Development of the Niger Basin Drought Monitor (NBDM) for Early Warning and Concurrent Tracking of Meteorological, Agricultural and Hydrological Droughts

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript develops an objective-weighted, multi-indicator framework (NBDM/DREM) for concurrent meteorological, agricultural, and hydrological droughts in the Niger Basin, extending an SPI-based system to a composite drought index (CDI). The topic is well suited to West Africa and has clear regional utility.

To strengthen the paper:

1.Clarify the core novelty relative to USDM/NADM/Chinese CDI/EDO: is it primarily the ADFM/VIC reanalysis–driven indicators, the community-vulnerability weights (0.278/0.322/0.400), or the localization of thresholds?

2.Add a brief section on “Limitations of bias correction and temporal transferability,” and include a small station-level table with MAE, NSE, and PBIAS before/after correction, since 1980–2001 factors are transferred to 2002–2016 despite shifts in seasonality.

3.In Related Work/Discussion, add a short comparison with recent precipitation/hydrology reconstruction studies using regional-scale intelligent optimization and topography-aware, multi-source correction frameworks (e.g., work reported in Communications Earth & Environment), to better position the approach as a regionalized implementation rather than an isolated attempt.

4.Justify why CDI thresholds anchored to the 1980s event were not jointly calibrated or cross-validated with more recent droughts (e.g., 2010–2013, 2023–2024).

5.Undertake a systematic language edit: shorten long sentences, remove repetition, and move external URLs to references or supplementary materials.

Author Response

1. Clarify the core novelty relative to USDM/NADM/Chinese CDI/EDO: is it primarily the ADFM/VIC reanalysis–driven indicators, the community-vulnerability weights (0.278/0.322/0.400), or the localization of thresholds? 

   Authors’ Response

   Yes, DREM model was developed using ADFM/VIC reanalysis–driven indicators. Nevertheless, it can also be driven by In-situ observations if required datasets are available. To accurately and effectively detect drought onset, the drought thresholds for defining it must be both impact and location specific. This is a well-known fact in several literatures. This underscored the reason for the community vulnerability weighting process and assigning of weights of 0.278/0.322/0.400 applied in developing the model to localize the thresholds.

   Also, a comparative analysis of outputs of DREM model relative to USDM and Chinese Modified CDI was carried out. To achieve this, we standardized the three indices first. Results showed that DREM needs lesser accumulation of precipitation deficit to detect an onset of drought relative other composite indices. Hence, it detects drought much earlier than other CDIs.

2. Add a brief section on “Limitations of bias correction and temporal transferability,” and include a small station-level table with MAE, NSE, and PBIAS before/after correction, since 1980–2001 factors are transferred to 2002–2016 despite shifts in seasonality. 

   Authors’ Response

         This concern has been addressed as shown in the manuscript, line 360 to 372.

3. In Related Work/Discussion, add a short comparison with recent precipitation/hydrology reconstruction studies using regional-scale intelligent optimization and topography-aware, multi-source correction frameworks (e.g., work reported in Communications Earth & Environment), to better position the approach as a regionalized implementation rather than an isolated attempt. Response:

   Authors’ Response

   The concern is noted, though, that it is currently beyond the scope of this study. We will try to consider it in our future study, because the scope of the research is wide and being handled in phases. Hence, the authors try not to deviate from the main goal of the research.

4. 1. Justify why CDI thresholds anchored to the 1980s event were not jointly calibrated or cross-validated with more recent droughts (e.g., 2010–2013, 2023–2024). 

      Authors’ Response

      West Africa experienced a back-to-back drought event both in the 1970s and the 1980s, which produced major humanitarian crises across the Sahel, but they differ in timing, cumulative severity, and socio-economic consequences. The 1980s droughts, occurring after a decade of poor rains, represented a more prolonged, cumulative, and in many places more severe phase, amplifying livelihood depletion, migration, and long-term economic setbacks. The colossal losses in terms of humanitarian and socio-economic impacts, as well as environmental and ecological impacts, were unfathomable relative to 1970s.  As mentioned in lines 550 to 554 of the manuscript, the 1980s drought episodes induced severe famine in the region. This justified why the CDI threshold anchored to the 1980s drought events. Since then, the region has never experienced such megadrought events. The next phase and focus of our future work are to extend the scope of the research to 2024.

