Submit to this Journal Review for this Journal Propose a Special Issue

Article Menu

Share Help Cite Discuss in SciProfiles

Open AccessEditor’s ChoiceArticle

Peer-Review Record

Spatial Evaluation of a Hydrological Model on Dominant Runoff Generation Processes Using Soil Hydrologic Maps

Hydrology 2023, 10(3), 55; https://doi.org/10.3390/hydrology10030055

by Hadis Mohajerani^1,*

, Mathias Jackel^1,2

, Zoé Salm¹, Tobias Schütz²

and Markus C. Casper¹

Reviewer 1: Anonymous

Reviewer 2:

Misagh Parhizkar

Reviewer 3: Anonymous

Hydrology 2023, 10(3), 55; https://doi.org/10.3390/hydrology10030055

Submission received: 6 January 2023 / Revised: 11 February 2023 / Accepted: 17 February 2023 / Published: 22 February 2023

Round 1

Reviewer 1 Report

Comments for author File: Comments.pdf

Author Response

Dear Reviewer

Thank you for your constructive remarks!

Find enclosed our responses point-by-point.

Kind regards,

Markus Casper, Hadis Mohajerani

***********************************************************************************

Reviewers' Comments to the Authors:

Reviewer 1

The study topic is good but additional discussion considering how they calibrated and validated their findings and the gap and new improvements they followed as compared to previous similar research need to be briefly explained to the scientific community. The authors at least need to justify how they validated the developed the results of this study. I suggest some minor changes in the manuscript:

To include sufficient background and relevant references in the introduction part to show the gaps, limitation of the related previous studies and the improvements in this study

Author response: Thank you for the suggestion. We have added the suggested content to the manuscript in the introduction part. The revised text reads as follows:

We incorporated some additional references on [introduction-line 47]: “…Such events may further influence the subsequent processes that govern the runoff re-sponse of a region, particularly in smaller catchments. For example, for water resources management and early flood warning, it is important to make quantitative measurement of the effects of urbanization on surface runoff. Such as, Das & Esraz-Ul-Zannat (2022) approached runoff fluctuations in an urban region using GIS and remote sensing technologies as well as the SCS-CN model. It found out that within the period of 15 years, the region experienced a significant growth of urban impervious areas and a notable decline in vegetated land cover, being the predominant drivers of surface runoff change. The rise in surface runoff was found positively correlated to the growth in urbanization and negatively correlated to the decreased vegetation cover. In another study, Ahmadi-Sani (2022) analyzed the effects of land use change on runoff production by using the SCS-CN approach, remote sensing data, and GIS tools, where runoff was predicted from precipitation, land use, and hydrological soil groups, using the SCS-CN model. According to another study by Lucas-Borja et al (2019), the influence of different land use covers on the soil hydraulic properties was investigated, and consequently, different soil hydrological behaviors to heavy storms were found and therefore different runoff production were observed. ”
In order to give more background on the topic, some parts of the discussion [line 427-437] were moved to the introduction [line 74]: “…This highlights the principle that the model evaluation should thus make use of all sources of data available in a catchment area [23,26]. …Incorporating the knowledge on DRPs into hydrological modeling, Antonetti et al. (2019) applied multiple DRP maps, and presented divergent catchment reactions in terms of DRPs to precipitation events for flash flood predictions. To assess the sensitivity of the hydrograph to the mapping approach, they implemented synthetic runoff simulations, and found out that simulations following the simplified procedures resulted in the strongest deviations from the reference map. Furthermore, in the Nahe catchment in Rhineland-Pfalz, Haag et al. (2016) [42] also integrated spatially distributed information on DRPs based on the classification of Scherrer and Naef (2003) [43] into LARSIM (Large Area Runoff Simulation Model; [44]), for operational flood forecast. They applied different soil parameterizations corresponding to DRPs in the catchment area…”
In order to give more background on the topic, some parts of the discussion [line 499-504] were moved to the introduction [line 79]: “…Thereupon, in the present study, we attempt to utilize an available soil hydrological map for the state of Rhineland-Palatinate (Western Germany) from which the DRPs in a land-scape unit would be identified. This map reflects different flow processes, which sound more plausible based on the site’s characteristics. The reference soil map (i.e., derived from data-mining-based digital soil mapping), in fact, integrated information about soil texture, soil aggregates, chemistry and bulk density that could be used to estimate the soil storage and infiltration capacity, which are eventually utilized to identify areas susceptible to various DRPs considering the topographic characteristics. These maps can then be advantageous to predict potential risk areas like for flooding, pesticide loss or soil erosion, and for model calibration [22]…”

