the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 27 Apr 2022
Dynamic root growth in response to depth-varying soil moisture availability: a rhizobox study
Abstract. Plant roots are highly adaptable, but their adaptability is not included in crop and land surface models. They rely on a simplified representation of root growth, independent of soil moisture availability. Data of subsurface processes and interactions, needed for model set-up and validation, are scarce. Here we investigated soil moisture driven root growth. To this end we installed subsurface drip lines and small soil moisture sensors (0.2 L measurement volume) inside rhizoboxes (45 x 45 x 10 cm). The development of the vertical soil moisture and root growth profiles are tracked with a high spatial and temporal resolution. The results confirm that root growth is predominantly driven by vertical soil moisture distribution, while influencing soil moisture at the same time. Besides support for the functional relationship between the soil moisture and the root density growth rate, the experiments suggest that vertical root extension only takes place if the soil moisture exceeds a threshold value at the root tip. We show that even a parsimonious one-dimensional water balance model, driven by the measured water input and output fluxes, can be convincingly improved by implementing root growth driven by soil moisture availability.
-
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
(737 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(737 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-104', Anonymous Referee #1, 21 May 2022
The authors investigated soil moisture-driven root growth using a rhizobox experiment. Their results confirm that vertical soil moisture distribution regulates the root growth profile, while at the same time, the soil moisture dynamics are impacted by the root growth. This study is mainly supported by the parsimonious one-dimensional water balance model they developed, with the functional relationship between the soil moisture and the root density growth rate being its key. The manuscript is well organized and well written. However, despite this reviewer's great interest in this study, he found there is still some room to extend this study for real case applications (if not at the global scale, then at the regional scale, or at least using a range of in-situ sites). This reviewer will motivate his comments as below:
- This reviewer agrees with the authors that the current LSM should consider the dynamic root growth model, which should take into account the impact of soil moisture. The parsimonious model as the author developed sounds ok, and the current toy example is good for raising the awareness of LSM community on this point. Nevertheless, this is not new. For example, NoahMP1 has a dynamic root growth module, which considers its dependence on soil moisture and soil temperature. The recently published STEMMUS-SCOPE2 model also considers the dynamic root growth as a function of air temperature, soil temperature (via water stress factor) and net assimilation.
With the above, this reviewer is trying to search for the unique contribution from this study. Perhaps the functional relationship between the soil moisture and the root density growth rate could be one significant contribution. On the other hand, this reviewer is not clear or the authors did not show how the current approach can be applied elsewhere in the field. Particularly, the most important parameters (e.g., u2/u3) are determined via sensitivity analysis, which makes this reviewer wonder how to obtain these parameters in the field, and further at regional/global scale?
- Please find attachment with some minor comments.
References:
1. Gayler, S., Wöhling, T., Grzeschik, M., Ingwersen, J., Wizemann, H.-D., Warrach-Sagi, K., Högy, P., Attinger, S., Streck, T., and Wulfmeyer, V. (2014), Incorporating dynamic root growth enhances the performance of Noah-MP at two contrasting winter wheat field sites, Water Resour. Res., 50, 1337– 1356, doi:10.1002/2013WR014634.2. Wang, Y., Zeng, Y., Yu, L., Yang, P., Van der Tol, C., Yu, Q., Lü, X., Cai, H., and Su, Z.: Integrated modeling of canopy photosynthesis, fluorescence, and the transfer of energy, mass, and momentum in the soil–plant–atmosphere continuum (STEMMUS–SCOPE v1.0.0), Geosci. Model Dev., 14, 1379–1407, https://doi.org/10.5194/gmd-14-1379-2021, 2021
-
AC1: 'Reply on RC1', Cynthia Maan, 21 Jul 2022
Dear reviewer,
Thank you for your constructive comments. Please find our reply to your remarks below
- The reviewer commented that there is still some room to extend the study for real case applications. We agree with the reviewer that there is room for extension to real case applications. However, we think that the results presented here – the parsimonious model with very direct relationships between the two state variables (soil moisture and root density) and the experimental setup for validation (including `constant’ and highly controlled flow rates through the drip lines) could already be ready to share with the public.