    

5. Undertake a systematic language edit: shorten long sentences, remove repetition, and move external URLs to references or supplementary materials. Response: 

   Authors’ Response

   Reviewer’s concern has been addressed.

    

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript presents the development of the Niger Basin Drought Monitor (NBDM), a hybrid drought resilience empirical model designed to integrate meteorological, agricultural and hydrological indicators for concurrent drought monitoring and early warning in the Niger River Basin.

 

Overall, the topic fits within the scope of Meteorology and the special issue on Special Issues “Early Career Scientists’ (ECS) Contributions to Meteorology (2025)” (https://www.mdpi.com/journal/meteorology/special_issues/0IFZ7EFZOC).

 

However, the research is methodologically thorough but has limited innovation. Most importantly, almost all the figures and overall presentation require significant improvement to meet publication standards. Therefore, major revisions are recommended.

 

Major issues:

 

First of all, I have some concerns regarding the Introduction section. The Introduction requires substantial improvement, in my opinion. Currently, the authors use 16 paragraphs (see Line 35-185) to "show" the background, which is excessive and makes it difficult for readers to identify the core research gaps. The Introduction should be revised to emphasize the specific limitations of existing Niger Basin drought monitoring approaches and the opportunities presented by recent technological advancements.

 

Then, the Results and Discussion section should be separated into distinct sections, which I strongly suggest! The Results section should focus solely on the findings from the developed methodology, while the Discussion should compare these results with previous studies, showing the strengths and limitations of the current work.

 

Last, I have another concern about percentage match. According to the iThenticate report, the percentage match is high. Although the authors have reduced it from 37% to 24%, it still exceeds the journal’s requirements. Further revisions are necessary.

 

 

Minor issues:

Below are a few specific comments that might be helpful:

 

1. Line 24: the acronym SPI is used without defined. Please provide its full form.
2. Figure 1 (Lines 52-55): several elements lack explanation, such as the meaning of “m”. The figure should also include a scale bar and a north arrow to improve interpretability.
3. Lines 188-189: “12oW and 15oE, and latitude 4o and 17oN”. A typo.
4. Line 189: Here, the claim that the basin covers “7.5 % of the continent of Africa” requires a supporting reference.
5. Line 196: “km2”. A typo.
6. Line 199: Figure 2 is of poor quality and appears to be a scanned image; it should be replaced with a digitally produced, high-resolution version.
7. Figures 2 and 3 are closely related and should be merged to improve readability and conserve space.
8. Table 1: the source AFDM website should include the full URL for accessibility.
9. Also Table 1: The spatial resolution in Table 1 is listed as “0.25” without units; it should be specified as “0.25° × 0.25°”.
10. Figure 4: It shows station distribution, would be more effectively presented as a table rather than a bar chart.
11. Line 262: “Bias Percent (PBias)” is incorrect; the standard term is “Percent Bias (PBIAS)”.
12. Line 270: The abbreviations “PDF” and “CDF” are not defined upon first use.
13. Figure 5 (Line 296): It has layout and scaling issues, and vertical text should be avoided for better readability.
14. Figure 6 (Line 309): It has scaling problems, and the figure caption does not explain the meaning of the different colors used.
15. Lines 355-356: the authors mention six (6) modules, but only two are described in detail in Lines 360-369. As for me, all modules should be thoroughly explained.
16. Figure 7: It does not fully in line with the module descriptions in Lines 366-369, particularly regarding the PET computation component.
17. Figures 8 and 9 do not contribute meaningfully to the methodology or results and should be removed.
18. Figure 12: the color gradation appears auto-generated and should be manually adjusted to better represent spatial patterns.
19. Line 513: “3.0” is a typo. It should be corrected to “3”.
20. Section 3.7, on calibration and validation, is placed at the end of the Results but should appear earlier, as validation typically precedes detailed results analysis, in my opinion.
21. Line 750-771: The Conclusions section should include quantitative statements summarizing key findings, such as performance metrics or validation success rates, to strengthen the take-home messages.
22. Lack of references cited in the past five years.

Author Response

Comment1: the research is methodologically thorough but has limited innovation. Most importantly, almost all the figures and overall presentations require significant improvement to meet publication standards. 

Authors’ Response

Concern has been addressed and reflected on the manuscript.  The figures’ sharpness and quality improved upon.

