Disentangling the scarcity of near-natural Iberian hydrological resources since 1980s: a multivariate-driven approach

Halifa-Marín, Amar; Torres-Vázquez, Miguel Ángel; Pravia-Sarabia, Enrique; Lemus-Cánovas, Marc; Montávez, Juan Pedro; Jiménez-Guerrero, Pedro

doi:https://doi.org/10.5194/hess-2021-565

Preprints

https://doi.org/10.5194/hess-2021-565

Preprints

30 Nov 2021

Status: this preprint is currently under review for the journal HESS.

Disentangling the scarcity of near-natural Iberian hydrological resources since 1980s: a multivariate-driven approach

Amar Halifa-Marín¹, Miguel Ángel Torres-Vázquez¹, Enrique Pravia-Sarabia¹, Marc Lemus-Cánovas², Juan Pedro Montávez¹, and Pedro Jiménez-Guerrero¹ Amar Halifa-Marín et al.

Show author details

¹Regional Atmospheric Modelling (MAR) Group, Regional Campus of International Excellence Campus Mare Nostrum (CEIR), University of Murcia, 30100 Murcia, Spain
²Climatology Group, Department of Geography, University of Barcelona, 08001 Barcelona, Spain

Received: 09 Nov 2021 – Discussion started: 30 Nov 2021

Abstract. A significant abrupt decrease of Winter Precipitation (WP) has been noticed in the Iberian Peninsula since the 1980s related to atmospheric drivers. This contribution assesses the long-term variability of water resources based on a multivariate-driven approach. For this purpose, the novel dataset of Near Natural Water Inflows to Reservoirs of Spain (NENWIRES) was created. Results confirm that Winter Water Inflows (WWI) have been modulated by the sudden decline in WP. These drastic reductions of WP/WWI were mainly controlled by the enhancement of the positive phase of the North Atlantic Oscillation (NAOi+). Nonetheless, our results also highlight the anthropogenic/physical causes contributing to the 1980s shift in the hydroclimate series. The rise of temperature, the cropland abandonment and forest extension provoked evapotranspiration gains and run-off weakening. NENWIRES most humid catchments registered the decrease of WWI promoted by NAOi+ persistence/frequency, while the land greening-up and ET rises explain the WWI losses in the Iberian semiarid environments. This contribution sheds some light on the recent debate about magnitude/drivers of streamflow declining over southern Europe. Therefore, it might help water planning with the goal of mitigating the climate change impacts affecting the water cycle.

Amar Halifa-Marín et al.

Status: final response (author comments only)

RC1:
'Interesting work, but some more clarification and quantification needed', Adriaan J. (Ryan) Teuling, 12 Feb 2022

Review of “Disentangling the scarcity of near-natural Iberian hydrological resources since 1980s: a multivariate-driven approach” by Amar Halifa-Marín et al.

The work by Halifa-Marín and co-workers addresses the causes of changes in water resources availability over the Iberian peninsula. This is highly relevant issue given the current stress on water resources in the region, and the fact that not all studies are in agreement on what drives these changes. Overall the study is well motivated, has clear research goals and hypotheses, and the results seem to be robust. However there are some issues that I believe need to be addressed before the manuscript can be considered for publication in HESS. These are discussed in more detail below.

Title

The current title seems to suggest that the focus is on understanding the scarcity of water resources, rather than its changes. This might be slightly misleading, because scarcity itself does not seem to be the focus of the study, but rather water resources availability. If the authors choose to keep scarcity in the title and manuscript, this term should be accompanied by a clear definition.

Data quality

The study is based on a dataset of water inflows into reservoirs as determined by the Spanish Ministry of Environment. This raises the question how these data were observed. From the text, the impression is given that these flows have been measured before they flow into the reservoir. However it seems much more likely that these are determined from the reservoir outflow while accounting for reservoir storage changes, and perhaps even using inverse modeling (have the flows been “naturalised”?). Of course this can have impacts on the inferred trends. More information should be given on the nature of the data, techniques used etc.

Natural flow

On line 94, the authors state that “we identify/focus on reservoirs where its water inflows nearly keep the natural flow regime.” Please define clearly what is meant by natural flow regime, and how this was checked. Clearly, these basins are still influenced by non-natural land use changes such as abandonment of agricultural land. I don’t necessarily disagree with the approach taken, but it is important to realize that basins without big water infrastructure and not necessarily near-natural.

Presentation

While generally the illustrations are of high quality, much of the results are presented in the form of maps. This risk here is that the manuscript reads more like a report than a scientific study, and at the same time it is much harder to draw quantitative (and objective) conclusions based on visual interpretation of spatial patterns. I suggest the authors reconsider the style of some of the figures, so that present key metrics that can be used to base the conclusions on.

