Frédéric Clette amp Laure Lefèvre Royal Observatory of Belgium WDC SILSO ESWW11 Nov 2014 SN and GN recalibration early preview 21112014 ESWW11 Nov 2014 2 ID: 564567
Download Presentation The PPT/PDF document "What do sunspots tell us about recent an..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
What do sunspots tell us about recent and past trends in solar activity ?
Frédéric Clette & Laure Lefèvre Royal Observatory of BelgiumWDC - SILSO
ESWW11 – Nov 2014Slide2
SN and GN recalibration:
early preview21/11/2014ESWW11 – Nov. 2014
2
Ongoing
revision
of inhomogeneities in the
Sunspot
Number
(WDC-SILSO)
and Group
Number
series
(
Hoyt
&
Schatten
1994, 1998)
SN Workshops, 2011-2015Slide3
SN and GN
recalibration: early preview
21/11/2014
ESWW11 – Nov. 2014
3
RGO trend
SN
weighting
RGO/
SOON
Specola
drift
Clette
,
Svalgaard
, Vaquero, Cliver, 2014
Space
Science
Reviews
,
Aug. 2014, Springer Online First , 69 pagesDOI 10.1007/s11214-014-0074-2Arxiv: http://arxiv.org/abs/1407.3231
4 main corrections
10 to 40%
Obtained independently
Based only on sunspot data
SN and GN agree within uncertainties back to the Dalton MinimumSlide4
Implications: low secular trends
21/11/2014ESWW11 – Nov. 2014
4
Corrected
SN & GN
series
agree
Secular
trend
is
largely
eliminated
< 5% /
century
Trend +15%/
century
Trend +40%/
century
SN correction:
SN /1.20
after
1947
GN correction:
GN * 1.37
before
1880
Soon after the Maunder Minimum, solar activity was similar to present levelsSlide5
21/11/2014
ESWW11 – Nov. 20145
Open solar
flux
(geomagnetic indices)
Similar conclusions: recent open magnetic field reconstructions show
only a weak
trend over last 180 years
(Lockwood, Living Reviews SP, 9/2013)Slide6
New SN series (red):
correction of SN scale drifts due to the Locarno pilot station
Changes in cycle 22
Implications: recent cycles
21/11/2014
ESWW11 – Nov. 2014
6
Cycle 22
Slightly
decreased (- 5 to10
%)
Maximum 22
second peak is higher, almost equal to first
peakSlide7
Changes in cycle 23
Implications: recent cycles
21/11/2014
ESWW11 – Nov. 2014
7
Cycle 23
Slightly increased (+ 5 to 10
%)
Maximum 23
second
peak becomes higher than the first peak.
Cycle 23 decline raised
by about
20
%
New maximum in 2002 instead of 2000Slide8
Implications: recent cycles
Still
significant disagreement during the decline of cycle 23
Indication of a real solar change
21/11/2014
ESWW11 – Nov. 2014
8
New SN series (red):
correction of SN scale drifts due to the Locarno pilot station
Deviation between
R
i
and F
10.7
in cycle 23Slide9
Implications: a variable number of spots/group (SN/GN ratio)
Reconstructed GN series (SILSO, SONNE):Both indices SN and GN calculated from the same data set
Changing
SN/GN
ratio :
Stable over cycles 19 to 22
Decline in cycle 23 and 24
Decrease of the average number of spots per group by ~30%
21/11/2014
ESWW11 – Nov. 2014
9
SONNE SSN
Locarno SSNSlide10
SN
and GN
contain
a
different
information about the
solar
cycle
A probe for
past
changes in the
solar
dynamo
?
