the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 19 Apr 2023
CHEEREIO 1.0: a versatile and user-friendly ensemble-based chemical data assimilation and emissions inversion platform for the GEOS-Chem chemical transport model
Drew C. Pendergrass
Daniel J. Jacob
Hannah Nesser
Daniel J. Varon
Melissa Sulprizio
Kazuyuki Miyazaki
Kevin W. Bowman
Abstract. We present a versatile, powerful, and user-friendly chemical data assimilation toolkit for simultaneously optimizing emissions and concentrations of chemical species based on atmospheric observations from satellites or suborbital platforms. The CHemistry and Emissions REanalysis Interface with Observations (CHEEREIO) exploits the GEOS-Chem chemical transport model and a localized ensemble transform Kalman filter algorithm (LETKF) to determine the Bayesian optimal (posterior) emissions and/or concentrations of a set of species based on observations and prior information, using an easy-to-modify configuration file with minimal changes to the GEOS-Chem or LETKF code base. The LETKF algorithm readily allows for non-linear chemistry and produces flow-dependent posterior error covariances from the ensemble simulation spread. The object-oriented Python-based design of CHEEREIO allows users to easily add new observation operators such as for satellites. CHEEREIO takes advantage of the HEMCO modular structure of input data management in GEOS-Chem to update emissions from the assimilation process independently from the GEOS-Chem code. It can seamlessly support GEOS-Chem version updates and is adaptable to other chemical transport models with similar modular input data structure. A postprocessing suite combines ensemble output into consolidated NetCDF files and supports a wide variety of diagnostic data and visualizations. We demonstrate CHEEREIO’s capabilities with an out-of-the-box application, assimilating global methane emissions and concentrations at weekly temporal resolution and 2°x2.5° spatial resolution for 2019 using TROPOMI satellite observations. CHEEREIO achieves a 50-fold improvement in computational performance compared to the equivalent analytical inversion of TROPOMI observations.
-
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
(3159 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(3159 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Drew C. Pendergrass et al.
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-616', Anonymous Referee #1, 05 Jun 2023
In this paper, the authors introduce CHEEREIO 1.0, a novel chemical data assimilation toolkit. This toolkit is designed to optimize emissions and concentrations of chemical species based on atmospheric observations from satellites or suborbital platforms. The toolkit utilizes the GEOS-Chem chemical transport model and a localized ensemble transform Kalman filter algorithm to determine the Bayesian optimal emissions and/or concentrations of a set of species. The Python-based design of CHEEREIO is commendable for its user-friendliness, allowing users to easily add new observation operators such as for satellites.
The toolkit demonstrates impressive computational performance, achieving a 50-fold improvement compared to the equivalent analytical inversion of TROPOMI observations. This advancement holds significant potential to enhance the efficiency of atmospheric chemistry research.
The paper is well-structured, with the authors providing a clear explanation of the methodology used.
Minor Comments:
1. Around line 540, the authors compare the computational efficiency of the toolkit with the analytical inversion. While the superior computational efficiency of the toolkit is evident, the comparison seems incomplete without a discussion on inversion accuracy. I would recommend the authors to include a comparison of the inversion accuracy.
2. The reduction in computational cost is mainly achieved by using a smaller ensemble size, primarily through the application of localization. It would be interesting to know if using the full rank ensemble size (as in the analytic solution) provides a superior solution. Also, can the analytical inversion also apply localization?
3. The authors briefly mention the limitations of CTM bias/error. However, the described CHEEREIO system does not seem to address errors in the model, such as perturbations in meteorological condition and physics. It would be beneficial if the authors could elaborate on how these potential sources of error are addressed in the CHEEREIO system.
