the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 04 Apr 2024
Propagation from Meteorological Drought to Hydrological Drought Using SPI and SPEI Combined with the Adapted Threshold Level Method
Abstract. The hydrological response within a region is often nonlinear, influenced by various factors such as physiographic features, anthropogenic activities, and climate change. Considering the inherent complexity, our study focuses on investigating the propagation from meteorological drought to hydrological drought in the Banabuiú, Castanhão, and Orós reservoirs, located in the State of Ceará, Brazil. To achieve this, we used the Standardized Precipitation Index (SPI) to identify meteorological drought events, which uses precipitation data, alongside the Standardized Precipitation Evapotranspiration Index (SPEI), incorporating potential evapotranspiration in its calculations. For assessing hydrological drought, we used the Adapted Threshold Level Method (ATLM), which calculates the water balance between the supply (represented available reservoir volume) and the demand (including withdrawals for various uses and evaporation losses from the watershed). For all three reservoirs, drought events were characterized by their frequency, duration, severity, magnitude, and drought recovery time across three aggregated time scales of 12, 24, and 36 months. To determine the propagation time from meteorological to hydrological drought, we calculated three indicators, corresponding to time differences between the onsets, peaks, and conclusions of the propagated drought events. Results indicated no significant differences in meteorological drought characteristics between SPI and SPEI methods. However, the SPEI showed higher values for meteorological drought in the Orós watershed. Analysis of drought propagation, employing different methods, revealed no defined pattern for the onset, peak, and end intervals across all tested combinations. Hence, we recommended utilizing all three indicators to enable a more comprehensive analysis of drought events propagated within the watersheds.
This preprint has been withdrawn.
-
Withdrawal notice
This preprint has been withdrawn.
-
Preprint
(3443 KB)
Interactive discussion
Status: closed
- RC1: 'Comment on egusphere-2024-813', Anonymous Referee #1, 29 Jun 2024
-
RC2: 'Comment on egusphere-2024-813', Anonymous Referee #2, 02 Jul 2024
Dear Authors,
You present an interesting comparative drought study in a large tropical dryland region, focusing on meteorological and reservoir droughts at three different time scales. Although the topic is scientifically relevant, you should clarify the novelty of your study and compare in detail your findings with literature. Discussion of results should be improved a lot. Please, find my comments and suggestions below.
Major comments:
- You presented the following gap of knowledge in drought research: Page 2, Lines: 45-52: “Although several studies have focused on explaining the propagation of meteorological drought to hydrological drought (Bevacqua et al., 2021; Guo et al., 2020; Huang et al., 2017; Wu et al., 2021a; Xu et al., 2019; Zhang et al., 2023), there remain some gaps requiring to be addressed. Sadhwani and Eldho (2024), for instance, investigated the characteristics and propagation time of meteorological-to-hydrological drought across different time scales using standardized indices, while considering future climate change scenarios. They concluded that identifying hydrological drought is more complex than meteorological drought due to the influence of diverse human activities, such as irrigation, industrial supply, urbanization, and watershed characteristics. This underscores the need for further research to quantify the effects of these factors on hydrological drought and its propagation.” However, you did not properly address this gap in the manuscript. What gap of knowledge did you address?
- Page 5, Lines 153-156: Please, explain in detail why you chose 12, 24, and 36 months as a time scale, relating to the specific characteristics of the study area, such as rainfall seasonality. It is not clear how and which sectors are related to SPI-12, SPI-24, and SPI-36.
- Page 6, Lines 165-188: The ATLM approach is interesting. However, it does not account for the interannual variability of lake evaporation and precipitation, which is rather high in the region. Also, you did not provide in the section 2.3.3, but I think the reservoir water demands, which are essential for the ATLM approach, are solely based on reservoir water allocation in 2022 (see 2.2 Data Collection). However, I expect the water demands changed a lot during about 30 years of analysis. My suggestion is to improve a lot of the needed data to apply the ATLM approach, or to use another approach focusing on the available hydrological variable of the reservoir. In the latter suggestion you can still compare the results with what was observed in water allocation constraints.
- Page 9, 2.3.5 Propagation of droughts: What about the hydrological droughts that began before the beginning of meteorological droughts or after the end of meteorological droughts? Meteorological droughts still propagate into hydrological droughts under those conditions.
