the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 14 Mar 2023
A modified version of RothC to model the direct and indirect effects of rice straw mulching on soil carbon dynamics, calibrated in a Mediterranean citrus orchard
Abstract. Mulching of agricultural soils has been identified as a viable solution to sequester carbon into the soil, increase soil health and fight desertification; as such, it is an interesting option for carbon farming in Mediterranean areas. Models are used to project the effects of agricultural practices on soil organic carbon in the future for various soil and climatic conditions, and to help policy makers and farmers assess the best way to implement carbon farming strategies. Here, we modify the widely used RothC model to include mulching practices and their direct and indirect effects on soil organic matter input, soil temperature changes, and soil hydraulic balance. We then calibrated and validated our modified RothC (RothC_MM) using the dataset collected in a field mulching experiment described in detail in a companying article, and used the validated RothC_MM to estimate the expected soil carbon sequestration by year 2050 due to mulching for the Valencian community (Spain). Our results show that RothC_MM improved the fit with experimental data with respect to basic RothC, and was able to predict SOC and CO2 observations taken in the field, and to model the effects of mulch on soil temperature and soil water content.
This preprint has been withdrawn.
-
Withdrawal notice
This preprint has been withdrawn.
-
Preprint
(1118 KB)
-
Supplement
(200 KB)
-
This preprint has been withdrawn.
- Preprint
(1118 KB) - Metadata XML
-
Supplement
(200 KB) - BibTeX
- EndNote
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-298', Anonymous Referee #1, 29 Mar 2023
The paper submitted by Pesce et al is an attempt to calibrate the RothC model over two mulching experiments in the Mediterranean region. The topic is of interest and fits with the scope of the journal but the study suffers from several methodological flaws making this study inappropriate for publication in its present form. I list them below:
- In general, a lot of information are missing making the evaluation of the work very difficult. For instance, the RothC_N version is not described and the reader needs to read the Farina et al. paper to understand the starting point of the version used here. Other problem, the author wrote that TSMD is calculated using pedotransfer function based on the Van genuchten equations. This is way too vague, much more details are needed here. In particular, the equations, the parameters values, how the pedotransfer functions were calibrated, etc. Another problematic description is how the MT parameter was calibrated. This is a key part of the paper and a much better descriptions is needed. Because of all the missing information I doubt that the study can be reproduced by someone else.
- Another point that must be clarified is the initialization procedure. The author wrote that they used a spinup to reach equilibrium (section 2.2.2) but then at line 244 they wrote that the proportion of the pools obtained with the spinup is multiplied by the observed stocks. It makes the initialization procedure quite confusing and a better description of what was done to initialize the model is absolutely needed. If the exact procedure is indeed a spinup to reach equilibrium, this is in contradiction with the statement in the introduction claiming at line 35 that cropland soils are loosing carbon.
- Something that also surprised me is written at line 284 where the authors explained that they had to increase the input up to 0.1 and 0.5 tC ha-1 for the mulch experiment. It suggests that the extra carbon inputs due to mulching are not measured and therefore the constrain on the stock given by the model is very poor.
- Another problematic approach is how they used the SOC data. If I understood well, they used SOC data obtained during 22 months to calibrate the model and the data from month 22 to 36 to evaluate. I think it is inappropriate because this period is way too short to detect any SOC changes and therefore the model calibration and evaluation are basically done using the same value.
- The calibration methods is not described enough. How it has been done, simple tunning, least square method, MCMC…?
- At line 296, where the gas chamber always on site? If yes how do the authors managed the temperature increase and the absence of evaporation within the chambers?
- What are the value of the C input in the model, where do they come from?
- About the future scenario, to which IPCC scenario does it corresponds? What is the scenario for C inputs?
- Why the starting point in the SOC data for the control and the mulch treatment are different?
- The author used different versions of the model but at the end they only present one. Or all the versions needs to be presented or only the final version needs to be described.
- When comparing the CO2 fluxes, you are comparing apples and oranges because the model is providing heterotrophic respiration and the data are given soil respiration including roots respiration. This is acknowledged by the authors at line 351 but I think it should be avoided and removed from the manuscript
- I disagree with the statement in the introduction saying that RothC is more simple than CENTURY (line 63). The approach in the two models is similar (pool based, first order kinetics, etc.). Much more complex model could have been found to support your statement in particular those which are driven by microbial mechanisms (Millenial, MIMICS, etc.).
- I also disagree with the statement at line 378-379. If you use site data to fit empirical parameters to then run local simulations, I don't think you really integrate any mechanisms related to mulching. You just tuned your model regarding local conditions and there is no guarantee that such a set of parameters would work in a different situation even under that same crop management and the same climate.
Citation: https://doi.org/10.5194/egusphere-2023-298-RC1 -
CC1: 'Reply on RC1', Simone Pesce, 24 Apr 2023
Dear Anonymous Referee #1, thank you for your time and effort in reviewing our manuscript. Your feedback will help improving the quality of our manuscript, and we appreciate your thoughtful and constructive comments. Below you will find some answer to your comment.
1- Dear reviewer, the RothC N model was not the focus of our study since it is a model developed in Farina et al 2013. Anyway, we will provide a better description of the model and pedotransfer function in the revised manuscript. Furthermore, we will include the Van Genuchten parameters in the supplementary material to enhance the clarity of our work. Regarding the second part of your question, we would like to direct your attention to our response to question N 7, where we have provided a detailed explanation
2- To clarify, the initial conditions come from the measured SOC stocks in the field. Unfortunately, however, there is no way to measure the RothC pools directly; as such, a spin-up run is required to estimate how to distribute the measured initial SOC into the four RothC pools. This step is crucial to accurately simulate the carbon dynamics in the soil and obtain reliable results. By doing so, we can better understand the fate of carbon in the soil and its impact on the environment. The spin up approach is a widely used method to initialize the RothC model (Pulcher et al 2022, Nemo et al 2017). We appreciate your feedback and will ensure that this process is described in detail in our manuscript.
