the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 12 Sep 2025
EcoTWIN 1.0: A Fully Distributed Tracer-Aided Ecohydrological Model Tracking Water, Isotopes, and Nutrients
Abstract. The value of stable water isotopes in constraining process representation in hydrological models is well acknowledged with numerous tracer-aided hydrological models developed in recent years, yet few have leveraged these benefits for more robust water quality modelling. Therefore, we introduce EcoTWIN, a fully distributed tracer-aided ecohydrological model that simultaneously tracks water, isotopic, and nutrient fluxes in an integrated C++ framework. A thorough validation was conducted by calibrating EcoTWIN against discharge, in-stream isotopes, and NO3-N concentrations (1980–2024) in 17 large-scale (103 – 105 km2) European catchments spanning a wide range of geographic and climatic gradients. Furthermore, three reanalysis products (ERA5 snow depth, MODIS evapotranspiration, and GRACE surface water anomaly) were employed to further validate the capacity of EcoTWIN to reproduce associated internal water fluxes without calibration. Results showed good model performance of both calibrated in-stream targets and uncalibrated internal fluxes in most catchments. Therefore, we conclude that EcoTWIN is a flexible, transferable modelling tool for prediction and process inference in terrestrial ecosystems ranging from boreal to subtropic climates. Constrained by tracer simulations, the model not only captures the celerity, but also the velocity of hydrological fluxes, thus providing spatio-temporally-explicit estimations of water ages and travel times. Such information provides opportunities to bridge catchment hydrology and water quality by linking travel times with biogeochemical processing times. We demonstrate this with a proof of concept using Damköhler Number in nitrogen modelling.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(9124 KB)
-
Supplement
(1573 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(9124 KB) - Metadata XML
-
Supplement
(1573 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-3941', Chang Liao, 20 Oct 2025
-
AC1: 'Reply on RC1', Songjun Wu, 27 Oct 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-3941/egusphere-2025-3941-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Songjun Wu, 27 Oct 2025
-
RC2: 'Comment on egusphere-2025-3941', Anonymous Referee #2, 12 Nov 2025
Reviewing of the manuscript open for discussion ‘EcoTWIN 1.0: A Fully Distributed Tracer-Aided Ecohydrological Model Tracking Water, Isotopes, and Nutrients’ by Songjun Wu et al. submitted to Geoscientific Model Development (Manuscript ID: egusphere-2025-3941).
The authors have developed/improved and ecohydrological model with advanced traits, tracking water, isotopic and nutrient fluxes. The reviewer considers that this model development work is worthwhile and beneficial to the modeling community and fall within well with the journal topic. There are a couple of issues that the reviewer suggests for the authors to refer, and they are listed in detail as follows:
- Almost all figures need to be replotted. Here I give some examples. Fig. 1: The texts are too small to read, and the subtitles need add some explanations for the figure, so that the reader is easy to follow it; Fig. 2: the coordinates need to be added, and some important landmarks or rivers information should better be added; the texts are to small to be identified; Fig. 3: the color for the color should be more identifiable, rather than orange and blue two colors; Fig. 4: the time series figures give no information to the reviewer; similar issues also for Fig. 5-7, and the reviewer will NOT list all issues for the figures. The authors should check them carefully and revise them all.
- The structure of this manuscript may be adjusted. Part 4.1 and 4.2 may be moved under the section of 3 Model calibration and validation? This should be one section to discuss the model advantages of accuracy compared with previous models? For example, providing specific skill metrics values.
- Provide some explanations for the reason why these 17 catchments are selected for this study.
- What is the difference between model calibration and validation? The reviewer could not quite understand herein, it seems the difference is defined by the different variables that are choose for the model-to-data comparison. Please explain it.
- More skill metrics should be defined and used for model performance. For example, Root Mean Square Error, Correlation Coefficient, Mean Value Difference etc. to evaluate the model performance. Then, give the statistics and compare them with previous other models.
- Some small mistakes or errors. Table 1, the table better not crossing two pages. Line 378, add the equation number and the meaning of it (e.g., what does l mean in the left of the equation). Be consistent use KEG or KGE, and define it and explain it (e.g., what are the value ranges, and the corresponding performance, excellent, good, normal, poor etc.)
- Line 451: ERA5 reanalysis products equal to observations? Please double check it. Line 531-532: ‘However, it increases severely increase the ……’ Delete one of the ‘increases’. Line 694: the first letter of the word ‘mediterranean’ should be initialized.
