the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 02 Jan 2023
Monitoring and quantifying CO2 emissions of isolated power plants from space
Abstract. Abstract: Top-down CO2 emission estimates based on satellite observations are potentially of great importance for independently verifying the accuracy of reported emissions and emission inventories. Difficulties in verifying these satellite-derived emissions arise from the fact that emission inventories often provide annual mean emissions while estimates from satellites are available only for a limited number of overpasses. Previous studies have derived CO2 emissions for power plants from OCO-2 and OCO-3 observations of their exhaust plumes, but the accuracy and the factors affecting these emissions are uncertain. We have selected only isolated power plants for this study, to avoid complications link to multiple sources in close proximity. We first compare the Gaussian plume model and cross-sectional flux methods for estimating CO2 emission of power plants. Then we examined the sensitivity of the emission estimates to possible choices for the wind field. For verification we have used power plant emissions that are reported on an hourly basis by the Environmental Protection Agency (EPA) in the United States. By using the OCO-2 and OCO-3 observations over the past four years we identified emission signals of isolated power plants and arrived at a total of 50 collocated cases involving 22 power plants. We correct for the time difference between the moment of the emission and the satellite observation. We found the wind field halfway the height of planetary boundary layer (PBL) yielded the best results. We found that the instantaneous satellite estimated emissions of these 50 cases and reported emissions display a weak correlation (R2=0.12). The correlation improves with averaging over multiple observations of the 22 power plants (R2=0.40). The method was subsequently applied to 106 power plants cases worldwide yielded a total emission of 1522 ± 501 Mt CO2/year, estimated to be about 17 % of the power sector emissions of our selected countries. The improved correlation highlights the potential for future planned satellite missions with a greatly improved coverage to monitor a significant fraction of global power plant emissions.
-
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
(1356 KB)
-
Supplement
(1026 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1356 KB) - Metadata XML
-
Supplement
(1026 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-1490', Ray Nassar, 24 Jan 2023
Lin et al. “Monitoring and quantifying CO2 emissions of isolated power plants from space” builds off previous work on quantifying power plant emissions using OCO-2 and OCO-3 observations together with models. It is good to see this effort toward development of a more systematic and automated method that leverages what has been demonstrated by others in past case studies. Furthermore, the comparison between the Gaussian plume method (GPM) and Integrated Mass Enhancement (IME) method is a useful investigation that highlights the importance of the satellite coverage and resolution and the different nature of CO2 and CH4 plumes since the conclusion differs from that based on high spatial resolution CH4 observations in the literature. Overall, this is a useful study that helps to bring the field a step closer to the implementation of an operational system for CO2 anthropogenic emission monitoring as planned for CO2M. Following some minor revisions related to the specific points below, I would recommend its publication.
Specific Points
Line 43-44: These are not really the primary references regarding the difficulty to achieve accurate and detailed consumption data
Line 63: Reuter et al. (2019) derived emission estimates for power plants, urban areas and wild fires
Line 66: Nassar et al. (2022) https://www.frontiersin.org/articles/10.3389/frsen.2022.1028240/full is a key OCO-3 example worth mentioning
Line 71: Schwandner et al. 2017 is not the best choice of reference. Although the paper mentions power plants, it really focuses on XCO2 enhancements in an urban area (later understood to be topography related biases), while the only emission estimate is of volcanic emissions from one cloudy overpass
Line 74: “manually-selected” is perhaps a better descriptor than “hand-picked” (slang)
Line 79: Intermittency of U.S. sources has previously been studied by Hill and Nassar (2019) https://doi.org/10.3390/rs11131608 and Velazco et al. (2011) www.atmos-meas-tech.net/4/2809/2011/, so these two past studies should be cited.
Line 97: “≤ 1.29 x 2.25 km2” (It is worth noting that this is the maximum footprint size, since it is usually smaller due to solar angle and viewing geometry)
Line 97: “~52” degrees is recommended since the value can be exceeded by a few tenths of a degree in some cases
Line 111: daily global coverage before loss of data due to clouds
Line 119: This EPA link has annual power plant emission data, but is it the correct link for the hourly data too?
Line 257: Nassar et al. 2021 used the assumed height of the chimney plus an assumed 250 m for typical plume rise above the stack height
Line 295: For clarify, it would be helpful to specify that the x-axis is labelled with YYMMDD.
Line 374: Should revise language about GeoCarb as it has recently been cancelled by NASA.
