Lieke Anna Melsen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Lieke Anna Melsen has a total of 247 co-authors. Here are the top co-authors by number of shared publications:
- Adriaan J. Teuling
Hydrology and Environmental Hydraulics Group, Wageningen University, Wageningen, the Netherlands
42 shared publications - Pieter Richards van Oel
Water resources Management Group, Wageningen University, Wageningen, The Netherlands
28 shared publications - Sarra Kchouk
Water resources Management Group, Wageningen University, Wageningen, The Netherlands
25 shared publications - David W. Walker
Water resources Management Group, Wageningen University, Wageningen, The Netherlands
19 shared publications - Germano Gondim Ribeiro Neto
School of Geographical Sciences, University of Bristol, Bristol, UK
17 shared publications - Louise Cavalcante
Public Administration and Policy Group, Wageningen University, Wageningen, The Netherlands
14 shared publications
Latest Publications
- Open Access
- 27 October 2026
- Published by Copernicus GmbH in Hydrology and Earth System Sciences
- Vol. 30 (20) , 6579-6597
Technical note: Machine learning metamodelling for global sensitivity analysis
Global sensitivity analysis (GSA) plays a central role in hydrologic modelling by supporting model understanding, diagnosis, and decision-making through the identification of influential and non-influential parameters and their interactions. Variance-based methods provide a rigorous framework for GSA but are often computationally expensive, as their estimation requires a large number of model evaluations. Metamodelling has therefore been widely adopted as a strategy to alleviate this issue, with recent advances in machine learning (ML) offering new opportunities to construct accurate and flexible surrogates for complex models. This technical note examines the practical relationship between Sobol' total-effect indices ( T i ) and feature importance measures derived from ML metamodels within a hydrologic modelling context. Building on theoretical results that link T i to permutation variable importance (PVI i ) under independence assumptions, we provide systematic numerical evidence using three hydrologic models of varying complexity (HBV, HyMod, and VIC) applied to three headwater catchments in northern Germany, together with three ML metamodels: a gradient boosting decision tree (GBDT), a neural network (NN), and a linear model (LM). The three metamodels were trained on Monte Carlo samples and used to estimate sensitivities through PVI i and SHapley Additive exPlanations (SHAP i ). The results demonstrate that GBDT and NN metamodels reliably reproduce both the ranking and relative magnitude of T i using PVI i across all hydrologic models, providing clear empirical support for the theoretical connection between the two measures. In contrast, the performance of LM-based estimates depends strongly on the degree of linearity in the underlying model response. Mean absolute SHAP i values exhibit a consistent monotonic relationship with T i and preserve parameter rankings, while sample-specific SHAP i values enable a distributed evaluation of sensitivities across both the parameter space and the target variable space. Overall, this study highlights ML metamodelling as a computationally efficient and conceptually sound framework for GSA in hydrologic modelling and beyond.
- 27 October 2026
- Published by Copernicus GmbH in Hydrology and Earth System Sciences
- Vol. 30 (20) , 6579-6597
Open AccessTechnical note: Machine learning metamodelling for global sensitivity analysis
Global sensitivity analysis (GSA) plays a central role in hydrologic modelling by supporting model understanding, diagnosis, and decision-making through the identification of influential and non-influential parameters and their interactions. Variance-based methods provide a rigorous framework for GSA but are often computationally expensive, as their estimation requires a large number of model evaluations. Metamodelling has therefore been widely adopted as a strategy to alleviate this issue, with recent advances in machine learning (ML) offering new opportunities to construct accurate and flexible surrogates for complex models. This technical note examines the practical relationship between Sobol' total-effect indices ( T i ) and feature importance measures derived from ML metamodels within a hydrologic modelling context. Building on theoretical results that link T i to permutation variable importance (PVI i ) under independence assumptions, we provide systematic numerical evidence using three hydrologic models of varying complexity (HBV, HyMod, and VIC) applied to three headwater catchments in northern Germany, together with three ML metamodels: a gradient boosting decision tree (GBDT), a neural network (NN), and a linear model (LM). The three metamodels were trained on Monte Carlo samples and used to estimate sensitivities through PVI i and SHapley Additive exPlanations (SHAP i ). The results demonstrate that GBDT and NN metamodels reliably reproduce both the ranking and relative magnitude of T i using PVI i across all hydrologic models, providing clear empirical support for the theoretical connection between the two measures. In contrast, the performance of LM-based estimates depends strongly on the degree of linearity in the underlying model response. Mean absolute SHAP i values exhibit a consistent monotonic relationship with T i and preserve parameter rankings, while sample-specific SHAP i values enable a distributed evaluation of sensitivities across both the parameter space and the target variable space. Overall, this study highlights ML metamodelling as a computationally efficient and conceptually sound framework for GSA in hydrologic modelling and beyond.