Citation: https://doi.org/10.5194/egusphere-2023-616-RC1 -
RC2: 'Comment on egusphere-2023-616', Anonymous Referee #2, 25 Jun 2023
The authors developed a user-friendly chemical data assimilation platform based on the GEOS-Chem model and the LETKF algorithm, which can optimize emissions and concentrations of different species constrained by satellite observations. The GEOS-Chem model is a powerful tool for atmospheric modeling widely used in the scientific community and the LETKF algorithm has been proven to work efficiently for both linear and nonlinear problems. I feel that the inversion toolkit developed in this paper is very important and useful, and I believe that this tool will promote the research on chemical data assimilation and emission inversions. This manuscript is well-structured and well-written. I suggest a minor revision after addressing my comments below.
- There are already a few open-source inversion tools that have been developed. I suggest that the authors briefly compare CHEEREIO and other platforms and discuss the differences and advantages of this new platform.
- I understand that due to length limitations, the authors provide detailed documentation of CHEEREIO online (cheereio.readthedocs.io), instead of presenting them within the manuscript. If possible, I suggest the authors explain a little bit more of the key parameters and settings of CHEEREIO, which could largely influence the emission inversion results. This can help the readers quickly understand the most important parts of the LETKF algorithm, as well as the most sensitive parameters that should be concerned about.
- In Sect. 4, the authors mentioned a comparison between the example application of CH4 emissions inversion with Qu et al. (2021), which optimized global CH4 emissions at the same horizontal resolution using the analytical approach. Is it possible to directly compare the posterior CH4 emissions with Qu et al. (2021) in the main text? This would be helpful to understand the differences between the LETKF algorithm and the analytical approach.
Citation: https://doi.org/10.5194/egusphere-2023-616-RC2 -
AC1: 'Comment on egusphere-2023-616', Drew Pendergrass, 10 Jul 2023
Response to reviewer 1.
In this paper, the authors introduce CHEEREIO 1.0, a novel chemical data assimilation toolkit. This toolkit is designed to optimize emissions and concentrations of chemical species based on atmospheric observations from satellites or suborbital platforms. The toolkit utilizes the GEOS-Chem chemical transport model and a localized ensemble transform Kalman filter algorithm to determine the Bayesian optimal emissions and/or concentrations of a set of species. The Python-based design of CHEEREIO is commendable for its user-friendliness, allowing users to easily add new observation operators such as for satellites.
The toolkit demonstrates impressive computational performance, achieving a 50-fold improvement compared to the equivalent analytical inversion of TROPOMI observations. This advancement holds significant potential to enhance the efficiency of atmospheric chemistry research.
The paper is well-structured, with the authors providing a clear explanation of the methodology used.
- Response: We thank the reviewer for their helpful feedback, which has clarified the methodology in the manuscript and offered potential directions for future research.
Minor Comments:
- Around line 540, the authors compare the computational efficiency of the toolkit with the analytical inversion. While the superior computational efficiency of the toolkit is evident, the comparison seems incomplete without a discussion on inversion accuracy. I would recommend the authors to include a comparison of the inversion accuracy.
- Response: In lines 574-81, we add a quantitative comparison with the posterior emissions and OH adjustment of Qu et al. [2021], concluding that the simulations are globally consistent with some regional differences. We briefly discuss possible reasons for posterior differences and indicate that further comparison between these methodologies would be of interest in future work.
- The reduction in computational cost is mainly achieved by using a smaller ensemble size, primarily through the application of localization. It would be interesting to know if using the full rank ensemble size (as in the analytic solution) provides a superior solution. Also, can the analytical inversion also apply localization?
- Response: We clarify in lines 160-3 that ensemble approaches exhibit “diminishing returns” as additional simulations are added, whereas the number of simulations required for the analytical approach is set by the size of the state vector. Additionally, we offer a short explanation for why analytical inversions do not need to apply localization in lines 179-80.
- The authors briefly mention the limitations of CTM bias/error. However, the described CHEEREIO system does not seem to address errors in the model, such as perturbations in meteorological condition and physics. It would be beneficial if the authors could elaborate on how these potential sources of error are addressed in the CHEEREIO system.
