Strength of TROPOMI observations on the retrieval of volcanic … · 2020. 5. 3. · Strength of...
1
Strength of TROPOMI observations on the retrieval of volcanic SO 2 emissions at high temporal resolution from space by inverse modeling Abhinna K. Behera 1 , Marie Boichu 1 , and François Thieuleux 1 1 Laboratoire d’Optique Atmosphérique, Université de Lille, 59650 Villeneuve d’Ascq, France. Abstract A refined knowledge of volcanic sulphur dioxide (SO 2 ) degassing at high temporal resolution, as SO 2 is oxidized to form sulfate aerosols in the atmosphere, is essential to advance further in understanding the consequence of volcanic emissions on air quality, ecosystems, and climate at a global scale. Concerning volcanological applications, infor- mation on SO 2 degassing remains critical for assessing the underlying magma dynamics and associated hazards. However, presently, nadir space-borne observations in the ultraviolet spectrum range, viz., OMI, GOME2, OMPS, and more recently launched TROPOMI hyperspectral sensor, can only provide snapshots of volcanic SO2 every 24 h. Here we determine hourly the SO 2 flux emissions as well as the altitude of injec- tion by applying an inverse modeling method [1]. The forward model involves the CHIMERE Eulerian regional chemistry-transport model [2]. The idea is to evaluate the advancements in utilizing high spatially and spectrally resolved measurements from TROPOMI spectrometer onboard ESA-Sentinel-5 Precursor. This investigation focuses on the 2018 eruption of Ambrym (16.24°S, 168.15°E), a top world-ranking vol- canic SO 2 emitter. It depicted a rapid cessation of a decade long SO 2 degassing and lava lake activity associated with the emplacement of a large injection of magma below the surface, crisscrossing the island of Ambrym, and spawning a submarine eruption [4]. TROPOMI assimilated CHIMERE simulation leads to more detailed information on the dispersal of the volcanic SO 2 cloud, and total mass as well as altitude of injection by the inverse modeling scheme. This provides the paroxysmal phase of the eruption occurred about 0430 UT on 15 December 2018, injecting ∼8 kt of SO 2 to the atmosphere at 12 km above sea level (asl). Ambrym volcano emitted ∼100 kt of SO 2 during 13-18 December 2018; most of the SO 2 mass was centered at 7 km altitude asl. Mostly, TROPOMI assimilated CHIMERE simula- tion depicts the weakest degassing of magnitude <100 t h -1 . At first order, SO 2 flux time series are similar using TROPOMI, GOME2, and OMPS data. Overall, TROPOMI determines better the dense parcels of SO 2 in the core of the volcanic plume than OMPS or GOME-2. As a consequence, model simulation assimilating TROPOMI observations describes better the dispersal of such dense parts of the SO2 cloud, of- ten co-located with ash-rich zones that should be avoided for aviation safety. Methodology • Inverse modeling method: Linear least-squares regression approach with non-negative constraint and no a priori [1] : d 0 Satellite Observation = G κ 2 Δ (Simulated) Forward Operator × ( m ) Model Parameter , where κ 2 is the Lagrangian multiplier and Δ is the Laplacian - second order Tikhonov regularisation. • Volcanic SO 2 cloud dispersal, including the physico-chemical pro- cesses, is simulated using CHIMERE (Eulerian) chemistry-transport model [2] at a large-scale during 13 to 18 December 2018. • WRF-ARW [3] model simulation is used for forcing the meteorological fields in CHIMERE. ERA-5 reanalysis with 0.25° spatial resolution is used hourly for the initialisation and nudging at the grid boundaries of WRF-ARW simulation. • 10 km × 10 km horizontal resolution is implemented to resolve well the volcanic SO 2 dispersal, and the vertical layer thickness is maintained within 25 m to 550 m that varies as a geometric sequence. • Gaseous tracers (passive, no aerosol chemistry) are injected each hour to the atmosphere at each kilometer in a range of 2 km to 12 km altitude asl during 2018121300 to 2018121800. • To reproduce well the complex horizontal wind-fields, sharply varying with altitude in the surroundings of the Australian continent, WRF- ARW simulation was run with a fine vertical resolution of ∼250 m since 00 UT 1 December 2018 giving enough spin-up time. Outlook • The high spatial and spectral resolution TROPOMI SO 2 data paves the way for the Eulerian