so2 and dust retrieval : status update adventures in so2 ...overview calbuco apr2015 sarychev...
TRANSCRIPT
1
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
SO2 and Dust Retrieval : Status UpdateAdventures in SO2 retrievals
Sergio DeSouza-Machado, Vince Realmuto
JCET/Physics, University of Maryland Baltimore CountyJet Propulsion Laboratory
AIRS Science Team MeetingOctober 2015Greenbelt, MD
2
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
Overview
Have delivered SARTA-scatter to JPL
Have placed generic match-up code to run model ECM/ERAcalcs with AIRS L1b data
Have put code to match AIRS L1b data to ascending vsdescending AIRS L3 climatology (very fast, less than 15 secs)
SO2 retrievals using Scott Hannon’s code and algorithm -linear interpolation
Have also implemented OEM retrievals were col WV isadjusted (1400 cm-1 channels), then SO2 retrievals doneassuming WV is "corrected"
Assumes SO2 plume is in upper troposphere, above WV(though can adjust height)
V. Realmuto providing test cases and other retrievals tocompare against
3
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
WV Jacobian (Tropical profile)
colorbar = �BT (K) for 10% change in gas amount in each layer
Black lines = STRONG SO2 channels; Blue lines = Weak channels
4
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
SO2 Jacobian (Tropical profile)
colorbar = �BT (K) for 10% change in gas amount in each layer
Black = STRONG SO2 channels; Blue = Weak channels"nominal" SO2 (0.11 du) buried under water, don’t see anything!!!
5
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
SO2 Jacobian x10 (Tropical profile)
colorbar = �BT (K) for 10% change in gas amount in each layer
Black lines = STRONG SO2 channels; Blue lines = Weak channelsSO2 (1.1 du) but still very tiny jacobian (0.0001K)!!!
6
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
Column Jacobians (with SO2 x 10,100) (Tropicalprofile)
Begin to see the signal pop out beyond noise at x10 (1 du) levels
7
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
� BT(SO2 mult, height) (Tropical profile)
AIRS NeDN in strong SO2 channels ⇠ 0.07 KSee how AIRS observed BT in strong SO2 channel(s) change as weput in slab SO2 perturbation in 2 km thick layer
Optimal SO2 height detection is in UT or higher
8
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
Averaging Kernel (Tropical profile)
Pretend entire SO2 profile is x10, compute jacobians at each layer,then compute AK = GK = (K0S�1
✏ K + S�1a )�1K0S�1
✏ K
Assumed 10% uncertainty in SO2 profile
9
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
Retrieval Outline
Colocate AIRS L1BWith model fi elds
Reset spres Stemp based on BT1231
See where BTD in strong SO2 lines
Is larger than 1.3 K
Predict SO2 multiplier
Reset SO2
1350 cm-1 SO2Multiplier
1400-1500 cm-1Water col mult
Excess SO2 burden
SCOTT OEM
10
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
Calbuco (S. Chile) 23 Apr 2015
-80 -75 -70 -65 -60
-40
-35
-30
-25
sergio
-1
-0.5
0
0.5
1
1.5
-80 -75 -70 -65 -60
-40
-35
-30
-25
scott
-1
-0.5
0
0.5
1
1.5
Left : OEM Right : RESETcolorbar is log10(du)g 186
11
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
Calbuco (S. Chile) 24 Apr 2015
-65 -60 -55 -50
-35
-30
-25
-20
sergio
-1
-0.5
0
0.5
1
1.5
-65 -60 -55 -50
-35
-30
-25
-20
scott
-1
-0.5
0
0.5
1
1.5
Left : OEM Right : RESETcolorbar is log10(du)g 177
12
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
Calbuco (S. Chile) 23-25 Apr 2015
Date
22 22.5 23 23.5 24 24.5 25 25.5 26
ma
ss S
O2
(kilo
ton
s)
0
50
100
150
200
250
sergio
scott
Added together couple or sogranules per day, most of SO2burden in one of them
Emission (kilotons)
Method model Date Date Date23 24 25
UMBC OEM L3 187 176 14UMBC OEM ECM 179 173 9
UMBC RESET ECM 179 182 11
13
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
Sarychev 2009/09/16 (g26,27)
Emission (kilotons) using ECMUMBC OEM AIRS 220.8
UMBC RESET AIRS 212.9
Plume Tracker MODIS 790Plume Tracker AIRS 640Prata-Bernando AIRS 440
Yang OMI 430
14
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
Manam PNG 31 July 2015 g044
130 135 140 145
-14
-12
-10
-8
-6
-4
-2
0
2
4
6
sergio
20
40
60
80
100
120 Emission (kilotons)ECM AIRS L3
OEM 9.1 12.5RESET 1.7 1.6
Plume Tracker MODIS 100-400Prata-Bernando AIRS 52
Yang OMI 2.67Yang OMPS 1.5
Krotkov OMI/OMPS 3.12
15
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
Kasatochi 2008/08
Left : Aug 08, g137 (du) Right Aug 10, g230 (du)
-180 -175 -170 -165 -160
45
46
47
48
49
50
51
52
53
54
55sergio
0
50
100
150
200
250
300
-170 -165 -160 -155 -150 -145 -140
40
45
50
55
60
sergio
20
40
60
80
100
120
DAY Granule OEM (kt) Linear (kt) OEM Linear OMI15km 15 km 10 km 10 km
8 137 352.99 129.85 15000 602 8508 232 89.28 62.44 18000 6849 9 56.03 45.31 6600 365 8709 127 291.4 159.54 7436 4719 128 211.2 184.44 40000 2279
10 229 4.72 4.06 150 23 134010 230 353.3 314.4 41000 3011
16
Overview Calbuco Apr2015 Sarychev 2009/06/16 (g26,27) Manam PNG July 2015 Kasatochi 2008/08 (g26,27)
Conclusions
implemented Scott Hannon’s "reset" (based on linear interp)also implemented OEM retrieval (with column water vaporburden retrieval done prior to SO2 column)compared against results for various eruptions; agree towithin order of magnitude (all results have large differencesamongst each other)
My OEM does a column WV adjustment before doing the OEMSO2 column retrieval, while Scott only did a linearized SO2column retrieval (so water could be incorrect)there are discrepancies between using easily available AIRS L3climatology for model fields versus eg ECMWF or ERA, for whichI have to wait for the model fields to come inretrievals between different methods and instruments showmuch variability, as they differ in their sensitivities to surfaceand atmospheric properties, compositions of plumes,clouds/ash in the wayAIRS retrievals available day and night, window channels couldbe used to report on presence of ash and/or clouds
Thanks to Scott Hannon and Larrabee Strow!