solar rotational modulations of ssi and correlations with...

Solar Rotational Modulations of SSI and Correlations with the Variability of TSI

1.  TSI (TIM, ACRIM III, VIRGO) 2.  SSI (SIM V19, SOLSTICE, TIMED/SEE, SATIRE-S)

Jae N. Lee1,2 , Robert F. Cahalan2,3, and Dong L. Wu2

1. Joint Center for Earth Systems Tech., University of Maryland, Baltimore County, MD 2. NASA Goddard Space Flight Center, Greenbelt, MD 3. Applied Physics Lab/Johns Hopkins University, Laurel, MD

Sun-‐Climate connec/on in solar rota/onal /me scale

Sun Climate Symposium Savannah, GA, 11/11/15 Lee, Cahalan, and Wu 2

where our data were continuous and almost completely clean,i.e., for a period of 220 days (7 months from 21 May 2007until 26 December 2007). Seven maxima and seven minimaare evident. The daily noise values yield a time series wherethe most dominant frequency is 1/27 day (Figure 3).[12] Because of the fact that these changes in the APD

occur only during daytime, approximately from sunrise tillsunset, we suspected that these were related to the Sun andto the solar rotation. To validate this we compared our datawith different solar indices data. However, after we gatheredall the solar indices (Solar Radiation and Climate Experi-ment, Solar spectral irradiance, 2007, available online http://lasp.colorado.edu/sorce/index.htm; NOAA National WeatherService, Space Weather Prediction Center, 2007, available at

http://www.swpc.noaa.gov/) (sunspot number, 10.7 cm radioflux, Lyman alpha flux) we found out that none of these solarindices, or any other measured quantity, revealed a contin-uous 27-day periodic fluctuation throughout this period.[13] Figure 4 represents our daily measured VLF APD

(2–5 kHz) together with three different solar indices, whichdo not always correlate well with each other. However,when we compare our measured index when all threeindices agree, we notice that there is a strong negativecorrelation between our VLF index and the solar ones. Thisimplies that during enhanced solar activity, i.e., with highervalues for the solar indices, we detect less atmosphericnoise, and vice versa when the solar indices are at minima.However, unlike all the solar indices presented, our ELF-

Figure 3. Power spectrum of the atmospheric signals between 2 and 5 kHz computed for 220 days ofdata. It is clear that the maximum value for the relative amplitude occur at the 1/27 frequency (vertical redline). Half-width values are ±2 days.

Figure 4. Daily VLF noise levels at 2–5 kHz versus different solar indices. When there is a goodagreement between all three solar indices (red, blue, and green), they all show negative correlation withour VLF index (black).

A10306 REUVENI AND PRICE: VLF AND THE 27-DAY SOLAR ROTATION

4 of 7

A10306

Africa�

Lightning,in,MC�

Pacific�

Cloud(varia5ons((OLR)�

America�

Atlan/c�

2003.8� 2004.7�

-‐  VLF lightening signal correlates with solar ac/vity only when the solar rota/onal signals are large. -‐  There are 27-‐day like signals in lightening ac/vity, but not always correlated with solar rota/onal signals.

Takahashi et al. (2014, SORCE mee/ng presenta/on) Reuveni and Price (2009)

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8TSI

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8 rotational variation 2003/2004: TSI vs Lyman-,

Lyman-,

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2TSI

-0.2

-0.1

0

0.1

0.2 rotational variation 2007: TSI vs Lyman-,

Lyman-,

Lightening

TSI and SSI

�  SSI is a distribu/on of TSI, SSI varia/on is wavelength dependent, and also /me scale dependent. Solar spectrum originates at different levels of the solar atmosphere.

�  With current satellite observa/ons, more than 100 solar rota/onal cycles of 27-‐day varia/on are covered (2003 – Aug. 2013).

�  Dis/nct solar rota/onal modula/ons of SSI at each wavelength can be iden/fied in terms of amplitude and phase.

�  The phase of the rota/onal mode obtained from SSI is compared with that of TSI.