Citation: https://doi.org/10.5194/egusphere-2025-3941-RC2 -
AC2: 'Reply on RC2', Songjun Wu, 21 Nov 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-3941/egusphere-2025-3941-AC2-supplement.pdf
-
RC3: 'Comment on egusphere-2025-3941', Anonymous Referee #2, 26 Nov 2025
The authors have responded some of my concern while it is still not clear to me what changes and improvement have been made during this round. The responses are either vaguely stated in the response letter or could not find the revisions in the updated manuscript. Additionally, the differences between the observations and ERA5 data are still not well explained. Most figures still need significant improvement. A much clear point-to-point response letter and revised manuscript should be added.
Citation: https://doi.org/10.5194/egusphere-2025-3941-RC3 -
AC3: 'Reply on RC3', Songjun Wu, 02 Dec 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-3941/egusphere-2025-3941-AC3-supplement.pdf
-
AC3: 'Reply on RC3', Songjun Wu, 02 Dec 2025
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-3941', Chang Liao, 20 Oct 2025
This manuscript describes a novel spatially distributed ecohydrologic model ECOTWIN. This model links hydrology with nutrient and isotopic modules. The capability has great potential to improve our understand the terrestrial and aquifer ecosystem interactions. Although similar capability has been attempted in several other studies with different focus such as dissolved organic carbon (DOC), the inclusion of lateral flow with nutrient and carbon cycle is still a challenge in most spatially distributed (eco)hydrological models.
There are some missing information in the current manuscript, such as (1) groundwater level; (2) resolution for evaluation; (3) chemical reaction. See details below. The presentation of some information can be improved, especially some figures are too small to be read.
Distributed, do you mean spatially distributed?Line 40:
Sentence stoped abruptly without giving reasons.
Line 43:
What is “somehow”
Line 44:
Where is 81% come from? Reference needed
Line 82:
On disk/memory change to online and offline coupling
Figure 1. Is soil zone part of vadose zone?
Line 192:
What if the water table is above the last soil layer?
Equation 16:
If the pond is not next to the river channel, why or how it contribute to the channel flow?
Similarly, in equation 18, if the terrestrial grid cell is far away from the river channel, how does the vadose zone contribute to the channel flow?
Figure 2:
Does not have coordinates. Such as longitude and latitude information.
What are the particular reason why these 17 watersheds were chosen.
Line 346:
Does the model need spin-up to reach steady state?
Line 399:
How do you consider the upstream - downstream dependency in discharge, some watersheds are connected based on Figure 2.
Figure 4:
Show line/color meaning in the Figure as well. Also missing time step information.
The subfigures are too small to read. Maybe move some into SI and only show a few here.
Figure 5. Is the time series watershed averaged? If so, the caption should mention that.
Figure 6:
The GRACE datasets resolution might be too coarse for comparison; there need to be some justification somewhere.
Check other spatial products as well, maybe a table listing all produce with original and resample resolution for reference.
Line 534:
Integrated C++ framework: C++ is a programming language, not a framework.
Citation: https://doi.org/10.5194/egusphere-2025-3941-RC1 -
AC1: 'Reply on RC1', Songjun Wu, 27 Oct 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-3941/egusphere-2025-3941-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Songjun Wu, 27 Oct 2025
-
RC2: 'Comment on egusphere-2025-3941', Anonymous Referee #2, 12 Nov 2025
Reviewing of the manuscript open for discussion ‘EcoTWIN 1.0: A Fully Distributed Tracer-Aided Ecohydrological Model Tracking Water, Isotopes, and Nutrients’ by Songjun Wu et al. submitted to Geoscientific Model Development (Manuscript ID: egusphere-2025-3941).
The authors have developed/improved and ecohydrological model with advanced traits, tracking water, isotopic and nutrient fluxes. The reviewer considers that this model development work is worthwhile and beneficial to the modeling community and fall within well with the journal topic. There are a couple of issues that the reviewer suggests for the authors to refer, and they are listed in detail as follows:
- Almost all figures need to be replotted. Here I give some examples. Fig. 1: The texts are too small to read, and the subtitles need add some explanations for the figure, so that the reader is easy to follow it; Fig. 2: the coordinates need to be added, and some important landmarks or rivers information should better be added; the texts are to small to be identified; Fig. 3: the color for the color should be more identifiable, rather than orange and blue two colors; Fig. 4: the time series figures give no information to the reviewer; similar issues also for Fig. 5-7, and the reviewer will NOT list all issues for the figures. The authors should check them carefully and revise them all.