Detailed comments

Line 73: “All these works highlighted the role of afforestation processes into the rise of potential evapotranspiration (ETP)” -> I don’t think this statement can be backed up by previous studies. ETP does increase in response to global warming, and afforestation increased actual ET – but how exactly would one lead to the other?

Figure 13: please show the regression lines, and explain the vertical lines.

Citation: https://doi.org/10.5194/hess-2021-565-RC1
- AC1:
  'Reply on RC1', Amar Halifa Marín, 14 Mar 2022
  GC: The work by Halifa-Marín and co-workers addresses the causes of changes in water resources availability over the Iberian Peninsula. This is highly relevant issue given the current stress on water resources in the region, and the fact that not all studies are in agreement on what drives these changes. Overall the study is well motivated, has clear research goals and hypotheses, and the results seem to be robust. However there are some issues that I believe need to be addressed before the manuscript can be considered for publication in HESS. These are discussed in more detail below.
  We will complete the work as suggested. Thank you for your consideration.
  
  C1: Title. The current title seems to suggest that the focus is on understanding the scarcity of water resources, rather than its changes. This might be slightly misleading, because scarcity itself does not seem to be the focus of the study, but rather water resources availability. If the authors choose to keep scarcity in the title and manuscript, this term should be accompanied by a clear definition.
  Right, thank you for the support. The title has been modified.
  
  C2: Data quality. The study is based on a dataset of water inflows into reservoirs as determined by the Spanish Ministry of Environment. This raises the question how these data were observed. From the text, the impression is given that these flows have been measured before they flow into the reservoir. However it seems much more likely that these are determined from the reservoir outflow while accounting for reservoir storage changes, and perhaps even using inverse modeling (have the flows been “naturalised”?). Of course this can have impacts on the inferred trends. More information should be given on the nature of the data, techniques used etc.
  Thank you for the recommendations. Further information about data (techniques used to estimate the water inflows) has been added in the manuscript and Appendix A.
  C3: Natural flow. On line 94, the authors state that “we identify/focus on reservoirs where its water inflows nearly keep the natural flow regime.” Please define clearly what is meant by natural flow regime, and how this was checked. Clearly, these basins are still influenced by non-natural land use changes such as abandonment of agricultural land. I don’t necessarily disagree with the approach taken, but it is important to realize that basins without big water infrastructure and not necessarily near-natural.
  Right, thanks. It has been modified/clarified in the manuscript. Line 328, Section 2.1.
  C4: Presentation. While generally the illustrations are of high quality, much of the results are presented in the form of maps. This risk here is that the manuscript reads more like a report than a scientific study, and at the same time it is much harder to draw quantitative (and objective) conclusions based on visual interpretation of spatial patterns. I suggest the authors reconsider the style of some of the figures, so that present key metrics that can be used to base the conclusions on.
  Thanks for the recommendation. The style of some of the figures has been modified in order to improve the interpretation of results/conclusions.
  Detailed comments.
  Line 73: “All these works highlighted the role of afforestation processes into the rise of potential evapotranspiration (ETP)” -> I don’t think this statement can be backed up by previous studies. ETP does increase in response to global warming, and afforestation increased actual ET – but how exactly would one lead to the other?
  
  Thanks for your recommendation. The sentence has been modified in the manuscript. Line 199.
  Figure 13: please show the regression lines, and explain the vertical lines.
  
  Thanks for the recommendation. All these suggesitons have been added to that figure (currently, Fig. 12).
  
  Citation: https://doi.org/10.5194/hess-2021-565-AC1
RC2:
'Comment on hess-2021-565', Anonymous Referee #2, 20 Feb 2022

The study presented in this manuscript investigated winter precipitation and reservoir inflows on the Iberian peninsula with the objective of attribution of recent declines in these inflows in particular to climate and landcover changes. While the topic is generally interesting to the region, the manuscript unfortunately does not present the material clearly enough for publication. The title indicates where more clarity and rigor is needed: neither is "resource scarcity" clearly defined and analysed, nor does it become clear what is meant by 'a multivariate-driven approach'. Multiple variables as drivers of change are studied, but not together in a multi-variate approach. An overall attribution approach is not clearly developed or at least not understandably presented. Overall, phrasing of the manuscript is often imprecise and will require a lot of improvement before this manuscript could potentially be considered for publication.