2. Comment 2: 

2. I have some concerns regarding the Introduction section. The Introduction requires substantial improvement, in my opinion. Currently, the authors use 16 paragraphs (see Line 35-185) to "show" the background, which is excessive and makes it difficult for readers to identify the core research gaps. The Introduction should be revised to emphasize the specific limitations of existing Niger Basin drought monitoring approaches and the opportunities presented by recent technological advancements Response: 

   Authors’ Response

   The highlighted concerns have been addressed and the introduction section revised accordingly.

3. Last, I have another concern about percentage match. According to the iThenticate report, the percentage match is high. Although the authors have reduced it from 37% to 24%, it still exceeds the journal’s requirements. Further revisions are necessary.  

   Authors’ Response

   The concerns have been addressed, and all relevant references cited accordingly

4. 1. Reviewer Comment: Line 24: the acronym SPI is used without defined. Please provide its full form.

   Authors’ Response: The acronym SPI has been defined as Standardized Precipitation Index (SPI)

5.  Reviewer Comment: Figure 1 (Lines 52-55): several elements lack explanation, such as the meaning of “m”. The figure should also include a scale bar and a north arrow to improve interpretability.

Authors’ response: Figure 1 shows the distribution of drought affected vulnerable population (in millions) between 2010 and 2012 in West Africa and the source of the figure is FAO, 2012. Since the figure was not an analysis carried out by the authors and did not have the original data, we therefore cannot modify it to add the scale bar and north arrow.

6.

3. Reviewer Comment: Lines 188-189: “12oW and 15oE, and latitude 4o and 17oN”. A typo.

Authors’ response: This has been corrected accordingly as longitudes 12^oW and 15^oE, and latitude 4^o and 17^oN

4. Reviewer Comment: Line 189: Here, the claim that the basin covers “7.5 % of the continent of Africa” requires a supporting reference.

Authors’ response: The supporting references to claim that Niger Basin covers 7.5% of the continent of Africa has been provided (i.e., Agricultural Water Use Final Report, 2003).

5. Reviewer Comment: Line 196: “km2”. A typo.

Authors’ response: The typo error has been fixed i.e., active catchment area of 1.5 million km².

6. Reviewer Comment: Line 199: Figure 2 is of poor quality and appears to be a scanned image; it should be replaced with a digitally produced, high-resolution version.

Authors’ response:  The authors do not have access to the original data of Figure 2. The source of the figure is Tarhule et al, 2014. Nevertheless, the sharpness of the image has been improved

7. Reviewer Comment: Figures 2 and 3 are closely related and should be merged to improve readability and conserve space

Authors’ response: Since the authors do not have access to the original dataset of Figure 2, it is difficult to merge figure 2 and 3 together into one image.

8. Reviewer Comment: Table 1: the source AFDM website should include the full URL for accessibility

Authors’ response: The source of the AFDM reanalysis dataset URL was provided in line 244 as http://stream.princeton.edu/AWCM/WEBPAGE/interface.php.

9. Reviewer Comment: Also, Table 1: The spatial resolution in Table 1 is listed as “0.25” without units; it should be specified as “0.25° × 0.25°”.

Authors’ response: The authors have provided the unit of the spatial resolution in Table 1, which is in degrees.

10. Reviewer Comment: Figure 4: It shows station distribution would be more effectively presented as a table rather than a bar chart.

Authors’ response: The authors chose to use bar charts because it allows quick visual comparison between different categories or groups than use of Table. Moreover, bar charts visually display differences between categories using bars, making it easier for people to quickly see which values are higher or lower. Also, visuals like bar charts are more attractive and hold attention better than tables filled with numbers.

11. Reviewer Comment: Line 262: “Bias Percent (PBias)” is incorrect; the standard term is “Percent Bias (PBIAS)”.

Authors’ response: The error has been fixed as Percent Bias (PBIAS)

12. Reviewer Comment: Line 270: The abbreviations “PDF” and “CDF” are not defined upon first use.

Authors’ response: The authors have provided definitions for “PDF” and “CDF”. The probability distribution functions (PDF) is broadly referred to as any function that describes how probabilities are distributed over possible outcomes of a random variable; while Cumulative Distribution Function (CDF) shows the total probability accumulated up to a certain value of a random variable or an integral of PDF from negative infinity up to point x

13. Reviewer Comment: Figure 5 (Line 296): It has layout and scaling issues, and vertical text should be avoided for better readability.

