the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 13 Oct 2023
Identifying episodic carbon monoxide emission events in the MOPITT measurement dataset
Abstract. The Measurements Of Pollution In The Troposphere (MOPITT) instrument aboard NASA’s Terra satellite has been measuring upwelling radiance in a nadir-viewing mode since March 2000. These radiance measurements are inverted to yield estimates of carbon monoxide (CO) profiles and total columns, providing the longest satellite record of this trace gas to date. The CO measurements from MOPITT have been used in a variety of ways, including trend analyses and the construction of CO budgets. However, their use is complicated by the influence of episodic emission events, which release large quantities of CO into the atmosphere with irregular timing, such as large sporadic wildfires of natural or anthropogenic origin. The chaotic nature of these events is a large source of variability in CO budgets and models, requiring that these events be well characterized in order to develop an improved understanding of the role they have in influencing tropospheric CO. This study describes the development of a multi-step algorithm that is used to identify large episodic emission events using daily-mean Level 2 (L2) MOPITT total column measurements gridded to 0.5° by 0.5° spatial resolution. The core component of this procedure involves empirically determining the expectation density function (EDF) that describes the departure of daily-mean CO observations from the baseline behaviour of CO, as described by its periodic components and trends. The EDFs employed are not assumed to be symmetric, but instead are constructed from a pair of superimposed normal distributions. Enhancement flag files are produced following this methodology, identifying the episodic events that show strong enhancement of CO outside of the range of expected CO behaviour, and are now made available for the period 3 March 2000 to 31 July 2022. The distribution and frequency of these flagged measurements over this 22-year period is analyzed, to illustrate the robustness of this method.
-
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
(2234 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2234 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-2219', Shyno Susan John, 17 Nov 2023
First, I would like to congratulate the authors for this excellent and tremendous amount of work.
Here are a few comments I would like to point out on this manuscript:
- Along with the influence of ENSO, the other key climate drivers like, Indian Ocean Dipole (IOD) and Madden-Julian Oscillation (MJO, especially in sub-seasonal time scale in southeast Asia) could have an impact on the trends in CO. Have you already considered or analyzed these effects?. It would be worth updating the MLR based on these trends from the above mentioned climate driver impacts.
- A summary table of events including the threshold value for the key regions (with the event classifications) in the study could have been useful.
- It would be good to add some information about the episodic total column CO intensities also in addition to the frequency and distribution.
- It would be relevant to include the method used to re-grid MOPITT data to 0.5 deg.
Citation: https://doi.org/10.5194/egusphere-2023-2219-RC1 -
RC2: 'Comment on egusphere-2023-2219', Anonymous Referee #2, 25 Nov 2023
The study presents statistical analysis of outliers in satellite-measured CO using the MOPITT instrument. The long term measurements from MOPITT (2000-2022) make it an ideal instrument for evaluating outliers over a long time period. The authors use level 2, thermal infrared retrievals, regridded to 0.5 degree resolution, and evaluate temporal properties for each 0.5 x 0.5 degree grid box. They use advanced statistical techniques to agnostically quantify outliers. Outliers are evaluated for persistence and strength and linked to known emission events, which are often, but not always, biomass burning. Overall the manuscript is very well written and the analysis rigorous. I have several comments to be addressed below.
------------------------------------------------------
Main Comments:
1. Temporal evaluation of outliers: The authors may have missed an opportunity to temporally evaluate outliers. Specifically, it would be interesting to test whether a temporal trend exists for data in Figure 4 (the number of days flagged as affected by outliers) or alternatively compare 2000-2011 values with 2012-2022 to see if there is a change in the number of outlier days between periods.
2. Comments regarding the EDF:
(i) Choice of threshold: The authors choose a threshold of 0.05 tolerance (95% confidence) to determine outliers. Were different thresholds explored? For example does a threshold of 90% confidence include more of the burning events. Does a threshold of 99% confidence reduce the number of outliers over Antarctica?
(ii) Line 162: Is the EDF bimodal everywhere, or are some grid points explained by a unimodal gaussian and others by bimodal?
(iii) Can anything be interpreted about the datapoints that are above the EDF, but below the threshold? For example at around residual values of -10 and between 40 and 50 in Figure 2. Are these used to determine how well the EDF fits the data?
(iv) Figure 2: Consider showing the equivalent unimodal gaussian fit as a supplement figure to demonstrate how it is inadequate.3. Line 196-197: Is there evidence of atmospheric transport from the Northern Territory to NSW for the November 6, 2002 outlier? I would suggest fires in Victoria or within NSW would be much more likely to contribute to NSW outliers. Fires began in Victoria in September 2002.
4. Paragraph starting 272: This could be due to the choice of colorbar, but I do not see evidence in Figure 5 for the patterns discussed about the Equatorial Africa region – specifically the westward transport is not evident to me in Figure 5.
Minor Comments/Suggestions:
L118 to L123: Consider adding in a discussion about how additional aerosol scenes, which have been previously mis-identified as clouds, allow for evaluation of more data points in the MOPITT record, especially for polluted scenes.
