave
Paul Natsuo Kishimoto
Alessio Mastrucci
Yusuke Satoh
Amanda Palazzo
Madeleine McPherson
Keywan Riahi
Volker Krey
Abstract. The Integrated Assessment Model (IAM) MESSAGEix-GLOBIOM developed by IIASA is widely used to analyse global change and socioeconomic development scenarios within the energy and land systems across different scales. However, until now, the representation of impacts from climate impacts and water systems within the IAM has been limited. We present a new nexus module for MESSAGEix-GLOBIOM that improves the representation of climate impacts and enables the analysis of interactions between population, economic growth, energy, land, and water resources in a dynamic system. The module uses a spatially resolved representation of water systems to retain hydrological information without compromising computational feasibility. It maps simplified water availability and key infrastructure assumptions with the energy and land systems. The results of this study inform on the transformation pathways required under climate change impacts and mitigation scenarios. The pathways include multi-sectoral indicators highlighting the importance of water as a constraint in energy and land-use decisions and the implications of global responses to limited water availability from different sources, suggesting possible shifts in the energy and land sectors.
- Preprint
(1150 KB) - Metadata XML
-
Supplement
(1325 KB) - BibTeX
- EndNote
Muhammad Awais et al.
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2023-258', Anonymous Referee #1, 03 Apr 2023
This paper describes the advancement of the MESSAGEix-GLOBIUM integrated assessment modeling framework through the integration of a new module, the nexus module. This module provides a detailed representation of the water sector with high spatial (basin level) resolution and describes the interaction of the water with the energy and land sector. This is a crucial component for a comprehensive evaluation of the sectoral interaction in the face of climate policy measures, climate change impacts and sustainable development goals, as outlined by the paper. Therefore, this constitutes an important step to allow the model to be used in highly policy relevant applications and I find this contribution important and fitting to GMD.
The paper describes the new module as well as exemplary applications, in particular an analysis of effects of climate change impacts in two policy scenarios. It However, it refrains from detailed analysis and interpretation of results, referring to planned future publications, thereby remaining quite vague. While the paper is well structured and gives a reasonable overview of the modeling approach. However, it would benefit from a critical revision regarding clarity and language. I go into details on this in the next section.
Specific comments
I think the clarity of the description could be improved in two ways. The authors seem to assume a certain level of knowledge of the readers about the MESSAGEix-GLOBIOM model and its applications. It is not clearly explained what the core application of the model is and how the new module extends that. In particular, it doesn’t become sufficiently clear what is assumed along given scenarios and what is subject to endogenous optimization decisions for readers without much previous knowledge. The term “optimization” is mentioned in line 137 without further explanation. Similarly, what is new regarding the representation of climate impacts and SDG constraints and what has been applied before (e.g. on the land-use and energy side)?
The introduction would benefit from some streamlining and a more precise wording and description. I refer to some examples of unclear sentences in the technical section below. It would be helpful to clearly distinguish between underlying scenarios (the SSPs, which are combined with the RCPs in climate policy analysis), how models implement them and what models analyze additionally with those scenarios as a backdrop (e.g. transformations to achieve the given climate policy goals). The introduction currently fails to embed the new advances properly in the IAM literature, e.g. distinguishing process-based from cost-benefit IAMs, describing which (if any) other process-based IAMs capture climate impacts/SDGs/the water sector and how this implementation differs from other approaches. The citations in the first paragraph are relevant but also quite dated. In line 52 it is mentioned that “Many IAMs consider adaptation costs in an aggregated spatial region” – I think most IAMs don’t consider adaptation costs at all (aside from fairly old approaches), if there are examples, please provide some references.
In the introduction and also in the discussion the possible application in the context of evaluating adaptation measures is emphasized. However, it doesn’t become clear how that could be done, i.e. just through endogenous model responses to climate impacts like increased irrigation or land-use change or also through explicit adaptation policy. From the introduction the reader expects more (technical/methodological) detail on this, even if an actual application is planned in a future paper.
There are a few things which don’t become clear regarding the implementation and would benefit from an improved description. Here are some examples. In the introduction (page 2) the need for higher spatial resolution of water is described. But that is also true for land (and handled like this in the EPIC-GLOBIOM connection) – maybe worth pointing this out? Further there is the issue of cooling water for thermal powerplants. That depends on water temperature but doesn’t use water up without feeding it back into rivers to my knowledge. Is that how it is considered here and where does the input data on water temperature come from? In Table 1 a bit more information on key assumptions behind some input data would be useful here (without having to look them up in all the other papers), e.g. one wonders how future water demand data are projected into the future. In the first paragraph on page 11 it is mentioned that an iteration between MESSAGE and the energy access model is not done because it would alter the GDP pathways. How much would they be altered? In the discussion of the environmental flow constraint a categorization of basins based on development status is mentioned (line 321) – is that static or changing in the future? In line 329 there is a distinction of basins with high and low adaptive capacity – do the ones with low capacity never have any targets or just later? How is groundwater treated in regions which already overuse it today?