Authors’ response: Figure 5 is a schematic representation of the conceptual framework of development of different types of droughts, which is normally, a simplified visual diagram used to explain or illustrate relationships, processes, or concepts in a study or analysis. It does not necessarily show every detail but focuses on the main components and how they connect logically or functionally. This is exactly what Figure 5 tries to demonstrate. Schematic representations are commonly allowed or used in scientific papers to show conceptual frameworks, models, or systems etc.  Hence, we do not think there is layout and scaling issues as mentioned by the reviewer

14. Reviewer Comment: Figure 6 (Line 309): It has scaling problems, and the figure caption does not explain the meaning of the different colors used.

Authors’ response: Like Figure 5, Figure 6 is a schematic representation of the working of the linear combination model used to integrate the different independent predictors indices highlighted in green color, the new linear predictor highlighted in orange color and the dependent predictand highlighted in blue color.

15. Reviewer Comment: Lines 355-356: the authors mention six (6) modules, but only two are described in detail in Lines 360-369. As for me, all modules should be thoroughly explained

Authors’ response: They were seven (7) modules, not six (6) as earlier mentioned and were described, but their numbering was distorted. This has been corrected accordingly as highlighted in the manuscript.

16. Reviewer Comment: Figure 7: It does not fully in line with the module descriptions in Lines 366-369, particularly regarding the PET computation component.

Authors’ response: The effective precipitation module (EP) in Figure 7, initially omitted has been added in the description. To compute the EP, the PET has to be computed first.

17. Reviewer Comment: Figures 8 and 9 do not contribute meaningfully to the methodology or results and should be removed.

Authors’ response: Figure 8 has a meaningful contribution in determining the characteristics of drought such as the onset, duration, severity, frequency and termination of drought a particular drought event.  With regards to Figure 9, the reviewer needs to understand that the importance of the user login interface in any software cannot be over-emphasized. The Login user interface is for inputting the user’s ID to ensure that user’s information and work done is secure. Hence, it has meaningful contribution in this study.

18. Reviewer Comment: Figure 12: the color gradation appears auto-generated and should be manually adjusted to better represent spatial patterns.

Authors’ response: The Figure 12 is a result of Arcmap GIS analysis. The spatial analysis was meant for easy identification of drought onset threshold at any location within the basin. GIS tools show or represent better spatial pattern than manual analysis as suggested by the reviewer.

19. Reviewer Comment: Line 513: “3.0” is a typo. It should be corrected to “3”.

Authors’ response: The numbering of the Results and Discussion section has been corrected as 3. and not 3.0

20. Reviewer Comment: Section 3.7, on calibration and validation, is placed at the end of the Results but should appear earlier, as validation typically precedes detailed results analysis, in my opinion.

Authors’ response: It is important the reviewer understands the authors’ concept of calibration and validation in this study. The authors defined validation as the measure of the agreement between CDI-defined droughts and historical events, expressed as hit/success rates. Hence, we needed to state and discuss the DREM-CDI model outputs first, then, followed by calibration and validation. Also, as stated in lines 176 to 182, this research addresses certain questions or concerns in the region, amongst which are how can decision-makers in the Niger Basin be aided to make holistic drought management decisions that can benefit relevant socio-economic sectors or diverse audience?  How reliable is the information from the system support tool used in making such drought management decisions that can profit diverse audience? Calibration and validation analysis were carried out to address the above second questions. This explained why they were placed at the end of the result, after the first question was addressed. So, the placement of the calibration validation analysis at the end of the results fit the narrative well.

21. Reviewer Comment: Line 750-771: The Conclusions section should include quantitative statements summarizing key findings, such as performance metrics or validation success rates, to strengthen the take-home messages.

Authors’ response: In the conclusion section, the key findings of the research were summarized. The fourth item has already addressed the above concerns. For instance, it was reported that based on the degree of success rate, the validation results revealed a success rate of range 67 to 100% based on past records of drought events captured by DREM CDI, and 62 to 77% based on ENSO-induced drought records captured by the CDI.

22. Reviewer Comment: Lack of references cited in the past five years.

Authors’ response

More recent relevant journal publications of the past five years have been cited

 

2.  

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript aims to provide a trial of the DREM model at the Niger basin for the early detection of drought. The writing is generally acceptable, but serious restructuring is required. The reviewer has the following comments:

1) Figure 1 can be easily converted into a table or remade as a graph. The current screenshot from the external source is not clear enough and contains many small texts.