3- Dear reviewer, we had to rise the carbon input to the soil in both bare soils and mulched soils, in both sites, by the same amount. Thus, it is safe to say that the problem is not in the estimated carbon input from the mulch, but rather the value we used for the carbon input to the soil from the citrus orchard (which came from literature). We will include this description in the discussion section of the revised manuscript.
4- Dear reviewer, we acknowledge that the period of the experiment was only three years. However, we preferred to split the dataset into calibration and validation data to check the accuracy and performance of the model, rather than performing calibration only. This approach allowed us to ensure that the model was reliable. If you prefer, we can use the entire dataset for calibration, but we believe that the current approach is more appropriate given the limited duration of the experiment. We discussed about this limitation of the study in line 447. Anyway, there are others soil modelling studies in literature that account a similar timeframe (Nieto et 2010, Mondini et al 2017).
5- Dear reviewer, the calibration method we applied is the least square method. We are going to clarify it in the next version of the manuscript, and in general we will describe more in depth the calibration procedure.
6- Dear reviewer, the methodology chosen for measuring the soil respiration has been based on closed chambers. In a closed chamber , the CO2 flux (F) is determined from concentration increase within the chamber’s headspace during a known period. In particular, the chamber used was a static one also known as non-steady-state, in which the CO2 coming from the soil accumulate inside. The soil CO2 measurements were conducted by placing soil respiration chambers on PVC collars that were inserted into the ground and remained at their position for 30 minutes. Two collars were placed in the centre of each treatment at 1 m from each other. The measurement was made every month and taken between 9:00 h and 13:00 h. More details about the methods used will be presented in a related manuscript (actually under review).
7- Dear reviewer, the first values we used as estimates were: for a citrus orchard in Mediterranean climate, that suggested in Mota et al. 2011; for the straw mulch, that indicated in Dossou-Yovo et al. 2016. After that, we adjusted the value of the citrus orchard input in order to fit SOC and respiration data, using a least square procedure, as suggested in Nemo 2017.
8- Dear reviewer, we used the IPCC air temperature increase scenario SSP1-1.9, which is the most optimistic one. However, as suggested by the reviewer #2, we will improve the manuscript using, instead, scenario CMIP6, in the revised manuscript.
9- Dear reviewer, we would like to clarify that the differences in initial SOC observed in our simulation are entirely due to the different values of SOC measured in the field at the beginning of the experiment. The field conditions can have a significant impact on the performance of the model, and we have taken this into account in our analysis.
10- Dear reviewer, we would like to clarify that the results of the different version are presented in Supplementary S1 (the supplementary figure could be improved, however). In the Materials and Methods section, we presented different versions of the model to account for the step-by-step modifications and to respond to the possible needs of other users. For examples, users in need for RothC working in Mediterranean climates with a shallow water table may find RothC_Med useful, and users in need for a RothC version accounting for Mulch in non-Mediterranean areas may find RothC_Mulch useful. We just happened to find ourselves in need for both modifications, and thus the RothC_MM version we implemented. If you wish, however, to clarify the issue in the manuscript, we could either (a) describe better why we considered 3 different modifications of RothC, or (b) mention only the RothC_MM version, and refer to the other 2 modifications as “steps in modifying RothC”. Let us know which of these two options seems better, and we will include it in the revised manuscript.
11- Dear reviewer, we were (and are) aware of the problem with measuring CO2 concentrations near trees, and we took measures to reduce the problem by (a) taking CO2 measurements as far as possible from the trees, and in soils without living grass; and (b) we used CO2 measurements only as an upper boundary constraints in the calibration (least square reducing the distance but with constraint of keeping the estimates smaller than the measured values). We will specify this in sections 2.2.1 and 4.1 to ensure that it is clear in the manuscript. Measuring CO2 concentrations can be challenging, and we took care to ensure that our measurements were as accurate as possible given the limitations of the equipment and the environment.
12- Dear reviewer, we acknowledge that words as “simple” and “complex” are too generic and can lead to misunderstandings. What we meant is that RothC is widely less data demanding than Century, and is also smaller as a model, with RothC modelling only soil C (and making a very simple bucket balance for soil water), while Century models different layers of soil, infiltration processes, vegetation dynamics, accumulation of snow, etc… MIMIC, from this point of view, is a “simple” model, in that it does not require a lot of parameters (the hard part is to estimate the parameters for the microbial functional groups in a particular soil). We appreciate your comment, and we will try to clarify this in the manuscript by avoid the use of generic terms as “simple” and “complex”, and trying to refer more accurately to data demand and processes modelled.
13- Dear reviewer, the two parameters we included in the RothC model to simulate the indirect effects of the straw mulch are not theoretical parameters used to fit a function, but rather are steeped in observations from the field (il literature, it is acknowledged that mulching reduces temperature changes in the soil by insulating it, and has effects on infiltration of rain water and reduction in run-off). It is possible to directly estimate these two parameters in the field using a soil temperature and moisture sensor, or to estimate their values using a soil hydrological and thermal model (for example, Hydrus1D). Thus, we do believe our RothC modification not only makes sense, but is easily adaptable to many other areas (insulating soil from cold snatches in the north, reducing run-off and increasing infiltration in equatorial areas, etc…).