- The reviewer also notes that our parsimonious model is not new, while other LS models (e.g. NoahMP and STEMMUS-SCOPE model) have a dynamic root growth module.
The authors agree that the proposed model is not the first one considering dynamic root growth. However, only few models take soilmoisture driven root growth into account. STEMMUS scope considers the effect of the total root zone soil moisture availability, but does not consider vertical variations in the soil moisture profile. The authors are aware (now) of only the coupled NoahMP/VOM-ROOT that considers the impact of the vertical soil moisture profile on the root growth profile. Last model makes another assumption about the bulk root growth.
Furthermore, the existing dynamic root growth models are unambiguous about first principles. E.g. should the bulk root growth rate be related to the biomass growth rate, as Adiko et al. suggest, to the difference between the water demand and the actual water uptake (Schymanski et al./ NoahMP/VOM-ROOT; growth of the root bulk in case of water shortage), or could it be related to the soil water status directly/only (as proposed here). It generally is a good idea to keep the model as simple as possible, while including the most dominant (proven) relationships. Therefore we propose a very simplified model, which nevertheless includes a dominant functional relationship between root growth and soil moisture.
Unique about this study is the following:
- In the last version of the model we propose to link the (local) root growth rate very directly to the local soil moisture, with no other (e.g. above ground) dependencies. This is a new proposal/hypothesis.
- In our experiment we maintain the water flow relatively constant. In other studies irrigation is concentrated in short periods of irrigation or rain, which makes it more difficult to distinguish between signals of the forcing and of the response of the plant. By applying constant `irrigation forcing’ it is easier to recognize the adjustment processes to constant irrigation forcing.
Also by simulating a single plant and perform direct measurements on both the soil moisture and root density growth rates simultaneously, i.e. it can give more insight in the direct interactions between the two variables. This information is missing when root densities/weights are measured after the soil moisture measurements. Also we measure both variables (roots densities and soil moisture) directly instead of indirect proxies (like evaporation or temperature) that depend on a translation by more complicated models with more assumptions.
- The reviewer is not clear or the authors did not show how the current approach can be applied elsewhere in the field. Particularly, the most important parameters (e.g., u2/u3) are determined via sensitivity analysis, which makes this reviewer wonder how to obtain these parameters in the field, and further at regional/global scale?
We propose two different model-versions for root growth that can be applied in different settings/situations.
The first one is only the relationship between (vertical varying) soil moisture and root density growth. This part of the model is easy to couple with a LSM, because the vertical extension velocity u1 is a parameter that is already included in LS models with a dynamic root growth module (however not yet including soil moisture driven root growth, like SWAP), and is therefore already estimated/determined for different plants and crops. The most easy way to implement our model in the most basic form (only the impact of vertical varying soil moisture) is to incorporate only equation 1 (possibly in combination with adding a threshold value at the root tip as we also propose), and leave the bulk (depth integrated) root growth of the LSMs intact.
The water extraction rate per centimeter of roots in saturated conditions u2 was only introduced to make a simple translation between the vertical root profiles and the (measureable) soil moisture. Most existing LSMs contain a much better described and well validated module to calculate water extraction from given root profile and soil moisture profiles. So this part of the model is not meant for incorporation or practical application.
In the second version we propose to link the root growth rate directly to the normalized soil moisture. This is a new proposal/hypothesis. We however fully agree that further measurements and experiments are needed to test this hypothesis, and to determine the root density growth rate in optimal conditions (saturation) u3 for different plant types. Note however, that we make use of only one and the same value to simulate the root growth in all time spans and at each level in Figure 8 . In practical applications, to implement this model in LSMs, it will be necessary to maximize the total root growth to (a part of) the total biomass growth.
Citation: https://doi.org/10.5194/egusphere-2022-104-AC1
-
RC2: 'Comment on egusphere-2022-104', Anonymous Referee #2, 28 May 2022
Dear Editor,
This work investigates the driven factor of root growth. It provide a simply but practical method to model root growth. Here the manuscript needs some improvements to enhance its clarity and readability. The specific comments are as follows.
Specific comments:
L 21-24: some models described the maximum rooting depth by a linearly increasing function with accumulated temperature.