The authors are advised to discuss their findings in detail and need to clearly show how they validated their main findings.

Author response: Thank you for pointing this out. However, we would like to mention that, here, we are not making any model calibration/validation. We intended to compare the pattern (of the simulated dominant runoff generation process) with the patterns (of dominant runoff generation process taken from the regionalized reference digital map), and to recommend our methodology as part of model evaluation (and ultimately integrated in the model validation processes). Here, we do not make any comparison (validation) to the measured data/ field measurements. Regionalized reference map for the dominant runoff generation process is created according to data-driven based digital soil mapping approach (Behrens et al. 2005). Digital soil mapping (or also called soil-landscape modeling) as a tool to generate spatial soil information provides solutions for the growing demand for high-resolution soil maps worldwide. Even in highly developed countries like Germany, digital soil mapping becomes essential due to the decreasing, time-consuming, and expensive field surveys which are no longer affordable by the soil surveys of the individual federal states (more information can be found in Behrens and Scholten (2006).

Behrens, T., & Scholten, T. (2006). Digital soil mapping in Germany—a review. Journal of Plant Nutrition and Soil Science, 169(3), 434-443).

Behrens, T.; Förster, H.; Scholten, T.; Steinrücken, U.; Spies, E.-D.; Goldschmitt, M. Digital Soil Mapping Using Artificial Neural Networks. J. Plant Nutr. Soil Sci. 2005, 168, 21–33.

In our study, we use this regionalized digital map to give us the estimate of the expected dominant runoff process in the landscape, and to consider its patterns to be compared with the simulated patterns of dominant runoff processes by the hydrological model. Nevertheless, spatial soil information obtained from this regionalized reference map is not the exact measured data (field observations) in the real system. In fact, the methodology intends to obtain the plausible results (plausible occurrence of the dominant runoff process on a specific geological location), rather than exact matching the simulations with the field measurements. The regionalized reference map used for the model evaluation is not a first-order digital soil map (which is classically surveyed), but this is a third-order digital soil map (are generated using pedotransfer functions, geostatistical or scorpan approaches).

The authors advised to include the future outlook of the paper for next researcher please indicate future direction

Author response: Thanks for the precise consideration. However, as suggested by the reviewer, we have already mentioned about the next steps and future directions being taken in the last part of the conclusion (page 19, line 584-589), which reads as follows:

“…The present study, nevertheless, will go on towards a comprehensive model calibration procedure considering multiple data sources simultaneously, with the specificity of incorporating the spatial patterns of satellite remote sensing data as well as reference DRP maps in the parameter estimation method to plausibly reproduce the dynamics of the various hydrological processes (e.g., evapotranspiration, soil water storage, and runoff).”

Reviewer 2 Report

Review Report on the Manuscript Number: hydrology-2177804 Title: Spatial evaluation of a hydrological model on dominant runoff generation processes using soil hydrologic maps

The manuscript investigates a process-oriented simulation of hydrological processes and a pattern-oriented evaluation by addressing the spatio-temporal variability of the runoff generation processes using a physically-based distributed hydrological model in the Kronweiler catchment in the Nahe valley of south-west of Germany. Generally, I think that the manuscript is well written and the topic seems to be appropriate for the journal of Hydrology. In the following, I suggest some possible improvements.

1. I recommend summarize more accurate and representative keywords. Like the soil hydrological maps and water resources management, just general keywords which are common, are they appropriate to be key words for this paper?