chemistry-transport models to identify and track better the dense SO 2 parcels as well as weakest degassing by the inversion scheme. The assimilation of TROPOMI observations should improve the quality of volcanic plume forecasts that are critical for aviation safety. • The complex wind-field, sharply changing with altitude, can impact the reconstruction of volcanic SO 2 clouds by the inverse scheme. Hence, the best available reanalysis data (e.g., ERA5) and the high- resolution meteorological model simulation are crucial to drive CHIMERE simulation. Inverted SO 2 flux time series SO 2 mass vs. altitude References [1] M. Boichu, L. Menut, D. Khvorostyanov, L. Clarisse, C. Clerbaux, S. Turquety, and P.-F. Coheur. Inverting for volcanic SO 2 flux at high temporal resolution using spaceborne plume imagery and chemistry-transport modelling: the 2010 Eyjafjallajökull eruption case study. Atmospheric Chemistry and Physics, 13(17):8569–8584, Sept. 2013. [2] S. Mailler, L. Menut, D. Khvorostyanov, M. Valari, F. Couvidat, G. Siour, S. Turquety, R. Briant, P. Tuccella, B. Bessagnet, A. Colette, L. Létinois, K. Markakis, and F. Meleux. CHIMERE-2017: from urban to hemispheric chemistry-transport modeling. Geoscientific Model Development, 10(6):2397– 2423, June 2017. [3] J. G. Powers, J. B. Klemp, W. C. Skamarock, C. A. Davis, J. Dudhia, D. O. Gill, J. L. Coen, D. J. Gochis, R. Ahmadov, S. E. Peckham, et al. The weather research and forecasting model: Overview, system efforts, and future directions. Bulletin of the American Meteorological Society, 98(8):1717–1737, 2017. [4] T. Shreve, R. Grandin, M. Boichu, E. Garaebiti, Y. Moussallam, V. Ballu, F. Delgado, F. Leclerc, M. Vallée, N. Henriot, et al. From prodigious volcanic degassing to caldera subsidence and quies- cence at ambrym (vanuatu): the influence of regional tectonics. Scientific reports, 9(1):1–13, 2019. Main features of the space-borne sensors Properties TROPOMI OMPS-NM OMI GOME2 Satellite ESA Sentinel-5 Precursor NOAA-Suomi NPP NASA EOS-Aura NOAA/ESA Metop-A+B Spectral ranges UV-VIS-NIR-SWIR UV UV-VIS UV-VIS-NIR Nadir spatial resolution 7 km × 3.5 km 50 km × 50 km 13 km × 24 km 80 km × 40 km SO 2 column density retrieval DOAS LF DOAS DOAS SO 2 detection limit (DU) 0.2 0.2 0.2 0.5 Time of overpass (local solar time) 13h30 13h30 13h30 9h30 L2 Data Source NASA GES DISC AC-SAF/DLR Simulation of dense volcanic SO 2 air parcels on 16 December 2018: Role of TROPOMI TROPOMI 01:37 UT Model OMPS 01:58 UT Model GOME-2 21:57 UT Model OMI 01:48 UT Model
Transcript of Strength of TROPOMI observations on the retrieval of volcanic … · 2020. 5. 3. · Strength of...
Strength of TROPOMI observations on the retrieval of volcanic SO2
emissions at high temporal resolution from space by inverse modelingAbhinna K. Behera1, Marie Boichu1, and François Thieuleux1
1Laboratoire d’Optique Atmosphérique, Université de Lille, 59650 Villeneuve d’Ascq, France.
AbstractA refined knowledge of volcanic sulphur dioxide (SO2) degassing
at high temporal resolution, as SO2 is oxidized to form sulfate aerosolsin the atmosphere, is essential to advance further in understandingthe consequence of volcanic emissions on air quality, ecosystems, andclimate at a global scale. Concerning volcanological applications, infor-mation on SO2 degassing remains critical for assessing the underlyingmagma dynamics and associated hazards. However, presently, nadirspace-borne observations in the ultraviolet spectrum range, viz., OMI,GOME2, OMPS, and more recently launched TROPOMI hyperspectralsensor, can only provide snapshots of volcanic SO2 every 24 h. Here wedetermine hourly the SO2 flux emissions as well as the altitude of injec-tion by applying an inverse modeling method [1]. The forward modelinvolves the CHIMERE Eulerian regional chemistry-transport model[2]. The idea is to evaluate the advancements in utilizing high spatiallyand spectrally resolved measurements from TROPOMI spectrometeronboard ESA-Sentinel-5 Precursor. This investigation focuses on the2018 eruption of Ambrym (16.24°S, 168.15°E), a top world-ranking vol-canic SO2 emitter. It depicted a rapid cessation of a decade long SO2
degassing and lava lake activity associated with the emplacement of alarge injection of magma below the surface, crisscrossing the island ofAmbrym, and spawning a submarine eruption [4].