- Section 3.1 Characterization of Meteorological Drought Using the SPI and SPEI Indices: Please, provide a table that summarizes the characteristics of all meteorological events. Those characteristics, such as duration and severity, were already defined by you in the methodology. You can provide it as supplementary material.
- Section 3.1 Characterization of Meteorological Drought Using the SPI and SPEI Indices: Please, compare in detail your findings with literature. The droughts in Ceará and in Northeast Brazil have been extensively studied. Highlight the novelty of your results.
- Section 3.1 Characterization of Meteorological Drought Using the SPI and SPEI Indices: The frequency of negative SPI values should be around 50% (Fig. 7). However, that is inherent to the SPI approach. Please, focus on the frequency of drought events as you defined in the methodology.
- Page 15, Lines 309-311: “This result may suggest that the influence of evapotranspiration on the SPEI calculation has a more pronounced impact on drought characteristics in the Orós watershed compared to Banabuiú and Castanhão.” Why?
- Section 3.1 Characterization of Meteorological Drought Using the SPI and SPEI Indices: It is not clear what you expect from the comparison between SPI and SPEI indices. What did you learn from this comparison?
- Section 3.1 Characterization of Meteorological Drought Using the SPI and SPEI Indices: Increasing the time scale from 12 to 36 months, is it not expected a smoothing of drought events? Again, what did you learn from this comparison?
- In general, I suggest separating the discussion from the results. In a new discussion section, you can focus on the main results of your work, then compare them with literature, present your study limitations, and propose further work.
- Page 16, Lines 329-332: “As the time factor increases, drought events tend to begin earlier, and the intervals between them shorten, ultimately leading to prolonged periods of hydrological drought across the historical series for the analyzed reservoirs, particularly noticeable in the Orós and Banabuiú Dams. Moreover, there is a notable escalation in the volume of water deficit over time. While this pattern is observed across all three reservoirs, it is most pronounced in the Orós and Banabuiú Dams.” Why was this pattern of hydrological drought completely different from the findings of meteorological droughts? Is it comparable SPI and SPEI as meteorological drought variables with ATLM-based variables for hydrological drought? What if you use a SPI-equivalent index for the reservoirs (see Van Langen et al. 2021).
- Figure 11: You should standardize the hydrological drought characteristics [(b) and (c)], because the reservoirs have different total capacities.
- Section 3.2 Characterization of Hydrological Drought Using the Adapted Threshold Level Method (ATLM): Here, I repeat the same suggestion for meteorological drought: Please, compare in detail your findings with literature. The droughts in Ceará and in Northeast Brazil have been extensively studied. Highlight the novelty of your results.
- 3.3. Propagation from meteorological drought to hydrological drought: It is confusing to me that now you need to compare meteorological and hydrological droughts at different time scales. If it was a goal to find the best combination of drought indices to better understand the drought propagation, varying the timescale values. I think only three timescales are not enough, and assuming 12, 24 and 36 months as timescales is quite arbitrary.
- How may the dense reservoir network found in your studied areas impact the propagation of meteorological droughts into hydrological droughts [see e.g. Van Langen et al. (2021) and Ribeiro Neto et al. (2022)]? How does your study help us to improve our understanding of this process?
Minor comments:
- Page 3, Lines 91-92: Please, review the sentence: “The region is monitored by 153 reservoirs are monitored…”
- Page 3, Line 94: “Consequently, the watersheds of the three main monitored reservoirs in this region were selected…” Why “consequently”?
- Page 4, Line 99: “The Orós Dam plays a crucial as a water source…” Do you mean “a crucial role”?
- Page 4, Lines: 102-104: “These reservoirs were selected due to their strategic significance to the state of Ceará, as they supply water for various purposes including drinking, agriculture, livestock, and other economic activities across multiple municipalities in the state (Gonçalves et al., 2023).” Please provide a more detailed description of water uses of these reservoirs, highlighting “numbers”
- Page 4, Line 106: “High evaporation rates…” How much?
- Fig. 1: Municipal boundaries are not visible, but I think they are not necessary; Change the color of Pluviometric stations on the map, maybe black; and have a look at the decimal separation of DEM limits.
