Embed Size (px)
Citation preview
Why does stratospheric water increase in models?
A. E. Dessler Texas A&M University
H. Ye, T. Wang, M.R. Schoeberl L.D. Oman, A.R. Douglass
A.H. Butler, K.H. Rosenlof, S.M. Davis, R.W. Portmann
30N-30S, 85 hPaannual average
3
3
Cold trap
3
Cold trap
Hypothesis: cold trap warms
Tropopause
355 K
380 K
400 K
14.5 km
16.5 km
19 km
X X X X X X X X X X
level of zero netradiative heating
Tropopause
355 K
380 K
400 K
14.5 km
16.5 km
19 km
X X X X X X X X X X
Tropopause
355 K
380 K
400 K
14.5 km
16.5 km
19 km
X X X X X X X X X X
Tropopause
355 K
380 K
400 K
14.5 km
16.5 km
19 km
X X X X X X X X X X
X X X X X X X X X X
Tropopause
355 K
380 K
400 K
14.5 km
16.5 km
19 km
X X X X X X X X X X
X X X X X X X X X X
Tropopause
355 K
380 K
400 K
14.5 km
16.5 km
19 km
X X X X X X X X X X
X X X X X X X X X X
X X X X X X X X X X
Horizontal view Ver/cal view
1 day
Horizontal view Ver/cal view
3 days
Horizontal view Ver/cal view
1 week
Horizontal view Ver/cal view
1 month
Horizontal view Ver/cal view
3 months
Parcels have been thinned out by a factor of 10 Expanded the al/tude scale
Horizontal view Ver/cal view
6 months
Parcels have been thinned out by a factor of 10
Horizontal view Ver/cal view
1 year
Parcels have been thinned out by a factor of 10
Horizontal view
Ver/cal view
Parcels have been thinned out by a factor of 10
12/31/2005
355 K
380 K
400 K
Lagrangian cold point
X
355 K
380 K
400 K
Lagrangian cold point
X
Schoeberl and Dessler, ACP, 2011Schoeberl et al., ACP, 2012, 2013Bowman, JGR, 1993; Bowman and Carrie, JAS, 2002
plan• use 6-hourly met data from GEOSCCM and
WACCM to drive our trajectory model
• predict entry-level H2O and compare to full model
• traj. model only contains TTL temperature variations
• compare to CCM prediction; differences caused by other processes
GEOSCCM
GEOSCCM
0.74
GEOSCCM
0.74
0.24
GEOSCCM
only one-third of the model’sincrease is due to warming TTL
0.74
0.24
WACCM
WACCM
about 80% of the model’sincrease is due to warming TTL
conclusions, I• both models have a bigger change in strat.
H2O than can be explained just by TTL temps
• these other processes explain 66% (GEOSCCM) and 20% (WACCM) of the increase in H2O
24
H2O
vm
r (re
lativ
e to
200
0-05
)GEOSCCM lower stratospheric tropical water vapor
Dessler, PNAS, 2013
24
H2O
vm
r (re
lativ
e to
200
0-05
)GEOSCCM lower stratospheric tropical water vapor
Dessler, PNAS, 2013
24
H2O
vm
r (re
lativ
e to
200
0-05
)GEOSCCM lower stratospheric tropical water vapor
Dessler, PNAS, 2013
24
H2O
vm
r (re
lativ
e to
200
0-05
)GEOSCCM lower stratospheric tropical water vapor
85-hPa tropical heating rates
Dessler, PNAS, 2013
24
H2O
vm
r (re
lativ
e to
200
0-05
)GEOSCCM lower stratospheric tropical water vapor
85-hPa tropical heating rates500-hPa tropical temps.
