template ATMOS 2015 Official - ESA SEOMseom.esa.int/atmos2015/files/presentation36.pdf · C. E....
1
C. E. Robert, C. Bingen, F. Vanhellemont, N. Mateshvili, D. Fussen, C. Tétard, E. Dekemper D. Pieroux BIRA-IASB, Brussels This research is supported by a Marie Curie CIG Grant funded by the European Union 7 th Framework Program under grant agreement n°293560. −200 −100 0 100 200 10 15 20 25 30 35 (aergom−osiris)/(osiris) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−osiris)/(osiris) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−osiris)/(osiris) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−ace)/(ace) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−ace)/(ace) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−ace)/(ace) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−poam3)/(poam3) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−poam3)/(poam3) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−poam3)/(poam3) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−sage3)/(sage3) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−sage3)/(sage3) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−sage3)/(sage3) [%] Altitude [km] Rel. Diff. 452 nm Rel. Diff. 525 nm Rel. Diff. 385 nm Rel. Diff. 520 nm Rel. Diff. 755 nm SAGE III Rel. Diff. 525 nm Rel. Diff. 675 nm Rel. Diff. 779 nm SAGE II POAM III Rel. Diff. 354 nm Rel. Diff. 603 nm Rel. Diff. 779 nm ACE Rel. Diff. 350 nm Rel. Diff. 550 nm Rel. Diff. 756 nm OSIRIS bright+cold mid−bright+cold dim+cold bright+mid−cold mid−bright+mid−cold dim+mid−cold bright+hot mid−bright+hot dim+hot Legend −200 −100 0 100 200 10 15 20 25 30 35 (aergom−sage2)/(sage2) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−sage2)/(sage2) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−osiris)/(osiris) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−osiris)/(osiris) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−osiris)/(osiris) [%] Altitude [km] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−ace)/(ace) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−ace)/(ace) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−ace)/(ace) [%] Altitude [km] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−poam3)/(poam3) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−poam3)/(poam3) [%] Altitude [km] −200 −100 0 100 200 10 15 20 25 30 35 (aergom−poam3)/(poam3) [%] Altitude [km] Rel. Diff. 603 nm Rel. Diff. 779 nm POAM III Rel. Diff. 779 nm Rel. Diff. 603 nm Rel. Diff. 525 nm Rel. Diff. 756 nm ACE OSIRIS Rel. Diff. 550 nm Rel. Diff. 350 nm Rel. Diff. 354 nm Legend sza 110−120 sza 120−130 sza 130−140 sza 140−150 sza 150−160 sza 160−170 sza 170−180 0 50 100 150 200 250 300 10 15 20 25 30 35 (aergom−gopr)/(gopr) [%] Altitude [km] 0 50 100 150 200 250 300 10 15 20 25 30 35 (aergom−osiris)/(osiris) [%] Altitude [km] 0 50 100 150 200 250 300 10 15 20 25 30 35 (aergom−ace)/(ace) [%] Altitude [km] 0 50 100 150 200 250 300 10 15 20 25 30 35 (aergom−sage2)/(sage2) [%] Altitude [km] 10 −6 10 −5 10 −4 10 −3 10 −2 10 15 20 25 30 35 Extinction [km−1] Altitude [km] 10 −6 10 −5 10 −4 10 −3 10 −2 10 15 20 25 30 35 Extinction [km−1] Altitude [km] −200 −150 −100 −50 0 50 100 150 200 10 15 20 25 30 35 (aergom−poam3)/(poam3) [%] Altitude [km] 354 nm 439 nm 602 nm 778 nm −200 −150 −100 −50 0 50 100 150 200 10 15 20 25 30 35 (aergom−sage3)/(sage3) [%] Altitude [km] −200 −150 −100 −50 0 50 100 150 200 10 15 20 25 30 35 (aergom−sage2)/(sage2) [%] Altitude [km] 452 nm 525 nm −200 −150 −100 −50 0 50 100 150 200 10 15 20 25 30 35 (aergom−ace)/(ace) [%] Altitude [km] 525 nm 603 nm 675 nm 779 nm ACE (sage3) 384 nm (aergom) 384 nm (sage3) 448 nm (aergom) 448 nm (sage3) 520 nm (aergom) 520 nm (sage3) 755 nm (aergom) 755 nm SAGE2 384 nm 448 nm 520 nm 755 nm SAGE3 POAM3 Relative Difference (Interquartile Mean) Absolute Extinction (Interquartile Mean) (poam3) 354 nm (aergom) 354 nm (poam3) 439 nm (aergom) 439 nm (poam3) 602 nm (aergom) 602 nm (poam3) 778 nm (aergom) 778 nm −200 −150 −100 −50 0 50 100 150 200 10 15 20 25 30 35 (aergom−osiris)/(osiris) [%] 350 nm 550 nm 756 nm OSIRIS 10 −6 10 −5 10 −4 10 −3 10 −2 10 15 20 25 30 35 