- Sometimes you use “Dam”, and sometimes you use “Reservoir”, I think you always mean “Reservoir”.
References
van Langen, S. C. H., Costa, A. C., Ribeiro Neto, G. G., & van Oel, P. R. (2021). Effect of a reservoir network on drought propagation in a semi-arid catchment in Brazil. Hydrological Sciences Journal, 66(10), 1567–1583. https://doi.org/10.1080/02626667.2021.1955891
Ribeiro Neto, G. G., Melsen, L. A., Martins, E. S. P. R., Walker, D. W., & van Oel, P. R. (2022). Drought Cycle Analysis to evaluate the influence of a Dense Network of small reservoirs on drought evolution. Water Resources Research, 58, e2021WR030799. https://doi.org/10.1029/2021WR030799
Citation: https://doi.org/10.5194/egusphere-2024-813-RC2
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-813', Anonymous Referee #1, 29 Jun 2024
Review of Medeiros and Silva (2024).
The manuscript by Medeiros and Silva investigates drought propagation (the transition from meteorological drought to hydrological drought) in the Brazilian semi-arid region. The authors use two different methods (SPI and SPEI) for defining meteorological drought and one method (ATLM) for defining hydrological drought, the latter using reservoir levels as the main variable.
I think the paper reads well, and the figures and analysis are well made and clear. My main issue is the lack of fundamental scientific advancement (what is the real objective of this paper besides combining different methods and comparing their results) or even an in-depth discussion.
I am surprised that the authors found out that (as per the abstract) the “Analysis of drought propagation, employing different methods, revealed no defined pattern for the onset peak, and end intervals across all tested combinations. Hence, we recommended utilizing all three indicators to enable a more comprehensive analysis of drought events propagated within the watersheds.”. At the same time, the authors acknowledge that such findings were expected according to recent literature.
As it is right now the paper reads as a report/case study and lacks a clear goal besides simply applying the method. I recommend rejection with the encouragement to resubmit, so that the authors have time to improve the manuscript. I also suggest breaking the results and discussion into two different sections so as to help the reader clearly understand the advances proposed in this paper.
Detailed comments:
L26: Isn’t this statement too broad? Is this happening everywhere? Is water supply affected only by extreme events? Are all extreme events tied to a reduction in water supply? Maybe be a bit more specific here.
L28: In the abstract there are different definition of droughts and already here drought is defined as an availability issue. But in the next paragraph you bring 4 other definitions. I am a bit confused.L32: Please expand this paragraph to explain briefly what each of these definitions mean. The following paragraphs all rely on these important definitions. Your study uses reservoirs as your target variable to compute hydrologic drought but is this the only way of estimating hydrologic drought? Please expand how hydrologic drought is inferred and connect it to your study (i.e. hydrology as reservoirs levels).
L37: Below average precipitation impairs soil moisture retention reads a bit too complex to something maybe simpler: reduction in soil moisture.
L68-74: I think this paragraph needs some modification. You’re starting it by talking about the importance of incorporating land use and demand in better quantifying hydrologic drought. Ok. But then you’re jumping on “instead of …” and talk about hydrologic indices that use inflows. I just can’t follow the logical connections between the first and second half of this paragraph. Additionally, TLM is barely defined in the context of your introduction, what it does and the differences between it and other methods. This is making this paragraph very confusing, especially because you’re moving toward defining the goals of the paper right after.
L76: I don’t think “therefore” fits here. This sentence has no connection with the previous paragraph.
*Why two different paragraphs? I think the paragraphs from L80-85 are just a continuation of L76-70, so both should go together.
*What is the main advantage of your method? Why are you proposing this method? Your introduction leaves too much room for vagueness. I can seem to follow why the reader should care. Try to emphasize what type of advantage/discovery/improvement comes from what you are proposing. As of now, the introduction reads more as a summary of what you have done, than a proper “sales pitch”.
L91: Try to consolidate all climate information in one paragraph. Here you mention semi-arid and two paragraphs below you describe the climate.
L92: Collapsed? Maybe went dry? When I read collapse, it reads as damage to the structure.
L92-93: Try to be succinct with the description of the watersheds. Is the information on the 2012 drought necessary here? Maybe you need a paragraph in the introduction talking about the relevance of these reservoirs for the research area.