Major concerns are:

1) There is no clear attribution methodology and a mix of assumptions regarding the type of time change: break points, but then correlation between variables (assuming linear relations over time), quantile-quantile tests, .... Many different indices are used, transformations and individual analyses are carried out. Unfortunately I did not understand how they together should arrive at a result regarding attribution of reservoir inflow change (abrupt) to different causes with different time change signals. A combined hypothesis and truly multi-variate ananlysis of cause-effect relations is missing.

2) Overall, phrasing of the manuscript is often imprecise. The introduction reviews many aspects not really addressed later and does not really follow a clear flow of thought. In the last paragraph of the introduction the scientific aim of the study does not become clear. The paragraph states three "secondary aims" but no primary aim. A dataset and method (mentioned before, but not clearly named objective or aim) cannot be a research aim, but are necessary materials and tools. Aims: 1) Attribution of what to what exactly? 2) Propagation OF ? the winter precip deficit INTO ? - wrong prepostitions make this impossible to understand. Into the annual net reservoir inflow? 3) What is meant with "confluence"? How can drivers flow together? This choice of terminology makes no sense.

Another example is line 221-222 where an aim is phrased in the method section (should be in the introduction), followed by "this approach..."...but which approach (method) is meant is unclear as none is named yet. The order of sentences is not logical to the reader, I am afraid, and this would make more sense as a concluding sentence at the end perhpas?.

3) As already noted by Rev. 1 the use of terminology should follow common climate risk research conventions.

- "water scarcity" is usually used from a human perspective of water demand exceeding water supply (usually for people, not rally used as a term for vegetation)

- p2 line 39 "vulnerability of freshwater" - only people, society, economy etc. can be "vulnerable" to climate change impacts such as e.g. a lack of water. I am not sure what is meant here but correct might be: ...study the vulnerability of the Iberian P. to a decline in freshwater supply (as an impact of coupled climatic and landcover change?). However this would imply that the social vulnerability is studied and that appears not to be the case - here the impact of climatic change on water supply is studied (not the social aspects of the demand, which might define vulnerability of certain sectors).

4) Data and Methods contain much unnecessary details about file formats and statistical analysis, but do not state what the methods do regarding an overall method of attribution of changes in y to changes in x. Many methods are missing - the results section provides many surprises for the reader.

5) In the 'Results' quite often 'drought' is mentioned, but how is drought defined and how do droughts (as temporary events rather than permanent changes) function within a hypothesis of cause-effect relations to be 'disentangled'?

6) The NAOI analysis maybe provides robust results, but why the conclusion that the NAOI shift has been "encouraged by T and ETP change" remains unclear to me. The results appear to introduce new analysis that has not been introduced clearly enough in the methods section. Readers need to be guided more which analysis tests which hypothesis and how this adds to the greater aim.

7) Similar to Rev. 1 I could not follow the forest-change argument. A review of knowledge on the issue and a clear process-cause-effect hypothesis building on it would help perhpas.

8) Further technical deficits include (not an exhaustive list) the wrong terminology in figure captions ("hydrograph" in Fig.2 is not a hydrograph for example), no discussion section, a conclusion list that does not contain congruent arguments.

Citation: https://doi.org/10.5194/hess-2021-565-RC2
- AC2:
  'Reply on RC2', Amar Halifa Marín, 14 Mar 2022
  The study presented in this manuscript investigated winter precipitation and reservoir inflows on the Iberian peninsula with the objective of attribution of recent declines in these inflows in particular to climate and landcover changes. While the topic is generally interesting to the region, the manuscript unfortunately does not present the material clearly enough for publication.
  Thank you very much for your kind words and for the invaluable time spent in a detailed and insightful reading of our manuscript.
  The title indicates where more clarity and rigor is needed: neither is "resource scarcity" clearly defined and analysed, nor does it become clear what is meant by 'a multivariate-driven approach'.
  Thank you for your comment. The title has been modified, as “A multivariate-driven approach disentangling the abrupt reduction of near-natural Iberian streamflow post-1980”, and we think that it adequately addressed the topic of our study, given that the manuscript aim/study the contribution of multiple variables to the water inflows (streamflow) declining.
  Multiple variables as drivers of change are studied, but not together in a multi-variate approach. An overall attribution approach is not clearly developed or at least not understandably presented. Overall, phrasing of the manuscript is often imprecise and will require a lot of improvement before this manuscript could potentially be considered for publication.
  A lot of changes have been applied to the manuscript, addressing your recommendations. We hope that we made significant improvements to the manuscript.
  Major concerns are:
  There is no clear attribution methodology and a mix of assumptions regarding the type of time change: break points, but then correlation between variables (assuming linear relations over time), quantile-quantile tests, .... Many different indices are used, transformations and individual analyses are carried out. Unfortunately I did not understand how they together should arrive at a result regarding attribution of reservoir inflow change (abrupt) to different causes with different time change signals. A combined hypothesis and truly multi-variate ananlysis of cause-effect relations is missing.
  