- Open Access
- 26 October 2026
- Published by Springer Nature in Nature Water
- p. 1
Author Correction: Fifty years of research have deepened uncertainties in global irrigation water withdrawals
- 26 October 2026
- Published by Springer Nature in Nature Water
- p. 1
Open AccessAuthor Correction: Fifty years of research have deepened uncertainties in global irrigation water withdrawals
- Open Access
- 6 October 2026
- Published by Copernicus GmbH in Geoscience Communication
- Vol. 9 (4) , 437-450
Field or virtual? Perceived impact of academic excursions through the eyes of (future) hydrologists
Field training such as excursions has always been an integral part of the education of geoscientists. The value of these excursions is frequently debated, related to the efforts and costs involved with their organization. Professionally prepared virtual excursions are sometimes presented as a possible alternative for field activities. This debate has been accelerated by the COVID-19 pandemic: the lockdown forced teachers to come up with virtual and online educational material, also for fieldwork and excursions. One of the courses at Wageningen University (the Netherlands) that was heavily affected by COVID-19 restrictions was the Master course Catchment and Climate Hydrology. An important and integral component of that course is an 8 d excursion to Iceland where relatively undisturbed catchments are studied in a country heavily affected by climate change. All students from the 2020 and 2021 cohort took the alternative professionally prepared virtual excursion. After the pandemic, in 2022, these students were offered to join an extra added excursion. About 50 % of that population was able to join. This configuration, involving student cohorts that participated only in the field excursion, only in the virtual excursion, or in both, provided an unique opportunity to assess student evaluations of field-based education and the potential of virtual alternatives. A questionnaire was sent to excursion participants to critically evaluate the impact of the excursion on their scientific development and their view on real versus virtual excursions for geoscientists. The time frame for the questionnaire ran from 2011 to 2025, including COVID-19 years 2020 and 2021. The response on the statement “Planning and/or fieldwork are an important part of my current job or study” is moderately agree , with > 50 % of the responses even with a response strongly to fully agree. This indicates that field training is very important for the majority of the junior and senior geoscientists. The study also showed that a multiple day excursion to Iceland strongly contributed to the understanding of undisturbed hydrologic systems and to the general scientific development of the participants. The survey clearly indicated that they appreciated the effort of making a virtual excursion, but that the impact of that virtual excursion on their scientific development was considerably lower than for the equivalent real excursion. The reasons indicated are manifold: they missed the active location-based learning in a country with undisturbed catchments, the inspirational and direct interaction with the teaching staff and linking up with their peers to create a professional network. Therefor the main conclusion of this inventory is that preparing virtual excursions for Master students can be worth the investment, but just as an extra tool in the course tool box, and certainly not as a replacement for real excursions.