L105: Is hot really an adjective used to classify climate? I just have never seen this.
L116: How was evaporation measured/estimated to obtain the climatological normal values?
L130: Please elaborate on what the seminar means. I don’t think I’ve heard this concept before so maybe other readers will have the same question. Additionally, are you referring to demands or withdrawals?
L158: I think the authors need to better explore in the introduction the reason why they’re using SPI and SPEI together. Is this approach common? What do you intend to gain from it? This ties back to my comment from a few paragraphs ago. Why are the authors proposing their method as it is?
L171: Losses due to demands or withdrawals? Please clarify if demands or withdrawals were considered, and provide more details on the “seminar”.
L465: If it is already known in the literature that the transition from meteorological drought to hydrological drought is often site specific, what does your study contribute to the topic?
Citation: https://doi.org/10.5194/egusphere-2024-813-RC1 -
RC2: 'Comment on egusphere-2024-813', Anonymous Referee #2, 02 Jul 2024
Dear Authors,
You present an interesting comparative drought study in a large tropical dryland region, focusing on meteorological and reservoir droughts at three different time scales. Although the topic is scientifically relevant, you should clarify the novelty of your study and compare in detail your findings with literature. Discussion of results should be improved a lot. Please, find my comments and suggestions below.
Major comments:
- You presented the following gap of knowledge in drought research: Page 2, Lines: 45-52: “Although several studies have focused on explaining the propagation of meteorological drought to hydrological drought (Bevacqua et al., 2021; Guo et al., 2020; Huang et al., 2017; Wu et al., 2021a; Xu et al., 2019; Zhang et al., 2023), there remain some gaps requiring to be addressed. Sadhwani and Eldho (2024), for instance, investigated the characteristics and propagation time of meteorological-to-hydrological drought across different time scales using standardized indices, while considering future climate change scenarios. They concluded that identifying hydrological drought is more complex than meteorological drought due to the influence of diverse human activities, such as irrigation, industrial supply, urbanization, and watershed characteristics. This underscores the need for further research to quantify the effects of these factors on hydrological drought and its propagation.” However, you did not properly address this gap in the manuscript. What gap of knowledge did you address?
- Page 5, Lines 153-156: Please, explain in detail why you chose 12, 24, and 36 months as a time scale, relating to the specific characteristics of the study area, such as rainfall seasonality. It is not clear how and which sectors are related to SPI-12, SPI-24, and SPI-36.
- Page 6, Lines 165-188: The ATLM approach is interesting. However, it does not account for the interannual variability of lake evaporation and precipitation, which is rather high in the region. Also, you did not provide in the section 2.3.3, but I think the reservoir water demands, which are essential for the ATLM approach, are solely based on reservoir water allocation in 2022 (see 2.2 Data Collection). However, I expect the water demands changed a lot during about 30 years of analysis. My suggestion is to improve a lot of the needed data to apply the ATLM approach, or to use another approach focusing on the available hydrological variable of the reservoir. In the latter suggestion you can still compare the results with what was observed in water allocation constraints.
- Page 9, 2.3.5 Propagation of droughts: What about the hydrological droughts that began before the beginning of meteorological droughts or after the end of meteorological droughts? Meteorological droughts still propagate into hydrological droughts under those conditions.
- Section 3.1 Characterization of Meteorological Drought Using the SPI and SPEI Indices: Please, provide a table that summarizes the characteristics of all meteorological events. Those characteristics, such as duration and severity, were already defined by you in the methodology. You can provide it as supplementary material.
- Section 3.1 Characterization of Meteorological Drought Using the SPI and SPEI Indices: Please, compare in detail your findings with literature. The droughts in Ceará and in Northeast Brazil have been extensively studied. Highlight the novelty of your results.
- Section 3.1 Characterization of Meteorological Drought Using the SPI and SPEI Indices: The frequency of negative SPI values should be around 50% (Fig. 7). However, that is inherent to the SPI approach. Please, focus on the frequency of drought events as you defined in the methodology.
- Page 15, Lines 309-311: “This result may suggest that the influence of evapotranspiration on the SPEI calculation has a more pronounced impact on drought characteristics in the Orós watershed compared to Banabuiú and Castanhão.” Why?