TROPOMI assimilated CHIMERE simulation leads to more detailedinformation on the dispersal of the volcanic SO2 cloud, and total massas well as altitude of injection by the inverse modeling scheme. Thisprovides the paroxysmal phase of the eruption occurred about 0430UT on 15 December 2018, injecting ∼8 kt of SO2 to the atmosphere at12 km above sea level (asl). Ambrym volcano emitted ∼100 kt of SO2
during 13-18 December 2018; most of the SO2 mass was centered at7 km altitude asl. Mostly, TROPOMI assimilated CHIMERE simula-tion depicts the weakest degassing of magnitude <100 t h
−1. At firstorder, SO2 flux time series are similar using TROPOMI, GOME2, andOMPS data. Overall, TROPOMI determines better the dense parcels ofSO2 in the core of the volcanic plume than OMPS or GOME-2. As aconsequence, model simulation assimilating TROPOMI observationsdescribes better the dispersal of such dense parts of the SO2 cloud, of-ten co-located with ash-rich zones that should be avoided for aviationsafety.
Methodology• Inverse modeling method: Linear least-squares regression approach
with non-negative constraint and no a priori [1] :(d
0
)
︸︷︷︸
Satellite Observation
=
(G
κ2∆
)
︸ ︷︷ ︸
(Simulated) Forward Operator
×
(m)
︸︷︷︸
Model Parameter
,
where κ2 is the Lagrangian multiplier and ∆ is the Laplacian - secondorder Tikhonov regularisation.
• Volcanic SO2 cloud dispersal, including the physico-chemical pro-cesses, is simulated using CHIMERE (Eulerian) chemistry-transportmodel [2] at a large-scale during 13 to 18 December 2018.
• WRF-ARW [3] model simulation is used for forcing the meteorologicalfields in CHIMERE. ERA-5 reanalysis with 0.25° spatial resolution isused hourly for the initialisation and nudging at the grid boundariesof WRF-ARW simulation.
• 10 km× 10 km horizontal resolution is implemented to resolve well thevolcanic SO2 dispersal, and the vertical layer thickness is maintainedwithin 25m to 550m that varies as a geometric sequence.
• Gaseous tracers (passive, no aerosol chemistry) are injected each hourto the atmosphere at each kilometer in a range of 2 km to 12 km altitudeasl during 2018121300 to 2018121800.
• To reproduce well the complex horizontal wind-fields, sharply varyingwith altitude in the surroundings of the Australian continent, WRF-ARW simulation was run with a fine vertical resolution of ∼250m since00 UT 1 December 2018 giving enough spin-up time.
Outlook• The high spatial and spectral resolution TROPOMI SO2 data paves
the way for the Eulerian chemistry-transport models to identify andtrack better the dense SO2 parcels as well as weakest degassing by theinversion scheme. The assimilation of TROPOMI observations shouldimprove the quality of volcanic plume forecasts that are critical foraviation safety.
• The complex wind-field, sharply changing with altitude, can impactthe reconstruction of volcanic SO2 clouds by the inverse scheme.Hence, the best available reanalysis data (e.g., ERA5) and the high-resolution meteorological model simulation are crucial to driveCHIMERE simulation.
Inverted SO2 flux time series
1300
30
1304
30
1308
30
1312
30
1316
30
1320
30
1400
30
1404
30
1408
30
1412
30
1416
30
1420
30
1500
30
1504
30
1508
30
1512
30
1516
30
1520
30
1600
30
1604
30
1608
30
1612
30
1616
30
1620
30
1700
30
1704
30
1708
30
1712
30
1716
30
1720
30
ddhhmm
101
102
103
104
Inve
rted
SO
2 Flu
x (to
n.hr
1 )
Log
-sca
le
OMI GOME2 OMPS TROPOMI
SO2 mass vs. altitude
2 3 4 5 6 Log10(SO2 mass) (ton)
0
2
4
6
8
10
12
14
16
Alt
itud
e(km
)
TROPOMI
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
KD
E
2 3 4 5 6 Log10(SO2 mass) (ton)
0
2
4
6
8
10
12
14
16
Alt
itud
e(km
)
OMPS
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
KD
E
2 3 4 5 6 Log10(SO2 mass) (ton)
0
2
4
6
8
10
12
14
16
Alt
itud
e(km
)
GOME2
0.01
0.03
0.05
0.07
0.09
0.11
0.13
0.15
0.17
KD
E
2 3 4 5 6 Log10(SO2 mass) (ton)
0
2
4
6
8
10
12
14
16
Alt
itud
e(km
)
OMI
0.01
0.03
0.05
0.07
0.09
0.11
0.13
0.15
0.17
KD
E
References[1] M. Boichu, L. Menut, D. Khvorostyanov, L. Clarisse, C. Clerbaux, S. Turquety, and P.-F. Coheur.
Inverting for volcanic SO2 flux at high temporal resolution using spaceborne plume imagery andchemistry-transport modelling: the 2010 Eyjafjallajökull eruption case study. Atmospheric Chemistryand Physics, 13(17):8569–8584, Sept. 2013.