The results section is very generic in its discussion, so it is not very helpful and leaves the interested reader with many questions. Maybe a detailed discussion of one application plus a generic list of possible future research questions to be answered with this framework would be better. It needs to be explained better what flow reliability in section 4 means and this application needs to be motivated. In the application of impacts and the energy sector the mentioning of mitigation is missing. How do the impacts interact with that, are there any impacts on bioenergy for example? Is the lower use of fossil energy mostly due to the cooling water issue or what other drivers are there? Why are the impacts in some regions like Central Asia larger under RCP2.6 than 6.0? In the water extraction example, brackish water appears for the first time. From the figure its use seems to go down but in the text a rise of its usage is mentioned. In the first paragraph on page 15 regional effects are discussed but regions are not named (“certain regions” in line 445). The discussion there doesn’t seem of the results but very generic (line 452: “impacts of climate change on water availability are likely to be negative” – the authors know what they put in from the biophysical modeling, right?). In line 458 the topic suddenly switches to yields.
Finally, the paper misses a critical discussion of assumptions. For example, why is the full ensemble of ISIMIP water models used for desalination potential but not for water availability? What about uncertainty from impact models? Is the uncertainty from climate models only used for water availability or also for the other parts like yield changes?
Minor technical points/clarifications:
Line 32/33: “… pose an extra threat to climate change risk” – maybe “additional to climate change risk”?
Line 38: “transition to ambitious global warming goals” – better “ambitious climate policy goals”
Line 50: “due to substantial challenges in technical implementation and representation” – representation of what?
Line 59: “need for a balanced integration” – what do you mean by balanced?
Line 65: You mention community resilience but that isn’t resolved in the model.
Line 114: “it simultaneously determines energy, land use..” – what do you mean by energy – energy demand, supply, mix of energy sources..?
Line 116: The acronym GHM is not explained.
Line 165: “using a set of configurations in the energy system” – what do you mean by that?
Caption Figure 2: typo: “the water system is modelled”
Line 207: the word “driving” seems to not fit
Line 252: I assume you refer to Figure 2 in the Gernaat paper – maybe make that clear?
Line 316: The sentence “The rivers’ environmental flows help protect … from achieving SDG target 6.6” – It is unclear what that means. Also, not all readers will be aware what specific sub-targets of the SDGs are, so they should be explained if mentioned. The sentence after starting with “We use the” is double.
Line 402: Some words seem to be missing.
Line 487: What is meant by “most be more tolerant of statistical
Citation: https://doi.org/10.5194/egusphere-2023-258-RC1 -
RC2: 'Comment on egusphere-2023-258', Page Kyle, 19 Apr 2023
Overall, the study marks an advancement in integrated assessment modeling, tightening the coupling between biophysical and economic systems and improving the capability to construct internally consistent scenarios. The paper summarizes a lot of good methodological work. Specifically the features added include climate impacts on agriculture, water, and energy, and an enhanced representation of the water sector. The paper would benefit from clearly articulating what is different in the study and modeling system here as compared with the prior cited model documentation (Krey et al. 2016). As well, most of the cited literature in the introduction is pretty old, and doesn't necessarily reflect the state of the art in integrated assessment modeling which has been focused on improving things like hydrology and climate impacts in the past 5 or so years. Most of the references here are from 2013-2017. Since the study addresses the sustainable development goals, a nod to the recent literature on using IAMs to quantify the SDGs would also be appropriate; I've provided some specific examples below.
The biggest issue that I have with the study is that the results that are shown are not especially meaningful indicators (among the set that could be shown), and are not explained or defended in the discourse of the results and discussion section, which right now is written largely as an extension of the introduction. A significant portion of the results and discussion is invested into background information about combined systems modeling, with commentary about why these modeling capabilities are novel and useful, but there's only sparse comparison between these results and the existing literature on similar indicator variables. Many of the key results are counter-intuitive for me, which is OK, but the problem is that the results aren't explained or defended. The result that most stands out to me here is that some of the highest levels of investment into water infrastructure globally, indicated in billions of dollars per year per water basin, are seen in mostly uninhabited Arctic basins such as Alaska, the Yukon, and Siberia, that generally have an over-abundance of freshwater. The authors clarify in the text that there aren't any inter-basin transfers, which was the only explanation for the result that I could think of (e.g., NAWAPA-type projects). Similarly, almost all of the results that are shown in the key figures (5 and 6) are indicated in units that make the logical comparisons in the figures largely meaningless. For example, Figure 5b is set up to compare the different global macro-regions, but the variable chosen is the total cost of water investment, in billions of dollars per year. The indicator isn't normalized for each region's total water supply, economic output, or population size. So, Eastern Europe sees the lowest investment costs among regions, but this is probably because it is comparatively small, and without that normalization there's not much that can be learned from the figure in terms of comparing the regions. That figure would probably make the most sense to construct as costs per unit of water supplied, but really that depends on what the figure is being used to demonstrate in the analysis. The authors stress that the results should be interpreted with caution (line 479), but what I'd prefer to see is for the figures to provide meaningful indicator variables, and then to compare the key results to the existing literature, where such literature exists. In some cases like the water investment quantities shown in Figure 5a, there might not be literature on the topic outside of this research team, and that's worth highlighting too.