Dessler, PNAS, 2013
25
H2O
vm
r (re
lativ
e to
200
0-05
)
26
H2O
vm
r (re
lativ
e to
200
0-05
)
GEOSCCM
R^2 = 0.95
R^2 = 0.83
Trajectory
GEOSCCM
BD term
∆T term
GEOSCCMTotal diff = 0.50
BD term
∆T term
GEOSCCM
0.09
Total diff = 0.50
BD term
∆T term
GEOSCCM
0.39
0.09
Total diff = 0.50
BD term
∆T term
CCMs tracks ice
CCMs tracks ice
traj. model to adds anvil ice to parcel; do not let parcel exceed 100% RH
30
GEOSCCM & trajectory, 85-hPa tropical annual avg.
31GEOSCCM & trajectory, 85-hPa tropical annual avg.
GEOSCCM & trajectory, 85-hPa tropical annual avg.
warmingtropopause
ice lofting
WACCM
conclusions, II• both models have a bigger change in strat.
H2O than can be explained just by TTL temps
• these other processes explain 15% (WACCM) and 60% (GEOSCCM) of the increase
• missing processes mainly correlate with tropospheric temperature
• ice lofting is a very plausible solution
35
MLSmonthly avg. anomalies82 hPa (18 km)30°N-30°S avg.
35
MLSmonthly avg. anomalies82 hPa (18 km)30°N-30°S avg.
interannual variability
36
36
• 10-year overlapping segments from the CCM and trajectory models
36
• 10-year overlapping segments from the CCM and trajectory models
• fit each segment: H2O = b BD + c ΔT
36
• 10-year overlapping segments from the CCM and trajectory models
• fit each segment: H2O = b BD + c ΔT
Model
H2O= -4.17 BD + 0.29 ΔT
Model
H2O= -4.17 BD + 0.29 ΔT
Model
Trajectory
H2O= -4.17 BD + 0.29 ΔT
H2O= -5.67 BD + 0.17 ΔT
Model
Trajectory
H2O= -4.17 BD + 0.29 ΔT
H2O= -5.67 BD + 0.17 ΔT
Model
Trajectory
H2O= -4.17 BD + 0.29 ΔT
H2O= -5.67 BD + 0.17 ΔT
Model
Trajectory
H2O= -4.17 BD + 0.29 ΔT
H2O= -5.67 BD + 0.17 ΔT
Model
Trajectory
38
38
ΔT coefficient = increase in water vapor per degree warming of the troposphere
monthly avg. anomalies82 hPa (18 km)30°N-30°S avg.
monthly avg. anomalies82 hPa (18 km)30°N-30°S avg.
monthly avg. anomalies82 hPa (18 km)30°N-30°S avg.
Dessler et al., JGR, 2014
monthly avg. anomalies82 hPa (18 km)30°N-30°S avg.
Dessler et al., JGR, 2014
40
GEOSCCM
40
ΔT coefficient = increase in water vapor per degree warming of the troposphere
GEOSCCM
41
GEOSCCM
41
ΔT coefficient = increase in water vapor per degree warming of the troposphere
GEOSCCM
conclusions, III• trend in ice lofting is responsible for a
significant part of the trend in strat. H2O in models over the 21st century
• signature of ice lofting can is (potentially) apparent in short-term climate variability in the model
• a similar signature is apparent in the MLS data
43
GEOSCCM
43
BD coefficient = increase in water vapor per unit change in BD circulation strength
GEOSCCM
44
R2 = 0.7
Dessler et al., PNAS, 2013
45Dessler et al., PNAS, 2013
45Dessler et al., PNAS, 2013
46
46
47
WACCM
47
ΔT coefficient = increase in water vapor per degree warming of the troposphere
WACCM
48
WACCM
48
BD coefficient = increase in water vapor per unit change in BD circulation strength
WACCM
WACCM
R^2 = 0.84
R^2 = 0.81
WACCM
WACCM
WACCM
WACCMTotal diff = 0.15
WACCM
0.02
Total diff = 0.15
WACCM
0.10
0.02
Total diff = 0.15
WACCM
0.10
0.02
Total diff = 0.15
QBO = 0.01