2) Lines 68-69: Please cite web resources using the correct format. Please note that there are multiple similar incidents throughout the manuscript.

3) Line 72: Please use the correct format to cite a specific paper in text. Please note that there are multiple similar incidents throughout the manuscript.

4) Line 133: Please state the cause for the paradigm shift. There must be some flaws in the conventional approaches.

5) Lines 165-175: Please provide a background of the application of multivariate drought analysis in West Africa before introducing DREM. There must be a history of applications before the authors decide to use DREM. Actually, the manuscript contains more than just DREM. Please establish well what you intend to do.

6) Figure 2: The quality of this figure is quite low. Texts cannot be recognized, and colors are not clean. Coordinates are not present either. The graph should be easy to remake using GIS without using screenshots from an external source.

7) Lines 188-189: Please make sure degrees are superscripted.

8) Please provide some climate norm data in Section 2.1.

9) Figure 3: texts not recognizable

10) Figures 5, 6: Resolution needs improvement.

11) Section 2.4: You utilized multiple models in this section, and it can be confusing. Please provide a short description of the models that will be used in this section and the relationship between them.

12) If DREM is implemented by software, basic information about the software, such as the manufacturer, must be stated.

13) Figures 12, 14: Please remake the figure so texts are recognizable.

14) Figure 12: The term “rainfall climatology” sounds strange. Why not just “rainfall”? The rainfall of which year?

15) Line 553: Is it “D1” (D-one) or “Dl” (D-L)? In the graph, it is written as D1 (D-one)?

16) Some of the indices were not mentioned before, like CDI? They should be described early in the literature review and methodology, not in the results section.

17) Line 659: Please break the sentence before “because” and delete “because” in the new sentence. Please also check the writing in the whole manuscript using the tools available today (such as Grammarly).

18) Section 3.7: Where is the calibration?

Author Response

1) Figure 1 can be easily converted into a table or remade as a graph. The current screenshot from the external source is not clear enough and contains many small texts.

Authors’ response

Figure I was sources from a literature, which as cited in the manuscript. The authors do not have access to the data set used in the analysis.

2) Lines 68-69: Please cite web resources using the correct format. Please note that there are multiple similar incidents throughout the manuscript.

Authors’ response

The concern has been addressed (Please, see manuscript)

3) Line 72: Please use the correct format to cite a specific paper in text. Please note that there are multiple similar incidents throughout the manuscript.

Authors’ response

The concern has been addressed (Please, see manuscript)

4) Line 133: Please state the cause for the paradigm shift. There must be some flaws in the conventional approaches.

Authors’ response

The concern has been addressed (please, see manuscript)

5) Lines 165-175: Please provide a background of the application of multivariate drought analysis in West Africa before introducing DREM. There must be a history of applications before the authors decide to use DREM. Actually, the manuscript contains more than just DREM. Please establish well what you intend to do.

Authors’ response

The concern has been addressed (Please, see manuscript)

 

6) Figure 2: The quality of this figure is quite low. Texts cannot be recognized, and colors are not clean. Coordinates are not present either. The graph should be easy to remake using GIS without using screenshots from an external source.

Authors’ response

The concern has been addressed and the quality of Figure 2 improved in terms of sharpness and quality (Please, see manuscript). Note also that Figure 2 was sourced from literature and author cited according. We do not have access to the original datasets

7) Lines 188-189: Please make sure degrees are superscripted.

Authors’ response

The concern has been addressed (Please, see manuscript)

8) Please provide some climate norm data in Section 2.1.

Authors’ response

The concern has been addressed (Please, see manuscript)

9) Figure 3: texts not recognizable

Authors’ response

The concern has been addressed (Please, see manuscript)

10) Figures 5, 6: Resolution needs improvement.

Authors’ response

The concern has been addressed (Please, see manuscript)

11) Section 2.4: You utilized multiple models in this section, and it can be confusing. Please provide a short description of the models that will be used in this section and the relationship between them.

Authors’ response

The concern has been addressed (Please, see manuscript)

12) If DREM is implemented by software, basic information about the software, such as the manufacturer, must be stated.

Authors’ response

The concern has been addressed (Please, see manuscript)

1 Authors’ response

The concern has been addressed (Please, see manuscript)

3) Figures 12, 14: Please remake the figure so texts are recognizable.