Line 375: CO2M is a Copernicus mission with ESA and EUMETSAT involvement
Citation: https://doi.org/10.5194/egusphere-2022-1490-RC1 -
AC1: 'Reply on RC1', Xiaojuan Lin, 27 Mar 2023
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2022-1490/egusphere-2022-1490-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Xiaojuan Lin, 27 Mar 2023
-
RC2: 'Comment on egusphere-2022-1490', Gerrit Kuhlmann, 31 Jan 2023
The study estimates CO2 emission of power plant using OCO-2 and OCO-3 observations using the Gaussian plume inversion and cross-sectional flux method with different input parameters. The methods are tested for U.S. power plant for which bottom-up reports of hourly emissions are available and afterwards applied globally. The paper well written but some aspects on the method are unclear. I would recommend publication following a revision based on the general and specific comments below:
- Background: For Gaussian plume model method please describe already in L153ff how you calculate the background. In L200, you write that the 90th percentile was used, which seems to be based on tests with different percentiles with the aim to minimize the difference between estimated and reported emissions (L220, Figure S2). The choice of the percentile (60-99th) will mostly result in a bias in the estimated emissions, which might be caused by the background, but can also be the result of other systematic errors in other parameters (e.g. effective wind speed). Therefore, how does this choice of the background agree with the background that you compute with the cross sectional flux method (L175)? Would a different background affect your conclusions on the best approach for computing the effective wind speed?
- Wind: The evaluation of the different wind products in your study is inconsistent. You use winds from ERA-5 (0.25°), MERRA-2 (0.5°x0.625°) and high-resolution ECWMF forecast (resolution not mentioned). You use the wind speed at the center of the PBL for the ECMWF forecast. However, for ERA-5 and MERRA-2, you take 10-m winds multiplied by the empirical scaling factor of 1.4 from Varon et al. (2018). When you compare the impact of the different wind products on the estimated emissions, it is not possible to identify if the different performances are caused by differences between the products or the different computation of the final product (scaling factor or wind at half PBL height). To analyse this better, I suggest comparing all datasets using both the 1.4-factor and the wind at half the height of the PBL. Note that the scaling factor of 1.4 is derived for CH4 plumes measured by high-resolution satellites, which are inherently different to CO2 plumes from power plants measured by OCO-2. Therefore, using the value might be not the best approach, even it is true that it has been used in previous studies "for convenience" (Reuter et al. 2019).
- Uncertainties: You seem to compute the uncertainties using an ensemble approach with a rather small number of members (3 for wind and 4 for background) for computing reasonable statistics (see also previous comment on the wind). How large are the uncertainties of wind speed in m/s and background in ppm? How do these uncertainty estimates compare to estimated uncertainties in previous studies? How large is the uncertainty of the fitting parameters for the background in Equation 3? Finally, how large would be the uncertainties from the assumption and simplification of your methods such as the assumption of steady state conditions?
Specific comments
L171ff: You write here that you fit equation 3 to obtain parameters, k, b, A and sigma. Then, you write that you subtract the background to compute the line density. However, your parameter A is already the line density, so I don't understand why you need to calculate it again.
L182: It is not clear to me how you compute the wind here. Do you rotate the wind vector so that it points from the source location to the maximum in the OCO swath?
L235: You write that the peak is well described by a Gaussian [curve] in Figure 1b. However, no curve is shown in the figure.
L240ff: You partly repeat the description of your method here, which seems unnecessary.
L261: Please discuss why WPBL provides better results than the other two options.
L265: Do you use the arithmetic average or the weighted average considering the uncertainty of the estimates?
L268/Figure 2: It surprises me that r² is so much higher for summing compared to averaging? Can you explain why this is the case?
L279: This relates back to my previous comment how you do calculate the normal wind for both methods. Are the difference between estimates and wind speed used in both method correlated? Another reason for deviations can be the method for computing the background.
Figure 5: The red line is somewhat misleading, because without reading the caption one would assume that you could estimate emissions for 8% of all tracks near power plant, while in truths it is only 0.05%. I would strongly suggest removing it to avoid confusion.
L340ff: Does this number of 1522 Mt/a correctly accounts for observing the same power plant in different years or does this never happens? I am asking because the percentage numbers for the individual years add up exactly to 17, which would not happen if you estimate for the same power plant more than once.
Figure 7b: It is difficult to see the bars for most countries. Maybe the figure can use a logarithmic scale on the y-axis.
L367f: The conclusion on the difference between cross-sectional flux and Gaussian plume method needs more explanations (see previous comment).