Extinction [km−1] Altitude [km] (osiris) 350 nm (aergom) 350 nm (osiris) 550 nm (aergom) 550 nm (osiris) 756 nm (aergom) 756 nm 10 −6 10 −5 10 −4 10 −3 10 −2 10 15 20 25 30 35 Extinction [km−1] Altitude [km] (ace) 525 nm (aergom) 525 nm (ace) 603 nm (aergom) 603 nm (ace) 675 nm (aergom) 675 nm (ace) 779 nm (aergom) 779 nm 10 −6 10 −5 10 −4 10 −3 10 −2 10 15 20 25 30 35 Extinction [km−1] Altitude [km] (sage2) 452 nm (aergom) 452 nm (sage2) 525 nm (aergom) 525 nm 10 −6 10 −5 10 −4 10 −3 10 −2 10 15 20 25 30 35 Extinction [km−1] Altitude [km] (gopr) 350 nm (aergom) 350 nm (gopr) 550 nm (aergom) 550 nm (gopr) 756 nm (aergom) 756 nm −200 −150 −100 −50 0 50 100 150 200 10 15 20 25 30 35 (aergom−gopr)/(gopr) [%] Altitude [km] 350 nm 550 nm 756 nm GOPR N=1017 N=3158 N=634 N=1364 N=13625 N=20000 Relative Difference Variability (Semi-Interquartile Range) 0 50 100 150 200 250 300 10 15 20 25 30 35 (aergom−sage3)/(sage3) [%] Altitude [km] 0 50 100 150 200 250 300 10 15 20 25 30 35 (aergom−poam3)/(poam3) [%] Altitude [km] Altitude [km] What is AerGom? Aergom is an improved stratospheric aerosol extinction retrieval algorithm for the GOMOS mission that uses a measurement technique based on stellar occultation. It’s important! The data derived from this algorithm are used internationally in the framework of the Aerosol Climate Change Initiative and the SPARC Data Initiative. AerGom Contact the author: [email protected] Intercomparisons Methodology In this work, we carried out a systematic comparison of the retrieved stratospheric aerosols coefficients by AerGom at various wavelengths with colocated observations (± 12h, 500 km) from multiple satellite instruments: SAGE II,SAGE III, POAM III,ACEMAESTRO, and OSIRIS. You can see the coverage of these satellite datasets in Fig. 1. We also performed comparisons with the official GOMOS processor (GOPR) v6.01. Why do it? These results are helpful because: they point out discrepancies between various datasets the bias determination can be used for merging multiplatform time series spanning decades it helps improve retrieval algorithms (either AerGom or others) Coverage AerGom SAGE 3 POAM 3 SAGE 2 ACE8MAESTRO OSIRIS Fig. 1 Latitude and temporal coverage for the various instruments used for the comparisons. The number of observations per month is calculated for a 10° latitude bin. Fig. 2 Relative difference interquartile mean (left panel), semiinterquartile range of the relative difference (central panel) and interquartile mean of the absolute aerosol extinction profiles (right panel) for each dataset at various wavelengths compared with colocated AerGom profiles. The dashed curves in the relative difference plots were calculated using GOPR instead of AerGom. Intercomparisons results The results of the intercomparisons are shown in Fig. 2. From these plots, one can make some general assertions: Agreement is typically within ± 50% for extinctions in the 400603 nm spectral range, and between 15 and 30 km tangent altitudes. The variability of the AerGom comparisons with other datasets is usually smaller then those made with GOPR. AerGom aerosol extinction profiles for λ > 700 nm present a strong negative bias above 2530 km with respect to other instruments, increasing towards higher altitudes. To carry out this study, we used the SAGE II, SAGE III, POAM III,ACEMAESTRO and OSIRISdatasets as reference (as they should not be influenced by these parameters), and looked for differences in the comparison results due to specific occultation parameters. For conciseness, we limit ourselves to the examination of two occultation parameterson the AerGom retrievals. Fig. 6 shows the impact of the SZA during the occultation, while Fig. 7 illustrates the effect of the star properties. Recent modifications to the retrieval algorithm gas cross sections, along with a proper convolution with the instrument function, seem to significantly improve the stratospheric aerosol extinction coefficients retrieved at larger wavelengths. Fig. 3 shows the progress made at 756 nm (using OSIRISas a reference) when using high resolution crosssections for O 3 , NO 2 and NO 3 . Note