  Thanks for your comment. Our work assesses the variability of wintertime Iberian streamflow in a long-term analysis. We quantified the trend magnitudes in order to analyse the recent evolution of streamflow in the target catchments (NENWIRES). This procedure allows us to confirm whether our data shows similar patterns of tendency which has been noticed in the literature. But our study particularly focused on the analysis of abrupt changes in water inflows (streamflow) post-1980 (echoing the call made in some references cited in the manuscript, for assess the potential propagation of the abrupt change registered in wintertime precipitation into streamflow series within the Iberian Peninsula). Given that our results noticed that water inflow changes cannot be essentially associated to precipitation variability in all NENWIRES series (target catchments), we thus evaluate the contribution of other variables. Nonetheless, we think that the Q-Q analysis and linear relationships also provide results necessary to establish our conclusions. Anyway, the drafting has been amended so as to render the text clearer with regard to readers.
  Overall, phrasing of the manuscript is often imprecise. The introduction reviews many aspects not really addressed later and does not really follow a clear flow of thought. In the last paragraph of the introduction the scientific aim of the study does not become clear. The paragraph states three "secondary aims" but no primary aim. A dataset and method (mentioned before, but not clearly named objective or aim) cannot be a research aim, but are necessary materials and tools. Aims: 1) Attribution of what to what exactly? 2) Propagation OF ? the winter precip deficit INTO ? - wrong prepostitions make this impossible to understand. Into the annual net reservoir inflow? 3) What is meant with "confluence"? How can drivers flow together? This choice of terminology makes no sense.
  
  Thanks for your comment. Phrasing of the manuscript has been modified attending these recommendations, as well as the aims of the manuscript currently are: “And this contribution aims at: 1) quantifying the tendency of near-natural water inflows (recent evolution); 2) assessing the potential propagation of the post-1980 abrupt WP reduction to the water inflows series; 3) analysing links between the abrupt changes detected on hydroclimate variables and the NAOi enhancement; 4) reporting the catchments where water inflows are not mainly modulated by WP/NAOi changes; 5) disentangling other factors that contributed to the water inflow changes; and 6) identifying the principal trigger of water inflow changes depending on climate conditions in basins”.
  Another example is line 221-222 where an aim is phrased in the method section (should be in the introduction), followed by "this approach..."...but which approach (method) is meant is unclear as none is named yet. The order of sentences is not logical to the reader, I am afraid, and this would make more sense as a concluding sentence at the end perhaps?.
  Right, we try to use a more accurate term in the new version of the manuscript. Thank you for pointing out.
  As already noted by Rev. 1 the use of terminology should follow common climate risk research conventions.
  
  - "water scarcity" is usually used from a human perspective of water demand exceeding water supply (usually for people, not rally used as a term for vegetation)
  We have tried to use a more accurate term in the revised manuscript. Thank you for pointing out.
  - p2 line 39 "vulnerability of freshwater" - only people, society, economy etc. can be "vulnerable" to climate change impacts such as e.g. a lack of water. I am not sure what is meant here but correct might be: ...study the vulnerability of the Iberian P. to a decline in freshwater supply (as an impact of coupled climatic and landcover change?). However, this would imply that the social vulnerability is studied and that appears not to be the case - here the impact of climatic change on water supply is studied (not the social aspects of the demand, which might define vulnerability of certain sectors).
  Right, thanks for your recommendation. This sentence has been amended in the manuscript.
  Data and Methods contain much unnecessary details about file formats and statistical analysis, but do not state what the methods do regarding an overall method of attribution of changes in y to changes in x. Many methods are missing - the results section provides many surprises for the reader.
  
  The Section 2.3 (analysis procedures) and Section 3 (results/discussion) have been amended in order to attend your suggestion. Thanks.
  In the 'Results' quite often 'drought' is mentioned, but how is drought defined and how do droughts (as temporary events rather than permanent changes) function within a hypothesis of cause-effect relations to be 'disentangled'?
  
  The revised manuscript mentions “permanent drought conditions” referring to SPEI6 and SPEI12 estimations. We assume that negative values of SPEI confirm “drought conditions” in the previous 6/12 months due to precipitation(temperature) was registered below(above) the normal records (climatology on the study period). It has been clarified in the Section 2.3 (analysis procedures) and Section 3 (results/discussion).
  The NAOI analysis maybe provides robust results, but why the conclusion that the NAOI shift has been "encouraged by T and ETP change" remains unclear to me. The results appear to introduce new analysis that has not been introduced clearly enough in the methods section. Readers need to be guided more which analysis tests which hypothesis and how this adds to the greater aim.
  