[2] S. Mailler, L. Menut, D. Khvorostyanov, M. Valari, F. Couvidat, G. Siour, S. Turquety, R. Briant,P. Tuccella, B. Bessagnet, A. Colette, L. Létinois, K. Markakis, and F. Meleux. CHIMERE-2017: fromurban to hemispheric chemistry-transport modeling. Geoscientific Model Development, 10(6):2397–2423, June 2017.
[3] J. G. Powers, J. B. Klemp, W. C. Skamarock, C. A. Davis, J. Dudhia, D. O. Gill, J. L. Coen, D. J.Gochis, R. Ahmadov, S. E. Peckham, et al. The weather research and forecasting model: Overview,system efforts, and future directions. Bulletin of the American Meteorological Society, 98(8):1717–1737,2017.
[4] T. Shreve, R. Grandin, M. Boichu, E. Garaebiti, Y. Moussallam, V. Ballu, F. Delgado, F. Leclerc,M. Vallée, N. Henriot, et al. From prodigious volcanic degassing to caldera subsidence and quies-cence at ambrym (vanuatu): the influence of regional tectonics. Scientific reports, 9(1):1–13, 2019.
Main features of the space-borne sensors
Properties TROPOMI OMPS-NM OMI GOME2
Satellite ESA Sentinel-5 Precursor NOAA-Suomi NPP NASA EOS-Aura NOAA/ESA Metop-A+BSpectral ranges UV-VIS-NIR-SWIR UV UV-VIS UV-VIS-NIR
Nadir spatial resolution 7 km× 3.5 km 50 km× 50 km 13 km× 24 km 80 km× 40 km
SO2 column density retrieval DOAS LF DOAS DOASSO2 detection limit (DU) 0.2 0.2 0.2 0.5
Time of overpass (local solar time) 13h30 13h30 13h30 9h30L2 Data Source NASA GES DISC AC-SAF/DLR
Simulation of dense volcanic SO2 air parcels on 16 December 2018: Role of TROPOMI
156°E 160°E 165°E 170°E 175°E 180° 175°W 170°W27°S
24°S
20°S
16°S
12°S
8°S
SO2 total vertical column: TROPOMITotal mass(SO2)=7.50e+10 gramDate: 2018-12-16
SO2 max=664.10 DUAt lon: 168.22, lat: -16.18Area =1.81e+06 km2
156°E 160°E 165°E 170°E 175°E 180° 175°W 170°W
CHIMERE (Prediction)Total mass(SO2)=5.41e+10 gramstd.dev of Residuals=2.27
SO2 max=204.77 DUAt lon: 168.22, lat: -16.18Area =5.66e+05 km2
0 1 10 100SO2 (D.U.)
TROPOMI
01:37 UT
Model
156°E 160°E 165°E 170°E 175°E 180° 175°W 170°W27°S
24°S
20°S
16°S
12°S
8°S
SO2 total vertical column: OMPSTotal mass(SO2)=4.87e+10 gramDate: 2018-12-16
SO2 max=30.93 DUAt lon: 168.62, lat: -15.83Area =6.05e+05 km2
156°E 160°E 165°E 170°E 175°E 180° 175°W 170°W
CHIMERE (Prediction)Total mass(SO2)=4.51e+10 gramstd.dev of Residuals=0.46
SO2 max=30.68 DUAt lon: 168.62, lat: -15.83Area =5.47e+05 km2
0 1 10 100SO2 (D.U.)
OMPS
01:58 UT
Model
156°E 160°E 165°E 170°E 175°E 180° 175°W 170°W27°S
24°S
20°S
16°S
12°S
8°S
SO2 total vertical column: GOME2Total mass(SO2)=1.07e+11 gramDate: 2018-12-16
SO2 max=11.51 DUAt lon: 172.26, lat: -15.35Area =2.34e+06 km2
156°E 160°E 165°E 170°E 175°E 180° 175°W 170°W
CHIMERE (Prediction)Total mass(SO2)=5.60e+10 gramstd.dev of Residuals=0.56
SO2 max=9.97 DUAt lon: 172.98, lat: -15.50Area =9.13e+05 km2
0 1 10 100SO2 (D.U.)
GOME-2
21:57 UT
Model
156°E 160°E 165°E 170°E 175°E 180° 175°W 170°W27°S
24°S
20°S
16°S
12°S
8°S
SO2 total vertical column: OMITotal mass(SO2)=1.65e+10 gramDate: 2018-12-16
SO2 max=13.41 DUAt lon: 173.57, lat: -13.49Area =3.02e+05 km2
156°E 160°E 165°E 170°E 175°E 180° 175°W 170°W
CHIMERE (Prediction)Total mass(SO2)=1.50e+10 gramstd.dev of Residuals=0.34
SO2 max=9.26 DUAt lon: 174.06, lat: -13.70Area =2.13e+05 km2
0 1 10 100SO2 (D.U.)
OMI
01:48 UT
Model