that these are preliminary results. AerGom (original) −200 −150 −100 −50 0 50 100 150 200 10 15 20 25 30 35 (aergom−osiris)/(osiris) [%] Altitude [km] 350 nm 550 nm 756 nm GOP R AerGom with high resolution cross-sections −200 −150 −100 −50 0 50 100 150 200 10 15 20 25 30 35 (aergom−osiris)/(osiris) [%] Altitude [km] Fig. 3 Comparisons of GOMOS profiles with colocated OSIRIS profiles using different versions of the AerGom retrievals: the original AerGom (left) and the latest version using highresolution gas crosssections (right). RESULTS Improvements ! −2 −1 0 1 x 10 11 10 20 30 40 50 AERGOM − NO 2 Number density [cm −3 ] Altitude [km] Good Bad Climatology −2 −1 0 1 2 3 x 10 9 10 20 30 40 50 Number density [cm −3 ] Altitude [km] AERGOM − NO 3 −2 0 2 4 6 x 10 12 10 20 30 40 50 Number density [cm −3 ] Altitude [km] AERGOM − O 3 −4 −2 0 2 x 10 −3 10 20 30 40 50 Extinction [km −1 ] Altitude [km] AERGOM − extinction @ 550nm AerGom is an improvement over GOMOS official processor, but it does have issues when performing retrievals with transmission data obtained using dim/cold stars (cf. Fig. 4), which getsthe gasand aerosols species completely wrong in some cases (socalled bad profiles). Fig. 5 shows the mean bad profiles compared with good and climatological profiles. These bad profiles can be easily flagged, but do limit the coverage of the AerGom datasets. Why show this? Because this finding prompted the consideration that some of the retrievals might be affected by occultation parameterssuch as star properties (temperature and magnitude), solar zenith angle (SZA) that could lead to straylight, and occultation obliquity which is an important factor in the imperfect correction of atmospheric scintillation. Do these occultation parameterssystematically affect the AerGom retrievals and if so, to what extent? The case of the Fig. 5 Median gas and aerosol extinction profiles for ‘good’ and ‘bad’ Aergom retrievals Fig. 4 Proportion of bad profiles vs star temperature and magnitude. outlandish profiles Effect of Occultation Parameters Star Properties Fig. 7 Relative differences between AerGom and various datasets for varying star categories. Again, the thick grey curve shows the median of the comparison using all data. See Table 1 for the definition of the various classes of stars in terms of temperature and magnitude. Star Temperature [ 10 3 K] Star Magnitude cold 0–6 bright 1.5 – 1.5 midcold 6 – 26 midbright 1.5 – 2.3 hot 26 – 40 dim 2.3 – 3 Solar Zenith Angles Fig. 6 Relative differences between AerGom and various datasets for different SZA during the occultation. The thick grey curve shows the median of the comparison using all data, the basis for comparison to see if all results are consistent. This work is important to understand the properties of the AerGom retrieval algorithm and even more crucial for data users, who should be aware of possible biases within the data. This detailed analysis leads to the following conclusions: The quality of the retrieval is mainly influenced by the star parametersthat directly impact the SNR of the measurement. The dominant parameter is the magnitude quantifying the strength of the star signal. The comparisons show that a threshold of M < 2.5 is suitable for high quality retrievals. For λ > 750 nm, Hot stars perform worse than cold stars and the recommended threshold is T < 26 x 10 3 K The second most important influence is the SZA. Using a threshold of 110° gives good quality results for extinction at λ < 750 nm, but one should use a threshold value of 130° for extinction at λ > 750 nm. The obliquity influences also the quality of the retrieval but to a lesser extent. Overall, a decrease of quality of the retrieval is only expected for a value of the obliquity above 40°, especially for λ > 750 nm. ➊ ➋ ➌ ➍ ➎ ➏ Results summary ➐ Table 1 Classes of star properties (as defined in this work) Assessment of GOMOS stratospheric aerosol extinction coefficients retrieved from the AerGom algorithm using contemporaneous satellite experiments
Transcript of template ATMOS 2015 Official - ESA SEOMseom.esa.int/atmos2015/files/presentation36.pdf · C. E....
