the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Independent validation of IASI/METOP-A LMD and RAL CH4 products using CAMS model, in situ profiles and ground-based FTIR measurements
Abstract. In this study, we carried out an independent validation of two methane retrieval algorithms using spectra from the Infrared Atmospheric Sounding Interferometer (IASI) onboard the Meteorological Operational satellite programme-A (MetOp-A) since 2006. Both algorithms, one developed by the Laboratoire de Météorologie Dynamique (LMD), called the non-linear inference scheme (NLISv8.3), the other by the Rutherford Appleton Laboratory (RAL), referred to as RALv2.0, provide longterm global CH4 concentrations using distinctively different retrieval approaches (Neural Network vs. Optimal Estimation, respectively). They also differ with respect to the vertical range covered, where LMD provides mid-tropospheric dry air mole fractions (mtCH4) and RAL provides mixing ratio profiles from which we can derive total column-averaged dry air mole fractions (XCH4) and potentially 2 partial column layers (qCH4).
We compared both CH4 products using the Copernicus Atmospheric Monitoring Service (CAMS) model, in situ profiles (range extended using CAMS model data) and ground-based Fourier transform infrared (FTIR) remote sensing measurements. The average difference in mtCH4 with respect to in situ profiles for LMD ranges between -0.3 and 10.9 ppb while for RAL the XCH4 difference ranges between -10.2 and -4.6 ppb. The standard deviation (stdv) of the observed differences between in situ and RAL retrievals is 14.5–23.0 ppb, which is consistently smaller than that between in situ and LMD retrievals about 15.2–30.6 ppb. By comparing with ground-based FTIR sites, the mean difference is within ±10 ppb for both RAL and LMD retrievals. However, the stdv of the differences at the ground-based FTIR stations show significantly lower values for RAL (11–16 ppb) than those for LMD (about 25 ppb).
The long-term stability and seasonal cycles of CH4 derived from the LMD and RAL products are further investigated and discussed. The seasonal variation of XCH4 derived from RAL is consistent with the seasonal variation observed by the groundbased FTIR measurements. However, the overall 2007–2015 XCH4 trend derived from RAL measurements is underestimated if not adjusted for an anomaly occurring on 16 May 2013 due to a L1 calibration change. For LMD, we see very good agreement at the (sub)tropics (<35° N–35° S), but notice deviations of the seasonal cycle (both in the amplitude and phase) and an underestimation of the long-term trend with respect to the RAL and reference data at higher latitude sites.
- Preprint
(12146 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on amt-2023-237', Anonymous Referee #1, 29 Jan 2024
1) General comments
This study compares the accuracy of two algorithms in determining methane (CH4) concentrations from IASI sensor data. The retrieval results with the two algorithms are compared with each other and with independent observations and model data using various methods.
It is crucial to validate IASI CH4 data comprehensively for scientific use, and this paper's scientific significance is high. Therefore, I believe this paper's content deserves to be published in AMT.
However, the structure and description of this paper are somewhat confusing. In particular, the intercomparison part in Section 4 needs a more organized structure. The "short summary" subsection should be included in each intercomparison subsection, with tables, to aid readers in comprehending the validation results.
Also, regarding the gap in the data for 2013, it is important to include clear communication in the data section that this issue exists. Then, a comparative validation distinguishing between before and after data, or at least an intercomparison between before and after RAL and LMD, should be performed to show the presence or absence of an impact.
In addition, several figures need more legends or are unreadable. Please improve visibility.
Considering the above, I recommend that this paper should be published after examining the issues and making necessary revisions.
2) Specific comments
・Page3 Line20 (P3L20): about IFOV of RAL measurements
Why use the IFOV with the highest brightness temperature among the four IFOVs instead of the average of the four?
・P8L19: about vertical sensitivity of LMD and RAL_LMDavk
You note that there is a difference in altitude concerning the sensitivity of the two, but how much difference would this be in mtCH4? Can mtCH4 be evaluated using, for example, climatological values?