L374f: Unfortunately, GeoCarb was recently canceled. CO2M is developed by ESA and EU with involvement from EUMETSAT and ECMWF. It is probably easiest to call it a "European mission". The Japanese GOSAT-GW should be mention here, as well.
Supplement: The resolution of the figures in the supplement is very low making it difficult to read the labels. In some cases, labels and units are missing (e.g., S9).
Citation: https://doi.org/10.5194/egusphere-2022-1490-RC2 -
AC2: 'Reply on RC2', Xiaojuan Lin, 27 Mar 2023
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2022-1490/egusphere-2022-1490-AC2-supplement.pdf
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-1490', Ray Nassar, 24 Jan 2023
Lin et al. “Monitoring and quantifying CO2 emissions of isolated power plants from space” builds off previous work on quantifying power plant emissions using OCO-2 and OCO-3 observations together with models. It is good to see this effort toward development of a more systematic and automated method that leverages what has been demonstrated by others in past case studies. Furthermore, the comparison between the Gaussian plume method (GPM) and Integrated Mass Enhancement (IME) method is a useful investigation that highlights the importance of the satellite coverage and resolution and the different nature of CO2 and CH4 plumes since the conclusion differs from that based on high spatial resolution CH4 observations in the literature. Overall, this is a useful study that helps to bring the field a step closer to the implementation of an operational system for CO2 anthropogenic emission monitoring as planned for CO2M. Following some minor revisions related to the specific points below, I would recommend its publication.
Specific Points
Line 43-44: These are not really the primary references regarding the difficulty to achieve accurate and detailed consumption data
Line 63: Reuter et al. (2019) derived emission estimates for power plants, urban areas and wild fires
Line 66: Nassar et al. (2022) https://www.frontiersin.org/articles/10.3389/frsen.2022.1028240/full is a key OCO-3 example worth mentioning
Line 71: Schwandner et al. 2017 is not the best choice of reference. Although the paper mentions power plants, it really focuses on XCO2 enhancements in an urban area (later understood to be topography related biases), while the only emission estimate is of volcanic emissions from one cloudy overpass
Line 74: “manually-selected” is perhaps a better descriptor than “hand-picked” (slang)
Line 79: Intermittency of U.S. sources has previously been studied by Hill and Nassar (2019) https://doi.org/10.3390/rs11131608 and Velazco et al. (2011) www.atmos-meas-tech.net/4/2809/2011/, so these two past studies should be cited.
Line 97: “≤ 1.29 x 2.25 km2” (It is worth noting that this is the maximum footprint size, since it is usually smaller due to solar angle and viewing geometry)
Line 97: “~52” degrees is recommended since the value can be exceeded by a few tenths of a degree in some cases
Line 111: daily global coverage before loss of data due to clouds
Line 119: This EPA link has annual power plant emission data, but is it the correct link for the hourly data too?
Line 257: Nassar et al. 2021 used the assumed height of the chimney plus an assumed 250 m for typical plume rise above the stack height
Line 295: For clarify, it would be helpful to specify that the x-axis is labelled with YYMMDD.
Line 374: Should revise language about GeoCarb as it has recently been cancelled by NASA.
Line 375: CO2M is a Copernicus mission with ESA and EUMETSAT involvement
Citation: https://doi.org/10.5194/egusphere-2022-1490-RC1 -
AC1: 'Reply on RC1', Xiaojuan Lin, 27 Mar 2023
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2022-1490/egusphere-2022-1490-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Xiaojuan Lin, 27 Mar 2023
-
RC2: 'Comment on egusphere-2022-1490', Gerrit Kuhlmann, 31 Jan 2023
The study estimates CO2 emission of power plant using OCO-2 and OCO-3 observations using the Gaussian plume inversion and cross-sectional flux method with different input parameters. The methods are tested for U.S. power plant for which bottom-up reports of hourly emissions are available and afterwards applied globally. The paper well written but some aspects on the method are unclear. I would recommend publication following a revision based on the general and specific comments below:
- Background: For Gaussian plume model method please describe already in L153ff how you calculate the background. In L200, you write that the 90th percentile was used, which seems to be based on tests with different percentiles with the aim to minimize the difference between estimated and reported emissions (L220, Figure S2). The choice of the percentile (60-99th) will mostly result in a bias in the estimated emissions, which might be caused by the background, but can also be the result of other systematic errors in other parameters (e.g. effective wind speed). Therefore, how does this choice of the background agree with the background that you compute with the cross sectional flux method (L175)? Would a different background affect your conclusions on the best approach for computing the effective wind speed?