C. E. Robert, C. Bingen, F. Vanhellemont, N. Mateshvili, D. Fussen, C. Tétard, E. Dekemper D. PierouxBIRA-IASB, Brussels
This research is supported by a Marie Curie CIG Grant funded by theEuropean Union 7th Framework Program under grant agreement n°293560.
0 1000 2000 3000 4000 500010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−osiris, λ=350nm
0 1000 2000 3000 4000 500010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−osiris, λ=550nm
0 1000 2000 3000 400010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−osiris, λ=756nm
bright+coldmid−bright+colddim+coldbright+mid−coldmid−bright+mid−colddim+mid−coldbright+hotmid−bright+hotdim+hot
−200 −100 0 100 20010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
*mean2575* , λ=350nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
*mean2575* , λ=550nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
*mean2575* , λ=756nm
0 100 200 300 400 50010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−ace, λ=525nm
0 100 200 300 40010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−ace, λ=603nm
0 100 200 300 40010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−ace, λ=675nm
0 100 200 30010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−ace, λ=779nm
bright+coldmid−bright+colddim+coldbright+mid−coldmid−bright+mid−colddim+mid−coldbright+hotmid−bright+hotdim+hot
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=525nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=603nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=675nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=779nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=525nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=603nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=675nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=779nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=525nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=603nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=675nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=779nm
0 50 100 150 20010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−poam3, λ=354nm
0 50 100 150 20010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−poam3, λ=440nm
0 50 100 150 20010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−poam3, λ=603nm
0 50 100 150 20010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−poam3, λ=779nm
bright+coldmid−bright+colddim+coldbright+mid−colddim+mid−coldmid−bright+hotdim+hot
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=354nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=440nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=603nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=779nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=354nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=440nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=603nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=779nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=354nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=440nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=603nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=779nm
0 200 400 600 80010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−sage3, λ=384.2687nm
0 200 400 600 80010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−sage3, λ=448.5182nm
0 200 400 600 80010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−sage3, λ=520.312nm
0 200 400 60010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−sage3, λ=755.3782nm
bright+coldmid−bright+colddim+coldbright+mid−coldmid−bright+mid−colddim+mid−coldbright+hotmid−bright+hotdim+hot
−200 −100 0 100 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*mean2575* , λ=384.2687nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*mean2575* , λ=448.5182nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*mean2575* , λ=520.312nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*mean2575* , λ=755.3782nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*mean2575* , λ=384.2687nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*mean2575* , λ=448.5182nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*mean2575* , λ=520.312nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*mean2575* , λ=755.3782nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*mean2575* , λ=384.2687nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*mean2575* , λ=448.5182nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*mean2575* , λ=520.312nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*mean2575* , λ=755.3782nm
0 50 100 150 200 25010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−sage2, λ=452nm
0 50 100 150 200 25010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−sage2, λ=525nm
bright+coldmid−bright+colddim+coldbright+mid−coldmid−bright+mid−colddim+mid−coldbright+hotmid−bright+hotdim+hot
Star)properties NEWRel.)Diff.) )452$nm Rel.)Diff.) )525$nm #)observations)[452)nm]
#)observations)[520)nm]Rel.)Diff.) )385$nm Rel.)Diff.) )520$nm Rel.)Diff.) )755$nm
SAGE)III
poam3
#)observations)[525)nm]Rel.)Diff.) )525$nm Rel.)Diff.) )675$nm Rel.)Diff.) )779$nm
#)observations)[440)nm]
SAGE)II
POAM)III
Rel.)Diff.) )354$nm Rel.)Diff.) )603$nm Rel.)Diff.) )779$nm
ACE
#)observations)[525)nm]Rel.)Diff.) )350$nm Rel.)Diff.) )550$nm Rel.)Diff.) )756$nm
OSIRIS