2. The Abstract is a little confusing for readers. The methods used in this study should be presented together.

3. Line 42-44, It would be good to add some new works about changes in land surface conditions. Please read and add all references as follows:

Lucas-Borja, M.E.; Zema, D.A.; Plaza-Álvarez, P.A.; Zupanc, V.; Baartman, J.; Sagra, J.; de las Heras, J. Effects of different land uses (abandoned farmland, intensive agriculture and forestland) on soil hydrological properties in Southern Spain. Water 2019, 11, 503.

Naser Ahmadi-Sani, Lida Razaghnia and Timo Pukkala. Effect of Land-Use Change on Runoff in Hyrcania. Land 2022, 11(2), 220.

Misagh Parhizkar , Mahmood Shabanpour, Manuel Esteban Lucas-Borja , Demetrio Antonio Zema, Siyue Li, Nobuaki Tanaka, Artemio Cerd. Effects of length and application rate of rice straw mulch on surface runoff and soil loss under laboratory simulated rainfall. International Journal of Sediment Research 36 (2021) 468e478.

Assessing and Modeling Soil Detachment Capacity by Overland Flow in Forest and Woodland of Northern Iran. Misagh Parhizkar, Mahmood Shabanpour, Mohammadreza Khaledian,…. Forests 2020. 11 (1), 65.

Assessing the impacts of land use-land cover changes on direct surface runoff: a remote sensing approach in Khulna City. Water Science & Technology. Palash Chandra Das, Esraz-Ul-Zannat. 2022.

4. The Materials and Methods section is too long and I think that this section could be reduced. I recommend combine 2.2 and 2.3 sections.

5. Line 98, Why did you choose this catchment? According to what? Investigation or other researches?

6. Line 437-441, it would be better if you could report the actual and comparable results, so that your readers can see similar findings, if the experimental conditions were comparable to your study.

7. Line 471 to 474, here you present interesting results, you could explore and discuss these results further.

8. Please explain relation and extension of results obtained from the study to natural conditions in larger scales (scaling).

9. Please check the format of the references carefully.

Author Response

Dear Reviewer

Thank you for your constructive remarks!

Find enclosed our responses point-by-point.

Kind regards,

Markus Casper, Hadis Mohajerani

***********************************************************************************

Reviewer 2

I recommend summarize more accurate and representative keywords. Like the soil hydrological maps and water resources management, just general keywords which are common, are they appropriate to be key words for this paper?

Author response: Thanks for the consideration. We have, accordingly, updated the keywords as follows:

data-mining based digital soil mapping

The Abstract is a little confusing for readers. The methods used in this study should be presented together

Author response: The Abstract has been rewritten (now 99 words)

The aim of this study is to simulate dominant runoff generation processes (DRPs) in a mesoscale catchment in south-western Germany with the physically-based distributed hydrological model WaSiM-ETH and to compare the resulting DRP patterns with a data-mining based digital soil map. The model is parameterized by using 11 Pedo-transfer Functions (PTFs) and driven by multiple synthetic rainfall events. For the pattern comparison, a multiple-component spatial performance metric (SPAEF) was applied. The simulated DRPs show a large variability in terms of land use, applied rainfall rates and the different PTFs which highly influence the rapid runoff generation under wet conditions.

Line 42-44, It would be good to add some new works about changes in land surface conditions. Please read and add all references as follows:

Author response: Thanks for the suggestion. We tried to incorporate the following recommended references based on their relevance to the topic of the current paper. They are mentioned in the introduction [line 47]:

Lucas-Borja, M. E., Zema, D. A., Plaza-Álvarez, P. A., Zupanc, V., Baartman, J., Sagra, J., ... & de las Heras, J. (2019). Effects of different land uses (abandoned farmland, intensive agriculture and forest) on soil hydrological properties in Southern Spain. Water, 11(3), 503.
Ahmadi-Sani, N., Razaghnia, L., & Pukkala, T. (2022). Effect of Land-Use Change on Runoff in Hyrcania. Land, 11(2), 220.
Das, P. C., & Esraz-Ul-Zannat, M. (2022). Assessing the impacts of land use–land cover changes on direct surface runoff: a remote sensing approach in Khulna City. Water Science and Technology, 85(10), 3122-3144.
These references are mentioned in the introduction [line 47]: “…For example, for water resources management and early flood warning, it is important to make quantitative measurement of the effects of urbanization on surface runoff. Such as, a recent study approached runoff fluctuations in an urban region using GIS and remote sensing technologies as well as the SCS-CN model [12]. It found out that within the period of 15 years, the region experienced a significant growth of urban impervious areas and a notable decline in vegetated land cover, being the predominant drivers of surface runoff change. The rise in surface runoff was found positively correlated to the growth in urbanization and negatively correlated to the decreased vegetation cover. Another study analyzed the effects of land use change on runoff production by using the SCS-CN approach, remote sensing data, and GIS tools, where runoff was predicted from precipitation, land use, and hydrological soil groups, using the SCS-CN model [13]. According to another study, the influence of different land use covers on the soil hydraulic properties was investigated, and consequently, different soil hydrological behaviors to heavy storms were found and therefore different runoff production were observed [14].”

The Materials and Methods section is too long and I think that this section could be reduced. I recommend combine 2.2 and 2.3 sections

Author response: The two sections were combined, and the text has been reduced.

Line 98, Why did you choose this catchment? According to what? Investigation or other researches?

Author response: the catchment has a long and reliable discharge record and high-resolution rainfall data. In addition, it has a high topographic gradient and (process) variability (flat mountain areas, steep slopes and alluvial plains).

Line 437-441, it would be better if you could report the actual and comparable results, so that your readers can see similar findings, if the experimental conditions were comparable to your study.

Author response: thanks for raising this question. It is important to remind that we have not done any experimental work to obtain the dominant runoff process on the specific point, so there was in fact no experimental condition that can be compared with other field findings. We only got some estimate derived from the regionalized reference map giving the expected dominant runoff process in the real system. This regionalized reference map contains a broad range of parameters such as slope, soil, geology, distance to the river etc. and translates this spatial information to estimate the soil hydraulic properties, and consequently, make plausible estimates of patterns of dominant runoff processes in various spatial scales. This regionalized map is a model too that uses artificial intelligence, and therefore, this is not a direct field measurement, but rather an approximation of reality. This then could allow for a spatial pattern evaluation of our simulated dominant runoff process. Accordingly, we have attempted to give a background and discuss the actual works relevant to the spatial pattern comparison in hydrological modeling.

Line 471 to 474, here you present interesting results, you could explore and discuss these results further

Author response: Since this is a simulation result and not a field observation, further discussion would be even more speculative.

Please explain relation and extension of results obtained from the study to natural conditions in larger scales (scaling)

Author response: Thanks for raising this question. Nevertheless, we would like to emphasize that what has been done here was scaling. This methodology makes it also possible to rely on measured data (but on the small scale). However, what we wanted to do was to extent it to the larger scale like a larger catchment. In the large catchments, you can only rely on data-driven digital maps and not on the field measurements. Therefore, this map has already extrapolated locally mapped information to larger scales (regionalization). Therefore, these maps are like interface between field observation (local expert knowledge) and the simulation results.

The access to the regionalized reference map for dominant runoff process and incorporating that in the proposed methodology allows for scaling the results to larger (or smaller) catchments. This digital soil map, however, doesn’t exactly give the precise measured values (experimental findings) in the field. The experimental catchments are usually way too smaller (around 1-2 km²) than our study area (about 64 km²), where field measurements are not easily possible. With the data-driven based digital soil mapping, however, the runoff processes can be regionalized using data mining technics. Depending on the modeling purpose (e.g, water management, water pollution, flood warning, water supply in the catchment areas of different sizes), the modelers are called to consider the plausibility of the simulated dominant runoff generation processes in the hydrological model validation procedures.

