the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
TECO-CNP Sv1.0: A coupled carbon-nitrogen-phosphorus model with data assimilation for subtropical forests
Abstract. Subtropical forests play a crucial role in global cycle, yet their carbon sink capacity is significantly constrained by phosphorus availability. Models that omit phosphorus dynamics risk overestimating carbon sinks, potentially undermining the scientific basis for carbon neutrality strategies. In this study, we developed TECO-CNP Sv1.0, a coupled carbon-nitrogen-phosphorus model based on the Terrestrial ECOsystem (TECO) model, explicitly capturing key biogeochemical interactions and nutrient-regulated carbon cycling. The model simulates how plant growth and carbon partitioning respond to both external soil nutrient availability and internal physiological constraints, enabling plant acclimation to varying nutrient conditions. Using observations from a phosphorus-limited subtropical forest in East China, we first evaluated model performance on estimating state variables with empirically calibrated parameters. Compared to the C-only and coupled C-N configurations, the CNP model better reproduced observed plant and soil C, N, and P pools. To systematically optimize model parameters and reduce uncertainties in predictions, we further incorporated a built-in data assimilation framework for parameter optimization. The CNP model with optimized parameters significantly improved carbon flux estimates, reducing root mean square errors and enhancing concordance correlation coefficients for gross primary productivity, ecosystem respiration, and net ecosystem exchange. By explicitly incorporating phosphorus dynamics and data assimilation, this study provides a more accurate and robust framework for predicting carbon sequestration in phosphorus-limited subtropical forests.
- Preprint
(1724 KB) - Metadata XML
-
Supplement
(148 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2025-1243', Anonymous Referee #1, 02 Jun 2025
-
AC1: 'Reply on RC1', Jianyang Xia, 27 Jul 2025
We appreciate the valuable feedback and insightful suggestions provided by all reviewers. The manuscript has been thoroughly revised in accordance with your recommendations, and we attach our response. Reviewer questions are presented in blue text, while our responses are provided in black text.
-
AC1: 'Reply on RC1', Jianyang Xia, 27 Jul 2025
-
RC2: 'Comment on egusphere-2025-1243', Anonymous Referee #2, 30 Jun 2025
This study developed a coupled carbon-nitrogen-phosphorus model, TECO-CNP Sv1.0, based on the Terrestrial ECOsystem (TECO) model. The developed model was used to simulate C, N, and P pools and fluxes in a phosphorus-limited subtropical forest site in East China. In addition, a parameter optimization algorithm was also incorporated into the model framework to improve the model's performance. Overall, the manuscript provides detailed information on the model structure, parameters, and performance. However, I still have some questions on the soil pool structure and calibration processes of the model.
Â
Four inorganic P pools, including labile P, sorbed P, secondary P, and occluded P, are set in TECO-CNP. This structure is different from other CNP models. For example, the labile P pool in ORCHIDEE-CNP includes both dissolved and sorbed P. In CLM-CNP, inorganic P pools include labile P (including solution P), secondary P, and occluded P. In a global P dataset developed by He et al. (2023) Biogeosciences, the soil inorganic P is divided into labile P, moderate P, and occluded P. Can you explain the differences in inorganic P pool structure among these models? I am confused about the definition of labile P pool. In addition, how did you initialize these inorganic P pools?
Â
What is the advantage of TECO-CNP compared with other CNP models?
Â
How were the soil P pools at the Tiantong site measured? Did you compare the measured soil P pool with other studies? They seem lower than other studies (Fig. 5c).
Â
The simulation results of C-only, CN, and CNP versions were compared in this study to prove the good performance of the CNP model. Did you calibrate these three versions individually? Was the same parameter optimization algorithm applied to all three versions? Or you just calibrate the only CNP model, and apply the same parameters to other versions. Did you simulate the C cycle in this site by using C-only or CN versions before the development of CNP? How well did these two models perform? Many parameters are constant values, such as Vre. I guess these parameters were not calibrated but derived from the literature. Were these parameters suitable for subtropical forest ecosystems?
Â
The vegetation and soil pools simulated by the three versions were compared in section 3.1. What about C and nutrient fluxes? In Fig.7, I cannot identify the simulated NEE by the three model configurations.