  Thanks for the recommendation. This sentence has been modified in the manuscript. We referred to the probable rise of maximum temperature caused by the increase of insolation. But we do not mention this sentence in the revised manuscript so as to render the text clearer with regard to readers.
  Similar to Rev. 1 I could not follow the forest-change argument. A review of knowledge on the issue and a clear process-cause-effect hypothesis building on it would help perhaps.
  
  Both recommendations were attended, and the text has been modified.
  Further technical deficits include (not an exhaustive list) the wrong terminology in figure captions ("hydrograph" in Fig.2 is not a hydrograph for example), no discussion section, a conclusion list that does not contain congruent arguments.
  
  Our hydrograph (Fig 2.) shows the contribution of monthly water inflows to annual records (as conventional hydrographs). A clarification has been added in the caption of Fig. 2. We have tried to use a more accurate terminology in the revised manuscript. Thank you for pointing out. Discussion is included in the Section 3 (Results and Discussion). The final remarks have been amended to reflect adequately our principal findings.
  
  Citation: https://doi.org/10.5194/hess-2021-565-AC2

Amar Halifa-Marín et al.

Data sets

Spanish catchments boundaries dataset Spanish Ministry of Environment (Ministerio para la Transición Ecológica y el Reto Demográfico, MITECO) https://www.miteco.gob.es/es/cartografia-y-sig/ide/descargas/agua/cuencas-y-subcuencas.aspx

HILDA: HIstoric Land Dynamics Assessment dataset Wageningen University and Research (WUR) https://www.wur.nl/en/Research-Results/Chair-groups/Environmental-Sciences/Laboratory-of-Geo-information-Science-and-Remote-Sensing/Models/Hilda/HILDA-data-downloads.htm

The North Atlantic Oscillation Index National Oceanic and Atmospheric Administration (NOAA) https://www.cpc.ncep.noaa.gov/products/precip/CWlink/pna/nao.shtml

NCEP/NCAR Reanalysis 1 climate data Physical Sciences Laboratory (PSL), NOAA https://psl.noaa.gov/data/gridded/data.ncep.reanalysis.html

Mid-resolution permeability of soils dataset Spanish Geological Survey (Instituto Geológico y Minero Español, IGME) http://mapas.igme.es/Servicios/default.aspx#IGME_Permeabilidad_1M

Observational monthly water inputs series Spanish Hydrological Repository (Centro de Experimentación de Obras Públicas, CEDEX) https://ceh.cedex.es/anuarioaforos/demarcaciones.asp

Precipitation/Temperature high-resolution gridded dataset Spanish Agency of Meteorology (Agencia Española de Meteorología, AEMET) http://www.aemet.es/es/serviciosclimaticos/cambio_climat/datos_diarios?w=2

Amar Halifa-Marín et al.

Viewed

Total article views: 499 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
396	88	15	499	8	7

HTML: 396
PDF: 88
XML: 15
Total: 499
BibTeX: 8
EndNote: 7

Views and downloads (calculated since 30 Nov 2021)

Month	HTML	PDF	XML	Total
Nov 2021	60	11	0	71
Dec 2021	144	25	2	171
Jan 2022	26	12	1	39
Feb 2022	61	10	4	75
Mar 2022	39	11	3	53
Apr 2022	29	11	3	43
May 2022	28	5	1	34
Jun 2022	9	3	1	13

Cumulative views and downloads (calculated since 30 Nov 2021)

Month	HTML	PDF	XML	Total
Nov 2021	60	11	0	71
Dec 2021	144	25	2	171
Jan 2022	26	12	1	39
Feb 2022	61	10	4	75
Mar 2022	39	11	3	53
Apr 2022	29	11	3	43
May 2022	28	5	1	34
Jun 2022	9	3	1	13

Viewed (geographical distribution)

Total article views: 478 (including HTML, PDF, and XML) Thereof 478 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 28 Jun 2022

Short summary

The near-natural Iberian water resources have suddenly decreased since the 1980s. These declines were promoted by the weakening (enhancement) of wintertime precipitation (NAO index) in the most humid areas, whereas the extension of afforestation processes (greening-up) played a crucial role in the semiarid areas. The future water management might require greater knowledge of North Atlantic climate variability and the afforestation control/reverse in the semiarid catchments.


Total:	0
HTML:	0
PDF:	0
XML:	0