Citation: https://doi.org/10.5194/egusphere-2023-298-CC1 -
AC1: 'Reply on RC1', Enrico Balugani, 04 May 2023
Dear Referee and Editors, I posted a response together with my PhD student Simone Pesce from his laptop, and it was registered as a "community comment". Thus, I copy here below our joint response to Referee #1.
Dear Anonymous Referee #1, thank you for your time and effort in reviewing our manuscript. Your feedback will help improving the quality of our manuscript, and we appreciate your thoughtful and constructive comments. Below you will find some answer to your comment.
1- Dear reviewer, the RothC N model was not the focus of our study since it is a model developed in Farina et al 2013. Anyway, we will provide a better description of the model and pedotransfer function in the revised manuscript. Furthermore, we will include the Van Genuchten parameters in the supplementary material to enhance the clarity of our work. Regarding the second part of your question, we would like to direct your attention to our response to question N 7, where we have provided a detailed explanation
2- To clarify, the initial conditions come from the measured SOC stocks in the field. Unfortunately, however, there is no way to measure the RothC pools directly; as such, a spin-up run is required to estimate how to distribute the measured initial SOC into the four RothC pools. This step is crucial to accurately simulate the carbon dynamics in the soil and obtain reliable results. By doing so, we can better understand the fate of carbon in the soil and its impact on the environment. The spin up approach is a widely used method to initialize the RothC model (Pulcher et al 2022, Nemo et al 2017). We appreciate your feedback and will ensure that this process is described in detail in our manuscript.
3- Dear reviewer, we had to rise the carbon input to the soil in both bare soils and mulched soils, in both sites, by the same amount. Thus, it is safe to say that the problem is not in the estimated carbon input from the mulch, but rather the value we used for the carbon input to the soil from the citrus orchard (which came from literature). We will include this description in the discussion section of the revised manuscript.
4- Dear reviewer, we acknowledge that the period of the experiment was only three years. However, we preferred to split the dataset into calibration and validation data to check the accuracy and performance of the model, rather than performing calibration only. This approach allowed us to ensure that the model was reliable. If you prefer, we can use the entire dataset for calibration, but we believe that the current approach is more appropriate given the limited duration of the experiment. We discussed about this limitation of the study in line 447. Anyway, there are others soil modelling studies in literature that account a similar timeframe (Nieto et 2010, Mondini et al 2017).
5- Dear reviewer, the calibration method we applied is the least square method. We are going to clarify it in the next version of the manuscript, and in general we will describe more in depth the calibration procedure.
6- Dear reviewer, the methodology chosen for measuring the soil respiration has been based on closed chambers. In a closed chamber , the CO2 flux (F) is determined from concentration increase within the chamber’s headspace during a known period. In particular, the chamber used was a static one also known as non-steady-state, in which the CO2 coming from the soil accumulate inside. The soil CO2 measurements were conducted by placing soil respiration chambers on PVC collars that were inserted into the ground and remained at their position for 30 minutes. Two collars were placed in the centre of each treatment at 1 m from each other. The measurement was made every month and taken between 9:00 h and 13:00 h. More details about the methods used will be presented in a related manuscript (actually under review).
7- Dear reviewer, the first values we used as estimates were: for a citrus orchard in Mediterranean climate, that suggested in Mota et al. 2011; for the straw mulch, that indicated in Dossou-Yovo et al. 2016. After that, we adjusted the value of the citrus orchard input in order to fit SOC and respiration data, using a least square procedure, as suggested in Nemo 2017.
8- Dear reviewer, we used the IPCC air temperature increase scenario SSP1-1.9, which is the most optimistic one. However, as suggested by the reviewer #2, we will improve the manuscript using, instead, scenario CMIP6, in the revised manuscript.
9- Dear reviewer, we would like to clarify that the differences in initial SOC observed in our simulation are entirely due to the different values of SOC measured in the field at the beginning of the experiment. The field conditions can have a significant impact on the performance of the model, and we have taken this into account in our analysis.
10- Dear reviewer, we would like to clarify that the results of the different version are presented in Supplementary S1 (the supplementary figure could be improved, however). In the Materials and Methods section, we presented different versions of the model to account for the step-by-step modifications and to respond to the possible needs of other users. For examples, users in need for RothC working in Mediterranean climates with a shallow water table may find RothC_Med useful, and users in need for a RothC version accounting for Mulch in non-Mediterranean areas may find RothC_Mulch useful. We just happened to find ourselves in need for both modifications, and thus the RothC_MM version we implemented. If you wish, however, to clarify the issue in the manuscript, we could either (a) describe better why we considered 3 different modifications of RothC, or (b) mention only the RothC_MM version, and refer to the other 2 modifications as “steps in modifying RothC”. Let us know which of these two options seems better, and we will include it in the revised manuscript.
11- Dear reviewer, we were (and are) aware of the problem with measuring CO2 concentrations near trees, and we took measures to reduce the problem by (a) taking CO2 measurements as far as possible from the trees, and in soils without living grass; and (b) we used CO2 measurements only as an upper boundary constraints in the calibration (least square reducing the distance but with constraint of keeping the estimates smaller than the measured values). We will specify this in sections 2.2.1 and 4.1 to ensure that it is clear in the manuscript. Measuring CO2 concentrations can be challenging, and we took care to ensure that our measurements were as accurate as possible given the limitations of the equipment and the environment.
12- Dear reviewer, we acknowledge that words as “simple” and “complex” are too generic and can lead to misunderstandings. What we meant is that RothC is widely less data demanding than Century, and is also smaller as a model, with RothC modelling only soil C (and making a very simple bucket balance for soil water), while Century models different layers of soil, infiltration processes, vegetation dynamics, accumulation of snow, etc… MIMIC, from this point of view, is a “simple” model, in that it does not require a lot of parameters (the hard part is to estimate the parameters for the microbial functional groups in a particular soil). We appreciate your comment, and we will try to clarify this in the manuscript by avoid the use of generic terms as “simple” and “complex”, and trying to refer more accurately to data demand and processes modelled.