- Section 3.1 Characterization of Meteorological Drought Using the SPI and SPEI Indices: It is not clear what you expect from the comparison between SPI and SPEI indices. What did you learn from this comparison?
- Section 3.1 Characterization of Meteorological Drought Using the SPI and SPEI Indices: Increasing the time scale from 12 to 36 months, is it not expected a smoothing of drought events? Again, what did you learn from this comparison?
- In general, I suggest separating the discussion from the results. In a new discussion section, you can focus on the main results of your work, then compare them with literature, present your study limitations, and propose further work.
- Page 16, Lines 329-332: “As the time factor increases, drought events tend to begin earlier, and the intervals between them shorten, ultimately leading to prolonged periods of hydrological drought across the historical series for the analyzed reservoirs, particularly noticeable in the Orós and Banabuiú Dams. Moreover, there is a notable escalation in the volume of water deficit over time. While this pattern is observed across all three reservoirs, it is most pronounced in the Orós and Banabuiú Dams.” Why was this pattern of hydrological drought completely different from the findings of meteorological droughts? Is it comparable SPI and SPEI as meteorological drought variables with ATLM-based variables for hydrological drought? What if you use a SPI-equivalent index for the reservoirs (see Van Langen et al. 2021).
- Figure 11: You should standardize the hydrological drought characteristics [(b) and (c)], because the reservoirs have different total capacities.
- Section 3.2 Characterization of Hydrological Drought Using the Adapted Threshold Level Method (ATLM): Here, I repeat the same suggestion for meteorological drought: Please, compare in detail your findings with literature. The droughts in Ceará and in Northeast Brazil have been extensively studied. Highlight the novelty of your results.
- 3.3. Propagation from meteorological drought to hydrological drought: It is confusing to me that now you need to compare meteorological and hydrological droughts at different time scales. If it was a goal to find the best combination of drought indices to better understand the drought propagation, varying the timescale values. I think only three timescales are not enough, and assuming 12, 24 and 36 months as timescales is quite arbitrary.
- How may the dense reservoir network found in your studied areas impact the propagation of meteorological droughts into hydrological droughts [see e.g. Van Langen et al. (2021) and Ribeiro Neto et al. (2022)]? How does your study help us to improve our understanding of this process?
Minor comments:
- Page 3, Lines 91-92: Please, review the sentence: “The region is monitored by 153 reservoirs are monitored…”
- Page 3, Line 94: “Consequently, the watersheds of the three main monitored reservoirs in this region were selected…” Why “consequently”?
- Page 4, Line 99: “The Orós Dam plays a crucial as a water source…” Do you mean “a crucial role”?
- Page 4, Lines: 102-104: “These reservoirs were selected due to their strategic significance to the state of Ceará, as they supply water for various purposes including drinking, agriculture, livestock, and other economic activities across multiple municipalities in the state (Gonçalves et al., 2023).” Please provide a more detailed description of water uses of these reservoirs, highlighting “numbers”
- Page 4, Line 106: “High evaporation rates…” How much?
- Fig. 1: Municipal boundaries are not visible, but I think they are not necessary; Change the color of Pluviometric stations on the map, maybe black; and have a look at the decimal separation of DEM limits.
- Sometimes you use “Dam”, and sometimes you use “Reservoir”, I think you always mean “Reservoir”.
References
van Langen, S. C. H., Costa, A. C., Ribeiro Neto, G. G., & van Oel, P. R. (2021). Effect of a reservoir network on drought propagation in a semi-arid catchment in Brazil. Hydrological Sciences Journal, 66(10), 1567–1583. https://doi.org/10.1080/02626667.2021.1955891
Ribeiro Neto, G. G., Melsen, L. A., Martins, E. S. P. R., Walker, D. W., & van Oel, P. R. (2022). Drought Cycle Analysis to evaluate the influence of a Dense Network of small reservoirs on drought evolution. Water Resources Research, 58, e2021WR030799. https://doi.org/10.1029/2021WR030799
Citation: https://doi.org/10.5194/egusphere-2024-813-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 585 | 235 | 32 | 852 | 29 | 61 |
- HTML: 585
- PDF: 235
- XML: 32
- Total: 852
- BibTeX: 29
- EndNote: 61
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Review of Medeiros and Silva (2024).