0 200 400 600 8005
10
15
20
25
30
35
Number of data points available
Alti
tude
[km
]
Data available: aergom−sage3, λ=384.2687nm
0 200 400 600 8005
10
15
20
25
30
35
Number of data points available
Alti
tude
[km
]
Data available: aergom−sage3, λ=448.5182nm
0 200 400 600 8005
10
15
20
25
30
35
Number of data points available
Alti
tude
[km
]
Data available: aergom−sage3, λ=520.312nm
0 200 400 6005
10
15
20
25
30
35
Number of data points available
Alti
tude
[km
]
Data available: aergom−sage3, λ=755.3782nm
bright+coldmid−bright+colddim+coldbright+mid−coldmid−bright+mid−colddim+mid−coldbright+hotmid−bright+hotdim+hot
Legend
−200 −100 0 100 20010
15
20
25
30
35
(aergom−sage2)/(sage2) [%]
Altit
ude
[km
]
*mean2575* , λ=452nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−sage2)/(sage2) [%]
Altit
ude
[km
]
*mean2575* , λ=525nm
0 1000 2000 3000 4000 500010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−osiris, λ=350nm
0 1000 2000 3000 4000 500010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−osiris, λ=550nm
0 1000 2000 3000 4000 500010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−osiris, λ=756nm
sza 110−120sza 120−130sza 130−140sza 140−150sza 150−160sza 160−170sza 170−180
−200 −100 0 100 20010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
*mean2575* , λ=350nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
*mean2575* , λ=550nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
*mean2575* , λ=756nm
0 100 200 300 400 50010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−ace, λ=525nm
0 100 200 300 400 50010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−ace, λ=603nm
0 100 200 300 400 50010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−ace, λ=675nm
0 100 200 300 40010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−ace, λ=779nm
sza 110−120sza 120−130sza 130−140sza 140−150sza 150−160sza 160−170
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=525nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=603nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=675nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=779nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=525nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=603nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=675nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*mean2575* , λ=779nm
0 100 200 300 40010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−poam3, λ=354nm
0 100 200 300 40010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−poam3, λ=440nm
0 100 200 300 40010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−poam3, λ=603nm
0 100 200 300 40010
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
aergom−poam3, λ=779nm
sza 110−120sza 120−130sza 130−140
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=354nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=440nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=603nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=779nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=354nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=440nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=603nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=779nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=354nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=440nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=603nm
−200 −100 0 100 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* , λ=779nm
SZA#)observations)[603)nm]Rel.)Diff.) )603$nm Rel.)Diff.) )779$nm
POAM)III
#)observations)[525)nm]Rel.)Diff.) )779$nmRel.)Diff.) )603$nmRel.)Diff.) )525$nm
Rel.)Diff.)756$nm #)observations)[550)nm]
ACE
OSIRIS
Rel.)Diff.)550$nmRel.)Diff.)350$nm
Rel.)Diff.) )354$nm
Legend
0 1000 2000 3000 4000 50005
10
15
20
25
30
35
Number of data points available
Altit
ude
[km
]
Data available: aergom−osiris, λ=350nm
sza 110−120sza 120−130sza 130−140sza 140−150sza 150−160sza 160−170sza 170−180
0 50 100 150 200 250 30010
15
20
25
30
35
(aergom−gopr)/(gopr) [%]
Altit
ude
[km
]
*siqr* aergom−gopr, good_aergom_data
0 50 100 150 200 250 30010
15
20
25
30
35
(aergom−gopr)/(gopr) [%]
Altit
ude
[km
]
*siqr* aergom−gopr, good_aergom_data
350 nm550 nm756 nm
0 50 100 150 200 250 30010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
*siqr* aergom−osiris, good_aergom_data
0 50 100 150 200 250 30010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
*siqr* aergom−osiris, good_aergom_data
350 nm550 nm756 nm
0 50 100 150 200 250 30010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*siqr* aergom−ace, good_aergom_data
0 50 100 150 200 250 30010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
*siqr* aergom−ace, good_aergom_data
525 nm603 nm675 nm779 nm
0 50 100 150 200 250 30010
15
20
25
30
35
(aergom−sage2)/(sage2) [%]
Altit
ude
[km
]
*siqr* aergom−sage2, good_aergom_data
0 50 100 150 200 250 30010
15
20
25
30
35
(aergom−sage2)/(sage2) [%]
Altit
ude
[km
]
*siqr* aergom−sage2, good_aergom_data
452 nm525 nm
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
*mean2575* aergom−sage3, good_aergom_data
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
*mean2575* aergom−sage3, good_aergom_data
(sage3) 384 nm(aergom) 384 nm(sage3) 448 nm(aergom) 448 nm(sage3) 520 nm(aergom) 520 nm(sage3) 755 nm(aergom) 755 nm
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