13- Dear reviewer, the two parameters we included in the RothC model to simulate the indirect effects of the straw mulch are not theoretical parameters used to fit a function, but rather are steeped in observations from the field (il literature, it is acknowledged that mulching reduces temperature changes in the soil by insulating it, and has effects on infiltration of rain water and reduction in run-off). It is possible to directly estimate these two parameters in the field using a soil temperature and moisture sensor, or to estimate their values using a soil hydrological and thermal model (for example, Hydrus1D). Thus, we do believe our RothC modification not only makes sense, but is easily adaptable to many other areas (insulating soil from cold snatches in the north, reducing run-off and increasing infiltration in equatorial areas, etc…).
Citation: https://doi.org/10.5194/egusphere-2023-298-AC1
-
RC2: 'Comment on egusphere-2023-298', Anonymous Referee #2, 03 Apr 2023
This study modified RothC to investigate the effect of mulching on SOC cycling. The effects were classified into two categories with the direct effect referring to the influence on C input and indirect effect referring to the influence on C decomposition by soil temperature and moisture variation. I have some major concerns on the manuscript.
(1) Section 2.3: The authors generated a climatic scenario of 1.5 degree increase by 2050 inferred from IPCC in this manuscript. Why not directly use the projected climates from CMIP6?
(2) Figure 1 shows that TSMD results generated by modified model fit the observations better compared to earlier versions of model. It is reasonable because the modified model incorporates more processes regarding soil water content. But performance of different models on SOC stock and CO2 emissions was absent in the paper.
(3) Lines 245-255: When comparing different models, the number of parameters in models should also be considered such as in the Akaike information criterion.
(4) Figures 4b, 4c and 4d projected SOC stock increase (some even up to more than a magnitude of 3 times). This is in contrary to widely reported predictions that SOC stock would decrease in a warming future. I think that more discussion is required on this surprising finding.
(5) As described in the title, this paper investigated both the direct and indirect effects of mulching on SOC dynamics. It would be good to quantify the relative contribution of each type of effect on the SOC dynamics.
Some minors:
Line 30: The abbreviation of carbon should already be denoted in Line 29.
Lines 90-91: The format of citation requires revisions.
Line 98: It should be ‘scales’.
Citation: https://doi.org/10.5194/egusphere-2023-298-RC2 -
CC2: 'Reply on RC2', Simone Pesce, 24 Apr 2023
Dear Anonymous Referee #2, thank you for your concise and useful feedback on our manuscript. Your comments have been very helpful in improving the quality of our work, and we appreciate the time and effort you have dedicated to reviewing our manuscript. We have carefully considered your suggestions and we will made the necessary revisions to the revised version of the manuscript. Below you will find some answer to your comment.
1- Dear reviewer, thanks for your suggestion. We are going to use the projected climates from CMIP6 in the next version if the manuscript.
2- Dear reviewer, we have included the analysis in Supplementary Figure S1, which is referenced in the manuscript on pages 282-283. If you believe that it would be beneficial to include the analysis in the main manuscript, we would be happy to do so.
3- Dear reviewer, thank you for your suggestion. We will perform an Akaike information criterion in the next version of the manuscript.
4- Dear reviewer, we want to underline that the predictions in our study are site-specific and account for the site-specific climate and soil characteristics. Figure 4c represents the Sueca control, where the soil quality is generally high due to its soil characteristics as Oxyaquic Xerofluvent, making the soil more resilient to climate change (Visconti et al under review). On the other hand, Figures 4b and 4d represent the SOC increases due to the mulching practice, which underlines the potential to mitigate or reverse the effect of climate warming, as in the case of Paiporta. We appreciate your feedback and hope that this response has addressed your concerns.
5- Dear reviewer, thank you for your feedback and advice. Your suggestions is very helpful, and we will include it in the revised version of the manuscript.
Citation: https://doi.org/10.5194/egusphere-2023-298-CC2 -
AC2: 'Reply on RC2', Enrico Balugani, 04 May 2023
Dear Referee, dear Editors, as for Referee #1 me and my PhD student Simone Pesce sent a response from his laptop, and the response was saved as "community comment". We are sorry about that; here below please find our joint answer to Referee #2 helpful comments!
Dear Anonymous Referee #2, thank you for your concise and useful feedback on our manuscript. Your comments have been very helpful in improving the quality of our work, and we appreciate the time and effort you have dedicated to reviewing our manuscript. We have carefully considered your suggestions and we will made the necessary revisions to the revised version of the manuscript. Below you will find some answer to your comment.
1- Dear reviewer, thanks for your suggestion. We are going to use the projected climates from CMIP6 in the next version if the manuscript.
2- Dear reviewer, we have included the analysis in Supplementary Figure S1, which is referenced in the manuscript on pages 282-283. If you believe that it would be beneficial to include the analysis in the main manuscript, we would be happy to do so.
3- Dear reviewer, thank you for your suggestion. We will perform an Akaike information criterion in the next version of the manuscript.
4- Dear reviewer, we want to underline that the predictions in our study are site-specific and account for the site-specific climate and soil characteristics. Figure 4c represents the Sueca control, where the soil quality is generally high due to its soil characteristics as Oxyaquic Xerofluvent, making the soil more resilient to climate change (Visconti et al under review). On the other hand, Figures 4b and 4d represent the SOC increases due to the mulching practice, which underlines the potential to mitigate or reverse the effect of climate warming, as in the case of Paiporta. We appreciate your feedback and hope that this response has addressed your concerns.