・P8L23: about grid averaging
Why did you use a 1-degree grid for intercomparison? This corresponds to a 100 km grid on a spatial scale, which might be too large, considering that the source of CH4 is more localized than CO2.
・P9L4 and P9L10-L17: about selection criteria
The time frame for simultaneous observation with IASI is currently at ±6 hours. However, I believe this interval may be too long, considering the horizontal and vertical transport of CH4. CH4 distribution can vary significantly over such a period, so it may not be accurate to consider the observation as genuinely simultaneous.
Additionally, comparing ground-based FTIRs to the satellite dataset is unfair due to the differences in time and longitude ranges for each comparison. Can the conditions be as consistent as possible?
・P9L5: about the mean of the satellite values
How many satellite data points are usually averaged? And how large is the variability (stdv) of these data?
・P10L21: about eq.(7)
Please tell me how this equation was derived. Also, please explain what C'r,R on the left side represents.
・P11L9, L24, and L28: about equations (9), (11), (12)
What does the right-hand side of equations (9), (11), and (12) represent, respectively? Please add an explanation.
・P13L5-L6: about internal consistency
Why was the internal consistency only checked in October 2014? Could there be differences depending on the season, especially summer and winter? Also, is there any impact of the 2013 gap?
・P13L30: IFOV selection
The authors claim that "RAL selects the best IFOV among 4 of them.", but in the previous section they mention "the one with the highest brightness temperature". Why is this the best?
・P14L16: about SAT
What is “SAT”? Does SAT mean RAL and/or LMD measurements?
・P15L13-L14: about the correlation between HIPPO and RAL measurements
What is the correlation between the two values obtained from the best fit?
・P15L15-L18: comparison with IAGOS
Why is the correlation coefficient with IAGOS lower than other comparisons? Is such a low correlation coefficient due to the spatial bias of IAGOS measurements?
・P17L24-P19L9: about subsection 4.7
This subsection repeats previous statements, making it redundant. It would be more useful to list each table in a related subsection.
・P19L10-P19L15: about discontinuity in RAL L2 data in mid-2013
Does the discontinuity in mid-2013 also affect the intercomparisons made in Section 4? For example, the comparison with AirCore and IAGOS was made using all data before and after the discontinuity. Would there be a difference in bias before and after?
Also, on page 23, the authors only compared partial columns between RAL and CAMS in 2012. Is there any difference in partial column bias before and after 2013?
・Figure 5, 6, 7, 12 and 13:
The legend should be relocated or resized to avoid overlapping the plots. In Figure 13, please add a legend.
・Figure 5, 6 and 7:
Could you please explain the meaning of the dotted lines in the correlation diagram?
3) Technical corrections
P2L9: andh => and
P4L23: (about 30 km; (Karion et al., 2010)) => (about 30 km) (Karion et al., 2010).
P5L3: CAMS model) as => CAMS model as
P7L9: in (Massart et al., 2014) => in Massart et al. (2014)
P9L20: According to (Rodgers and Connor, 2003) => According to Rodgers and Connor (2003),
P10L4: “ci” in bold should be plain.
P10L7: respectively; => respectively,
P10L7: XCH4 => XCH4. (need period)
P15L23, L28 and L29: XCH4 => qCH4
Table 1: ParkFalls => Park Falls?
Figure 1: Rikubetsu should be “TCCON (yellow)”, not “NDACC (green)”.
Figure 4: “The last three rows show …” maybe “The last two rows show …”.
Figure 6: In the x- and y-axis labels, “XCH4” should be “qCH4”.
Citation: https://doi.org/10.5194/amt-2023-237-RC1 - AC1: 'Reply on RC1', Bart Dils, 09 Mar 2024
-
RC2: 'Comment on amt-2023-237', Anonymous Referee #2, 29 Jan 2024
Comment on “Independent validation of IASI/METOP-A LMD and RAL CH4 products using CAMS model, in situ profiles and ground-based FTIR measurements” by Bart Dils et al.