*mean2575* aergom−poam3, good_aergom_data
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
*mean2575* aergom−poam3, good_aergom_data
(poam3) 354 nm(aergom) 354 nm(poam3) 439 nm(aergom) 439 nm(poam3) 602 nm(aergom) 602 nm(poam3) 778 nm(aergom) 778 nm
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* aergom−poam3, good_aergom_data
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*mean2575* aergom−poam3, good_aergom_data
354 nm439 nm602 nm778 nm
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*mean2575* aergom−sage3, good_aergom_data
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*mean2575* aergom−sage3, good_aergom_data
384 nm448 nm520 nm755 nm
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−sage2)/(sage2) [%]
Altit
ude
[km
]
*mean2575* aergom−sage2, good_aergom_data
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−sage2)/(sage2) [%]
Altit
ude
[km
]
*mean2575* aergom−sage2, good_aergom_data
452 nm525 nm
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]Al
titud
e [k
m]
aergom−ace comparison
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−ace)/(ace) [%]
Altit
ude
[km
]
aergom−ace comparison
525 nm603 nm675 nm779 nm
ACE
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−sage3 comparison
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−sage3 comparison
(sage3) 384 nm(aergom) 384 nm(sage3) 448 nm(aergom) 448 nm(sage3) 520 nm(aergom) 520 nm(sage3) 755 nm(aergom) 755 nm
SAGE2
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
aergom−sage3 comparison
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
aergom−sage3 comparison
384 nm448 nm520 nm755 nm
SAGE3
POAM3
Relative Difference(Interquartile Mean)
Absolute Extinction(Interquartile Mean)
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−poam3 comparison
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−poam3 comparison
(poam3) 354 nm(aergom) 354 nm(poam3) 439 nm(aergom) 439 nm(poam3) 602 nm(aergom) 602 nm(poam3) 778 nm(aergom) 778 nm
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
aergom−osiris comparison
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
aergom−osiris comparison
350 nm550 nm756 nm
OSIRIS
WITH%GOPR!!
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−osiris comparison
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−osiris comparison
(osiris) 350 nm(aergom) 350 nm(osiris) 550 nm(aergom) 550 nm(osiris) 756 nm(aergom) 756 nm
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−osiris comparison
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−osiris comparison
(osiris) 350 nm(aergom) 350 nm(osiris) 550 nm(aergom) 550 nm(osiris) 756 nm(aergom) 756 nm
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−ace comparison
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−ace comparison
(ace) 525 nm(aergom) 525 nm(ace) 603 nm(aergom) 603 nm(ace) 675 nm(aergom) 675 nm(ace) 779 nm(aergom) 779 nm
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Alti
tude
[km
]
aergom−ace comparison
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Alti
tude
[km
]
aergom−ace comparison
(ace) 525 nm(aergom) 525 nm(ace) 603 nm(aergom) 603 nm(ace) 675 nm(aergom) 675 nm(ace) 779 nm(aergom) 779 nm
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−sage2 comparison
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−sage2 comparison
(sage2) 452 nm(aergom) 452 nm(sage2) 525 nm(aergom) 525 nm
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−sage2 comparison
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−sage2 comparison
(sage2) 452 nm(aergom) 452 nm(sage2) 525 nm(aergom) 525 nm
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−gopr comparison
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−gopr comparison
(gopr) 350 nm(aergom) 350 nm(gopr) 550 nm(aergom) 550 nm(gopr) 756 nm(aergom) 756 nm
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−gopr comparison
10−6 10−5 10−4 10−3 10−210
15
20
25
30
35
Extinction [km−1]
Altit
ude
[km
]
aergom−gopr comparison
(gopr) 350 nm(aergom) 350 nm(gopr) 550 nm(aergom) 550 nm(gopr) 756 nm(aergom) 756 nm
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−gopr)/(gopr) [%]
Altit
ude
[km
]
aergom−gopr comparison
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−gopr)/(gopr) [%]
Altit
ude
[km
]
aergom−gopr comparison
350 nm550 nm756 nm
GOPR
N=1017
N=3158
N=634
N=1364
N=13625
N=20000
Relative Difference Variability(Semi-Interquartile Range)
0 50 100 150 200 250 30010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*siqr* aergom−sage3, good_aergom_data
0 50 100 150 200 250 30010
15
20
25
30
35
(aergom−sage3)/(sage3) [%]
Altit
ude
[km
]
*siqr* aergom−sage3, good_aergom_data
384 nm448 nm520 nm755 nm
0 50 100 150 200 250 30010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*siqr* aergom−poam3, good_aergom_data
0 50 100 150 200 250 30010
15
20
25
30
35
(aergom−poam3)/(poam3) [%]
Altit
ude
[km
]
*siqr* aergom−poam3, good_aergom_data
354 nm439 nm602 nm778 nm
0 50 100 150 200 250 30010
15
20
25
30
35
(aergom−gopr)/(gopr) [%]
Alti
tude
[km
]
*siqr* aergom−gopr, good_aergom_data
0 50 100 150 200 250 30010
15
20
25
30
35
(aergom−gopr)/(gopr) [%]
Alti
tude
[km
]
*siqr* aergom−gopr, good_aergom_data
350 nm550 nm756 nm
What is AerGom?Aergom is an improved stratospheric aerosol extinction retrieval algorithm for the GOMOS mission that uses a measurement technique based on stellar occultation.