5- Dear reviewer, thank you for your feedback and advice. Your suggestions is very helpful, and we will include it in the revised version of the manuscript.
Citation: https://doi.org/10.5194/egusphere-2023-298-AC2
-
CC2: 'Reply on RC2', Simone Pesce, 24 Apr 2023
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-298', Anonymous Referee #1, 29 Mar 2023
The paper submitted by Pesce et al is an attempt to calibrate the RothC model over two mulching experiments in the Mediterranean region. The topic is of interest and fits with the scope of the journal but the study suffers from several methodological flaws making this study inappropriate for publication in its present form. I list them below:
- In general, a lot of information are missing making the evaluation of the work very difficult. For instance, the RothC_N version is not described and the reader needs to read the Farina et al. paper to understand the starting point of the version used here. Other problem, the author wrote that TSMD is calculated using pedotransfer function based on the Van genuchten equations. This is way too vague, much more details are needed here. In particular, the equations, the parameters values, how the pedotransfer functions were calibrated, etc. Another problematic description is how the MT parameter was calibrated. This is a key part of the paper and a much better descriptions is needed. Because of all the missing information I doubt that the study can be reproduced by someone else.
- Another point that must be clarified is the initialization procedure. The author wrote that they used a spinup to reach equilibrium (section 2.2.2) but then at line 244 they wrote that the proportion of the pools obtained with the spinup is multiplied by the observed stocks. It makes the initialization procedure quite confusing and a better description of what was done to initialize the model is absolutely needed. If the exact procedure is indeed a spinup to reach equilibrium, this is in contradiction with the statement in the introduction claiming at line 35 that cropland soils are loosing carbon.
- Something that also surprised me is written at line 284 where the authors explained that they had to increase the input up to 0.1 and 0.5 tC ha-1 for the mulch experiment. It suggests that the extra carbon inputs due to mulching are not measured and therefore the constrain on the stock given by the model is very poor.
- Another problematic approach is how they used the SOC data. If I understood well, they used SOC data obtained during 22 months to calibrate the model and the data from month 22 to 36 to evaluate. I think it is inappropriate because this period is way too short to detect any SOC changes and therefore the model calibration and evaluation are basically done using the same value.
- The calibration methods is not described enough. How it has been done, simple tunning, least square method, MCMC…?
- At line 296, where the gas chamber always on site? If yes how do the authors managed the temperature increase and the absence of evaporation within the chambers?
- What are the value of the C input in the model, where do they come from?
- About the future scenario, to which IPCC scenario does it corresponds? What is the scenario for C inputs?
- Why the starting point in the SOC data for the control and the mulch treatment are different?
- The author used different versions of the model but at the end they only present one. Or all the versions needs to be presented or only the final version needs to be described.
- When comparing the CO2 fluxes, you are comparing apples and oranges because the model is providing heterotrophic respiration and the data are given soil respiration including roots respiration. This is acknowledged by the authors at line 351 but I think it should be avoided and removed from the manuscript
- I disagree with the statement in the introduction saying that RothC is more simple than CENTURY (line 63). The approach in the two models is similar (pool based, first order kinetics, etc.). Much more complex model could have been found to support your statement in particular those which are driven by microbial mechanisms (Millenial, MIMICS, etc.).
- I also disagree with the statement at line 378-379. If you use site data to fit empirical parameters to then run local simulations, I don't think you really integrate any mechanisms related to mulching. You just tuned your model regarding local conditions and there is no guarantee that such a set of parameters would work in a different situation even under that same crop management and the same climate.
Citation: https://doi.org/10.5194/egusphere-2023-298-RC1 -
CC1: 'Reply on RC1', Simone Pesce, 24 Apr 2023
Dear Anonymous Referee #1, thank you for your time and effort in reviewing our manuscript. Your feedback will help improving the quality of our manuscript, and we appreciate your thoughtful and constructive comments. Below you will find some answer to your comment.
1- Dear reviewer, the RothC N model was not the focus of our study since it is a model developed in Farina et al 2013. Anyway, we will provide a better description of the model and pedotransfer function in the revised manuscript. Furthermore, we will include the Van Genuchten parameters in the supplementary material to enhance the clarity of our work. Regarding the second part of your question, we would like to direct your attention to our response to question N 7, where we have provided a detailed explanation
2- To clarify, the initial conditions come from the measured SOC stocks in the field. Unfortunately, however, there is no way to measure the RothC pools directly; as such, a spin-up run is required to estimate how to distribute the measured initial SOC into the four RothC pools. This step is crucial to accurately simulate the carbon dynamics in the soil and obtain reliable results. By doing so, we can better understand the fate of carbon in the soil and its impact on the environment. The spin up approach is a widely used method to initialize the RothC model (Pulcher et al 2022, Nemo et al 2017). We appreciate your feedback and will ensure that this process is described in detail in our manuscript.
3- Dear reviewer, we had to rise the carbon input to the soil in both bare soils and mulched soils, in both sites, by the same amount. Thus, it is safe to say that the problem is not in the estimated carbon input from the mulch, but rather the value we used for the carbon input to the soil from the citrus orchard (which came from literature). We will include this description in the discussion section of the revised manuscript.
4- Dear reviewer, we acknowledge that the period of the experiment was only three years. However, we preferred to split the dataset into calibration and validation data to check the accuracy and performance of the model, rather than performing calibration only. This approach allowed us to ensure that the model was reliable. If you prefer, we can use the entire dataset for calibration, but we believe that the current approach is more appropriate given the limited duration of the experiment. We discussed about this limitation of the study in line 447. Anyway, there are others soil modelling studies in literature that account a similar timeframe (Nieto et 2010, Mondini et al 2017).