L 47: As I know, Some crop (e.g. Spacsys and STICS) and land surface models (e.g. CLM 5.0) have implemented dynamic root growth.
L 85: equation (1): the theta_n in bottom right should be followed by 'dz'
L 142-144: It should be the reason why exponential root profile is widely used in crop models and land surface models.
Fig. 7 and Fig. 8: “rooting dencity” should be “root density”.
Citation: https://doi.org/10.5194/egusphere-2022-104-RC2 -
AC2: 'Reply on RC2', Cynthia Maan, 21 Jul 2022
Dear reviewer,
Thank you for your constructive comments.
- The reviewer comments that some models described the maximum rooting depth by a linearly increasing function with accumulated temperature.
The authors agree that temperature can be a dominant driving factor. We propose to add the following line after “In most crop … i.e., both independently of soil moisture.” :
”Some exceptional models treat root growth more dynamically by relating root growth to soil related parameters as accumulated temperature or root zone soil moisture. Models that take the vertical profiles of soil moisture into account, however, are scarce.”
2. As I know, some crop (e.g. Spacsys and STICS) and land surface models (e.g. CLM 5.0) have implemented dynamic root growth
These models have indeed a dynamic root growth component, but do not include a dependency on the vertical variation in soil moisture.
3. The relative insensitive soil moisture should be the reason why exponential root profile is widely used in crop models and land surface models
Our data indeed suggests that roots do not only `follow moisture’, but the moisture also `finds roots’ if roots are locally not present (by diffusion), reducing the resulting error. However, the experimental setup in an enclosed box favores this process unrealistically. At a larger scale the ‘root follows moisture routine’ is expected to have a more pronounced effect on the soil moisture.
Citation: https://doi.org/10.5194/egusphere-2022-104-AC2
-
AC2: 'Reply on RC2', Cynthia Maan, 21 Jul 2022
-
RC3: 'Review of egusphere-2022-104', Stan Schymanski, 31 May 2022
Dear editor,
The manuscript by Maan et al. describes an experiment where a maize plant was grown in a transparent rhizobox and root growth was monitored along with soil moisture in the vertical soil profile. The authors compare different model simulations with the observations, one with a prescribed exponential root distribution, and several simulations based on dynamic root distributions with different levels of observation data assimilated into the simulations. The authors find that root growth rates follow the soil moisture distribution, and that this could be simulated by assuming that root growth rate is a function of soil moisture and only happens at a normalized water content greater than 0.075.
The results presented are interesting and potentially insightful, but unfortunately, the paper is not well organized and lacks details and clarity. Therefore I found the results difficult to interpret and the conclusions difficult to verify. The paper consists of essentially three separate methods and results sections, one for each model version. It is not clear which parameters were calibrated in each model run and how, e.g. whether in 2.4 only u3 was calibrated, or also u2, and on what data precisely. It is also not clear in which soil depth how much irrigation was applied when, and whether the vertical soil moisture differences were induced by differential irrigation or root water uptake. The main conclusion is that root growth is strongly determined by soil moisture, with high soil moisture promoting root growth. However, since irrigation was "adjusted in steps to follow the plants growth and demand for water" (L68), and the soil moisture increases over time in the deeper soil (Fig. 2C), I am a bit confused in how far the vertical soil moisture distribution was controlled and if it really triggered differential root growth. More information about the experimental strategy would be helpful.
I would propose to consolidate the methods and results in two separate sections as per standard convention, and include sufficient details about the experimental and modelling methods to enable reproduction of the experiment and anlysis. The authors promise to publish the data for producing the plots and results in the future, and provided links for the referees, but unfortunately, the links did not work for me, so I was not able to assess if the promised material will be indeed adequate to enable reproducibility of results.
I added detailed comments in an annotated pdf-file, as they are too many to list here. Some of the equations provided are incorrect, but I cannot tell if they were correct in the analysis. For linguistic glitches, I just highlighted bits of text in yellow. I hope that my review will help to revise the manuscript and add more clarity about the methods and results.