We intended to compare the pattern (of the simulated dominant runoff generation process) with the patterns (of dominant runoff generation process taken from the regionalized reference digital map), and to recommend our methodology as part of model evaluation (and ultimately integrated in the model validation processes). Here, we do not make any comparison (validation) to the measured data/ field measurements. Regionalized reference map for the dominant runoff generation process is created by a data-driven digital soil mapping approach. Digital soil mapping (or also called soil-landscape modeling) as a tool to generate spatial soil information provides solutions for the growing demand for high-resolution soil maps worldwide (Behrens et al., 2005). Even in highly developed countries like Germany, digital soil mapping becomes essential due to the decreasing, time-consuming, and expensive field surveys which are no longer affordable by the soil surveys of the individual federal states (more information can be found in Behrens and Scholten (2006).

Behrens, T., & Scholten, T. (2006). Digital soil mapping in Germany—a review. Journal of Plant Nutrition and Soil Science, 169(3), 434-443).

Behrens, T.; Förster, H.; Scholten, T.; Steinrücken, U.; Spies, E.-D.; Goldschmitt, M. Digital Soil Mapping Using Artificial Neural Networks. J. Plant Nutr. Soil Sci. 2005, 168, 21–33.

Please check the format of the references carefully.

Author response: Thanks for reminding this. We changed the format of the references to the template of MPDI journals

Reviewer 3 Report

Review of the manuscript:

Spatial evaluation of a hydrological model on dominant runoff generation processes using soil hydrologic maps.

Submitted by Hadis Mohajerani, Mathias Jackel, Zoé Salm, Tobias Schütz and Markus C. Casper to Hydrology (MDPI)

General comments

This manuscript relates a sensitivity study to compare the spatial distribution of dominant runoff generation processes as simulated by a distributed hydrological model on one side and as expert knowledge-based, using soil hydrologic maps on the other side. The set of simulations have been performed using various combinations of pedo-transfert functions and a range of synthetic rainfall events. The comparison include the use of an objective function defined across space.

The subject is relevant and worth being explored and reported. Unfortunately, the writing is not easy, and the methodology is unclear relative to several points. Much of the long Discussion section should appear in the Introduction to give a better idea of similar studies.

There is a shift in the meaning of “observed” toward “observed on a reference map”. This shift should be avoided as the ambiguous “perceived reality”.

Specific comments

Abstract

The writing is a bit heavy. Squeezed in the first 5 lines of the abstract are 4 occurrences of “process”, 3 of “spatial” and 2 of “pattern”. These words appear again in the remaining lines. “… to perform a process-oriented simulation oh hydrological processes …” sounds like a tautologism. Same for “... Runoff pattern varied from smaller to larger amounts of …”.

What could be the meaning of “we assimilated a … performance metric.” (Also Line 82)

1. Introduction

Line 27: “… in the community of hydrological modelers …”

2. Materials and method

2.1

Lines 106-111: at first sight, grassland cover a larger area than cropland; do you count both as arable land?

Figure 1. Error in the label.

A situation map in Germany could be useful.

2.2

Lines 161-162: 1 to 7 and8 to 10

2.3

Lines 204-218: Is this map part of the same project? Is there field work in the Kroneweiler catchment?

2.6

The better the similarity between the patterns, the closer the efficiency metric value approaches unity.

Line 302 maybe “except” instead of “expect”

Line 332: mix-up with the figure caption

4.4 Error labelling this section. Is it “4.”?

Figure 6: Is the reference map a static one? No simulation with a 100 mm/7h rainfall event?

Line 451: steeper slope

Author Response

Dear Reviewer,

Thank you for your constructive remarks!

Find enclosed our responses point-by-point.

Kind regards,

Markus Casper, Hadis Mohajerani

********************************************************************************

Reviewer 3

General comments:

The subject is relevant and worth being explored and reported. Unfortunately, the writing is not easy, and the methodology is unclear relative to several points. Much of the long Discussion section should appear in the Introduction to give a better idea of similar studies.