Authors’ response

The concern has been addressed (Please, see manuscript)

14) Figure 12: The term “rainfall climatology” sounds strange. Why not just “rainfall”? The rainfall of which year?

Authors’ response

The concern has been addressed (Please, see manuscript). We are referring to rainfall of over 3o years

15) Line 553: Is it “D1” (D-one) or “Dl” (D-L)? In the graph, it is written as D1 (D-one)?

Authors’ response

The concern has been addressed (Please, see manuscript). It is “D1” (D-one) , not “Dl” (D-L)

16) Some of the indices were not mentioned before, like CDI. They should be described early in the literature review and methodology, not in the results section.

Authors’ response

The concern has been addressed (Please, see manuscript)

17) Line 659: Please break the sentence before “because” and delete “because” in the new sentence. Please also check the writing in the whole manuscript using the tools available today (such as Grammarly).

Authors’ response

The concern has been addressed (Please, see manuscript)

18) Section 3.7: Where is the calibration?

Authors’ response

The concern has been addressed (Please, see manuscript). The section has been corrected to read as Validation of Performance of NBDM-CDI

 

Author Response File: Author Response.docx

Reviewer 4 Report

Comments and Suggestions for Authors

Dear Authors,

The manuscript entitled “Development of the Niger Basin Drought Monitor (NBDM) for Early Warning and Concurrent Tracking of Meteorological, Agricultural, and Hydrological Droughts” offers a fascinating contribution to the topic of drought monitoring in the Niger Basin. This topic is of high relevancy considering that drought events in West Africa are increasingly common, thereby making this topic a high priority in today’s scientific research needs. It is laudable that the authors sought to develop a model that incorporates various drought indicators in a way that is in line with today’s high priorities on Climate Services.

It is apparent that the manuscript shows a distinct and well-organized structure in the methodological approach, specifically in the discussion of the empirical model of DREM and process of weighting various drought indicators. Moreover, the combination of various data sources, spanning from various datasets of reanalysis to in-situ data, is a positive point of this analysis that inherently leads to a higher quality of results. Nevertheless, the process of validation seems not extensive enough in terms of statistical analyses in order to appropriately document the efficiency and accuracy of the proposed system. It seems important to incorporate a more extensive analysis process, including comparison of model results with independent observational data as well as well-documented drought indicators like SPI, SPEI, and PDSI.

Firstly, in discussions of uncertainties associated with the use of reanalysis data as well as bias correction techniques, improvements can be proposed. More clarification is needed on limitations of data employed in the article and consideration of possible effects of the identified bias on model performance in respective sub-basins of the Niger River. By considering a sensitivity analysis, uncertainties in the article will be evaluated in terms of influence of identified parameters on drought thresholds as well as the composite drought index. Moreover, more clarification is needed in the choice of weighting techniques.

It is also important that the conceptual clarity of the model is improved. "Empirical model" is a term that could be misleading if care is not taken in defining it in relation to the conceptual components of the model. It is important that a clear distinction is made between those components that are empirically calibrated in relation to observed data and the conceptual framework on which various components of the hydrological cycle are linked.

It is recommended that the conclusion portion of the article needs to be revised in a way that specifically emphasizes the primary outcomes of the proposed methods, as well as a consideration of the limitations of the presented research. Claims of operational maturity need to be moderated if they cannot be sufficiently justified. It will also be valuable to discuss possible directions of future studies that could involve temporal scaling of the analysis, use of satellite vegetation layers or evapotranspiration products, as well as a test of the NBDM model in a variety of climatic situations to assess the model transferability. In sum, this article presents a promising and original contribution to the topic of integrated drought monitoring in West African regions. With important revisions that improve statistical verification, analysis of uncertainties, as well as conceptual specificity, this article could be a valuable contribution to the field of climatology as well as early warning systems in the Niger Basin.

Sincerely,

Author Response

Reviewer 4

1. The process of validation does not seem extensive enough in terms of statistical analyses to appropriately document the efficiency and accuracy of the proposed system. It seems important to incorporate a more extensive analysis process, including comparison of model results with independent observational data as well as well-documented drought indicators like SPI, SPEI, and PDSI.

Authors’ response

The concern has been addressed through the inclusion of result of boxplot analysis of the various in the independent observational data as shown in Figure 15 (Please, see manuscript)

2. Discussions of uncertainties associated with the use of reanalysis data as well as bias correction techniques, improvements can be proposed. More clarification is needed on limitations of data employed in the article and consideration of possible effects of the identified bias on model performance in respective sub-basins of the Niger River. By considering a sensitivity analysis, uncertainties in the article will be evaluated in terms of influence of identified parameters on drought thresholds as well as the composite drought index. Moreover, more clarification is needed in the choice of weighting techniques.