5- Dear reviewer, the calibration method we applied is the least square method. We are going to clarify it in the next version of the manuscript, and in general we will describe more in depth the calibration procedure.
6- Dear reviewer, the methodology chosen for measuring the soil respiration has been based on closed chambers. In a closed chamber , the CO2 flux (F) is determined from concentration increase within the chamber’s headspace during a known period. In particular, the chamber used was a static one also known as non-steady-state, in which the CO2 coming from the soil accumulate inside. The soil CO2 measurements were conducted by placing soil respiration chambers on PVC collars that were inserted into the ground and remained at their position for 30 minutes. Two collars were placed in the centre of each treatment at 1 m from each other. The measurement was made every month and taken between 9:00 h and 13:00 h. More details about the methods used will be presented in a related manuscript (actually under review).
7- Dear reviewer, the first values we used as estimates were: for a citrus orchard in Mediterranean climate, that suggested in Mota et al. 2011; for the straw mulch, that indicated in Dossou-Yovo et al. 2016. After that, we adjusted the value of the citrus orchard input in order to fit SOC and respiration data, using a least square procedure, as suggested in Nemo 2017.
8- Dear reviewer, we used the IPCC air temperature increase scenario SSP1-1.9, which is the most optimistic one. However, as suggested by the reviewer #2, we will improve the manuscript using, instead, scenario CMIP6, in the revised manuscript.
9- Dear reviewer, we would like to clarify that the differences in initial SOC observed in our simulation are entirely due to the different values of SOC measured in the field at the beginning of the experiment. The field conditions can have a significant impact on the performance of the model, and we have taken this into account in our analysis.
10- Dear reviewer, we would like to clarify that the results of the different version are presented in Supplementary S1 (the supplementary figure could be improved, however). In the Materials and Methods section, we presented different versions of the model to account for the step-by-step modifications and to respond to the possible needs of other users. For examples, users in need for RothC working in Mediterranean climates with a shallow water table may find RothC_Med useful, and users in need for a RothC version accounting for Mulch in non-Mediterranean areas may find RothC_Mulch useful. We just happened to find ourselves in need for both modifications, and thus the RothC_MM version we implemented. If you wish, however, to clarify the issue in the manuscript, we could either (a) describe better why we considered 3 different modifications of RothC, or (b) mention only the RothC_MM version, and refer to the other 2 modifications as “steps in modifying RothC”. Let us know which of these two options seems better, and we will include it in the revised manuscript.
11- Dear reviewer, we were (and are) aware of the problem with measuring CO2 concentrations near trees, and we took measures to reduce the problem by (a) taking CO2 measurements as far as possible from the trees, and in soils without living grass; and (b) we used CO2 measurements only as an upper boundary constraints in the calibration (least square reducing the distance but with constraint of keeping the estimates smaller than the measured values). We will specify this in sections 2.2.1 and 4.1 to ensure that it is clear in the manuscript. Measuring CO2 concentrations can be challenging, and we took care to ensure that our measurements were as accurate as possible given the limitations of the equipment and the environment.
12- Dear reviewer, we acknowledge that words as “simple” and “complex” are too generic and can lead to misunderstandings. What we meant is that RothC is widely less data demanding than Century, and is also smaller as a model, with RothC modelling only soil C (and making a very simple bucket balance for soil water), while Century models different layers of soil, infiltration processes, vegetation dynamics, accumulation of snow, etc… MIMIC, from this point of view, is a “simple” model, in that it does not require a lot of parameters (the hard part is to estimate the parameters for the microbial functional groups in a particular soil). We appreciate your comment, and we will try to clarify this in the manuscript by avoid the use of generic terms as “simple” and “complex”, and trying to refer more accurately to data demand and processes modelled.
13- Dear reviewer, the two parameters we included in the RothC model to simulate the indirect effects of the straw mulch are not theoretical parameters used to fit a function, but rather are steeped in observations from the field (il literature, it is acknowledged that mulching reduces temperature changes in the soil by insulating it, and has effects on infiltration of rain water and reduction in run-off). It is possible to directly estimate these two parameters in the field using a soil temperature and moisture sensor, or to estimate their values using a soil hydrological and thermal model (for example, Hydrus1D). Thus, we do believe our RothC modification not only makes sense, but is easily adaptable to many other areas (insulating soil from cold snatches in the north, reducing run-off and increasing infiltration in equatorial areas, etc…).
Citation: https://doi.org/10.5194/egusphere-2023-298-CC1 -
AC1: 'Reply on RC1', Enrico Balugani, 04 May 2023
Dear Referee and Editors, I posted a response together with my PhD student Simone Pesce from his laptop, and it was registered as a "community comment". Thus, I copy here below our joint response to Referee #1.
Dear Anonymous Referee #1, thank you for your time and effort in reviewing our manuscript. Your feedback will help improving the quality of our manuscript, and we appreciate your thoughtful and constructive comments. Below you will find some answer to your comment.
1- Dear reviewer, the RothC N model was not the focus of our study since it is a model developed in Farina et al 2013. Anyway, we will provide a better description of the model and pedotransfer function in the revised manuscript. Furthermore, we will include the Van Genuchten parameters in the supplementary material to enhance the clarity of our work. Regarding the second part of your question, we would like to direct your attention to our response to question N 7, where we have provided a detailed explanation
2- To clarify, the initial conditions come from the measured SOC stocks in the field. Unfortunately, however, there is no way to measure the RothC pools directly; as such, a spin-up run is required to estimate how to distribute the measured initial SOC into the four RothC pools. This step is crucial to accurately simulate the carbon dynamics in the soil and obtain reliable results. By doing so, we can better understand the fate of carbon in the soil and its impact on the environment. The spin up approach is a widely used method to initialize the RothC model (Pulcher et al 2022, Nemo et al 2017). We appreciate your feedback and will ensure that this process is described in detail in our manuscript.