General:
This paper describes the validations of two methane products from IASI/METOP-A satellite sensor. This work is important, but I couldn’t understand why the authors think that it is enough to use RAL data without correction. They pointed out there is a discontinuity in mid-2013 and compared some trends in Section 5.4. The corrected trend (5.6 ppb/yr) is significantly higher than the original one (4.2 ppb/yr) and the difference is larger than the standard deviation of original one. I think the trend analyses in Section 5.1 and 5.2 should be done using corrected data, or at least separately for the periods before and after mid-2013. Furthermore, the discontinuity should affect the results of validation in Section 4 but there is no explanation. The validation of RAL should also be done using corrected data, or separately for the periods before and after mid-2013.
The paper should be published after major revisions.
Comments and questions:
Abstract
p1, l17
The long-term stability --> The long-term trend
Section 1
p2, l5 and 8
IPCC, 2013 --> It is better to refer newer report (IPCC, 2021)
Section 2
p6, l8
largly --> largely
p7 2.4
It is better to describe the original spatial resolution of CAMS because it was averaged onto a 1-degree latitude and 1-degree longitude grid before comparison.
Section 3
p9, l8-9
Please discuss on the impact of this difference to the validation results in Section 4.7.
Section 4
p13, l4
What is ‘pixel’? Is it mostly the same as IFOV? Please explain it in Section 2.
p14, l28
June --> July
p15, l24
… and IAGOS … : IAGOS isn’t used for Figure 6. This is misleading.
l28-29
XCH4 generally …: What does this sentence mean? Are there any relations to the validation results?
p16, l32
Figure 8 --> Figure 8 (left)
p17, l7
Figure 8 --> Figure 8 (right)
l10
Tule --> Thule
l11-13
Why this sentence is put here and the content is different from that described in l2-3.
p18, l3-6
… temporally …: Does HIPPO observation limited to some season? What figure in the top row of Figure 4 should be referred? This sentence is too vague.
Section 5
p20, l11
It is found that … --> At land regions, it is found that ...
l33-p21, l1
What about Maido?
p21, l25-27
What is the definition of ‘similar’ and ‘different’? Why Wollongong is ‘similar’ but Lauder isn’t categorized?
Section 6
p23, l3
Section 4.1 --> Section 4.3.1
Section 7
p26, l8
The long-term stability --> The long-term trend
Many figures
The legends covered some of data plot. Move them not to cover the data.
Size of the figure is too small.
Figure 1
There are many mistakes. For example,
I couldn’t find ‘AirCore’ mark at Sodankylä.
I found ‘NDACC’ mark at Ny-Ålesund but Ny-Ålesund isn’t listed in Table 2.
I found ‘NDACC’ mark at Rikubetsu but Rikubetsu is listed in Table 1 (TCCON site).
Please check carefully.
Figure 2 caption
The solid data --> The solid line
Figure 5, 6, and 7
The correlation plots should be written with HIPPO values in horizontal axis. The values to be validated (RAL or LMD) should be in the vertical axis.
Figure 5 caption
the scatter plot between the RAL and HIPPO XCH4.
--> the scatter plot between the RAL and HIPPO XCH4 (right).
Please add the explanation of the gray bar in the left figure, black bar, black solid line, and pink dashed line in the right figure.
Figure 6 caption
Same as the comments for Figure 5.
Figure 7 legend
GB --> AIR
Figure 14
There is no error bar for Rikubetsu.
Citation: https://doi.org/10.5194/amt-2023-237-RC2 - AC2: 'Reply on RC2', Bart Dils, 09 Mar 2024
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
181 | 55 | 16 | 252 | 11 | 12 |
- HTML: 181
- PDF: 55
- XML: 16
- Total: 252
- BibTeX: 11
- EndNote: 12
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1