-
AC3: 'Reply on RC3', Cynthia Maan, 22 Jul 2022
Dear reviewer, dear dr. Schymanski
Thank you for your remarks and suggestions, your constructive comments are very helpful. Please find our reply below as well as in the attachment.
- The reviewer found the results difficult to interpret and the conclusions difficult to verify. `The paper consists of essentially three separate methods and results sections, one for each model version. It is not clear which parameters were calibrated in each model run and how, e.g. whether in 2.4 only u3 was calibrated or also u2, and on what data precisely. “
In each section, one extra parameter is calibrated, while the other is taken from the previous step:
- In section 2.2 (‘root follows moisture’ model) the vertical extension velocity u1 is calibrated on the measured root growth profile/distribution. Note however that the results are not that sensitive to this parameter (figure 4)
- In section 2.3 (soil moisture and water uptake model) The water extraction rate per centimeter of roots in saturated conditions u2 is calibrated based on the soil moisture data, while u1 is taken from section 2.2 (based on the root growth data)
- In section 2.4 (the prognostic model) the root density growth in saturated (assumed to be ideal) conditions u3 is calibrated on the root density data. u1 (calibrated against the same root growth data) and u2 are taken directly from the previous sections and thus not calibrated for the `updated’ version of the model. We fully agree that it would be necessary to repeat these experiments for different datasets, in order to get more confidence about the parameter values.
This will be clarified in the MS.
2. It is also not clear in which soil depth how much irrigation was applied when, and whether the vertical soil moisture differences were induced by differential irrigation or root water uptake.
The authors will make the levels of irrigation visible in figure 2. The irrigation level was increased in steps (but normally remained constant in the order of days). We will indicate the days at which the irrigation level was adjusted in figure 2. At each moment only one level is irrigated; so the flow rate at the applied level equals the total flow rate which is indicated in figure 2A. We noticed that the dotted lines in the first panels in figure 2 C were missing by omission, this will be corrected in the manuscript. During the first 14days water was supplied at 0cm, on day 15 the level was adjusted to 10 cm depth (the roots already appeared around the 10 cm line around day 8, but until the 17th there was no significant growth in the 10cm-20cm interval. We will make the sequence of steps, and the implication for causality, better visible in the MS.
3. The reviewer is a bit confused in how far the vertical soil moisture distribution was controlled and if it really triggered differential root growth. More information about the experiment strategy would be helpful.
The vertical soil moisture differences were induced by a combination of differential irrigation and differential water uptake. The irrigation supply, controlled at (relatively constant) low rates during the experiment, is the truly independent system driver (however not constant). The irrigation depth was generally adjusted only after the rooting depth seemed to stabilize (so the soil moisture was the first factor that changed, followed by active root growth). The spatial correlation between the water content and the root density growth rates (figure 2) suggests a causal relationship. By adding more information about the sequence of events, we will be able to clarify this.
4.The reviewer proposes to consolidate the methods and results in two separate sections as per standard convention, and include sufficient details about the experimental and modeling methods to enable reproduction of the experiment and analysis.
The authors agree with the reviewer that sufficient details should be given to enable reproduction. The review process helped the authors to identify the missing information in the paper, which we will add to the MS. The authors are able to provide sufficient details to enable reproduction of the experiment and analysis. However, the authors think that sufficient clarity can be given without changing the structure of the paper.
5. The authors promised to publish the data for producing the plots and results in the future, and provided links for the referees, but unfortunately, the links did not work for the reviewer,, so he was not able to assess if the promised material will be indeed adequate to enable reproducibility of results.
The authors are very sorry that the data has not been visible before. We will have an active DOI soon, but the data is now `already' accessible via the following link: https://figshare.com/s/f206405c95ebfbbf9bf7
6. The reviewer notes that some of the equations provided are incorrect, but that he cannot tell if they were correct in the analysis.
The authors revised the equations in the code, and were delighted to find that the indicated errors were omissions in the text (not in the code). However, the remarks led us to discover another omission in the code; the `effective saturation van Genuchten’ (equation 2) was used instead of `soil wetness’ in equation 6 and 7. This has been corrected and the results were analyzed. Fortunately, the resulting differences in the graphs are small and do not affect the main results and conclusions of the paper. All the figures and calibration values that are affected (figure 5,6,8) will be adjusted in the manuscript.