Authors’ response

Concerns have been addressed by providing more clarifications to unravel the uncertainties associated with the use of reanalysis data as well as bias correction techniques, improvements can be proposed. Also, more clarification has been provided concerning the limitations of reanalysis data employed in the article and consideration of possible effects of the identified bias on model performance in respective sub-basins (See, manuscript).  The suggestion to consider sensitivity analysis is well noted and will be considered in the next phase of our study. The authors agree with reviewer that more clarification is needed in the choice of weighting techniques.

The weighting techniques were selected because drought thresholds must be impact-specific and location-specific, as no universal operational drought definition exists. To address this, we applied the concept of Burden of Drought Disaster (BDD) to estimate the relative importance of the three biophysical drought types, namely, meteorological, agricultural, and hydrological drought experienced in the study region during the 1980s drought-induced famine, being a worst-case scenario of drought episodes in the region. This approach allowed us to identify which drought type imposed the greatest societal burden. Our framework adapts the World Health Organization’s (WHO) Disability-Adjusted Life Years (DALYs) concept, which measures the burden of disease disasters (BDD) on society, to guide the weighting of drought impacts in the empirical model.

3. It is also important that the conceptual clarity of the model is improved. "Empirical model" is a term that could be misleading if care is not taken in defining it in relation to the conceptual components of the model. It is important that a clear distinction is made between those components that are empirically calibrated in relation to observed data and the conceptual framework on which various components of the hydrological cycle are linked.

Authors’ response

The authors are aware of this, because each of the independent input indicators is a model itself. Hence, in conceptualizing DREM model framework the authors were conscious of the presence such complexities in designing DREM.

It is recommended that the conclusion portion of the article needs to be revised in a way that specifically emphasizes the primary outcomes of the proposed methods, as well as a consideration of the limitations of the presented research. Claims of operational maturity need to be moderated if they cannot be sufficiently justified. It will also be valuable to discuss possible directions of future studies that could involve temporal scaling of the analysis, use of satellite vegetation layers or evapotranspiration products, as well as a test of the NBDM model in a variety of climatic situations to assess the model transferability. In sum, this article presents a promising and original contribution to the topic of integrated drought monitoring in West African regions. With important revisions that improve statistical verification, analysis of uncertainties, as well as conceptual specificity, this article could be a valuable contribution to the field of climatology as well as early warning systems in the Niger Basin.

Authors’ response

The reviewer’s concerns have been addressed, including revising of the conclusion section.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you for your thorough revisions. The manuscript has clearly improved in structure and clarity, and most of my earlier concerns have been satisfactorily addressed. However, one key issue remains.

Your drought monitor is fully driven by gridded precipitation and hydrological forcings (ADFM/VIC) and their bias correction, but the manuscript still does not position these forcings relative to recent regional, multi-source, topography-aware precipitation reconstruction work. This is important because uncertainty in the forcings directly propagates into drought detection and early warning, and readers need to see that you are aware of current approaches to reducing such data uncertainty.

Therefore, I still recommend that you briefly discuss and cite a representative study such as the Communications Earth & Environment paper on regional-scale intelligent optimization and topography-informed multi-source precipitation correction. That work explicitly targets uncertainty reduction in satellite/reanalysis precipitation fields in data-scarce, complex regions and is conceptually close and complementary to your framework: their method improves input data quality, while your DREM/NBDM translates those (uncertain) forcings into impact-based drought indicators. A short paragraph in Related Work or Discussion noting this complementarity and citing that paper would fully resolve my remaining concern.

Author Response

Your drought monitor is fully driven by gridded precipitation and hydrological forcings (ADFM/VIC) and their bias correction, but the manuscript still does not position these forcings relative to recent regional, multi-source, topography-aware precipitation reconstruction work. This is important because uncertainty in the forcings directly propagates into drought detection and early warning, and readers need to see that you are aware of current approaches to reducing such data uncertainty.

