Enhanced Baseflow Separation in Rural Catchments: Event-Specific Calibration of Recursive Digital Filters with Tracer-Derived Data

Helfer, Fernanda; Bernardi, Felipe; de Barros, Claudia Alessandra Peixoto; Allasia, Daniel Gustavo; Minella, Jean Paolo Gomes; Tassi, Rutinéia; Scariot, Néverton

doi:https://doi.org/10.5194/egusphere-2025-244

Preprints

https://doi.org/10.5194/egusphere-2025-244

Preprints

26 Mar 2025

| 26 Mar 2025

Enhanced Baseflow Separation in Rural Catchments: Event-Specific Calibration of Recursive Digital Filters with Tracer-Derived Data

Fernanda Helfer, Felipe Bernardi, Claudia Alessandra Peixoto de Barros, Daniel Gustavo Allasia, Jean Paolo Gomes Minella, Rutinéia Tassi, and Néverton Scariot

Abstract. This study investigates the performance of baseflow separation methods in a small rural catchment, emphasizing the calibration of three Recursive Digital Filters (RDFs): Eckhardt, Lyne and Hollick (LH), and Chapman and Maxwell (CM). By integrating dissolved silica concentration as a reference tracer, the study refines the parameterization of BFI_max in the Eckhardt’s filter and Beta in the LH filter. An innovative event-specific calibration methodology was applied, where rainfall events were categorized by intensity to tailor filter parameters accordingly. Results indicate that the Eckhardt’s filter, when calibrated dynamically per event magnitude, yields the most accurate baseflow estimates, closely aligning with observed data. The event-based calibration significantly enhanced accuracy, particularly for the Eckhardt’s and LH filters, compared to a general calibration method. The CM filter, despite generating reasonable hydrograph shapes, consistently underestimated baseflow due to its fixed parameters. These findings highlight the necessity of customized calibration strategies for improved baseflow separation and underscore the superior performance of the Eckhardt’s filter when integrated with event-specific calibrations. This research offers practical insights for hydrologists aiming to optimize baseflow modeling in rural catchments, contributing to improved water resource management and conservation.

Received: 20 Jan 2025 – Discussion started: 26 Mar 2025

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Fernanda Helfer, Felipe Bernardi, Claudia Alessandra Peixoto de Barros, Daniel Gustavo Allasia, Jean Paolo Gomes Minella, Rutinéia Tassi, and Néverton Scariot

Status: closed

RC1:
'Comment on egusphere-2025-244', Anonymous Referee #1, 28 Apr 2025

General Comments:

This manuscript presents a well-executed and innovative study on baseflow separation in small rural catchments, focusing on the calibration of three Recursive Digital Filters (Eckhardt, Lyne and Hollick, and Chapman and Maxwell). The integration of dissolved silica as a tracer and the event-specific calibration approach are both novel and valuable contributions to the field.
The writing is clear, and engaging, making the technical content accessible and enjoyable to read. The study’s methodology is robust, and the results convincingly demonstrate the advantages of dynamic, event-based calibration-particularly for the Eckhardt filter, which outperforms the others in accuracy when parameters are tailored to rainfall event intensity.
Specific Comments:

While the manuscript is strong overall, there are areas where it could be further improved. The authors provide a thorough explanation of performance metrics such as NSE and RMSE, but this level of detail may be unnecessary for the target audience, who are likely already familiar with these standard evaluation tools. Streamlining these sections would help maintain the manuscript’s focus and momentum. The presentation of results relies heavily on numerical values in the text. Incorporating more plots and figures would greatly enhance the reader’s ability to interpret and appreciate the findings.
This is a timely contribution with clear novelty in its event-specific calibration strategy and use of chemical tracers. The manuscript is exceptionally well-written and methodologically sound. To maximize its impact, I recommend reducing the over-explanation of standard metrics and enhancing the presentation of results with more visual aids. These improvements would make the findings even more accessible and persuasive to a broad hydrology audience.