- 6 October 2026
- Published by Copernicus GmbH in Geoscience Communication
- Vol. 9 (4) , 437-450
Open AccessField or virtual? Perceived impact of academic excursions through the eyes of (future) hydrologists
Field training such as excursions has always been an integral part of the education of geoscientists. The value of these excursions is frequently debated, related to the efforts and costs involved with their organization. Professionally prepared virtual excursions are sometimes presented as a possible alternative for field activities. This debate has been accelerated by the COVID-19 pandemic: the lockdown forced teachers to come up with virtual and online educational material, also for fieldwork and excursions. One of the courses at Wageningen University (the Netherlands) that was heavily affected by COVID-19 restrictions was the Master course Catchment and Climate Hydrology. An important and integral component of that course is an 8 d excursion to Iceland where relatively undisturbed catchments are studied in a country heavily affected by climate change. All students from the 2020 and 2021 cohort took the alternative professionally prepared virtual excursion. After the pandemic, in 2022, these students were offered to join an extra added excursion. About 50 % of that population was able to join. This configuration, involving student cohorts that participated only in the field excursion, only in the virtual excursion, or in both, provided an unique opportunity to assess student evaluations of field-based education and the potential of virtual alternatives. A questionnaire was sent to excursion participants to critically evaluate the impact of the excursion on their scientific development and their view on real versus virtual excursions for geoscientists. The time frame for the questionnaire ran from 2011 to 2025, including COVID-19 years 2020 and 2021. The response on the statement “Planning and/or fieldwork are an important part of my current job or study” is moderately agree , with > 50 % of the responses even with a response strongly to fully agree. This indicates that field training is very important for the majority of the junior and senior geoscientists. The study also showed that a multiple day excursion to Iceland strongly contributed to the understanding of undisturbed hydrologic systems and to the general scientific development of the participants. The survey clearly indicated that they appreciated the effort of making a virtual excursion, but that the impact of that virtual excursion on their scientific development was considerably lower than for the equivalent real excursion. The reasons indicated are manifold: they missed the active location-based learning in a country with undisturbed catchments, the inspirational and direct interaction with the teaching staff and linking up with their peers to create a professional network. Therefor the main conclusion of this inventory is that preparing virtual excursions for Master students can be worth the investment, but just as an extra tool in the course tool box, and certainly not as a replacement for real excursions.
- Open Access
- 14 September 2026
- Published by Springer Nature in Nature Water
- Vol. 4 (10) , 1343-1354
Fifty years of research have deepened uncertainties in global irrigation water withdrawals
For five decades, refining models and datasets has been regarded as central to reducing uncertainty in global irrigation water withdrawal estimates. However, whether these efforts have succeeded remains unclear. Here we compile 1,621 global estimates from 65 studies published between 1974 and 2025 and evaluate them across 360 analytical pathways, spanning a comprehensive range of plausible statistical designs. Our results indicate that uncertainty in global irrigation water withdrawals has not declined: only 18% of pathways show any reduction over time, falling to 10% once projections to 2100 are excluded. Paths split between an unstable trend (42%) and an increase in uncertainty (35%), and where a net direction exists, it is an increase by more than two to one. Progress may therefore require the field to renounce the goal of reducing uncertainty and instead double down on what it has historically done best: expanding it, until uncertainty reaches the boundary of what current knowledge of global irrigation cannot exclude.
- 14 September 2026
- Published by Springer Nature in Nature Water
- Vol. 4 (10) , 1343-1354
Open AccessFifty years of research have deepened uncertainties in global irrigation water withdrawals
For five decades, refining models and datasets has been regarded as central to reducing uncertainty in global irrigation water withdrawal estimates. However, whether these efforts have succeeded remains unclear. Here we compile 1,621 global estimates from 65 studies published between 1974 and 2025 and evaluate them across 360 analytical pathways, spanning a comprehensive range of plausible statistical designs. Our results indicate that uncertainty in global irrigation water withdrawals has not declined: only 18% of pathways show any reduction over time, falling to 10% once projections to 2100 are excluded. Paths split between an unstable trend (42%) and an increase in uncertainty (35%), and where a net direction exists, it is an increase by more than two to one. Progress may therefore require the field to renounce the goal of reducing uncertainty and instead double down on what it has historically done best: expanding it, until uncertainty reaches the boundary of what current knowledge of global irrigation cannot exclude.
- Open Access
- 1 September 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-26
Reply on RC3
Global sensitivity analysis (GSA) plays a central role in hydrologic modelling by supporting model understanding, diagnosis, and decision-making through the identification of influential and non-influential parameters and their interactions. Variance-based methods provide a rigorous framework for GSA but are often computationally expensive, as their estimation requires a large number of model evaluations. Metamodelling has therefore been widely adopted as a strategy to alleviate this issue, with recent advances in machine learning (ML) offering new opportunities to construct accurate and flexible surrogates for complex models. This technical note examines the practical relationship between Sobol’ total-effect indices (Ti) and feature importance measures derived from ML metamodels within a hydrologic modelling context. Building on theoretical results that link Ti to permutation variable importance (PVIi) under independence assumptions, we provide systematic numerical evidence using three conceptual hydrologic models of varying complexity (HBV, HyMod, and VIC) applied to three headwater catchments in northern Germany, together with three ML metamodels: a random forest (RF), a neural network (NN), and a linear model (LM). The three metamodels were trained on Monte Carlo samples and used to estimate sensitivities through PVIi and SHapley Additive exPlanations (SHAPi). The results demonstrate that RF and NN metamodels reliably reproduce both the ranking and relative magnitude of Ti using PVIi across all hydrologic models, providing clear empirical support for the theoretical connection between the two measures. In contrast, the performance of LM-based estimates depends strongly on the degree of linearity in the underlying model response. Mean absolute SHAPi values exhibit a consistent monotonic relationship with Ti and preserve parameter rankings, while sample-specific SHAPi values enable a distributed evaluation of sensitivities across both the parameter space and the target variable space. Overall, this study highlights ML metamodelling as a computationally efficient and conceptually sound framework for GSA in hydrologic modelling and beyond.