- Response: We add an explanation in lines 444-8 that users can represent model transport errors by using the residual error method and equation 8 in the paper, although this is unable to correct systematic bias from meteorology or chemistry. We indicate that future expansion of CHEEREIO to coupled chemistry-weather models would fully address the CTM limitations the reviewer raises.
Response to reviewer 2.
The authors developed a user-friendly chemical data assimilation platform based on the GEOS-Chem model and the LETKF algorithm, which can optimize emissions and concentrations of different species constrained by satellite observations. The GEOS-Chem model is a powerful tool for atmospheric modeling widely used in the scientific community and the LETKF algorithm has been proven to work efficiently for both linear and nonlinear problems. I feel that the inversion toolkit developed in this paper is very important and useful, and I believe that this tool will promote the research on chemical data assimilation and emission inversions. This manuscript is well-structured and well-written. I suggest a minor revision after addressing my comments below.
- Response: We thank the reviewer for their feedback, which has better contextualized the results and offered a more intuitive view of the algorithm behavior to the reader.
- There are already a few open-source inversion tools that have been developed. I suggest that the authors briefly compare CHEEREIO and other platforms and discuss the differences and advantages of this new platform.
- Response: We added a discussion of representative open-source inversion tools in lines 121-6 and contextualized CHEEREIO’s contributions.
- I understand that due to length limitations, the authors provide detailed documentation of CHEEREIO online (cheereio.readthedocs.io), instead of presenting them within the manuscript. If possible, I suggest the authors explain a little bit more of the key parameters and settings of CHEEREIO, which could largely influence the emission inversion results. This can help the readers quickly understand the most important parts of the LETKF algorithm, as well as the most sensitive parameters that should be concerned about.
- Response: We added the localization radius as a key parameter to Table 1, and offered a brief overview of the most critical parameters in lines 294-9.
- In Sect. 4, the authors mentioned a comparison between the example application of CH4 emissions inversion with Qu et al. (2021), which optimized global CH4 emissions at the same horizontal resolution using the analytical approach. Is it possible to directly compare the posterior CH4 emissions with Qu et al. (2021) in the main text? This would be helpful to understand the differences between the LETKF algorithm and the analytical approach.
- Response: As mentioned in our response to Reviewer 1, in lines 574-81, we add a quantitative comparison with the posterior emissions and OH adjustment of Qu et al. [2021], concluding that the simulations are globally consistent with some regional differences. We briefly discuss possible reasons for posterior differences and indicate that further comparison between these methodologies would be of interest in future work.
Citation: https://doi.org/10.5194/egusphere-2023-616-AC1
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-616', Anonymous Referee #1, 05 Jun 2023
In this paper, the authors introduce CHEEREIO 1.0, a novel chemical data assimilation toolkit. This toolkit is designed to optimize emissions and concentrations of chemical species based on atmospheric observations from satellites or suborbital platforms. The toolkit utilizes the GEOS-Chem chemical transport model and a localized ensemble transform Kalman filter algorithm to determine the Bayesian optimal emissions and/or concentrations of a set of species. The Python-based design of CHEEREIO is commendable for its user-friendliness, allowing users to easily add new observation operators such as for satellites.
The toolkit demonstrates impressive computational performance, achieving a 50-fold improvement compared to the equivalent analytical inversion of TROPOMI observations. This advancement holds significant potential to enhance the efficiency of atmospheric chemistry research.
The paper is well-structured, with the authors providing a clear explanation of the methodology used.
Minor Comments:
1. Around line 540, the authors compare the computational efficiency of the toolkit with the analytical inversion. While the superior computational efficiency of the toolkit is evident, the comparison seems incomplete without a discussion on inversion accuracy. I would recommend the authors to include a comparison of the inversion accuracy.
2. The reduction in computational cost is mainly achieved by using a smaller ensemble size, primarily through the application of localization. It would be interesting to know if using the full rank ensemble size (as in the analytic solution) provides a superior solution. Also, can the analytical inversion also apply localization?
3. The authors briefly mention the limitations of CTM bias/error. However, the described CHEEREIO system does not seem to address errors in the model, such as perturbations in meteorological condition and physics. It would be beneficial if the authors could elaborate on how these potential sources of error are addressed in the CHEEREIO system.