Citation: https://doi.org/10.5194/egusphere-2025-244-RC1
- AC1: 'Reply on RC1', Fernanda Helfer, 29 Apr 2025
  
  We thank the reviewer for their constructive feedback.
  Regarding the explanation of standard performance metrics (e.g., NSE and RMSE): We acknowledge the point that the target audience is likely familiar with these metrics. So, to streamline the manuscript and maintain its momentum, we will condense the descriptions of NSE, RMSE, and related metrics, with the focus shifting to their interpretation within the context of our event-specific calibration strategy.
  Regarding the reliance on numerical values in the results section: We agree that more visual aids would enhance clarity and reader engagement. We will revise the results section (section 3) to incorporate additional figures. Specifically, we will convert Tables 9 and 10 into comparative plots.
  
  Citation: https://doi.org/10.5194/egusphere-2025-244-AC1
RC2:
'Comment on egusphere-2025-244', Anonymous Referee #2, 05 May 2025

General Comments
The manuscript describes an analysis in which three baseflow separation methods (specifically types of recursive digital filters) are tested against a silica-based estimate of baseflow for a small watershed in Brazil. The authors compare RDFs fit against the entire dataset as well as RDFs fit to events falling into one of three magnitude categories. The analysis shows that the Eckhardt filter tends to provide better results across events and that the use of categorized parameters can provide some improvement in estimating baseflow across a range of event types.
The work presented is thorough and is explained with a commendable level of detail. The manuscript is organized nicely and is well written, making it easy to quickly understand what was done and the nature of the results. The authors provide a good explanation of the methods and the figures and tables support the interpretation and conclusions given.

One criticism I might offer is that this work is based on and potentially limited in relevance to a single (and comparatively small) watershed.

The manuscript would be more impactful if the value of the analysis is situated in a more generalized context - what about the results would be useful to a researcher or practitioner needing to understand or estimate baseflow somewhere else? I would suggest a potential expansion of the discussion and interpretation to address this broader context a bit more.
Specific and Technical Comments
- While the authors describe the physical characteristics of the watershed, a more specific explanation of the conceptual model of the watershed would be helpful in providing a basis for interpretation of the baseflow separation results. For example, it would be helpful to discuss what the Si data (or other previous work in the watershed) suggests about how baseflow contributes to streamflow. This can be used to provide context for the different methods and their interpretation too.
- What does the difference in performance from the tiered/categorized RDF parameter fits tell us about (potential) mechanisms in the watershed? This is an interesting result and I would appreciate a little more discussion on the topic.
- The description indicates the area receives relatively consistent precipitation, perhaps more than many other places in the world. This would suggest fairly constant GW levels and perhaps less variation in "old" water. Yet the results show a lot of variability in shape and timing of baseflow. I'm curious how this is to be interpreted, and whether the authors view this tier-type approach as something than can be useful in disentangling seasonal variation in groundwater-surface water connection? Again, some additional discussion addressing this would be appreciated
- Much of the information shown in error metrics tables is repetitive between the two sections of analysis and results. I'd suggest condensing some of this and/or moving some of it to an appendix/supplement.
- Table 1 (page 8) - Units for Qmax and Qmin are different. Check flow units for consistency.
- Figure 2 and Figure 5 - I appreciate the information density on these plots. However, the line widths on the RDF results (colored lines) came through rather faintly on the PDF I reviewed, making it difficult to readily differentiate among the results. Additionally the figure seemed a bit blurry. To the extent that the line widths and resolution could be adjusted, these figures will be much more effective.
- Line 558: "events 2,9, and 9 showed..." - I suspect the second "9" was in error here.

Citation: https://doi.org/10.5194/egusphere-2025-244-RC2
- AC2: 'Reply on RC2', Fernanda Helfer, 08 May 2025
  
  We sincerely thank the reviewer for their careful reading of our manuscript and for the constructive comments provided. We greatly appreciate the time and effort dedicated to improving the quality and clarity of our work. Below, we provide a brief response to each point raised. We will implement all suggested changes in the revised version of the manuscript.
  