Author response: We appreciate this comment. To address that, we made some changes in the text that can be read as follows:

In order to draw a clearer picture of the method we implement and to describe the way that regional runoff maps (data-driven-based digital mapping) are derived and employed to serve the modeling purposes, we have improved the methodology description in various parts of the paper (introduction and method section).
To address the comment, and in order to give more background on the topic, some parts of the discussion [line 427-437] were moved to the introduction [line 74]: “…This highlights the principle that the model evaluation should thus make use of all sources of data available in a catchment area [23,26]. …Incorporating the knowledge on DRPs into hydrological modeling, Antonetti et al. (2019) applied multiple DRP maps, and presented divergent catchment reactions in terms of DRPs to precipitation events for flash flood predictions. To assess the sensitivity of the hydrograph to the mapping approach, they implemented synthetic runoff simulations, and found out that simulations following the simplified procedures resulted in the strongest deviations from the reference map. Furthermore, in the Nahe catchment in Rhineland-Pfalz, Haag et al. (2016) [42] also integrated spatially distributed information on DRPs based on the classification of Scherrer and Naef (2003) [43] into LARSIM (Large Area Runoff Simulation Model; [44]), for operational flood forecast. They applied different soil parameterizations corresponding to DRPs in the catchment area…”
To address the comment, and In order to give more background on the topic, some parts of the discussion [line 499-504] were moved to the introduction [line 79]: “…Thereupon, in the present study, we attempt to utilize an available soil hydrological map for the state of Rhineland-Palatinate (Western Germany) from which the DRPs in a land-scape unit would be identified. This map reflects different flow processes, which sound more plausible based on the site’s characteristics. The reference soil map (i.e., derived from data-mining-based digital soil mapping), in fact, integrated information about soil texture, soil aggregates, chemistry and bulk density that could be used to estimate the soil storage and infiltration capacity, which are eventually utilized to identify areas susceptible to various DRPs considering the topographic characteristics. These maps can then be advantageous to predict potential risk areas like for flooding, pesticide loss or soil erosion, and for model calibration [22]…”

There is a shift in the meaning of “observed” toward “observed on a reference map”. This shift should be avoided as the ambiguous “perceived reality”.

Author response: thanks for your remark. We understand that using these terms might have caused hindrance to the description of methodology and results. Nevertheless, we tried to make them clear and coherent in the entire text. What we really meant was to explain the fact that although the regionalized reference map is NOT the measured values taken through the direct field work (no direct observation), but we would consider the spatial patterns of dominant runoff generation processes taken from this map as the plausible expected dominant runoff patterns in the area that can be compared to corresponding simulated patterns (by hydrological model). The regionalized reference map used for the model evaluation is not a first-order digital soil map (which is classically surveyed), but this is a third-order digital soil map (are generated using pedotransfer functions, geostatistical or artificial intelligence approaches). Doing this patterns comparison and using the SPEAF metric to quantify the spatial similarities would promote a perspective for hydrological model evaluation on the spatial patterns. We adapted our wording throughout the text.

Specific comments

Abstract:

The writing is a bit heavy. Squeezed in the first 5 lines of the abstract are 4 occurrences of “process”, 3 of “spatial” and 2 of “pattern”. These words appear again in the remaining lines. “… to perform a process-oriented simulation oh hydrological processes …” sounds like a tautologism. Same for “... Runoff pattern varied from smaller to larger amounts of …”.

Author response: The Abstract has been rewritten entirely.

“The aim of this study is to simulate dominant runoff generation processes (DRPs) in a mesoscale catchment in south-western Germany with the physically-based distributed hydrological model WaSiM-ETH and to compare the resulting DRP patterns with a data-mining based digital soil map. The model is parameterized by using 11 Pedo-transfer Functions (PTFs) and driven by multiple synthetic rainfall events. For the pattern comparison, a multiple-component spatial performance metric (SPAEF) was applied. The simulated DRPs show a large variability in terms of land use, applied rainfall rates and the different PTFs which highly influence the rapid runoff generation under wet conditions.”

What could be the meaning of “we assimilated a … performance metric.” (Also Line 82)

Author response: this sentence intended to say that we took in all the information on DRP patterns obtained from hydrological model and those from the regional refence map and adopted a metric to compare those patterns. This would be perhaps more expressive if we change it to: “we adopted a…performance metric.”. We made this change in the text too.

Introduction:

Line 27: “… in the community of hydrological modelers …”