Citation: https://doi.org/10.5194/egusphere-2023-616-RC1 -
RC2: 'Comment on egusphere-2023-616', Anonymous Referee #2, 25 Jun 2023
The authors developed a user-friendly chemical data assimilation platform based on the GEOS-Chem model and the LETKF algorithm, which can optimize emissions and concentrations of different species constrained by satellite observations. The GEOS-Chem model is a powerful tool for atmospheric modeling widely used in the scientific community and the LETKF algorithm has been proven to work efficiently for both linear and nonlinear problems. I feel that the inversion toolkit developed in this paper is very important and useful, and I believe that this tool will promote the research on chemical data assimilation and emission inversions. This manuscript is well-structured and well-written. I suggest a minor revision after addressing my comments below.
- There are already a few open-source inversion tools that have been developed. I suggest that the authors briefly compare CHEEREIO and other platforms and discuss the differences and advantages of this new platform.
- I understand that due to length limitations, the authors provide detailed documentation of CHEEREIO online (cheereio.readthedocs.io), instead of presenting them within the manuscript. If possible, I suggest the authors explain a little bit more of the key parameters and settings of CHEEREIO, which could largely influence the emission inversion results. This can help the readers quickly understand the most important parts of the LETKF algorithm, as well as the most sensitive parameters that should be concerned about.
- In Sect. 4, the authors mentioned a comparison between the example application of CH4 emissions inversion with Qu et al. (2021), which optimized global CH4 emissions at the same horizontal resolution using the analytical approach. Is it possible to directly compare the posterior CH4 emissions with Qu et al. (2021) in the main text? This would be helpful to understand the differences between the LETKF algorithm and the analytical approach.
Citation: https://doi.org/10.5194/egusphere-2023-616-RC2 -
AC1: 'Comment on egusphere-2023-616', Drew Pendergrass, 10 Jul 2023
Response to reviewer 1.
In this paper, the authors introduce CHEEREIO 1.0, a novel chemical data assimilation toolkit. This toolkit is designed to optimize emissions and concentrations of chemical species based on atmospheric observations from satellites or suborbital platforms. The toolkit utilizes the GEOS-Chem chemical transport model and a localized ensemble transform Kalman filter algorithm to determine the Bayesian optimal emissions and/or concentrations of a set of species. The Python-based design of CHEEREIO is commendable for its user-friendliness, allowing users to easily add new observation operators such as for satellites.
The toolkit demonstrates impressive computational performance, achieving a 50-fold improvement compared to the equivalent analytical inversion of TROPOMI observations. This advancement holds significant potential to enhance the efficiency of atmospheric chemistry research.
The paper is well-structured, with the authors providing a clear explanation of the methodology used.
- Response: We thank the reviewer for their helpful feedback, which has clarified the methodology in the manuscript and offered potential directions for future research.
Minor Comments:
- Around line 540, the authors compare the computational efficiency of the toolkit with the analytical inversion. While the superior computational efficiency of the toolkit is evident, the comparison seems incomplete without a discussion on inversion accuracy. I would recommend the authors to include a comparison of the inversion accuracy.
- Response: In lines 574-81, we add a quantitative comparison with the posterior emissions and OH adjustment of Qu et al. [2021], concluding that the simulations are globally consistent with some regional differences. We briefly discuss possible reasons for posterior differences and indicate that further comparison between these methodologies would be of interest in future work.
- The reduction in computational cost is mainly achieved by using a smaller ensemble size, primarily through the application of localization. It would be interesting to know if using the full rank ensemble size (as in the analytic solution) provides a superior solution. Also, can the analytical inversion also apply localization?
- Response: We clarify in lines 160-3 that ensemble approaches exhibit “diminishing returns” as additional simulations are added, whereas the number of simulations required for the analytical approach is set by the size of the state vector. Additionally, we offer a short explanation for why analytical inversions do not need to apply localization in lines 179-80.
