Solar Spectral Irradiance (SSI)
changes, atmospheric effects?
J. Fontenla
NorthWest Research Associates
and
LASP-University of Colorado
The topic: SSI changes,
and does it matter?
• Solar Physics issues:
–
–
–
–
Solar atmosphere structure and SSI
Non-LTE radiative transfer
Solar magnetic (sunspot) cycle
Magnetic effects on the solar atmospheric layers
• Atmospheric issues:
–
–
–
–
Photochemistry
Heating of various layers of the Earth atmosphere
Ocean currents and energy transport
Effects of all the above on circulation
The solar research on SSI modeling
• Emitted intensity spectrum =>solar atmospheres
• 1970s: HSRA, BCA, Gingerich, Peytremann, Holweger & Muller
• Avrett et al. 1981 VAL non-LTE, Athay & Thomas non-LTE and
chromosphere, Mihalas, Kurucz LTE stellar models, 1980s
• Transition region and Lyα, Fontenla et al. 1993 FAL, 1990s
•
•
•
•
•
Solar atmospheres => SSI calculations
Solanki/Unruh 1998, 3 component, LTE
RISE models 1999, 6 components, full/approx NLTE
Shapiro/Krivova 2011, full NLTE in few species/levels
SRPM models 2011, 9 components, resolved over the disk, full
NLTE in 50 species/over 13,000 levels/over 170,000
atomic/ionic lines and over 550,000 molecular lines (LTE).
SSI spectral features and
atmospheric regions
Fontenla, Avrett, & Loeser 1991, FAL 2, The Astrophysical Journal, 377:712-725
Transition region: EUV/FUV
emission lines
Upper chromosphere: deep
absorption line cores and UV
emission lines
Lower chromosphere: NUV,
visible, IR absorption lines
Photosphere: visible,IR
continuum and weak
absorption lines
Solar Surface Features
A-weak internetwork (new)
B-internetwork (changed C)
D-network (new)
E-active network (changed F)
H-normal plage (new)
P-bright plage (changed P)
Q-very hot plage (new)
S-sunspot umbra (temp)
R-sunspot penumbra (new/temp)
Models of solar atmospheric features
Fontenla et al. 2006,The
Astrophysical Journal,
639:441–458,
Models cross at ~6500 K,
in NLTE.
9000
Feature C model 1001
Feature H model 1004
Feature P model 1005
8000
Temperature (K)
Solanki & Unruh
1998, Astron.
Astrophys. 329,
747-753, LTE.
No crossing in
these models, SSI
computed in LTE
from FAL P with
modifications.
Fontenla et al. 2011, JGR, 116,
full NLTE, Tmin very different
7000
6000
5000
4000
5
4
10
10
Pressure (dyne cm-2)
Fontenla et al. 2011, JGR, 116, D20108
Photospheric and chromospheric layers
Transition-region and Coronal layers
3.2
CA
CB
CD
CF
CH
CP
CQ
Temperature (MK)
2.8
2.4
2.0
1.6
1.2
0.8
0.4
0
40
80
120
160
200
240
Height (Mm)
Contributions to Quiet-Sun TSI (1360 W m-2):
•Photosphere: ~1351 W m-2
•Chromosphere: ~8 W m-2 (power >> TSI observed changes)
•Corona+Transition-region: ~70 mW m-2
Computed vs observed SSI
Some observations considered for
SRPM set of atmospheric models
Features continuum contrast varies with wavelength and
heliocentric angle, corresponds to the slope of T vs p, SRPM
model set used detailed radiance observations
Sanchez Cuberes et al. 2002, The
Astrophysical Journal, 570:886–899
Topka et al. 1997, The Astrophysical
Journal, 484:479-486
San Fernando Observatory
Ground-based radiance observations confirm that ARs are dim in the visible,
over the solar cycle plage near the limb do not increase the visible SSI.