Citation: https://doi.org/10.5194/egusphere-2022-104-AC3
-
AC3: 'Reply on RC3', Cynthia Maan, 22 Jul 2022
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-104', Anonymous Referee #1, 21 May 2022
The authors investigated soil moisture-driven root growth using a rhizobox experiment. Their results confirm that vertical soil moisture distribution regulates the root growth profile, while at the same time, the soil moisture dynamics are impacted by the root growth. This study is mainly supported by the parsimonious one-dimensional water balance model they developed, with the functional relationship between the soil moisture and the root density growth rate being its key. The manuscript is well organized and well written. However, despite this reviewer's great interest in this study, he found there is still some room to extend this study for real case applications (if not at the global scale, then at the regional scale, or at least using a range of in-situ sites). This reviewer will motivate his comments as below:
- This reviewer agrees with the authors that the current LSM should consider the dynamic root growth model, which should take into account the impact of soil moisture. The parsimonious model as the author developed sounds ok, and the current toy example is good for raising the awareness of LSM community on this point. Nevertheless, this is not new. For example, NoahMP1 has a dynamic root growth module, which considers its dependence on soil moisture and soil temperature. The recently published STEMMUS-SCOPE2 model also considers the dynamic root growth as a function of air temperature, soil temperature (via water stress factor) and net assimilation.
With the above, this reviewer is trying to search for the unique contribution from this study. Perhaps the functional relationship between the soil moisture and the root density growth rate could be one significant contribution. On the other hand, this reviewer is not clear or the authors did not show how the current approach can be applied elsewhere in the field. Particularly, the most important parameters (e.g., u2/u3) are determined via sensitivity analysis, which makes this reviewer wonder how to obtain these parameters in the field, and further at regional/global scale?
- Please find attachment with some minor comments.
References:
1. Gayler, S., Wöhling, T., Grzeschik, M., Ingwersen, J., Wizemann, H.-D., Warrach-Sagi, K., Högy, P., Attinger, S., Streck, T., and Wulfmeyer, V. (2014), Incorporating dynamic root growth enhances the performance of Noah-MP at two contrasting winter wheat field sites, Water Resour. Res., 50, 1337– 1356, doi:10.1002/2013WR014634.2. Wang, Y., Zeng, Y., Yu, L., Yang, P., Van der Tol, C., Yu, Q., Lü, X., Cai, H., and Su, Z.: Integrated modeling of canopy photosynthesis, fluorescence, and the transfer of energy, mass, and momentum in the soil–plant–atmosphere continuum (STEMMUS–SCOPE v1.0.0), Geosci. Model Dev., 14, 1379–1407, https://doi.org/10.5194/gmd-14-1379-2021, 2021
-
AC1: 'Reply on RC1', Cynthia Maan, 21 Jul 2022
Dear reviewer,
Thank you for your constructive comments. Please find our reply to your remarks below
- The reviewer commented that there is still some room to extend the study for real case applications. We agree with the reviewer that there is room for extension to real case applications. However, we think that the results presented here – the parsimonious model with very direct relationships between the two state variables (soil moisture and root density) and the experimental setup for validation (including `constant’ and highly controlled flow rates through the drip lines) could already be ready to share with the public.
- The reviewer also notes that our parsimonious model is not new, while other LS models (e.g. NoahMP and STEMMUS-SCOPE model) have a dynamic root growth module.
The authors agree that the proposed model is not the first one considering dynamic root growth. However, only few models take soilmoisture driven root growth into account. STEMMUS scope considers the effect of the total root zone soil moisture availability, but does not consider vertical variations in the soil moisture profile. The authors are aware (now) of only the coupled NoahMP/VOM-ROOT that considers the impact of the vertical soil moisture profile on the root growth profile. Last model makes another assumption about the bulk root growth.