- 1 September 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-26
Open AccessReply on RC3
Global sensitivity analysis (GSA) plays a central role in hydrologic modelling by supporting model understanding, diagnosis, and decision-making through the identification of influential and non-influential parameters and their interactions. Variance-based methods provide a rigorous framework for GSA but are often computationally expensive, as their estimation requires a large number of model evaluations. Metamodelling has therefore been widely adopted as a strategy to alleviate this issue, with recent advances in machine learning (ML) offering new opportunities to construct accurate and flexible surrogates for complex models. This technical note examines the practical relationship between Sobol’ total-effect indices (Ti) and feature importance measures derived from ML metamodels within a hydrologic modelling context. Building on theoretical results that link Ti to permutation variable importance (PVIi) under independence assumptions, we provide systematic numerical evidence using three conceptual hydrologic models of varying complexity (HBV, HyMod, and VIC) applied to three headwater catchments in northern Germany, together with three ML metamodels: a random forest (RF), a neural network (NN), and a linear model (LM). The three metamodels were trained on Monte Carlo samples and used to estimate sensitivities through PVIi and SHapley Additive exPlanations (SHAPi). The results demonstrate that RF and NN metamodels reliably reproduce both the ranking and relative magnitude of Ti using PVIi across all hydrologic models, providing clear empirical support for the theoretical connection between the two measures. In contrast, the performance of LM-based estimates depends strongly on the degree of linearity in the underlying model response. Mean absolute SHAPi values exhibit a consistent monotonic relationship with Ti and preserve parameter rankings, while sample-specific SHAPi values enable a distributed evaluation of sensitivities across both the parameter space and the target variable space. Overall, this study highlights ML metamodelling as a computationally efficient and conceptually sound framework for GSA in hydrologic modelling and beyond.
- Open Access
- 9 August 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-26
Comment on egusphere-2026-1787
Global sensitivity analysis (GSA) plays a central role in hydrologic modelling by supporting model understanding, diagnosis, and decision-making through the identification of influential and non-influential parameters and their interactions. Variance-based methods provide a rigorous framework for GSA but are often computationally expensive, as their estimation requires a large number of model evaluations. Metamodelling has therefore been widely adopted as a strategy to alleviate this issue, with recent advances in machine learning (ML) offering new opportunities to construct accurate and flexible surrogates for complex models. This technical note examines the practical relationship between Sobol’ total-effect indices (Ti) and feature importance measures derived from ML metamodels within a hydrologic modelling context. Building on theoretical results that link Ti to permutation variable importance (PVIi) under independence assumptions, we provide systematic numerical evidence using three conceptual hydrologic models of varying complexity (HBV, HyMod, and VIC) applied to three headwater catchments in northern Germany, together with three ML metamodels: a random forest (RF), a neural network (NN), and a linear model (LM). The three metamodels were trained on Monte Carlo samples and used to estimate sensitivities through PVIi and SHapley Additive exPlanations (SHAPi). The results demonstrate that RF and NN metamodels reliably reproduce both the ranking and relative magnitude of Ti using PVIi across all hydrologic models, providing clear empirical support for the theoretical connection between the two measures. In contrast, the performance of LM-based estimates depends strongly on the degree of linearity in the underlying model response. Mean absolute SHAPi values exhibit a consistent monotonic relationship with Ti and preserve parameter rankings, while sample-specific SHAPi values enable a distributed evaluation of sensitivities across both the parameter space and the target variable space. Overall, this study highlights ML metamodelling as a computationally efficient and conceptually sound framework for GSA in hydrologic modelling and beyond.