Preminger et al. 2011, ApJ
Ca II K
Red
Blue
Controversy1: Calculated SSI behavior
corr
nocorr
- corr
- nocorr
nrlssi
- nrlssi
0.1
0.01
SSI/SSI
Solanki & Unruh 1998,
Astron. Astrophys. 329, 747-753
1E-3
1E-4
1E-5
100
1000
10000
Wavelength (nm)
Relative changes between Solar Cycle 23 peak/min
that I am using for WACCM4 simulation runs.
Nocorr – Fontenla et al 2011, SRPM + PSPT images
Corr - same as above with a correction to match TSI
NRLSSI – WACCM4 default.
“Lean_1610-2140_ann_c100405”
According to this paper: «The
dotted curve shows the
observed relative irradiance
variation for λ < 400 nm between
solar activity minimum and
maximum vs. wavelength,
compiled by Lean et al. (1997)
and extrapolated to longer
wavelengths by Lean (1991). »
“nocorr” SSI (wavelength,time)
Lower- and upper-chromosphere
bright/dark fine structure, 1-D models only a first
approximation to the net medium-resolution
Lower chromosphere
Extension of the granulation structure.
Some localized energy dissipation in the
walls of downdrafts.
Upper-chromosphere
Loops and mechanical dissipation
3D Radiation Transport & NLTE
Computed for photospheric convection simulation snapshot
with data from Stein & Nordlund 2005
C I 5381
500 nm
1200 nm
1600 nm
SRPM 306
Stein & Nordlund
SRPM 306 * 0.95
400
mostly
convective
transport
mostly radiative
transport
6000
Height (km)
8000
Temperature (K)
800 nm
200
SRPM 306
Stein & Nordlund
SRPM 306 + 30 km
0
4000
1 10
3
CN band
Mg I 4572
1 10
4
1 10
5
Pressure
(dyne
Height
(km) cm^-2)
1 10
6
1 10
3
1 10
4
1 10
Pressure (dyne cm^-2)
5
1 10
6
Network and its change over the
cycle, what is “quiet-Sun”?
In the so-called “quiet-Sun”,
i.e. locations where no obvious
AR are present, the intensity
distribution of the network is
observed to change with the
solar cycle (maybe not strictly
in phase with the sunspot
index).
1
A
B
D F
H
P
Relative ar ea
0.1
0.01
3
110
4
110
peak
low
Intensity distribution at the disk center
1
1.2
Intensity
This has
implications for
SSI and for TSI.
But available
images lack
reliable absolute
calibration. Day
to day matching
was done with
the median.
Feature, model, TSI
A
1101
1374.60
B
1001
1382.19
D
1002
1388.15
F
1003
1391.44
H
1004
1400.86
P
1005
1419.14
S
1006
265.97
R
1007
1103.82
Q
1008
1428.82
Time series of solar spectral variability from SORCE/SIM
Controversy2: NUV Observations
Various instruments claiming reliable calibration for long term
Most instruments show variation of about ~50/1000~5% except for SUSIM.
Only SUSIM measured one peak, since UARS/SOLSTICE hardware failed in 2000
Both SORCE instruments show ~6% variations; their decreasing SSI turned
around to increasing as SC 24 started in ~2010, but later data is not shown here.
NUV effects on O3
Calculations were carried out by Merkel et al (see GRL38, L13802 2011), using
SORCE data extrapolated in time. These are done with WACCM3 in static SSI runs.
Other authors also made simplified calculations showing important differences.
I am carrying out transient WACCM4 (NCAR Community Earth System Model 1.0.3 )
runs with coupled atmosphere, ocean, land, and ice. O3 is included but so are many
other processes.
nocorr
corr
nrl
delta.Flux/Flux_min (%/100 F10.7 units)
8
6
4
2
0
180
200
220
240
260
280
300
320
W avelength (nm)
340
360
380
400
CESM (WACCM4) for SSI study
• Transient runs 1955-2005 including all
observed forcing. Imposing observed QBO.