Implications: a variable number of spots/group (SN/GN ratio)
10
21/11/2014
ESWW11 – Nov. 2014
Tlatov 2013
Apparent
secular
variations of the
ratio SN/GN
(
Tlatov
2013)
:
NS/NG increase for stronger cycles ?Ratio of the original SN/GN series: different sets of observations. Slide11
Vanishing small spots
21/11/2014ESWW11 – Nov. 2014
11Slide12
Cycle 23 small-spot deficit
Exploitation of detailed sunspot catalogs: DPD, Debrecen; NOAA/SOON (Lefèvre &
Clette
2011, 2012, 2013)
Scale-dependent small-spot deficit in cycle 23
:
Deficit of small groups (A & B types):
Ratio cycle 23 / cycle 22 ~ 50%
Deficit of small spots inside all groups:
Ratio cycle 23 / cycle 22 < 75%
Starts in 1998, significant only after 2000
21/11/2014
ESWW11 – Nov. 2014
Small vs Large groups
Clette & Lefèvre 2012
12
Small vs Large spotsSlide13
Cycle 23 small
-spot deficit
ESWW11 – Nov. 2014
Similar size-dependent trends found by
Kilcik
et al. 2013
:
U
pgrade of
Kilcik
et al. 2011 results
Based on
Learmonth
data
Small spots
A,B types: factor ~2 deficit Intermediate C type:
moderate decreaseLarge D,E,F,H groups: no difference between SC23 and 24.
A,BC
D,E,F
H_____ Sunspot counts
--------- Sunspot group counts_____ International SN Ri
21/11/201413Slide14
Similar results from
Nagovitsyn et al. 20134 classes based on their area (multimodal distribution)Only the smallest spots (A<17 msh) show a decline in SC23
Number of largest spots increases
Intermediate sizes: no change
Cycle 23
small
-spot
deficit
ESWW11 – Nov. 2014
21/11/2014
14
SS: small < 17
msh
SL: 17
msh
<S<58
msh
LS: 58
msh
<S<174 mshLL: > 174
mshSlide15
New catalog validation
21/11/2014ESWW11 – Nov. 2014
15Slide16
Sunspot catalog cross-validation
Need to assess the homogeneity of the primary DPD sunspot catalog:Several data sources: ground based
stations (80
% in Hungary)
Resolution
: 1 to 2”/pixel
(instrument,
seeing)Sunspot groups and individual
sunspots (unique identification)
Cross-analysis with the
new STARA
MDI sunspot
catalog
(
F.Watson, NSO)
:SOHO MDI continuum imagesResolution: 2”/pixelIndividual
sunspots (not tracked)Two versions built separately: Whole spot (penumbral area)
Separate umbral areas
21/11/2014ESWW11 – Nov. 201416Slide17
Main Sources of mismatch
DPD spot classification:“Too” detailed: sunspot classes without equivalent in other catalogs
:
P
enumbrae without umbrae
S
mall
umbral
kernels inside common penumbrae
Dropped in sunspot area comparisons
Time difference between daily DPD and STARA observations:
STARA
version 1: mean
∆
t=10 h
Image closest to daily magnetogram
Poor match for small spots due to spot evolution over 10 hours.STARA version 2:mean ∆t=
2-3 hWe use only version 2
21/11/2014ESWW11 – Nov. 2014
17
-20 -10 0 10 20
∆
t DPD-STARA (hours)NNSlide18
Sunspot group matching: movie
ESWW11 – Nov. 201418
21/11/2014Slide19
Matching in distance and size
21/11/2014ESWW11 – Nov. 201419
All
UP > 10msh
UP> 100
msh
UP> 500
msh
Sunspot position: very accurate match
Mean difference:
0.35°
max. ~5° (due to splitting of large complex spots)
Sunspot
areas
(
umbra+penumbra
): good match
No bias: mean ratio = 1
RMS dispersion increases for small sunspot areas:
10msh<A<100msh: σ= 20%Main causes of
larger differences in small spot areas:DPD-STARA time difference: small spots evolve fasterMDI lower spatial resolution (2 “): pixel quantizationSlide20
Catalog matching: conclusions
Total number of individual spots: 57500Matching in DPD & STARA: 93 %Non-matching: 7 % (mainly small short-lived spots)Sunspot areas: the accuracy of DPD sunspot areas is the best available among existing catalogs
The comparison with STARA confirms the accuracy of areas
for A >
30msh
For A <
30msh, only DPD: results still rest on the intrinsic stability of the DPD catalog construction
Sunspot counts: no systematic variation over
time
found in the DPD
catalog
(
T. Baranyi, private
communication)The in/exclusion of DPD-specific classes does not influence the time variations found in the small-spot population.