The following are my specific comments:
Section 2: in the description here, there is no distinction drawn between withdrawals and consumption. It should be clarified which one is used (seems to be withdrawals), and the representation of return flows should also be described. Power plants using once-through flow cooling systems don't drive water scarcity, though a withdrawals-focused accounting framework could find such a result if return flows aren't tracked.
Line 150: Does the model consider drip irrigation or other technologies that could reduce the irrigation water intensity of irrigated crop production? Water use efficiency is introduced as a concept in the discussion (line 459), and would seem to be one of the few ways of simultaneously meeting SDGs 2 and 6, but there isn't any description in the methods.
Line 179: "water withdrawals for irrigation, energy, and cooling" - throughout the text, the term "cooling" is used interchangeably for buildings air conditioning and for thermo-electric power plant cooling. Sometimes it isn't clear if both are intended, and large buildings sometimes use water-based "chillers" for air conditioning. The industrial sector also uses water for process cooling and it isn't clear whether that's classified as such here. Please clarify what sectors/processes are intended whenever the term "cooling" is used, if it's not obvious from context.
Section 3.1, general: Please provide a brief commentary on how the basin-region crosswalk is handled. There are basins that supply water to multiple regions; for example the Nile supplies both SSA and MENA. Are the basins disaggregated in the model, such that there are e.g. basins for Nile-SSA and Nile-MENA? In that case, how is the apportioning done? Does the water supply in Nile-MENA only include renewable water from within Nile-MENA, or do its supplies include the runoff in the river, that wasn't consumed by Nile-SSA? Alternatively, do the relevant portions of each region simply share a single water resource base?
Section 3.1, general: Please provide a brief commentary on how the energy used by the water sector is handled. This seems like it should be somewhat complex in the modeling system used, as the energy and water modules are run separately, and Table 1 indicates that water distribution and wastewater treatment energy footprints are used, but I didn't see a description otherwise.
Lines 185-195: Not all renewable water input to a basin that is in excess of the base environmental flow requirement is available for abstraction (withdrawal), as a portion comes during floods that exceed the capacity of water impoundments. Table 1 states that "the outputs [of runoff and groundwater recharge] are temporally processed for further use" but it isn't clear what this means. This topic comes back up in the paper in line 455 so it seems to have been considered. Is there any reduction in the "Fr" to account for this aspect of renewable water supply?
Line 230: Why is GDP a predictor for desalination capacity? I know a reference is provided but the relationship is not intuitive, and should be explained here.
Line 238: Why is desalination potential influenced by climate? It seems that the infrastructural investment should be a function of climate, but this doesn't influence the desalination potential of any basin, which should be unlimited for coastal basins, and perhaps based on saline aquifer volumes in endorheic basins. This might just be a case of terminology, and "desalination potential" needs to be defined (normally a resource "potential" means the upper limit of production of the resource).
Section 3.3 (SDG section): there's no mention of the internal conflicts within and between the SDG's analyzed, and how those are resolved in the scenario design. One pertinent example for this study is that SDG6 simultaneously calls for reducing water impoundments in order to restore natural aquatic environments (6.6), whereas many of the other SDGs would require improvements in flood control to provide irrigation water for high and stable crop yields (SDG2), and to protect farms and infrastructures (urban, transportation, industrial; SDGs 8-11) from flood-related damages. How this study balances such competing goals should be described in the scenario design.
Lines 355-360: The authors make reference to conducting a regional downscaling case study of Zambia in the methods, but then the results don't have anything about Zambia, and the spatial maps don't have it disaggregated. Perhaps this part should be dropped from the methods altogether? I'm not sure what was intended to describe this in the methods but not the results.
Lines 436-454: I found these 20 lines in particular to involve lots of re-stating of obvious stuff that anyways isn't established by this study. The point being made is pretty simple: climate change impacts on precipitation patterns and therefore renewable water availability are heterogeneous both geographically and temporally, and this drives the behaviors in the integrated scenarios.