- 9 August 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-26
Open AccessComment on egusphere-2026-1787
Global sensitivity analysis (GSA) plays a central role in hydrologic modelling by supporting model understanding, diagnosis, and decision-making through the identification of influential and non-influential parameters and their interactions. Variance-based methods provide a rigorous framework for GSA but are often computationally expensive, as their estimation requires a large number of model evaluations. Metamodelling has therefore been widely adopted as a strategy to alleviate this issue, with recent advances in machine learning (ML) offering new opportunities to construct accurate and flexible surrogates for complex models. This technical note examines the practical relationship between Sobol’ total-effect indices (Ti) and feature importance measures derived from ML metamodels within a hydrologic modelling context. Building on theoretical results that link Ti to permutation variable importance (PVIi) under independence assumptions, we provide systematic numerical evidence using three conceptual hydrologic models of varying complexity (HBV, HyMod, and VIC) applied to three headwater catchments in northern Germany, together with three ML metamodels: a random forest (RF), a neural network (NN), and a linear model (LM). The three metamodels were trained on Monte Carlo samples and used to estimate sensitivities through PVIi and SHapley Additive exPlanations (SHAPi). The results demonstrate that RF and NN metamodels reliably reproduce both the ranking and relative magnitude of Ti using PVIi across all hydrologic models, providing clear empirical support for the theoretical connection between the two measures. In contrast, the performance of LM-based estimates depends strongly on the degree of linearity in the underlying model response. Mean absolute SHAPi values exhibit a consistent monotonic relationship with Ti and preserve parameter rankings, while sample-specific SHAPi values enable a distributed evaluation of sensitivities across both the parameter space and the target variable space. Overall, this study highlights ML metamodelling as a computationally efficient and conceptually sound framework for GSA in hydrologic modelling and beyond.
- Open Access
- 22 July 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-26
Reply on RC1
Global sensitivity analysis (GSA) plays a central role in hydrologic modelling by supporting model understanding, diagnosis, and decision-making through the identification of influential and non-influential parameters and their interactions. Variance-based methods provide a rigorous framework for GSA but are often computationally expensive, as their estimation requires a large number of model evaluations. Metamodelling has therefore been widely adopted as a strategy to alleviate this issue, with recent advances in machine learning (ML) offering new opportunities to construct accurate and flexible surrogates for complex models. This technical note examines the practical relationship between Sobol’ total-effect indices (Ti) and feature importance measures derived from ML metamodels within a hydrologic modelling context. Building on theoretical results that link Ti to permutation variable importance (PVIi) under independence assumptions, we provide systematic numerical evidence using three conceptual hydrologic models of varying complexity (HBV, HyMod, and VIC) applied to three headwater catchments in northern Germany, together with three ML metamodels: a random forest (RF), a neural network (NN), and a linear model (LM). The three metamodels were trained on Monte Carlo samples and used to estimate sensitivities through PVIi and SHapley Additive exPlanations (SHAPi). The results demonstrate that RF and NN metamodels reliably reproduce both the ranking and relative magnitude of Ti using PVIi across all hydrologic models, providing clear empirical support for the theoretical connection between the two measures. In contrast, the performance of LM-based estimates depends strongly on the degree of linearity in the underlying model response. Mean absolute SHAPi values exhibit a consistent monotonic relationship with Ti and preserve parameter rankings, while sample-specific SHAPi values enable a distributed evaluation of sensitivities across both the parameter space and the target variable space. Overall, this study highlights ML metamodelling as a computationally efficient and conceptually sound framework for GSA in hydrologic modelling and beyond.