- The authors briefly mention the limitations of CTM bias/error. However, the described CHEEREIO system does not seem to address errors in the model, such as perturbations in meteorological condition and physics. It would be beneficial if the authors could elaborate on how these potential sources of error are addressed in the CHEEREIO system.
- Response: We add an explanation in lines 444-8 that users can represent model transport errors by using the residual error method and equation 8 in the paper, although this is unable to correct systematic bias from meteorology or chemistry. We indicate that future expansion of CHEEREIO to coupled chemistry-weather models would fully address the CTM limitations the reviewer raises.
Response to reviewer 2.
The authors developed a user-friendly chemical data assimilation platform based on the GEOS-Chem model and the LETKF algorithm, which can optimize emissions and concentrations of different species constrained by satellite observations. The GEOS-Chem model is a powerful tool for atmospheric modeling widely used in the scientific community and the LETKF algorithm has been proven to work efficiently for both linear and nonlinear problems. I feel that the inversion toolkit developed in this paper is very important and useful, and I believe that this tool will promote the research on chemical data assimilation and emission inversions. This manuscript is well-structured and well-written. I suggest a minor revision after addressing my comments below.
- Response: We thank the reviewer for their feedback, which has better contextualized the results and offered a more intuitive view of the algorithm behavior to the reader.
- There are already a few open-source inversion tools that have been developed. I suggest that the authors briefly compare CHEEREIO and other platforms and discuss the differences and advantages of this new platform.
- Response: We added a discussion of representative open-source inversion tools in lines 121-6 and contextualized CHEEREIO’s contributions.
- I understand that due to length limitations, the authors provide detailed documentation of CHEEREIO online (cheereio.readthedocs.io), instead of presenting them within the manuscript. If possible, I suggest the authors explain a little bit more of the key parameters and settings of CHEEREIO, which could largely influence the emission inversion results. This can help the readers quickly understand the most important parts of the LETKF algorithm, as well as the most sensitive parameters that should be concerned about.
- Response: We added the localization radius as a key parameter to Table 1, and offered a brief overview of the most critical parameters in lines 294-9.
- In Sect. 4, the authors mentioned a comparison between the example application of CH4 emissions inversion with Qu et al. (2021), which optimized global CH4 emissions at the same horizontal resolution using the analytical approach. Is it possible to directly compare the posterior CH4 emissions with Qu et al. (2021) in the main text? This would be helpful to understand the differences between the LETKF algorithm and the analytical approach.
- Response: As mentioned in our response to Reviewer 1, in lines 574-81, we add a quantitative comparison with the posterior emissions and OH adjustment of Qu et al. [2021], concluding that the simulations are globally consistent with some regional differences. We briefly discuss possible reasons for posterior differences and indicate that further comparison between these methodologies would be of interest in future work.
Citation: https://doi.org/10.5194/egusphere-2023-616-AC1
Peer review completion
Journal article(s) based on this preprint
Drew C. Pendergrass et al.
Data sets
Replication Data for: CHEEREIO 1.0: a versatile and user-friendly ensemble-based chemical data assimilation and emissions inversion platform for the GEOS-Chem chemical transport model Drew C. Pendergrass, Daniel J. Jacob, Hannah Nesser, Daniel J. Varon, Melissa Sulprizio, Kazuyuki Miyazaki, and Kevin W. Bowman https://doi.org/10.5281/zenodo.7806312
Model code and software
CHEEREIO Drew C. Pendergrass, Daniel J. Jacob, Hannah Nesser, Daniel J. Varon, Melissa Sulprizio, Kazuyuki Miyazaki, and Kevin W. Bowman https://doi.org/10.5281/zenodo.7781437
Drew C. Pendergrass et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 365 | 156 | 14 | 535 | 5 | 5 |
- HTML: 365
- PDF: 156
- XML: 14
- Total: 535
- BibTeX: 5
- EndNote: 5
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Cited
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(3159 KB) - Metadata XML