L138: Please clarify the 15 day moving window multi-year average. Is it a 15-day moving-window climatology, 2000-2022?
Figure 1 caption: Specify the regression fit is to account for ENSO influence.
Figure 2 caption: add in (r$_{th}$) after “threshold value”.
L159: I suggest to provide an example of the interquartile range for the data in Figure 2.
L160: Clarify that data for each 0.5 degree grid point will have a different bin width.
L217 to L224: How were the sources that contributed to the observed threshold values determined or identified?
Citation: https://doi.org/10.5194/egusphere-2023-2219-RC2 -
AC1: 'Response to Reviewers', Paul Jeffery, 26 Jan 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-2219/egusphere-2023-2219-AC1-supplement.pdf
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-2219', Shyno Susan John, 17 Nov 2023
First, I would like to congratulate the authors for this excellent and tremendous amount of work.
Here are a few comments I would like to point out on this manuscript:
- Along with the influence of ENSO, the other key climate drivers like, Indian Ocean Dipole (IOD) and Madden-Julian Oscillation (MJO, especially in sub-seasonal time scale in southeast Asia) could have an impact on the trends in CO. Have you already considered or analyzed these effects?. It would be worth updating the MLR based on these trends from the above mentioned climate driver impacts.
- A summary table of events including the threshold value for the key regions (with the event classifications) in the study could have been useful.
- It would be good to add some information about the episodic total column CO intensities also in addition to the frequency and distribution.
- It would be relevant to include the method used to re-grid MOPITT data to 0.5 deg.
Citation: https://doi.org/10.5194/egusphere-2023-2219-RC1 -
RC2: 'Comment on egusphere-2023-2219', Anonymous Referee #2, 25 Nov 2023
The study presents statistical analysis of outliers in satellite-measured CO using the MOPITT instrument. The long term measurements from MOPITT (2000-2022) make it an ideal instrument for evaluating outliers over a long time period. The authors use level 2, thermal infrared retrievals, regridded to 0.5 degree resolution, and evaluate temporal properties for each 0.5 x 0.5 degree grid box. They use advanced statistical techniques to agnostically quantify outliers. Outliers are evaluated for persistence and strength and linked to known emission events, which are often, but not always, biomass burning. Overall the manuscript is very well written and the analysis rigorous. I have several comments to be addressed below.
------------------------------------------------------
Main Comments:
1. Temporal evaluation of outliers: The authors may have missed an opportunity to temporally evaluate outliers. Specifically, it would be interesting to test whether a temporal trend exists for data in Figure 4 (the number of days flagged as affected by outliers) or alternatively compare 2000-2011 values with 2012-2022 to see if there is a change in the number of outlier days between periods.
2. Comments regarding the EDF:
(i) Choice of threshold: The authors choose a threshold of 0.05 tolerance (95% confidence) to determine outliers. Were different thresholds explored? For example does a threshold of 90% confidence include more of the burning events. Does a threshold of 99% confidence reduce the number of outliers over Antarctica?
(ii) Line 162: Is the EDF bimodal everywhere, or are some grid points explained by a unimodal gaussian and others by bimodal?
(iii) Can anything be interpreted about the datapoints that are above the EDF, but below the threshold? For example at around residual values of -10 and between 40 and 50 in Figure 2. Are these used to determine how well the EDF fits the data?
(iv) Figure 2: Consider showing the equivalent unimodal gaussian fit as a supplement figure to demonstrate how it is inadequate.3. Line 196-197: Is there evidence of atmospheric transport from the Northern Territory to NSW for the November 6, 2002 outlier? I would suggest fires in Victoria or within NSW would be much more likely to contribute to NSW outliers. Fires began in Victoria in September 2002.
4. Paragraph starting 272: This could be due to the choice of colorbar, but I do not see evidence in Figure 5 for the patterns discussed about the Equatorial Africa region – specifically the westward transport is not evident to me in Figure 5.
Minor Comments/Suggestions:
L118 to L123: Consider adding in a discussion about how additional aerosol scenes, which have been previously mis-identified as clouds, allow for evaluation of more data points in the MOPITT record, especially for polluted scenes.
L138: Please clarify the 15 day moving window multi-year average. Is it a 15-day moving-window climatology, 2000-2022?
Figure 1 caption: Specify the regression fit is to account for ENSO influence.
Figure 2 caption: add in (r$_{th}$) after “threshold value”.
L159: I suggest to provide an example of the interquartile range for the data in Figure 2.
L160: Clarify that data for each 0.5 degree grid point will have a different bin width.
L217 to L224: How were the sources that contributed to the observed threshold values determined or identified?
Citation: https://doi.org/10.5194/egusphere-2023-2219-RC2 -
AC1: 'Response to Reviewers', Paul Jeffery, 26 Jan 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-2219/egusphere-2023-2219-AC1-supplement.pdf
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 222 | 91 | 19 | 332 | 12 | 9 |
- HTML: 222
- PDF: 91
- XML: 19
- Total: 332
- BibTeX: 12
- EndNote: 9
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Paul S. Jeffery
James R. Drummond
Jiansheng Zou
Kaley A. Walker
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2234 KB) - Metadata XML