21/11/2014ESWW11 – Nov. 2014
20Slide21
Variations of Other sunspot properties
21/11/2014ESWW11 – Nov. 2014
21Slide22
Decline
of core magnetic fieldsAverage core magnetic field in umbra (FeI line: 1565 nm,
Kitt
Peak)
21/11/2014
22
Solar cycle modulation
(
Nagovitsyn
et al., 2012)
NB: only the strongest field each day.
ESWW11 – Nov. 2014
Linear
decline
-
40
G/
year
(Penn & Livingston 2011, 2013)
Most
recent
data (2014): BABO (
Penn & Livingston)
, MDI/HMI
(F. Watson)
Decline
has
stopped
in cycle 24, but no
solar cycle modulationSlide23
Growth/decay rates of active regions
Study of group growth and decline (Javaraiah 2011):Data: Greenwich photographic catalog, USAF/SOON catalog
In
cycle 23:
Lower growth rate
Decay rate increasing
Scarcity
of groups with A < 37 msh (Javaraiah 2013)
Coherent with sunspot deficit
Weaker growth rates similar to moderate cycles 12 to14
21/11/2014
ESWW11 – Nov. 2014
23
Growth
Decay
Javaraiah
2011Slide24
Implications: a shallow dynamo?
Solar-cycle modulation of high-
frequency
p-modes
(
Basu
et al. 2013)
: BISON dataTop layers
(r
>
0.997
r
ʘ
): deviate after
1998Deeper layer: deviation during entire cycle 23
Thinning of the subsurface magnetic field
layer (< 20000 km)
21/11/2014ESWW11 – Nov. 2014
Basu et al. 2013
24
Dynamo
models
:
Are
there
two
dynamo components,
deep
and
shallow ? Does the near-surface shear layer play an independant
role?Babcock-Leighton
near-surface flux diffusion mechanism (Muñoz-Jaramillo
et al. 2010, 2011)Role of a near-surface shear layer (A. Brandenburg 2005, Brnadenburg et al., 2013)Near-surface magnetic flux aggregation mechanism (K. Schatten 2009, Rempel, et al 2009)Slide25
Implications for the TSI – SSI reconstructions
Vanishing small spots = irradiance excess ?Less sunspots (B < 1500 G spot formation threshold)Lower sunspot blocking (visible + IR)Additional contribution to
plage
and network
Excess in near-UV
,
microwaves (and
solar wind ?)Possible cause of the divergence between sunspot indices and solar proxies
21/11/2014
ESWW11 – Nov. 2014
25
Can this
“reversal”
process reduce
the effective irradiance decrease expected at low solar activity?
(Grand Minima?)
Base level in solar flux close to the last SC23-24 minimum
(
Schrijver
et al. 2011)
Proxies cannot be based on a simple linear extrapolation of recent high solar cycles (scale-dependant, lifetimes)Slide26
Conclusions
Multiple evidence of a global change in small-scale sunspot magnetic fieldsStill unclear if the current change is:A steady evolution towards a new activity regimeA larger deviation in a global solar cycle modulationLong-
term
variations of the
number
of spots per group
Results supported by
In-depth validation DPD sunspot catalog versus the MDI/STARA
catalog
Recalibrated SN and GN series
Limited
rise
of
average
solar
activity since the Maunder minimumConcept of a Grand Maximum in the 20
th century is questioned
Release of new the SN and
GN series by mid-201521/11/2014
ESWW11 – Nov. 201426Slide27
More information available at …
21/11/2014ESWW11 – Nov. 2014
27
http://sidc.be/silso
http://
ssnworkshop.wikia.com/wiki/Home
http://haso.unex.es/
Sunspot
Number
Workshops
Historical
Archive of
Sunspot
Observations
WDC –
SILSO
Sunspot
Index
and
Long-term Solar ObservationsSlide28
21/11/2014
ESWW11 – Nov. 201428Slide29
Long-term integration
: a moderate Modern Maximum ?Time-integrated responses to the solar
input:
Cosmogenic
isotopes:
deposition
processes
(ice, sediments)
Earth
climate