- The structure of this manuscript may be adjusted. Part 4.1 and 4.2 may be moved under the section of 3 Model calibration and validation? This should be one section to discuss the model advantages of accuracy compared with previous models? For example, providing specific skill metrics values.
- Provide some explanations for the reason why these 17 catchments are selected for this study.
- What is the difference between model calibration and validation? The reviewer could not quite understand herein, it seems the difference is defined by the different variables that are choose for the model-to-data comparison. Please explain it.
- More skill metrics should be defined and used for model performance. For example, Root Mean Square Error, Correlation Coefficient, Mean Value Difference etc. to evaluate the model performance. Then, give the statistics and compare them with previous other models.
- Some small mistakes or errors. Table 1, the table better not crossing two pages. Line 378, add the equation number and the meaning of it (e.g., what does l mean in the left of the equation). Be consistent use KEG or KGE, and define it and explain it (e.g., what are the value ranges, and the corresponding performance, excellent, good, normal, poor etc.)
- Line 451: ERA5 reanalysis products equal to observations? Please double check it. Line 531-532: ‘However, it increases severely increase the ……’ Delete one of the ‘increases’. Line 694: the first letter of the word ‘mediterranean’ should be initialized.
Citation: https://doi.org/10.5194/egusphere-2025-3941-RC2 -
AC2: 'Reply on RC2', Songjun Wu, 21 Nov 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-3941/egusphere-2025-3941-AC2-supplement.pdf
-
RC3: 'Comment on egusphere-2025-3941', Anonymous Referee #2, 26 Nov 2025
The authors have responded some of my concern while it is still not clear to me what changes and improvement have been made during this round. The responses are either vaguely stated in the response letter or could not find the revisions in the updated manuscript. Additionally, the differences between the observations and ERA5 data are still not well explained. Most figures still need significant improvement. A much clear point-to-point response letter and revised manuscript should be added.
Citation: https://doi.org/10.5194/egusphere-2025-3941-RC3 -
AC3: 'Reply on RC3', Songjun Wu, 02 Dec 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-3941/egusphere-2025-3941-AC3-supplement.pdf
-
AC3: 'Reply on RC3', Songjun Wu, 02 Dec 2025
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 1,430 | 193 | 41 | 1,664 | 64 | 25 | 32 |
- HTML: 1,430
- PDF: 193
- XML: 41
- Total: 1,664
- Supplement: 64
- BibTeX: 25
- EndNote: 32
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Doerthe Tetzlaff
Chris Soulsby
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(9124 KB) - Metadata XML
-
Supplement
(1573 KB) - BibTeX
- EndNote
- Final revised paper
This manuscript describes a novel spatially distributed ecohydrologic model ECOTWIN. This model links hydrology with nutrient and isotopic modules. The capability has great potential to improve our understand the terrestrial and aquifer ecosystem interactions. Although similar capability has been attempted in several other studies with different focus such as dissolved organic carbon (DOC), the inclusion of lateral flow with nutrient and carbon cycle is still a challenge in most spatially distributed (eco)hydrological models.
There are some missing information in the current manuscript, such as (1) groundwater level; (2) resolution for evaluation; (3) chemical reaction. See details below. The presentation of some information can be improved, especially some figures are too small to be read.
Distributed, do you mean spatially distributed?
Line 40:
Sentence stoped abruptly without giving reasons.
Line 43:
What is “somehow”
Line 44:
Where is 81% come from? Reference needed
Line 82:
On disk/memory change to online and offline coupling
Figure 1. Is soil zone part of vadose zone?
Line 192:
What if the water table is above the last soil layer?
Equation 16:
If the pond is not next to the river channel, why or how it contribute to the channel flow?
Similarly, in equation 18, if the terrestrial grid cell is far away from the river channel, how does the vadose zone contribute to the channel flow?
Figure 2:
Does not have coordinates. Such as longitude and latitude information.
What are the particular reason why these 17 watersheds were chosen.
Line 346:
Does the model need spin-up to reach steady state?
Line 399:
How do you consider the upstream - downstream dependency in discharge, some watersheds are connected based on Figure 2.
Figure 4:
Show line/color meaning in the Figure as well. Also missing time step information.
The subfigures are too small to read. Maybe move some into SI and only show a few here.
Figure 5. Is the time series watershed averaged? If so, the caption should mention that.
Figure 6:
The GRACE datasets resolution might be too coarse for comparison; there need to be some justification somewhere.
Check other spatial products as well, maybe a table listing all produce with original and resample resolution for reference.
Line 534:
Integrated C++ framework: C++ is a programming language, not a framework.