- Wind: The evaluation of the different wind products in your study is inconsistent. You use winds from ERA-5 (0.25°), MERRA-2 (0.5°x0.625°) and high-resolution ECWMF forecast (resolution not mentioned). You use the wind speed at the center of the PBL for the ECMWF forecast. However, for ERA-5 and MERRA-2, you take 10-m winds multiplied by the empirical scaling factor of 1.4 from Varon et al. (2018). When you compare the impact of the different wind products on the estimated emissions, it is not possible to identify if the different performances are caused by differences between the products or the different computation of the final product (scaling factor or wind at half PBL height). To analyse this better, I suggest comparing all datasets using both the 1.4-factor and the wind at half the height of the PBL. Note that the scaling factor of 1.4 is derived for CH4 plumes measured by high-resolution satellites, which are inherently different to CO2 plumes from power plants measured by OCO-2. Therefore, using the value might be not the best approach, even it is true that it has been used in previous studies "for convenience" (Reuter et al. 2019).
- Uncertainties: You seem to compute the uncertainties using an ensemble approach with a rather small number of members (3 for wind and 4 for background) for computing reasonable statistics (see also previous comment on the wind). How large are the uncertainties of wind speed in m/s and background in ppm? How do these uncertainty estimates compare to estimated uncertainties in previous studies? How large is the uncertainty of the fitting parameters for the background in Equation 3? Finally, how large would be the uncertainties from the assumption and simplification of your methods such as the assumption of steady state conditions?
Specific comments
L171ff: You write here that you fit equation 3 to obtain parameters, k, b, A and sigma. Then, you write that you subtract the background to compute the line density. However, your parameter A is already the line density, so I don't understand why you need to calculate it again.
L182: It is not clear to me how you compute the wind here. Do you rotate the wind vector so that it points from the source location to the maximum in the OCO swath?
L235: You write that the peak is well described by a Gaussian [curve] in Figure 1b. However, no curve is shown in the figure.
L240ff: You partly repeat the description of your method here, which seems unnecessary.
L261: Please discuss why WPBL provides better results than the other two options.
L265: Do you use the arithmetic average or the weighted average considering the uncertainty of the estimates?
L268/Figure 2: It surprises me that r² is so much higher for summing compared to averaging? Can you explain why this is the case?
L279: This relates back to my previous comment how you do calculate the normal wind for both methods. Are the difference between estimates and wind speed used in both method correlated? Another reason for deviations can be the method for computing the background.
Figure 5: The red line is somewhat misleading, because without reading the caption one would assume that you could estimate emissions for 8% of all tracks near power plant, while in truths it is only 0.05%. I would strongly suggest removing it to avoid confusion.
L340ff: Does this number of 1522 Mt/a correctly accounts for observing the same power plant in different years or does this never happens? I am asking because the percentage numbers for the individual years add up exactly to 17, which would not happen if you estimate for the same power plant more than once.
Figure 7b: It is difficult to see the bars for most countries. Maybe the figure can use a logarithmic scale on the y-axis.
L367f: The conclusion on the difference between cross-sectional flux and Gaussian plume method needs more explanations (see previous comment).
L374f: Unfortunately, GeoCarb was recently canceled. CO2M is developed by ESA and EU with involvement from EUMETSAT and ECMWF. It is probably easiest to call it a "European mission". The Japanese GOSAT-GW should be mention here, as well.
Supplement: The resolution of the figures in the supplement is very low making it difficult to read the labels. In some cases, labels and units are missing (e.g., S9).
Citation: https://doi.org/10.5194/egusphere-2022-1490-RC2 -
AC2: 'Reply on RC2', Xiaojuan Lin, 27 Mar 2023
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2022-1490/egusphere-2022-1490-AC2-supplement.pdf
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 1,032 | 378 | 26 | 1,436 | 66 | 9 | 18 |
- HTML: 1,032
- PDF: 378
- XML: 26
- Total: 1,436
- Supplement: 66
- BibTeX: 9
- EndNote: 18
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| China | 1 | 450 | 30 |
| United States of America | 2 | 372 | 25 |
| Netherlands | 3 | 94 | 6 |
| France | 4 | 67 | 4 |
| United Kingdom | 5 | 66 | 4 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 450
Xiaojuan Lin
Ronald van der A
Jos de Laat
Henk Eskes
Frédéric Chevallier
Philippe Ciais
Zhu Deng
Yuanhao Geng
Xuanren Song
Xiliang Ni
Da Huo
Xinyu Dou
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1356 KB) - Metadata XML
-
Supplement
(1026 KB) - BibTeX
- EndNote
- Final revised paper