Â
Minor
L491-493. Please list the equations of P loss from SOM pools.
Equation 54. What is the meaning of Pl
L607. Please correct the reference of Xu et al.
Fig 8. What is the meaning of the posterior distribution of parameters? Do they change with time?
Table. How did you identify these target parameters? Did you conduct a sensitivity analysis?
Â
Reference: He, Xianjin, et al. "Global patterns and drivers of phosphorus fractions in natural soils." Biogeosciences 20.19 (2023): 4147-4163.
Citation: https://doi.org/10.5194/egusphere-2025-1243-RC2 -
AC2: 'Reply on RC2', Jianyang Xia, 27 Jul 2025
We appreciate the valuable feedback and insightful suggestions provided by all reviewers. The manuscript has been thoroughly revised in accordance with your recommendations, and we attach our detailed response. Reviewer questions are presented in blue text, while our responses are provided in black text.
-
AC2: 'Reply on RC2', Jianyang Xia, 27 Jul 2025
Status: closed
-
RC1: 'Comment on egusphere-2025-1243', Anonymous Referee #1, 02 Jun 2025
Â
This is a nicely written and well-executed work implementing phosphorus cycle and data assimilation framework into a process-based ecosystem model (TECO). The novelty of this work lie in the model development and its coupling with data assimilation. By comparing CNP model against observations and their respective C-only and CN coupled models, the authors show a superior performance of the newly developed model.
Â
Overall, I enjoy reading the manuscript, and support its publication. There are several occasions where I think some justifications/modifications would further improve the quality of the manuscript. Below I list my main questions/concerns.
Â
There are certain processes where N and P interact. For example, some models consider N to have an effect on soil P biochemical mineralization rate (e.g. ORCHIDEE, ELM, etc.). It does not seem that the authors have adopted these NP interacting processes in their model. Furthermore, others consider N and P to have joint effect on C processes, such as photosynthesis. In this work, it seems that nutrient effect on photosynthesis is realized via downregulation of leaf surface area. There have been some empirical relationships derived on how N and P affects photosynthetic traits (e.g. Vcmax, Jmax; Ellsworth et al., 2022; Walker et al,, 2014), and these relationships have been incorporated into models. What is the authors’ consideration on not following these conventional approaches? Â
Solution P is part of labile P, and some work suggests the need to explicitly model solution P in addition to labile P (e.g. Reed et al., 2015). In this work, how does the author consider this suggestion and what is the rationale for only simulating labile P?
It seems that the data assimilation framework was only applied to the CNP model, and then the authors reported CN and C models to overestimate observations. I find this logic to be a bit problematic. Without data assimilation, does CNP model still achieve better match with observations? Alternatively, how does C-only model coupled with data assimilation perform relative to observations? If it can achieve similar performance as compared to CNP model, what benefits of having a CNP model?
Citation: https://doi.org/10.5194/egusphere-2025-1243-RC1 -
AC1: 'Reply on RC1', Jianyang Xia, 27 Jul 2025
We appreciate the valuable feedback and insightful suggestions provided by all reviewers. The manuscript has been thoroughly revised in accordance with your recommendations, and we attach our response. Reviewer questions are presented in blue text, while our responses are provided in black text.
-
AC1: 'Reply on RC1', Jianyang Xia, 27 Jul 2025
-
RC2: 'Comment on egusphere-2025-1243', Anonymous Referee #2, 30 Jun 2025
This study developed a coupled carbon-nitrogen-phosphorus model, TECO-CNP Sv1.0, based on the Terrestrial ECOsystem (TECO) model. The developed model was used to simulate C, N, and P pools and fluxes in a phosphorus-limited subtropical forest site in East China. In addition, a parameter optimization algorithm was also incorporated into the model framework to improve the model's performance. Overall, the manuscript provides detailed information on the model structure, parameters, and performance. However, I still have some questions on the soil pool structure and calibration processes of the model.