  General Comment: Situating the findings in a broader context to increase the manuscript’s impact.
  
  Our Response: We will expand the Discussion and Conclusions sections to emphasize how the findings can inform baseflow estimation in other catchments beyond the study area.
  
  Specific and Technical Comments
  Conceptual model of the watershed: The reviewer requested a clearer conceptual model to support the interpretation of the results.
  
  Our Response: We will add a more detailed explanation of the catchment’s conceptual hydrological model, including insights from the silica data, to clarify how baseflow contributes to streamflow.
  Interpretation of the tiered RDF parameters: Discuss what the performance differences tell us about hydrological mechanisms.
  
  Our Response: We will add a discussion in Section 3.5 explaining how variations in BFImax and Beta reflect different flow generation processes and hydrological responses across event magnitudes.
  Seasonality and variability despite consistent precipitation: The reviewer requested more discussion on seasonal variation and the applicability of the tiered approach.
  
  Our Response: We will expand the Discussion to explain how seasonal groundwater dynamics and event-driven variations contribute to the observed variability, and how the event-based calibration approach can help disentangle these effects.
  Condensing repetitive tables: Move some repetitive error metric tables to the Supplementary Material.
  
  Our Response: We will attempt to condense the results in the main text and/or move detailed performance tables to the Supplementary Material to improve the manuscript’s readability.
  Units in Table 1: Inconsistency in flow units (Qmin and Qmax).
  Our Response: We will correct the units in Table 1 to ensure consistency.
  
  Figure 2 and Figure 5 – line widths and resolution: Faint lines and blurred figures.
  
  Our Response: We have reviewed printed versions of the manuscript and confirmed that all lines are clearly visible and the figures are not blurry. The faint appearance may have been due to the resolution of the version provided for review. Nevertheless, we will enhance line thickness and improve figure resolution in the revised manuscript to avoid any possible issues.
  Typographical error (Line 558): The reviewer pointed out a repeated event number ("9").
  
  Our Response: We will correct this typographical error.
  
  Citation: https://doi.org/10.5194/egusphere-2025-244-AC2
AC3: 'Comment on egusphere-2025-244', Fernanda Helfer, 13 May 2025

Dear referees and editor,
We have now implemented all the suggestions provided by the referees. Find below our final point-by-point reply document. We look forward to submitting the revised version of the manuscript.

Citation: https://doi.org/10.5194/egusphere-2025-244-AC3

Status: closed

RC1:
'Comment on egusphere-2025-244', Anonymous Referee #1, 28 Apr 2025

General Comments:

This manuscript presents a well-executed and innovative study on baseflow separation in small rural catchments, focusing on the calibration of three Recursive Digital Filters (Eckhardt, Lyne and Hollick, and Chapman and Maxwell). The integration of dissolved silica as a tracer and the event-specific calibration approach are both novel and valuable contributions to the field.
The writing is clear, and engaging, making the technical content accessible and enjoyable to read. The study’s methodology is robust, and the results convincingly demonstrate the advantages of dynamic, event-based calibration-particularly for the Eckhardt filter, which outperforms the others in accuracy when parameters are tailored to rainfall event intensity.
Specific Comments:

While the manuscript is strong overall, there are areas where it could be further improved. The authors provide a thorough explanation of performance metrics such as NSE and RMSE, but this level of detail may be unnecessary for the target audience, who are likely already familiar with these standard evaluation tools. Streamlining these sections would help maintain the manuscript’s focus and momentum. The presentation of results relies heavily on numerical values in the text. Incorporating more plots and figures would greatly enhance the reader’s ability to interpret and appreciate the findings.
This is a timely contribution with clear novelty in its event-specific calibration strategy and use of chemical tracers. The manuscript is exceptionally well-written and methodologically sound. To maximize its impact, I recommend reducing the over-explanation of standard metrics and enhancing the presentation of results with more visual aids. These improvements would make the findings even more accessible and persuasive to a broad hydrology audience.