It’s important!The data derived from this algorithm are used internationally in the framework of the Aerosol Climate Change Initiative and the SPARC Data Initiative.
Everything must be half the final size !!! I will scale it up!!!
AerGom
Contact the author: [email protected]
IntercomparisonsMethodologyIn this work, we carried out a systematic comparison of the retrieved stratospheric aerosols coefficients by AerGom at various wavelengths with colocatedobservations (± 12h, 500 km) from multiple satellite instruments: SAGE II, SAGE III, POAM III, ACE-‐MAESTRO, and OSIRIS. You can see the coverage of these satellite datasets in Fig. 1. We also performed comparisons with the official GOMOS processor (GOPR) v6.01.
Why do it?These results are helpful because:
›❯ they point out discrepancies between various datasets
›❯ the bias determination can be used for merging multi-‐platform time series spanning decades
›❯ it helps improve retrieval algorithms (either AerGomor others)
CoverageAerGom
SAGE$3
POAM$3
SAGE$2
ACE8MAESTRO
OSIRIS
Fig. 1 Latitude and temporal coverage for the various instruments used for the comparisons. The number of observations per month is calculated for a 10° latitude bin.
Fig. 2 Relative difference interquartile mean (left panel), semi-‐interquartile range of the relative difference (central panel) and interquartile mean of the absolute aerosol extinction profiles (right panel) for each dataset at various wavelengths compared with colocated AerGom profiles. The dashed curves in the relative difference plots were calculated using GOPR instead of AerGom.
Intercomparisons results
The results of the intercomparisons are shown in Fig. 2. From these plots, one can make some general assertions:
›❯ Agreement is typically within ± 50% for extinctions in the 400-‐603 nm spectral range, and between 15 and 30 km tangent altitudes.
›❯ The variability of the AerGom comparisons with other datasets is usually smaller then those made with GOPR.
›❯ AerGom aerosol extinction profiles for λ > 700 nm present a strong negative bias above 25-‐30 km with respect to other instruments, increasing towards higher altitudes.
To carry out this study, we used the SAGE II, SAGE III, POAM III, ACE-‐MAESTRO and OSIRIS datasets as reference (as they should not be influenced by these parameters), and looked for differences in the comparison results due to specific occultation parameters.
For conciseness, we limit ourselves to the examination of two occultation parameters on the AerGom retrievals. Fig. 6 shows the impact of the SZA during the occultation, while Fig. 7 illustrates the effect of the star properties.
Recent modifications to the retrieval algorithm gas cross-‐sections, along with a proper convolution with the instrument function, seem to significantly improve the stratospheric aerosol extinction coefficients retrieved at larger wavelengths. Fig. 3 shows the progress made at 756 nm (using OSIRIS as a reference) when using high resolution cross-‐sections for O3, NO2 and NO3. Note that these are preliminary results.