- 22 July 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-26
Open AccessReply on RC1
Global sensitivity analysis (GSA) plays a central role in hydrologic modelling by supporting model understanding, diagnosis, and decision-making through the identification of influential and non-influential parameters and their interactions. Variance-based methods provide a rigorous framework for GSA but are often computationally expensive, as their estimation requires a large number of model evaluations. Metamodelling has therefore been widely adopted as a strategy to alleviate this issue, with recent advances in machine learning (ML) offering new opportunities to construct accurate and flexible surrogates for complex models. This technical note examines the practical relationship between Sobol’ total-effect indices (Ti) and feature importance measures derived from ML metamodels within a hydrologic modelling context. Building on theoretical results that link Ti to permutation variable importance (PVIi) under independence assumptions, we provide systematic numerical evidence using three conceptual hydrologic models of varying complexity (HBV, HyMod, and VIC) applied to three headwater catchments in northern Germany, together with three ML metamodels: a random forest (RF), a neural network (NN), and a linear model (LM). The three metamodels were trained on Monte Carlo samples and used to estimate sensitivities through PVIi and SHapley Additive exPlanations (SHAPi). The results demonstrate that RF and NN metamodels reliably reproduce both the ranking and relative magnitude of Ti using PVIi across all hydrologic models, providing clear empirical support for the theoretical connection between the two measures. In contrast, the performance of LM-based estimates depends strongly on the degree of linearity in the underlying model response. Mean absolute SHAPi values exhibit a consistent monotonic relationship with Ti and preserve parameter rankings, while sample-specific SHAPi values enable a distributed evaluation of sensitivities across both the parameter space and the target variable space. Overall, this study highlights ML metamodelling as a computationally efficient and conceptually sound framework for GSA in hydrologic modelling and beyond.
- Open Access
- 22 July 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-26
Reply on RC2
Global sensitivity analysis (GSA) plays a central role in hydrologic modelling by supporting model understanding, diagnosis, and decision-making through the identification of influential and non-influential parameters and their interactions. Variance-based methods provide a rigorous framework for GSA but are often computationally expensive, as their estimation requires a large number of model evaluations. Metamodelling has therefore been widely adopted as a strategy to alleviate this issue, with recent advances in machine learning (ML) offering new opportunities to construct accurate and flexible surrogates for complex models. This technical note examines the practical relationship between Sobol’ total-effect indices (Ti) and feature importance measures derived from ML metamodels within a hydrologic modelling context. Building on theoretical results that link Ti to permutation variable importance (PVIi) under independence assumptions, we provide systematic numerical evidence using three conceptual hydrologic models of varying complexity (HBV, HyMod, and VIC) applied to three headwater catchments in northern Germany, together with three ML metamodels: a random forest (RF), a neural network (NN), and a linear model (LM). The three metamodels were trained on Monte Carlo samples and used to estimate sensitivities through PVIi and SHapley Additive exPlanations (SHAPi). The results demonstrate that RF and NN metamodels reliably reproduce both the ranking and relative magnitude of Ti using PVIi across all hydrologic models, providing clear empirical support for the theoretical connection between the two measures. In contrast, the performance of LM-based estimates depends strongly on the degree of linearity in the underlying model response. Mean absolute SHAPi values exhibit a consistent monotonic relationship with Ti and preserve parameter rankings, while sample-specific SHAPi values enable a distributed evaluation of sensitivities across both the parameter space and the target variable space. Overall, this study highlights ML metamodelling as a computationally efficient and conceptually sound framework for GSA in hydrologic modelling and beyond.
- 22 July 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-26
Open AccessReply on RC2
Global sensitivity analysis (GSA) plays a central role in hydrologic modelling by supporting model understanding, diagnosis, and decision-making through the identification of influential and non-influential parameters and their interactions. Variance-based methods provide a rigorous framework for GSA but are often computationally expensive, as their estimation requires a large number of model evaluations. Metamodelling has therefore been widely adopted as a strategy to alleviate this issue, with recent advances in machine learning (ML) offering new opportunities to construct accurate and flexible surrogates for complex models. This technical note examines the practical relationship between Sobol’ total-effect indices (Ti) and feature importance measures derived from ML metamodels within a hydrologic modelling context. Building on theoretical results that link Ti to permutation variable importance (PVIi) under independence assumptions, we provide systematic numerical evidence using three conceptual hydrologic models of varying complexity (HBV, HyMod, and VIC) applied to three headwater catchments in northern Germany, together with three ML metamodels: a random forest (RF), a neural network (NN), and a linear model (LM). The three metamodels were trained on Monte Carlo samples and used to estimate sensitivities through PVIi and SHapley Additive exPlanations (SHAPi). The results demonstrate that RF and NN metamodels reliably reproduce both the ranking and relative magnitude of Ti using PVIi across all hydrologic models, providing clear empirical support for the theoretical connection between the two measures. In contrast, the performance of LM-based estimates depends strongly on the degree of linearity in the underlying model response. Mean absolute SHAPi values exhibit a consistent monotonic relationship with Ti and preserve parameter rankings, while sample-specific SHAPi values enable a distributed evaluation of sensitivities across both the parameter space and the target variable space. Overall, this study highlights ML metamodelling as a computationally efficient and conceptually sound framework for GSA in hydrologic modelling and beyond.