Citation: https://doi.org/10.5194/egusphere-2025-244-RC1
- AC1: 'Reply on RC1', Fernanda Helfer, 29 Apr 2025
  
  We thank the reviewer for their constructive feedback.
  Regarding the explanation of standard performance metrics (e.g., NSE and RMSE): We acknowledge the point that the target audience is likely familiar with these metrics. So, to streamline the manuscript and maintain its momentum, we will condense the descriptions of NSE, RMSE, and related metrics, with the focus shifting to their interpretation within the context of our event-specific calibration strategy.
  Regarding the reliance on numerical values in the results section: We agree that more visual aids would enhance clarity and reader engagement. We will revise the results section (section 3) to incorporate additional figures. Specifically, we will convert Tables 9 and 10 into comparative plots.
  
  Citation: https://doi.org/10.5194/egusphere-2025-244-AC1
RC2:
'Comment on egusphere-2025-244', Anonymous Referee #2, 05 May 2025

General Comments
The manuscript describes an analysis in which three baseflow separation methods (specifically types of recursive digital filters) are tested against a silica-based estimate of baseflow for a small watershed in Brazil. The authors compare RDFs fit against the entire dataset as well as RDFs fit to events falling into one of three magnitude categories. The analysis shows that the Eckhardt filter tends to provide better results across events and that the use of categorized parameters can provide some improvement in estimating baseflow across a range of event types.
The work presented is thorough and is explained with a commendable level of detail. The manuscript is organized nicely and is well written, making it easy to quickly understand what was done and the nature of the results. The authors provide a good explanation of the methods and the figures and tables support the interpretation and conclusions given.

One criticism I might offer is that this work is based on and potentially limited in relevance to a single (and comparatively small) watershed.

The manuscript would be more impactful if the value of the analysis is situated in a more generalized context - what about the results would be useful to a researcher or practitioner needing to understand or estimate baseflow somewhere else? I would suggest a potential expansion of the discussion and interpretation to address this broader context a bit more.
Specific and Technical Comments
- While the authors describe the physical characteristics of the watershed, a more specific explanation of the conceptual model of the watershed would be helpful in providing a basis for interpretation of the baseflow separation results. For example, it would be helpful to discuss what the Si data (or other previous work in the watershed) suggests about how baseflow contributes to streamflow. This can be used to provide context for the different methods and their interpretation too.
- What does the difference in performance from the tiered/categorized RDF parameter fits tell us about (potential) mechanisms in the watershed? This is an interesting result and I would appreciate a little more discussion on the topic.
- The description indicates the area receives relatively consistent precipitation, perhaps more than many other places in the world. This would suggest fairly constant GW levels and perhaps less variation in "old" water. Yet the results show a lot of variability in shape and timing of baseflow. I'm curious how this is to be interpreted, and whether the authors view this tier-type approach as something than can be useful in disentangling seasonal variation in groundwater-surface water connection? Again, some additional discussion addressing this would be appreciated
- Much of the information shown in error metrics tables is repetitive between the two sections of analysis and results. I'd suggest condensing some of this and/or moving some of it to an appendix/supplement.
- Table 1 (page 8) - Units for Qmax and Qmin are different. Check flow units for consistency.
- Figure 2 and Figure 5 - I appreciate the information density on these plots. However, the line widths on the RDF results (colored lines) came through rather faintly on the PDF I reviewed, making it difficult to readily differentiate among the results. Additionally the figure seemed a bit blurry. To the extent that the line widths and resolution could be adjusted, these figures will be much more effective.
- Line 558: "events 2,9, and 9 showed..." - I suspect the second "9" was in error here.

Citation: https://doi.org/10.5194/egusphere-2025-244-RC2
- AC2: 'Reply on RC2', Fernanda Helfer, 08 May 2025
  
  We sincerely thank the reviewer for their careful reading of our manuscript and for the constructive comments provided. We greatly appreciate the time and effort dedicated to improving the quality and clarity of our work. Below, we provide a brief response to each point raised. We will implement all suggested changes in the revised version of the manuscript.
  