: thermal
inertia
of
oceansGaussian running mean over 22
years (2 solar cycles):21/11/2014
ESWW11 – Nov. 201429
Original series
Ratio Max cycles 3-19 = 1.27
Ratio 22-yr envelope = 1.30
Corrected series
Ratio Max cycles 3-19 = 1.08
Ratio 22-yr envelope = 1.17
The
clustering
of high
solar
cycles
reduces
the time-
integrated
effects
of the corrections by 50%Slide30
Sunspot areas versus F10.7
Same F10.7 excess versus SOON sunspot areas in the late part of cycle 23 (Hathaway 2010, 2013)
21/11/2014
ESWW11 – Nov. 2014
30
F10.7 vs SSN
F10.7 vs
sunspot
areaSlide31
Cycle 23 small
-spot deficit
De
Toma
et al., 2013
Contradictory results from
de
Toma
et al. (2013)
:
Based on San Fernando Obs. Images
Small spots: no decline
Large spots: decrease in cycle 23
but
Spatial resolution
of
the CFDT1 instrument is
too low for analysis of the smallest spots (>5”/pixel).Smallest spots <30
msh are not properly detected.21/11/2014
ESWW11 – Nov. 201431Slide32
DPD: an example
A very detailed dataset
NOAA 7815
Common penumbra
Master Spot corresponding to this common penumbra
Penumbra with no umbra (U=0)
Not very contrasted
21/11/2014
ESWW11 – Nov. 2014
32Slide33
Area differences
The main difference between MDI images and DPD images is due to the lower resolution of MDI
images
Area
difference of two circles with radius r and
r+dr
is
dA
/A
≈
2dr/r
Bias appears when A is close or below the pixels size.
Gyori
et al., 2004
(
STARA- DPD
)/DPD
50000
points
(SOHO-DPD)/DPD 100 points
Confirmed by this study:
much
larger statistics (50000 spots instead of < 100
)Slide34
Sunspot number versus other solar indices and fluxes
Very high correlation with photospheric parameters (R2 >0.95): RG, RA
,
R
Boulder
, Area,
Mx
(Bachmann et al. 2004, Rybansky et al. 2005, Wilson and Hathaway 2006, Tapping et al. 2007,
Bertello
et al. 2010,
Hempelmann
and Weber 2012)
Measure of the global emergence rate of (
toroidal
) magnetic flux (Petrovay 2010,
Stenflo 2012)Chromospheric and mixed indices (TSI
, CaII-K, MgII):Good but lower correlations:Non-linear relationTime lags (magn
. flux dispersion)Different physics !21/11/2014
ESWW11 – Nov. 201434
Mean
mag. flux
Sunspot Nb.Stenflo, 2012Solanki
&
Fligge
1999Slide35
Sunspot number versus other solar indices and fluxes
Blind Source Separation applied solar radio and UV indices:Method: Bayesian positive source separation (Moussaoui et al. 2006)21/11/2014
ESWW11 – Nov. 2014
35
3 clusters of indices, each dominated by one of 3 sources :
Photosphere (SN, GN)
Chromospheric
(DSA,
MgII
, Ly
α
, radio
λ
> 10cm)
Coronal (radio
λ
< 10cm)
T.
Dudok
de Wit, SSN2 Workshop, 2012
Corona
Chromosphere
Chromosphere
Photosphere
Network,
plages
Radio
gyroresonance
Active regionsSlide36
Cycle 24 in the long-
term SN perspectiveCycle 24 is among the late cycles
Tie-point (
R
i
=13) at the end of preceding cycle
2 main cycle “families”:
Steep – strong
(max > 130)
Slow –
weak
(max < 80)
Best
fit
with
cycles 12, 14, 15, 16 Tie-point (Ri=40) in the rising phase of the cyclesReturn to an average activity regime
like in the late 19th and early 20th century
21/11/2014ESWW11 – Nov. 2014
36Slide37
Cycle
24 best matches:Cycle 14 [1902-1913]: Rmax = 64.2
Cycle 15 [1913-1923]:
R
max
= 105.4
!
3 features of moderate cycles:Plateau (duration up to 3 years)Multiple sharp peaks
Late maximum
(~highest random peak)
Cycle decline by mid-2015
Next minimum in 2019-2020
21/11/2014
ESWW11 – Nov. 2014
Cycle 15
37
Cycle 14
Cycle 24 in the long-
term
SN perspective