3- Dear reviewer, we had to rise the carbon input to the soil in both bare soils and mulched soils, in both sites, by the same amount. Thus, it is safe to say that the problem is not in the estimated carbon input from the mulch, but rather the value we used for the carbon input to the soil from the citrus orchard (which came from literature). We will include this description in the discussion section of the revised manuscript.
4- Dear reviewer, we acknowledge that the period of the experiment was only three years. However, we preferred to split the dataset into calibration and validation data to check the accuracy and performance of the model, rather than performing calibration only. This approach allowed us to ensure that the model was reliable. If you prefer, we can use the entire dataset for calibration, but we believe that the current approach is more appropriate given the limited duration of the experiment. We discussed about this limitation of the study in line 447. Anyway, there are others soil modelling studies in literature that account a similar timeframe (Nieto et 2010, Mondini et al 2017).
5- Dear reviewer, the calibration method we applied is the least square method. We are going to clarify it in the next version of the manuscript, and in general we will describe more in depth the calibration procedure.
6- Dear reviewer, the methodology chosen for measuring the soil respiration has been based on closed chambers. In a closed chamber , the CO2 flux (F) is determined from concentration increase within the chamber’s headspace during a known period. In particular, the chamber used was a static one also known as non-steady-state, in which the CO2 coming from the soil accumulate inside. The soil CO2 measurements were conducted by placing soil respiration chambers on PVC collars that were inserted into the ground and remained at their position for 30 minutes. Two collars were placed in the centre of each treatment at 1 m from each other. The measurement was made every month and taken between 9:00 h and 13:00 h. More details about the methods used will be presented in a related manuscript (actually under review).
7- Dear reviewer, the first values we used as estimates were: for a citrus orchard in Mediterranean climate, that suggested in Mota et al. 2011; for the straw mulch, that indicated in Dossou-Yovo et al. 2016. After that, we adjusted the value of the citrus orchard input in order to fit SOC and respiration data, using a least square procedure, as suggested in Nemo 2017.
8- Dear reviewer, we used the IPCC air temperature increase scenario SSP1-1.9, which is the most optimistic one. However, as suggested by the reviewer #2, we will improve the manuscript using, instead, scenario CMIP6, in the revised manuscript.
9- Dear reviewer, we would like to clarify that the differences in initial SOC observed in our simulation are entirely due to the different values of SOC measured in the field at the beginning of the experiment. The field conditions can have a significant impact on the performance of the model, and we have taken this into account in our analysis.
10- Dear reviewer, we would like to clarify that the results of the different version are presented in Supplementary S1 (the supplementary figure could be improved, however). In the Materials and Methods section, we presented different versions of the model to account for the step-by-step modifications and to respond to the possible needs of other users. For examples, users in need for RothC working in Mediterranean climates with a shallow water table may find RothC_Med useful, and users in need for a RothC version accounting for Mulch in non-Mediterranean areas may find RothC_Mulch useful. We just happened to find ourselves in need for both modifications, and thus the RothC_MM version we implemented. If you wish, however, to clarify the issue in the manuscript, we could either (a) describe better why we considered 3 different modifications of RothC, or (b) mention only the RothC_MM version, and refer to the other 2 modifications as “steps in modifying RothC”. Let us know which of these two options seems better, and we will include it in the revised manuscript.
11- Dear reviewer, we were (and are) aware of the problem with measuring CO2 concentrations near trees, and we took measures to reduce the problem by (a) taking CO2 measurements as far as possible from the trees, and in soils without living grass; and (b) we used CO2 measurements only as an upper boundary constraints in the calibration (least square reducing the distance but with constraint of keeping the estimates smaller than the measured values). We will specify this in sections 2.2.1 and 4.1 to ensure that it is clear in the manuscript. Measuring CO2 concentrations can be challenging, and we took care to ensure that our measurements were as accurate as possible given the limitations of the equipment and the environment.
12- Dear reviewer, we acknowledge that words as “simple” and “complex” are too generic and can lead to misunderstandings. What we meant is that RothC is widely less data demanding than Century, and is also smaller as a model, with RothC modelling only soil C (and making a very simple bucket balance for soil water), while Century models different layers of soil, infiltration processes, vegetation dynamics, accumulation of snow, etc… MIMIC, from this point of view, is a “simple” model, in that it does not require a lot of parameters (the hard part is to estimate the parameters for the microbial functional groups in a particular soil). We appreciate your comment, and we will try to clarify this in the manuscript by avoid the use of generic terms as “simple” and “complex”, and trying to refer more accurately to data demand and processes modelled.
13- Dear reviewer, the two parameters we included in the RothC model to simulate the indirect effects of the straw mulch are not theoretical parameters used to fit a function, but rather are steeped in observations from the field (il literature, it is acknowledged that mulching reduces temperature changes in the soil by insulating it, and has effects on infiltration of rain water and reduction in run-off). It is possible to directly estimate these two parameters in the field using a soil temperature and moisture sensor, or to estimate their values using a soil hydrological and thermal model (for example, Hydrus1D). Thus, we do believe our RothC modification not only makes sense, but is easily adaptable to many other areas (insulating soil from cold snatches in the north, reducing run-off and increasing infiltration in equatorial areas, etc…).