Furthermore, the existing dynamic root growth models are unambiguous about first principles. E.g. should the bulk root growth rate be related to the biomass growth rate, as Adiko et al. suggest, to the difference between the water demand and the actual water uptake (Schymanski et al./ NoahMP/VOM-ROOT; growth of the root bulk in case of water shortage), or could it be related to the soil water status directly/only (as proposed here). It generally is a good idea to keep the model as simple as possible, while including the most dominant (proven) relationships. Therefore we propose a very simplified model, which nevertheless includes a dominant functional relationship between root growth and soil moisture.
Unique about this study is the following:
- In the last version of the model we propose to link the (local) root growth rate very directly to the local soil moisture, with no other (e.g. above ground) dependencies. This is a new proposal/hypothesis.
- In our experiment we maintain the water flow relatively constant. In other studies irrigation is concentrated in short periods of irrigation or rain, which makes it more difficult to distinguish between signals of the forcing and of the response of the plant. By applying constant `irrigation forcing’ it is easier to recognize the adjustment processes to constant irrigation forcing.
Also by simulating a single plant and perform direct measurements on both the soil moisture and root density growth rates simultaneously, i.e. it can give more insight in the direct interactions between the two variables. This information is missing when root densities/weights are measured after the soil moisture measurements. Also we measure both variables (roots densities and soil moisture) directly instead of indirect proxies (like evaporation or temperature) that depend on a translation by more complicated models with more assumptions.
- The reviewer is not clear or the authors did not show how the current approach can be applied elsewhere in the field. Particularly, the most important parameters (e.g., u2/u3) are determined via sensitivity analysis, which makes this reviewer wonder how to obtain these parameters in the field, and further at regional/global scale?
We propose two different model-versions for root growth that can be applied in different settings/situations.
The first one is only the relationship between (vertical varying) soil moisture and root density growth. This part of the model is easy to couple with a LSM, because the vertical extension velocity u1 is a parameter that is already included in LS models with a dynamic root growth module (however not yet including soil moisture driven root growth, like SWAP), and is therefore already estimated/determined for different plants and crops. The most easy way to implement our model in the most basic form (only the impact of vertical varying soil moisture) is to incorporate only equation 1 (possibly in combination with adding a threshold value at the root tip as we also propose), and leave the bulk (depth integrated) root growth of the LSMs intact.
The water extraction rate per centimeter of roots in saturated conditions u2 was only introduced to make a simple translation between the vertical root profiles and the (measureable) soil moisture. Most existing LSMs contain a much better described and well validated module to calculate water extraction from given root profile and soil moisture profiles. So this part of the model is not meant for incorporation or practical application.
In the second version we propose to link the root growth rate directly to the normalized soil moisture. This is a new proposal/hypothesis. We however fully agree that further measurements and experiments are needed to test this hypothesis, and to determine the root density growth rate in optimal conditions (saturation) u3 for different plant types. Note however, that we make use of only one and the same value to simulate the root growth in all time spans and at each level in Figure 8 . In practical applications, to implement this model in LSMs, it will be necessary to maximize the total root growth to (a part of) the total biomass growth.
Citation: https://doi.org/10.5194/egusphere-2022-104-AC1
-
RC2: 'Comment on egusphere-2022-104', Anonymous Referee #2, 28 May 2022
Dear Editor,
This work investigates the driven factor of root growth. It provide a simply but practical method to model root growth. Here the manuscript needs some improvements to enhance its clarity and readability. The specific comments are as follows.
Specific comments:
L 21-24: some models described the maximum rooting depth by a linearly increasing function with accumulated temperature.
L 47: As I know, Some crop (e.g. Spacsys and STICS) and land surface models (e.g. CLM 5.0) have implemented dynamic root growth.
L 85: equation (1): the theta_n in bottom right should be followed by 'dz'
L 142-144: It should be the reason why exponential root profile is widely used in crop models and land surface models.
Fig. 7 and Fig. 8: “rooting dencity” should be “root density”.
Citation: https://doi.org/10.5194/egusphere-2022-104-RC2 -
AC2: 'Reply on RC2', Cynthia Maan, 21 Jul 2022
Dear reviewer,
Thank you for your constructive comments.
- The reviewer comments that some models described the maximum rooting depth by a linearly increasing function with accumulated temperature.