Line 455: please see comment about lines 185-195; it isn't clear to me in the methods whether such "unavailable" water is deducted from the water supply in the model. It's kind of a tricky thing to model because the unavailable fraction is a function of infrastructure which in this modeling scheme is endogenous.
Line 467: "there have been numerous publications on integrating SDG dimensions into Integrated Assessment Models" - please provide appropriate references.
Line 467: "this study stands out due to its novel approach of combining SDG policies with climate goals and impacts and evaluating their effectiveness in understanding the climate adaptation narrative" - This research question has been discussed in Moallemi et al. (2022, One Earth) and in Soergel et al. (2021, Nature Climate Change). The approach used in this study still adds to the portfolio of methods and research questions that have been published about the nexus between the SDGs and climate change mitigation in process-based integrated assessment models, but the authors should make reference to the recent literature and clarify what the present study adds. Also worth reading are van Vuuren et al. (2022, One Earth) and van Soest et al. (2019, Global Transitions).line ~480: "the results of this study should be compared with those of other model evaluations"
I agree! The Results & Discussion section would be a great place to conduct such a comparison. All of the key results shown in Figures 5 and 6 should be compared against existing literature. A lot of the value added of this study is the integration of multiple different systems, but most of the variables have also been assessed in other studies.Figure 6 - These are strange and counter-intuitive results (for me, anyway) that should be described, and also should be indexed against the total electric supply in each place. On its own, 100 TWh of electricity in some region in some year far into the future is a pretty abstract thing; even if a reader happens to know the base-year electricity supply in these different macro-regions, the values far into the future for some scenario are not necessarily known. Within the data shown, one interesting thing is the net change in total electricity demand. There are particularly large increases in "other non-fossil" generation which in some cases more than counter-balance the reduction in fossil generation that is due to climate impacts. Why is this?
If the climate-driven increase in total electricity supply is demand-driven, is this all because of additional air conditioning demand? If so, how do the results here compare with the existing literature, e.g. Clarke et al. (2018, Energy Economics) or van Ruijven et al. (2019, Nature Communications)? These quantities of net electric demand increase (driven by climate impacts, all else equal) seem large at a glance, but again without the sectoral decomposition and presentation of baseline electric demand for cooling, it's hard to interpret, and anyways I can't recall exactly what the literature estimates for climate-driven growth in electricity demands for air conditioning in buildings.
Also there should be observational studies on the climatic sensitivity of fossil generation to temperature changes, and just at a glance these results seem to be more climatically sensitive than what I'd expect. Regardless, my expectations shouldn't matter; the results should be compared with the literature. I believe Michelle van Vliet has published a few papers on the topic that could be useful as a starting point.
Another question that figure 6 brings up is whether these changes reflect changes in investment over time, versus capacity factors (i.e., operational differences for a similar capital base). The increase of ~80 TWh in North America hydro stands out here; is this from pushing capacity factors up due to increased water flow through the existing hydropower installed capacity, or is this from new hydropower investment that is encouraged by increased river flows and/or increased market prices for electricity?
Citation: https://doi.org/10.5194/egusphere-2023-258-RC2 -
AC1: 'Comment on egusphere-2023-258 - Response to Reviewers' Comments', Muhammad Awais, 09 Jun 2023
We are very grateful to the reviewers for their insightful feedback and comments on our paper. Their observations and suggestions have not only strengthened the validity of our research but also greatly improved the clarity of our findings. As evidence of our thorough response to the feedback, we have enclosed a document with this response. The attached file contains the reviewers' comments highlighted in bold, our responses in regular text, and any changes made to the paper in italics. This strategy has been implemented to provide a clear and systematic understanding of the modifications made in response to the reviews. We believe that this comprehensive strategy demonstrates our dedication to addressing the reviewers' comments thoroughly and transparently.
Muhammad Awais et al.
Model code and software
MESSAGEix-GLOBIOM Nexus Module: Model & Datasets Muhammad Awais, Adriano Vinca, Edward Byers, Stefan Frank, Oliver Fricko, Esther Boere, Peter Burek, Miguel Poblete Cazenave, Paul Natsuo Kishimoto, Alessio Mastrucci, Yusuke Satoh, Amanda Palazzo, Madeleine McPherson, Keywan Riahi, and Volker Krey https://doi.org/10.5281/zenodo.7687578
Muhammad Awais et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 495 | 316 | 17 | 828 | 48 | 11 | 11 |
- HTML: 495
- PDF: 316
- XML: 17
- Total: 828
- Supplement: 48
- BibTeX: 11
- EndNote: 11
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1