Therefore, I still recommend that you briefly discuss and cite a representative study such as the Communications Earth & Environment paper on regional-scale intelligent optimization and topography-informed multi-source precipitation correction. That work explicitly targets uncertainty reduction in satellite/reanalysis precipitation fields in data-scarce, complex regions and is conceptually close and complementary to your framework: their method improves input data quality, while your DREM/NBDM translates those (uncertain) forcings into impact-based drought indicators. A short paragraph in Related Work or Discussion noting this complementarity and citing that paper would fully resolve my remaining concern.

Authors Response: This concern has been addressed as shown in the manuscript in blue color, lines 283 to 306 and lines 330 – 340.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

All my concerns in the last round have been addressed. Also the references cited issue has been fixed as well.

Therefore, I recommend this manuscript be accepted for the special issue.

Author Response

Comments: Related to Figures and general concerns

Author’s Response: All my concerns have been addressed including improvement in the figures.  All new and additional information is in blue color letters in the manuscript.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

1) Figure 1: If the quality of figures from external sources is not good, the authors must remake the figure. If that is not possible, you need to take it down. Besides, the citation format for the figure is wrong, and the source format in "references" is missing key information.

2) Figure 2: Please utilize a GIS software to make this figure.

3) Many of the points that the authors claim to be corrected still remain uncorrected. For example, line 74 still contains "As [17] notes...", degrees are not superscripted at many places, tiny texts in Figures 3, 12, and 14, resolution issues in Figures 5-6, required citation for the software (section 3.1), "rainfall climatology" in Figure 12, grammar issues, and more.

Author Response

Reviewer 3 Comments:

1) Figure 1: If the quality of figures from external sources is not good, the authors must remake the figure. If that is not possible, you need to take it down. Besides, the citation format for the figure is wrong, and the source format in "references" is missing key information

Author’s Response: All my concerns have been addressed including improvement in figure1.  All new and additional information is in blue color letters in the manuscript.

2) Figure 2: Please utilize GIS software to make this figure.

Author’s Response: All my concerns have been addressed including improvement in figure1.

3) Many of the points that the authors claim to be corrected still remain uncorrected. For example, line 74 still contains "As [17] notes...", degrees are not superscripted at many places, tiny texts in Figures 3, 12, and 14, resolution issues in Figures 5-6, required citation for the software (section 3.1), "rainfall climatology" in Figure 12, grammar issues, and more.

Author’s Response: All my concerns have been addressed including improvement in figure1.  All new and additional information is in blue color letters in the manuscript.

With regards to the phrase or term "rainfall climatology. The authors assert that it is acceptable scientific phrase used in many scientific literature including the World Meteorological Organization (WMO)(Reviewer should see the citations on lines 677 – 681 of manuscript in blue color). Hence, we do not think or see the grammar issue there. Thanks for your understanding.

Author Response File: Author Response.docx

Reviewer 4 Report

Comments and Suggestions for Authors

Dear Authors,

In this new version of the manuscript, it is apparent that all the comments raised by the reviewers in the first stage of assessment have been carefully considered. Indeed, improvements achieved here are extensive, significantly increasing the scientific merits of the submitted work. A great improvement in the validation framework, which further enhances the credibility of the Niger Basin Drought Monitoring (NBDM) system as being able to provide a coherent presentation of the different aspects of drought conditions in the Niger Basin, is achieved through the use of independent datasets, valid drought variables, as well as sound performance measures. An improvement in the presentation of the uncertainties associated with reanalysis datasets, together with their correction, so that their effects on modeling can be clearly considered, also takes place. Finally, a better distinction between the empirical and conceptual aspects of the method, thus allowing easier reading of the methodological framework, as well as proper placement within this framework of the Drought Risk Evaluation Model (DREM) within the NBDM system, is achieved. Finally, establishing that the justification of the weights represents a further advance, as the introduction of a justification related to the economic burden associated with drought helps further increase the scientific merits of the composite indicator, through lending additional robustness. Also, improved is the conclusions paragraph, as realism, together with a more concise, realistic, and further-oriented presentation, is achieved, taking into account limitations as well as further directions. Overall, this reflects a serious effort aimed at addressing all comments raised initially. In any case, this new version of the manuscript is robust, well-organized, with high scientific merits, thus eligible for the required publications with some trifling text corrections.

Sincerely,

   

Author Response

Reviewer 4 Comments: Regards to further improvement in the quality of the paper

Author’s Response: All my concerns have been addressed including improvement in figure1.  All new and additional information is in blue color letters in the manuscript.

 

Author Response File: Author Response.docx