The manuscript by Medeiros and Silva investigates drought propagation (the transition from meteorological drought to hydrological drought) in the Brazilian semi-arid region. The authors use two different methods (SPI and SPEI) for defining meteorological drought and one method (ATLM) for defining hydrological drought, the latter using reservoir levels as the main variable.
I think the paper reads well, and the figures and analysis are well made and clear. My main issue is the lack of fundamental scientific advancement (what is the real objective of this paper besides combining different methods and comparing their results) or even an in-depth discussion.
I am surprised that the authors found out that (as per the abstract) the “Analysis of drought propagation, employing different methods, revealed no defined pattern for the onset peak, and end intervals across all tested combinations. Hence, we recommended utilizing all three indicators to enable a more comprehensive analysis of drought events propagated within the watersheds.”. At the same time, the authors acknowledge that such findings were expected according to recent literature.
As it is right now the paper reads as a report/case study and lacks a clear goal besides simply applying the method. I recommend rejection with the encouragement to resubmit, so that the authors have time to improve the manuscript. I also suggest breaking the results and discussion into two different sections so as to help the reader clearly understand the advances proposed in this paper.
Detailed comments:
L26: Isn’t this statement too broad? Is this happening everywhere? Is water supply affected only by extreme events? Are all extreme events tied to a reduction in water supply? Maybe be a bit more specific here.
L28: In the abstract there are different definition of droughts and already here drought is defined as an availability issue. But in the next paragraph you bring 4 other definitions. I am a bit confused.
L32: Please expand this paragraph to explain briefly what each of these definitions mean. The following paragraphs all rely on these important definitions. Your study uses reservoirs as your target variable to compute hydrologic drought but is this the only way of estimating hydrologic drought? Please expand how hydrologic drought is inferred and connect it to your study (i.e. hydrology as reservoirs levels).
L37: Below average precipitation impairs soil moisture retention reads a bit too complex to something maybe simpler: reduction in soil moisture.
L68-74: I think this paragraph needs some modification. You’re starting it by talking about the importance of incorporating land use and demand in better quantifying hydrologic drought. Ok. But then you’re jumping on “instead of …” and talk about hydrologic indices that use inflows. I just can’t follow the logical connections between the first and second half of this paragraph. Additionally, TLM is barely defined in the context of your introduction, what it does and the differences between it and other methods. This is making this paragraph very confusing, especially because you’re moving toward defining the goals of the paper right after.
L76: I don’t think “therefore” fits here. This sentence has no connection with the previous paragraph.
*Why two different paragraphs? I think the paragraphs from L80-85 are just a continuation of L76-70, so both should go together.
*What is the main advantage of your method? Why are you proposing this method? Your introduction leaves too much room for vagueness. I can seem to follow why the reader should care. Try to emphasize what type of advantage/discovery/improvement comes from what you are proposing. As of now, the introduction reads more as a summary of what you have done, than a proper “sales pitch”.
L91: Try to consolidate all climate information in one paragraph. Here you mention semi-arid and two paragraphs below you describe the climate.
L92: Collapsed? Maybe went dry? When I read collapse, it reads as damage to the structure.
L92-93: Try to be succinct with the description of the watersheds. Is the information on the 2012 drought necessary here? Maybe you need a paragraph in the introduction talking about the relevance of these reservoirs for the research area.
L105: Is hot really an adjective used to classify climate? I just have never seen this.
L116: How was evaporation measured/estimated to obtain the climatological normal values?
L130: Please elaborate on what the seminar means. I don’t think I’ve heard this concept before so maybe other readers will have the same question. Additionally, are you referring to demands or withdrawals?
L158: I think the authors need to better explore in the introduction the reason why they’re using SPI and SPEI together. Is this approach common? What do you intend to gain from it? This ties back to my comment from a few paragraphs ago. Why are the authors proposing their method as it is?
L171: Losses due to demands or withdrawals? Please clarify if demands or withdrawals were considered, and provide more details on the “seminar”.
L465: If it is already known in the literature that the transition from meteorological drought to hydrological drought is often site specific, what does your study contribute to the topic?