Â
Four inorganic P pools, including labile P, sorbed P, secondary P, and occluded P, are set in TECO-CNP. This structure is different from other CNP models. For example, the labile P pool in ORCHIDEE-CNP includes both dissolved and sorbed P. In CLM-CNP, inorganic P pools include labile P (including solution P), secondary P, and occluded P. In a global P dataset developed by He et al. (2023) Biogeosciences, the soil inorganic P is divided into labile P, moderate P, and occluded P. Can you explain the differences in inorganic P pool structure among these models? I am confused about the definition of labile P pool. In addition, how did you initialize these inorganic P pools?
Â
What is the advantage of TECO-CNP compared with other CNP models?
Â
How were the soil P pools at the Tiantong site measured? Did you compare the measured soil P pool with other studies? They seem lower than other studies (Fig. 5c).
Â
The simulation results of C-only, CN, and CNP versions were compared in this study to prove the good performance of the CNP model. Did you calibrate these three versions individually? Was the same parameter optimization algorithm applied to all three versions? Or you just calibrate the only CNP model, and apply the same parameters to other versions. Did you simulate the C cycle in this site by using C-only or CN versions before the development of CNP? How well did these two models perform? Many parameters are constant values, such as Vre. I guess these parameters were not calibrated but derived from the literature. Were these parameters suitable for subtropical forest ecosystems?
Â
The vegetation and soil pools simulated by the three versions were compared in section 3.1. What about C and nutrient fluxes? In Fig.7, I cannot identify the simulated NEE by the three model configurations.
Â
Minor
L491-493. Please list the equations of P loss from SOM pools.
Equation 54. What is the meaning of Pl
L607. Please correct the reference of Xu et al.
Fig 8. What is the meaning of the posterior distribution of parameters? Do they change with time?
Table. How did you identify these target parameters? Did you conduct a sensitivity analysis?
Â
Reference: He, Xianjin, et al. "Global patterns and drivers of phosphorus fractions in natural soils." Biogeosciences 20.19 (2023): 4147-4163.
Citation: https://doi.org/10.5194/egusphere-2025-1243-RC2 -
AC2: 'Reply on RC2', Jianyang Xia, 27 Jul 2025
We appreciate the valuable feedback and insightful suggestions provided by all reviewers. The manuscript has been thoroughly revised in accordance with your recommendations, and we attach our detailed response. Reviewer questions are presented in blue text, while our responses are provided in black text.
-
AC2: 'Reply on RC2', Jianyang Xia, 27 Jul 2025
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
1,979 | 83 | 18 | 2,080 | 35 | 26 | 39 |
- HTML: 1,979
- PDF: 83
- XML: 18
- Total: 2,080
- Supplement: 35
- BibTeX: 26
- EndNote: 39
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1
Â
This is a nicely written and well-executed work implementing phosphorus cycle and data assimilation framework into a process-based ecosystem model (TECO). The novelty of this work lie in the model development and its coupling with data assimilation. By comparing CNP model against observations and their respective C-only and CN coupled models, the authors show a superior performance of the newly developed model.
Â
Overall, I enjoy reading the manuscript, and support its publication. There are several occasions where I think some justifications/modifications would further improve the quality of the manuscript. Below I list my main questions/concerns.
Â
There are certain processes where N and P interact. For example, some models consider N to have an effect on soil P biochemical mineralization rate (e.g. ORCHIDEE, ELM, etc.). It does not seem that the authors have adopted these NP interacting processes in their model. Furthermore, others consider N and P to have joint effect on C processes, such as photosynthesis. In this work, it seems that nutrient effect on photosynthesis is realized via downregulation of leaf surface area. There have been some empirical relationships derived on how N and P affects photosynthetic traits (e.g. Vcmax, Jmax; Ellsworth et al., 2022; Walker et al,, 2014), and these relationships have been incorporated into models. What is the authors’ consideration on not following these conventional approaches? Â
Solution P is part of labile P, and some work suggests the need to explicitly model solution P in addition to labile P (e.g. Reed et al., 2015). In this work, how does the author consider this suggestion and what is the rationale for only simulating labile P?
It seems that the data assimilation framework was only applied to the CNP model, and then the authors reported CN and C models to overestimate observations. I find this logic to be a bit problematic. Without data assimilation, does CNP model still achieve better match with observations? Alternatively, how does C-only model coupled with data assimilation perform relative to observations? If it can achieve similar performance as compared to CNP model, what benefits of having a CNP model?