AerGom (original)
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
*mean2575* aergom−osiris, good_aergom_data
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
*mean2575* aergom−osiris, good_aergom_data
350 nm550 nm756 nm
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
*mean2575* aergom−osiris, good_aergom_data
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
*mean2575* aergom−osiris, good_aergom_data
350 nm550 nm756 nm
GOPR
AerGom with high resolution cross-sections
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
*mean2575* aergom−osiris, good_aergom_data
−200 −150 −100 −50 0 50 100 150 20010
15
20
25
30
35
(aergom−osiris)/(osiris) [%]
Altit
ude
[km
]
*mean2575* aergom−osiris, good_aergom_data
350 nm550 nm756 nm
Fig. 3 Comparisons of GOMOS profiles with colocated OSIRIS profiles using different versions of the AerGom retrievals: the original AerGom (left) and the latest version using high-‐resolution gas cross-‐sections (right).
RESULTS
Improvements !
−2 −1 0 1x 1011
10
20
30
40
50AERGOM − NO2
Number density [cm−3]
Altit
ude
[km
]
GoodBadClimatology
−2 −1 0 1 2 3x 109
10
20
30
40
50
Number density [cm−3]
Altit
ude
[km
]
AERGOM − NO3
−2 0 2 4 6x 1012
10
20
30
40
50
Number density [cm−3]
Altit
ude
[km
]
AERGOM − O3
−4 −2 0 2x 10−3
10
20
30
40
50
Extinction [km−1]
Altit
ude
[km
]
AERGOM − extinction @ 550nm
AerGom is an improvement over GOMOS official processor, but it does have issues when performing retrievals with transmission data obtained using dim/cold stars (cf. Fig. 4), which gets the gas and aerosols species completely wrong in some cases (so-‐called bad profiles). Fig. 5 shows the mean bad profiles compared with good and climatological profiles. These bad profiles can be easily flagged, but do limit the coverage of the AerGom datasets.
Why show this?Because this finding prompted the consideration that some of the retrievals might be affected by occultation parameters such as star properties (temperature and magnitude), solar zenith angle (SZA) that could lead to straylight, and occultation obliquity which is an important factor in the imperfect correction of atmospheric scintillation. Do these occultation parameters systematically affect the AerGom retrievals and if so, to what extent?
The case of the
Fig. 5 Median gas and aerosol extinction profiles for ‘good’ and ‘bad’ Aergomretrievals
!Fig. 4 Proportion of bad profiles vs star temperature and magnitude.
outlandish profiles
Effect of Occultation Parameters Star Properties
Fig. 7 Relative differences between AerGom and various datasets for varying star categories. Again, the thick grey curve shows the median of the comparison using all data. See Table 1 for the definition of the various classes of stars in terms of temperature and magnitude.
Star Temperature [ 103 K ] Star Magnitude
cold 0 – 6 bright -‐1.5 – 1.5
mid-‐cold 6 – 26 mid-‐bright 1.5 – 2.3
hot 26 – 40 dim 2.3 – 3
Solar Zenith Angles
Fig. 6 Relative differences between AerGom and various datasets for different SZA during the occultation. The thick grey curve shows the median of the comparison using all data, the basis for comparison to see if all results are consistent.
This work is important to understand the properties of the AerGom retrieval algorithm and even more crucial for data users, who should be aware of possible biases within the data. This detailed analysis leads to the following conclusions:
›❯ The quality of the retrieval is mainly influenced by the star parameters that directly impact the SNR of the measurement. The dominant parameter is the magnitude quantifying the strength of the star signal. The comparisons show that a threshold of M < 2.5 is suitable for high quality retrievals. For λ > 750 nm, Hot stars perform worse than cold stars and the recommended threshold is T < 26 x 103 K
›❯ The second most important influence is the SZA. Using a threshold of 110° gives good quality results for extinction at λ < 750 nm, but one should use a threshold value of 130° for extinction at λ > 750 nm.
›❯ The obliquity influences also the quality of the retrieval but to a lesser extent. Overall, a decrease of quality of the retrieval is only expected for a value of the obliquity above 40°, especially for λ > 750 nm.
➊
➋
➌
➍ ➎
➏ Results summary➐
Table 1 Classes of star properties (as defined in this work)
Assessment of GOMOS stratospheric aerosol extinction coefficients retrieved from the AerGom algorithm using contemporaneous satellite experiments
![OFFICIAL GAZETTE OF THE HELLENIC REPUBLIC · ΕΠΙΣΗΜΗ ΜΕΤΑΦΡΑΣΗ traduction officielle official translation amtliche Übersetzung Φ09222/6416 [emblem] official gazette](https://static.fdocument.org/doc/165x107/5fcc1c5ad22df16cce77e303/official-gazette-of-the-hellenic-republic-oe-oe-traduction.jpg)