  General Comment: Situating the findings in a broader context to increase the manuscript’s impact.
  
  Our Response: We will expand the Discussion and Conclusions sections to emphasize how the findings can inform baseflow estimation in other catchments beyond the study area.
  
  Specific and Technical Comments
  Conceptual model of the watershed: The reviewer requested a clearer conceptual model to support the interpretation of the results.
  
  Our Response: We will add a more detailed explanation of the catchment’s conceptual hydrological model, including insights from the silica data, to clarify how baseflow contributes to streamflow.
  Interpretation of the tiered RDF parameters: Discuss what the performance differences tell us about hydrological mechanisms.
  
  Our Response: We will add a discussion in Section 3.5 explaining how variations in BFImax and Beta reflect different flow generation processes and hydrological responses across event magnitudes.
  Seasonality and variability despite consistent precipitation: The reviewer requested more discussion on seasonal variation and the applicability of the tiered approach.
  
  Our Response: We will expand the Discussion to explain how seasonal groundwater dynamics and event-driven variations contribute to the observed variability, and how the event-based calibration approach can help disentangle these effects.
  Condensing repetitive tables: Move some repetitive error metric tables to the Supplementary Material.
  
  Our Response: We will attempt to condense the results in the main text and/or move detailed performance tables to the Supplementary Material to improve the manuscript’s readability.
  Units in Table 1: Inconsistency in flow units (Qmin and Qmax).
  Our Response: We will correct the units in Table 1 to ensure consistency.
  
  Figure 2 and Figure 5 – line widths and resolution: Faint lines and blurred figures.
  
  Our Response: We have reviewed printed versions of the manuscript and confirmed that all lines are clearly visible and the figures are not blurry. The faint appearance may have been due to the resolution of the version provided for review. Nevertheless, we will enhance line thickness and improve figure resolution in the revised manuscript to avoid any possible issues.
  Typographical error (Line 558): The reviewer pointed out a repeated event number ("9").
  
  Our Response: We will correct this typographical error.
  
  Citation: https://doi.org/10.5194/egusphere-2025-244-AC2
AC3: 'Comment on egusphere-2025-244', Fernanda Helfer, 13 May 2025

Dear referees and editor,
We have now implemented all the suggestions provided by the referees. Find below our final point-by-point reply document. We look forward to submitting the revised version of the manuscript.

Citation: https://doi.org/10.5194/egusphere-2025-244-AC3

Fernanda Helfer, Felipe Bernardi, Claudia Alessandra Peixoto de Barros, Daniel Gustavo Allasia, Jean Paolo Gomes Minella, Rutinéia Tassi, and Néverton Scariot

Viewed

Total article views: 587 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
378	181	28	587	14	33

HTML: 378
PDF: 181
XML: 28
Total: 587
BibTeX: 14
EndNote: 33

Views and downloads (calculated since 26 Mar 2025)

Month	HTML	PDF	XML	Total
Mar 2025	38	8	2	48
Apr 2025	49	18	8	75
May 2025	44	27	4	75
Jun 2025	42	12	4	58
Jul 2025	30	98	1	129
Aug 2025	136	17	8	161
Sep 2025	39	1	1	41

Cumulative views and downloads (calculated since 26 Mar 2025)

Month	HTML	PDF	XML	Total
Mar 2025	38	8	2	48
Apr 2025	49	18	8	75
May 2025	44	27	4	75
Jun 2025	42	12	4	58
Jul 2025	30	98	1	129
Aug 2025	136	17	8	161
Sep 2025	39	1	1	41

Viewed (geographical distribution)

Total article views: 575 (including HTML, PDF, and XML) Thereof 575 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 07 Sep 2025

Short summary

We explored how water flows in small rural streams to improve tools for better managing water resources. Using a new method, we adjusted existing models to consider the size of rainfall events, showing that water movement patterns change depending on the rain’s intensity. This approach makes predictions more accurate and helps scientists and managers understand water availability and protect ecosystems.


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