The authors agree that temperature can be a dominant driving factor. We propose to add the following line after “In most crop … i.e., both independently of soil moisture.” :
”Some exceptional models treat root growth more dynamically by relating root growth to soil related parameters as accumulated temperature or root zone soil moisture. Models that take the vertical profiles of soil moisture into account, however, are scarce.”
2. As I know, some crop (e.g. Spacsys and STICS) and land surface models (e.g. CLM 5.0) have implemented dynamic root growth
These models have indeed a dynamic root growth component, but do not include a dependency on the vertical variation in soil moisture.
3. The relative insensitive soil moisture should be the reason why exponential root profile is widely used in crop models and land surface models
Our data indeed suggests that roots do not only `follow moisture’, but the moisture also `finds roots’ if roots are locally not present (by diffusion), reducing the resulting error. However, the experimental setup in an enclosed box favores this process unrealistically. At a larger scale the ‘root follows moisture routine’ is expected to have a more pronounced effect on the soil moisture.
Citation: https://doi.org/10.5194/egusphere-2022-104-AC2
-
AC2: 'Reply on RC2', Cynthia Maan, 21 Jul 2022
-
RC3: 'Review of egusphere-2022-104', Stan Schymanski, 31 May 2022
Dear editor,
The manuscript by Maan et al. describes an experiment where a maize plant was grown in a transparent rhizobox and root growth was monitored along with soil moisture in the vertical soil profile. The authors compare different model simulations with the observations, one with a prescribed exponential root distribution, and several simulations based on dynamic root distributions with different levels of observation data assimilated into the simulations. The authors find that root growth rates follow the soil moisture distribution, and that this could be simulated by assuming that root growth rate is a function of soil moisture and only happens at a normalized water content greater than 0.075.
The results presented are interesting and potentially insightful, but unfortunately, the paper is not well organized and lacks details and clarity. Therefore I found the results difficult to interpret and the conclusions difficult to verify. The paper consists of essentially three separate methods and results sections, one for each model version. It is not clear which parameters were calibrated in each model run and how, e.g. whether in 2.4 only u3 was calibrated, or also u2, and on what data precisely. It is also not clear in which soil depth how much irrigation was applied when, and whether the vertical soil moisture differences were induced by differential irrigation or root water uptake. The main conclusion is that root growth is strongly determined by soil moisture, with high soil moisture promoting root growth. However, since irrigation was "adjusted in steps to follow the plants growth and demand for water" (L68), and the soil moisture increases over time in the deeper soil (Fig. 2C), I am a bit confused in how far the vertical soil moisture distribution was controlled and if it really triggered differential root growth. More information about the experimental strategy would be helpful.
I would propose to consolidate the methods and results in two separate sections as per standard convention, and include sufficient details about the experimental and modelling methods to enable reproduction of the experiment and anlysis. The authors promise to publish the data for producing the plots and results in the future, and provided links for the referees, but unfortunately, the links did not work for me, so I was not able to assess if the promised material will be indeed adequate to enable reproducibility of results.
I added detailed comments in an annotated pdf-file, as they are too many to list here. Some of the equations provided are incorrect, but I cannot tell if they were correct in the analysis. For linguistic glitches, I just highlighted bits of text in yellow. I hope that my review will help to revise the manuscript and add more clarity about the methods and results.
-
AC3: 'Reply on RC3', Cynthia Maan, 22 Jul 2022
Dear reviewer, dear dr. Schymanski
Thank you for your remarks and suggestions, your constructive comments are very helpful. Please find our reply below as well as in the attachment.
- The reviewer found the results difficult to interpret and the conclusions difficult to verify. `The paper consists of essentially three separate methods and results sections, one for each model version. It is not clear which parameters were calibrated in each model run and how, e.g. whether in 2.4 only u3 was calibrated or also u2, and on what data precisely. “
In each section, one extra parameter is calibrated, while the other is taken from the previous step:
- In section 2.2 (‘root follows moisture’ model) the vertical extension velocity u1 is calibrated on the measured root growth profile/distribution. Note however that the results are not that sensitive to this parameter (figure 4)
- In section 2.3 (soil moisture and water uptake model) The water extraction rate per centimeter of roots in saturated conditions u2 is calibrated based on the soil moisture data, while u1 is taken from section 2.2 (based on the root growth data)
- In section 2.4 (the prognostic model) the root density growth in saturated (assumed to be ideal) conditions u3 is calibrated on the root density data. u1 (calibrated against the same root growth data) and u2 are taken directly from the previous sections and thus not calibrated for the `updated’ version of the model. We fully agree that it would be necessary to repeat these experiments for different datasets, in order to get more confidence about the parameter values.