• SSI: wavelength < 120 nm uses F10.7 proxy
• SSI: 120 nm < wavelength < 100 μm:
– “const” uses time independent low activity SSI
– “nocorr” from SRPM + PSPT & Meudon images,
repeats SC23 (with stretching)
– “corr” same as above but with a correction to
match observed TSI, still under development
– “nrlssi” using the default SSI in CESM, from Lean
SSI “nocorr” model of SC23, vs NRLSSI
1990
1060
Irradiance
8
7
6
SRP M
Lean*0.972
1055
1985
Irradiance
SRP M
Lean
9
Irradiance
448.5 nm
368.5 nm
121.5 nm
10
1050
1980
1045
1975
SRP M
Lean*0.9 78
5
4
2000
1040
2005
2000
1598
1596
796
795
SRP M
Lean*0.97
SRP M
Lean*1.006
794
X
 3 .59 2 1 0
3592.5 nm
X
2000
2005
Year
3
0 1600
 6 .18 5 1 0
6185.5 nm
TSI
13.02
SRP M
Lean*0.995
13.01
2000
-1
1.583
1362
1.582
1360
SRPM
Lean
1.581
2005
Year
SRP M
Lean*0.995
Irradiance
Irradiance
Irradiance
13.025
-2
SRP M
Lean*1.018
1.584
13.03
13.015
252.4
252
2005
Year
3
0 1200
252.6
252.2
2000
2005
Year
X
Year
252.8
Irradiance
Irradiance
Irradiance
1600
2005
1593.5 nm
797
1602
2000
0 670
2000
 9 42 .5
942.5 nm
648.5 nm
1592
X
Year
Year
1594
1970
2005
1.58
1358
2000
2000
2005
Year
-2
2005
Year
0 8
Follow some preliminary results
• Only from one complete run of each case, the 3
years near the minimum, over 4 solar cycles were
averaged and compared.
• The same was done for and 3 years near the
maximum over 4 solar cycles.
• Then, the averages of 12 years near min and 12
years near max were subtracted to show the
effect of SSI change.
• The maps shown below are for the DJF season,
the JJA patterns are different.
• The zonal means are annual.
• More instances are running to form an ensemble.
However, Earth behavior is only one instance.
Preliminary WACCM surface results
Cloud fraction
delta.CLOUDT nocorr
delta.CLOUDT const
delta.TS const
delta.PS const
delta.CLOUDT nrlssi
Surfacte temperature
delta.TS nocorr
delta.TS nrlssi
Surface pressure
delta.PS nocorr
delta.PS nrlssi
latitude
longitude
ENSO and “natural” variability issues
Does the SSI choice affect these?
More “realizations” are needed
How to cancel volcanic effects?
ENSO 3.4 region (DJF)
4
Volcanic eruptions are a big issue:
Mt. St. Helens 1980
El Chichon 1982
Pinatubo 1991
delta.T (K)
2
0
2
4
1960
1970
delta.PS (hPa)
delta.PS (hPa)
2000
10
100
0
 100
 300
1990
Year
ENSO
SOI (DJF)
Alleutian Pressure (DJF)
 200
1980
const
corr*
nocorr
nrlssi
1960
5
const
corr*
nocorr
nrlssi
0
5
1970
1980
Year
1990
2000
1960
1970
1980
Year
1990
2000
Zonal mean T and H2O changes
“const” displays changes that are not due to the SSI choice, difference of difference
can eliminate some but is affected by the “noise” in both “nocorr” and const”
const
nocorr
nocorr-const
-2
Log10(pressure)
Log10(pressure)
-1
0
1
-3
-2.000
-1.600
-1.200
-2
-0.8000
-3
-2.000
-1.600
-1.200
-2
-0.8000
-0.4000
0.000
-1
0.4000
0.8000
1.200
0
1.600
2.000
-0.4000
0.000
-1
0.4000
0.8000
1.200
0
1.600
2.000
Log10(pressure)
-3
1
2
3
3
-80
-60
-40
-20
0
20
40
60
80
T (K)
3
-80
-60
-40
-20
Latitude
0
20
40
60
80
-80
-60
-40
-20
Latitude
-2
Log10(pressure)
-1
0
1
0
20
40
60
80
Latitude
-3
-0.02000
-0.01600
-0.01200
-2
-0.008000
-0.004000
0.000
-1
0.004000
0.008000
0.01200
0
0.01600
0.02000
-3
-0.02000
-0.02000