- Open Access
- 30 June 2026
- Published by Elsevier in Journal of Hydrology
- Vol. 677
Setting the standard? Comparison of the implementation of decision-support modeling standards in Australia and the Netherlands
- 30 June 2026
- Published by Elsevier in Journal of Hydrology
- Vol. 677
Open AccessSetting the standard? Comparison of the implementation of decision-support modeling standards in Australia and the Netherlands
- Open Access
- 25 June 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-22
Reply on RC2
Field training such as excursions have always been an integral part of the education of geoscientists. The value of these excursions is frequently debated, related to the efforts and costs involved with their organization. Professionally prepared virtual excursions are sometimes presented as a possible alternative for field activities. This debate has been accelerated by the COVID-19 pandemic: the lockdown forced teachers to come up with virtual and online education material, also for fieldwork and excursions. One of the courses at Wageningen University (the Netherlands) heavily affected by COVID-19 restrictions was the Master course Catchment and Climate Hydrology. An important and integral component of that course is an 8-day excursion to Iceland where relatively undisturbed catchments are studied in a country heavily affected by climate change. All students from the 2020 and 2021 cohort took the alternative professionally prepared virtual excursion. After the pandemic, in 2022, these students did get the offer to join an extra added excursion. About 50 % of that population was able to join. This configuration, involving student cohorts that participated only in the field excursion, only in the virtual excursion, or in both, provided an unique opportunity to assess student evaluations of field-based education and the potential of virtual alternatives. A questionnaire was sent to excursion participants to critically evaluate the impact of the excursion on their scientific development and their view on real versus virtual excursions for geoscientists. The time frame for the questionnaire runs from 2011 to 2025, including COVID-19 years 2020 and 2021. The participants clearly indicate that they appreciated the effort of making a virtual excursion, but that the impact of that virtual excursion on their scientific development was a lot lower than for the equivalent real excursion. The reasons indicated are manifold: they missed the inspirational interaction with the teaching staff and linking up with their peers to create a professional network. Therefor the main conclusion of this inventory is that preparing virtual excursions for Master students can be worth the investment, but just as an extra tool in the course tool box, but not as a replacement for real excursions.
- 25 June 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-22
Open AccessReply on RC2
Field training such as excursions have always been an integral part of the education of geoscientists. The value of these excursions is frequently debated, related to the efforts and costs involved with their organization. Professionally prepared virtual excursions are sometimes presented as a possible alternative for field activities. This debate has been accelerated by the COVID-19 pandemic: the lockdown forced teachers to come up with virtual and online education material, also for fieldwork and excursions. One of the courses at Wageningen University (the Netherlands) heavily affected by COVID-19 restrictions was the Master course Catchment and Climate Hydrology. An important and integral component of that course is an 8-day excursion to Iceland where relatively undisturbed catchments are studied in a country heavily affected by climate change. All students from the 2020 and 2021 cohort took the alternative professionally prepared virtual excursion. After the pandemic, in 2022, these students did get the offer to join an extra added excursion. About 50 % of that population was able to join. This configuration, involving student cohorts that participated only in the field excursion, only in the virtual excursion, or in both, provided an unique opportunity to assess student evaluations of field-based education and the potential of virtual alternatives. A questionnaire was sent to excursion participants to critically evaluate the impact of the excursion on their scientific development and their view on real versus virtual excursions for geoscientists. The time frame for the questionnaire runs from 2011 to 2025, including COVID-19 years 2020 and 2021. The participants clearly indicate that they appreciated the effort of making a virtual excursion, but that the impact of that virtual excursion on their scientific development was a lot lower than for the equivalent real excursion. The reasons indicated are manifold: they missed the inspirational interaction with the teaching staff and linking up with their peers to create a professional network. Therefor the main conclusion of this inventory is that preparing virtual excursions for Master students can be worth the investment, but just as an extra tool in the course tool box, but not as a replacement for real excursions.