Citation: https://doi.org/10.5194/egusphere-2023-298-AC1
-
RC2: 'Comment on egusphere-2023-298', Anonymous Referee #2, 03 Apr 2023
This study modified RothC to investigate the effect of mulching on SOC cycling. The effects were classified into two categories with the direct effect referring to the influence on C input and indirect effect referring to the influence on C decomposition by soil temperature and moisture variation. I have some major concerns on the manuscript.
(1) Section 2.3: The authors generated a climatic scenario of 1.5 degree increase by 2050 inferred from IPCC in this manuscript. Why not directly use the projected climates from CMIP6?
(2) Figure 1 shows that TSMD results generated by modified model fit the observations better compared to earlier versions of model. It is reasonable because the modified model incorporates more processes regarding soil water content. But performance of different models on SOC stock and CO2 emissions was absent in the paper.
(3) Lines 245-255: When comparing different models, the number of parameters in models should also be considered such as in the Akaike information criterion.
(4) Figures 4b, 4c and 4d projected SOC stock increase (some even up to more than a magnitude of 3 times). This is in contrary to widely reported predictions that SOC stock would decrease in a warming future. I think that more discussion is required on this surprising finding.
(5) As described in the title, this paper investigated both the direct and indirect effects of mulching on SOC dynamics. It would be good to quantify the relative contribution of each type of effect on the SOC dynamics.
Some minors:
Line 30: The abbreviation of carbon should already be denoted in Line 29.
Lines 90-91: The format of citation requires revisions.
Line 98: It should be ‘scales’.
Citation: https://doi.org/10.5194/egusphere-2023-298-RC2 -
CC2: 'Reply on RC2', Simone Pesce, 24 Apr 2023
Dear Anonymous Referee #2, thank you for your concise and useful feedback on our manuscript. Your comments have been very helpful in improving the quality of our work, and we appreciate the time and effort you have dedicated to reviewing our manuscript. We have carefully considered your suggestions and we will made the necessary revisions to the revised version of the manuscript. Below you will find some answer to your comment.
1- Dear reviewer, thanks for your suggestion. We are going to use the projected climates from CMIP6 in the next version if the manuscript.
2- Dear reviewer, we have included the analysis in Supplementary Figure S1, which is referenced in the manuscript on pages 282-283. If you believe that it would be beneficial to include the analysis in the main manuscript, we would be happy to do so.
3- Dear reviewer, thank you for your suggestion. We will perform an Akaike information criterion in the next version of the manuscript.
4- Dear reviewer, we want to underline that the predictions in our study are site-specific and account for the site-specific climate and soil characteristics. Figure 4c represents the Sueca control, where the soil quality is generally high due to its soil characteristics as Oxyaquic Xerofluvent, making the soil more resilient to climate change (Visconti et al under review). On the other hand, Figures 4b and 4d represent the SOC increases due to the mulching practice, which underlines the potential to mitigate or reverse the effect of climate warming, as in the case of Paiporta. We appreciate your feedback and hope that this response has addressed your concerns.
5- Dear reviewer, thank you for your feedback and advice. Your suggestions is very helpful, and we will include it in the revised version of the manuscript.
Citation: https://doi.org/10.5194/egusphere-2023-298-CC2 -
AC2: 'Reply on RC2', Enrico Balugani, 04 May 2023
Dear Referee, dear Editors, as for Referee #1 me and my PhD student Simone Pesce sent a response from his laptop, and the response was saved as "community comment". We are sorry about that; here below please find our joint answer to Referee #2 helpful comments!
Dear Anonymous Referee #2, thank you for your concise and useful feedback on our manuscript. Your comments have been very helpful in improving the quality of our work, and we appreciate the time and effort you have dedicated to reviewing our manuscript. We have carefully considered your suggestions and we will made the necessary revisions to the revised version of the manuscript. Below you will find some answer to your comment.
1- Dear reviewer, thanks for your suggestion. We are going to use the projected climates from CMIP6 in the next version if the manuscript.
2- Dear reviewer, we have included the analysis in Supplementary Figure S1, which is referenced in the manuscript on pages 282-283. If you believe that it would be beneficial to include the analysis in the main manuscript, we would be happy to do so.
3- Dear reviewer, thank you for your suggestion. We will perform an Akaike information criterion in the next version of the manuscript.
4- Dear reviewer, we want to underline that the predictions in our study are site-specific and account for the site-specific climate and soil characteristics. Figure 4c represents the Sueca control, where the soil quality is generally high due to its soil characteristics as Oxyaquic Xerofluvent, making the soil more resilient to climate change (Visconti et al under review). On the other hand, Figures 4b and 4d represent the SOC increases due to the mulching practice, which underlines the potential to mitigate or reverse the effect of climate warming, as in the case of Paiporta. We appreciate your feedback and hope that this response has addressed your concerns.
5- Dear reviewer, thank you for your feedback and advice. Your suggestions is very helpful, and we will include it in the revised version of the manuscript.
Citation: https://doi.org/10.5194/egusphere-2023-298-AC2
-
CC2: 'Reply on RC2', Simone Pesce, 24 Apr 2023
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 663 | 358 | 38 | 1,059 | 84 | 16 | 16 |
- HTML: 663
- PDF: 358
- XML: 38
- Total: 1,059
- Supplement: 84
- BibTeX: 16
- EndNote: 16
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 322 | 30 |
| China | 2 | 184 | 17 |
| Italy | 3 | 62 | 5 |
| Germany | 4 | 59 | 5 |
| France | 5 | 52 | 5 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 322
Simone Pesce
Enrico Balugani
José Miguel de Paz
Diego Marazza
Fernando Visconti
This preprint has been withdrawn.
- Preprint
(1118 KB) - Metadata XML
-
Supplement
(200 KB) - BibTeX
- EndNote