-0.01600
-0.01200
-0.008000
-0.004000
0.000
0.004000
0.008000
0.01200
0.01600
0.02000
-0.01600
-0.01200
-2
-0.008000
-0.004000
0.000
-1
0.004000
Log10(pressure)
-3
Log10(pressure)
1
2
2
-2.000
-1.600
-1.200
-0.8000
-0.4000
0.000
0.4000
0.8000
1.200
1.600
2.000
0.008000
0.01200
0
0.01600
0.02000
1
1
2
2
2
3
3
3
H2O
(relat)
-80
-60
-40
-20
0
Latitude
20
40
60
80
-80
-60
-40
-20
0
Latitude
20
40
60
80
-80
-60
-40
-20
0
Latitude
20
40
60
80
Issues analyzing simulation results to
separate SSI effects from other effects
tropical (±25 deg) annual differences between peak and min years
300
14
10000
12
tropos
1000
stratos
mesos
H2O (ppm)
thermos
O3 (ppm)
T (K)
10
8
6
100
10
4
1
2
0.1
200
0
-4
-2
0
2
-4
-2
Log10(pressure)
0
4
-4
0
2
4
O3
nocorr
const
nocorr-const
8
6
4
H2O
nocorr
const
nocorr-const
2
Delta.H2O
deltarel.O3 (%)
T
nocorr
const
nocorr-const
2
-2
Log10(pressure)
12
10
delta.T (K)
2
Log10(pressure)
0
2
0
0
-2
-2
-4
-2
0
Log10(pressure)
2
-4
-2
0
Log10(pressure)
2
-4
-2
0
Log10(pressure)
2
Other WACCM Simulation Interesting Results
Nocorr-Const
NRLSSI-Const
= DifDif.FSDS
Shows the downwelling solar
shortwave flux at the surface
increases of ~30 W m-2 at the
Pacific Ocean Warm Pool region
at solar max times. But was
shown before that the surface
temperature does not increase
much there. Ocean effects,
Kuroshio stream, moderate T?
The NRLSSI dif. of dif. have also
some of this behavior on the Pacific
Ocean Warm Pool but the details
are quite different. Also, NRLSSI
results show several differences in
other regions, e.g. the patterns in
Mexico Pacific area, Brasil Atlantic,
and Madagascar Indian Ocean
which are not shown by “nocorr”
results
Ocean gyres
Ocean currents couple to atmospheric winds and tropospheric energy
transport. This is represented in CESM1.0.3/WACCM4 by the integration of
the atmospheric model (CAM2) with the deep ocean model (POP).
The tropospheric and ocean phenomena are very tangled!
Analysis is very complex but these simulations contain a wealth of data
which could nail down the physical processes induced by SSI changes.
That is, if one could also figure out other forcing and variability.
Future work
• All the maps shown above is for DJF season, similar ones for JJA
season were done.
• Still improving “corr” SSI case by reprocessing images into “corr2”,
hope to have it by end of year 2012 .
• Performing runs for more instances of all SRPM cases, necessary to
separate natural variability and SSI effects 4 instances are the
target.
• Analysis of the data for ocean and other components of CESM
model remains to be done.
• Comparison between CESM 1.0.3 and MODTRAN@ atmospheric
radiative heating/cooling to be carried out to evaluate spectral model
and resolution effects.
• FUV/EUV SSI ongoing modeling for replacing F10.7 proxy and for
forecast of thermospheric neutral density and ionization.
Dark active regions
A nice example on
2/3/2007 shows two large
magnetic active regions
sunspots side-by-side
and one is associated
with a lot of
chromospheric and
coronal heating but the
other is not showing
much heating.
The magnetic flux of the
sunspots is not too
different but the bright
region is bipolar and more
complex.