This will be clarified in the MS.
2. It is also not clear in which soil depth how much irrigation was applied when, and whether the vertical soil moisture differences were induced by differential irrigation or root water uptake.
The authors will make the levels of irrigation visible in figure 2. The irrigation level was increased in steps (but normally remained constant in the order of days). We will indicate the days at which the irrigation level was adjusted in figure 2. At each moment only one level is irrigated; so the flow rate at the applied level equals the total flow rate which is indicated in figure 2A. We noticed that the dotted lines in the first panels in figure 2 C were missing by omission, this will be corrected in the manuscript. During the first 14days water was supplied at 0cm, on day 15 the level was adjusted to 10 cm depth (the roots already appeared around the 10 cm line around day 8, but until the 17th there was no significant growth in the 10cm-20cm interval. We will make the sequence of steps, and the implication for causality, better visible in the MS.
3. The reviewer is a bit confused in how far the vertical soil moisture distribution was controlled and if it really triggered differential root growth. More information about the experiment strategy would be helpful.
The vertical soil moisture differences were induced by a combination of differential irrigation and differential water uptake. The irrigation supply, controlled at (relatively constant) low rates during the experiment, is the truly independent system driver (however not constant). The irrigation depth was generally adjusted only after the rooting depth seemed to stabilize (so the soil moisture was the first factor that changed, followed by active root growth). The spatial correlation between the water content and the root density growth rates (figure 2) suggests a causal relationship. By adding more information about the sequence of events, we will be able to clarify this.
4.The reviewer proposes to consolidate the methods and results in two separate sections as per standard convention, and include sufficient details about the experimental and modeling methods to enable reproduction of the experiment and analysis.
The authors agree with the reviewer that sufficient details should be given to enable reproduction. The review process helped the authors to identify the missing information in the paper, which we will add to the MS. The authors are able to provide sufficient details to enable reproduction of the experiment and analysis. However, the authors think that sufficient clarity can be given without changing the structure of the paper.
5. The authors promised to publish the data for producing the plots and results in the future, and provided links for the referees, but unfortunately, the links did not work for the reviewer,, so he was not able to assess if the promised material will be indeed adequate to enable reproducibility of results.
The authors are very sorry that the data has not been visible before. We will have an active DOI soon, but the data is now `already' accessible via the following link: https://figshare.com/s/f206405c95ebfbbf9bf7
6. The reviewer notes that some of the equations provided are incorrect, but that he cannot tell if they were correct in the analysis.
The authors revised the equations in the code, and were delighted to find that the indicated errors were omissions in the text (not in the code). However, the remarks led us to discover another omission in the code; the `effective saturation van Genuchten’ (equation 2) was used instead of `soil wetness’ in equation 6 and 7. This has been corrected and the results were analyzed. Fortunately, the resulting differences in the graphs are small and do not affect the main results and conclusions of the paper. All the figures and calibration values that are affected (figure 5,6,8) will be adjusted in the manuscript.
Citation: https://doi.org/10.5194/egusphere-2022-104-AC3
-
AC3: 'Reply on RC3', Cynthia Maan, 22 Jul 2022
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 390 | 127 | 15 | 532 | 15 | 16 |
- HTML: 390
- PDF: 127
- XML: 15
- Total: 532
- BibTeX: 15
- EndNote: 16
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 132 | 28 |
| Netherlands | 2 | 68 | 14 |
| China | 3 | 48 | 10 |
| Germany | 4 | 35 | 7 |
| India | 5 | 19 | 4 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 132
Cited
Debora Cynthia Maan
Marie-Claire ten Veldhuis
Bas Johannes Henricus van de Wiel
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(737 KB) - Metadata XML