Lee, Cahalan, and Wu 3 Sun Climate Symposium

Savannah, GA, 11/11/15

Sta/s/cal analysis : EEMD (Ensemble Empirical Mode Decomposi/on)

[Huang et al., 1998; Ruzmaikin et al., 2007; Barnhart and Eichinger, 2011]

Fourier Transform

EEMD Hilbert-‐Huang Transform with /me variable amplitudes and phases

Lee, Cahalan, and Wu 4 Sun Climate Symposium Savannah, GA, 11/11/15

Low frequency varia/on

High frequency varia/on

The 27-‐Day Rota/onal Varia/ons in TSI Observa/ons: from SORCE/TIM, ACRIMSAT/ACRIM III, and SOHO/VIRGO [Lee et al., 2015]

Year04 05 06 07 08 09 10 11 12 13 14 15

Tota

l Sol

ar Ir

radi

ance

(TSI

) at 1

AU

(W/m

2 )

1357

1358

1359

1360

1361

1362

1363SORCE Total Solar Irradiance - daily average

declining SC23

missing

rising SC24

minimum SC23

maximum



-‐  The rota/onal signals are well

captured with three independent space-‐borne TSI measurements.

-‐  They agree well with each other, with large amplitudes in declining and rising phases, but with small amplitudes in minimum phase.

-‐  How about SSI?

A few examples : How rota/onal varia/ons of SSI are related with that of TSI?

•  The modes from two observa/ons of Lyman-‐α line (SORCE/SOLSTICE and TIMED/SEE) agree well in amplitudes and phases.

•  The modes of Lyman-‐α line are not always in phase with TIM/TSI modes.

•  TSI varia/ons are affected by solar proton event (SPE), but UV varia/ons are not.

Lee, Cahalan, and Wu

6

ATSI (t) =ITSI (t)− ITSIσ (ITSI )

AUV (t) =IUV (t)− IUVσ (IUV )

−10

−5

0

5

10

−5

0

5

−5

0

52004

−5

0

5

−5

0

52005

−5

0

5

−5

0

52006

−5

0

5

−5

0

52007

−10

−5

0

5

10rotational variation : TSI vs UV (121nm)

2003

Rota/onal Varia/on : TSI vs Lyman-‐α (121 nm)

•  Standardized amplitudes are es/mated by removing the mean and by dividing with σ.

,

TIM/TSI TIMED/SEE SOLSTICE

-‐ Rota/onal varia/ons of TSI and Lyman-‐α are ohen out-‐phase during high solar ac/vity period, but are in-‐phase during solar minimum (2008-‐2010).

7

−5

0

5

−5

0

5

−5

0

5

−5

0

5

−5

0

5

−5

0

5

−5

0

5rotational variation : TSI vs UV

2008

−5

0

52009

−5

0

52010

−5

0

52011

−5

0

52012

−5

0

52013

Rota/onal Varia/on : TSI vs Lyman-‐α (121 nm)

Sun Climate Symposium Savannah, GA, 11/11/15 Lee, Cahalan, and Wu

TIM/TSI TIMED/SEE SOLSTICE

Similar but TSI with MUV (240nm) - They are ohen out-‐phase during high solar ac/vity period. -‐ Modes from SIM and SOLSTICE match well.

−10

−5

0

5

10

−10

−5

0

5


2003

−5

0

5

−5

0

52004

−5

0

5

−5

0

52005

−5

0

5

−5

0

52006

−5

0

5

−5

0

52007

8

Rota/onal Varia/on : TSI vs MUV (240 nm)


TIM/TSI SIM SOLSTICE

-‐ TSI and MUV are in phase during minimum phase, but ohen out-‐phase during solar ac/vity rising period.

−5

0

5

−5

0


2008

−5

0

5

−5

0

52009

−5

0

5

−5

0

52010

−5

0

5

−5

0

52011

−5

0

5

−5

0

52012

−5

0

5

−5

0

52013

9

Rota/onal Varia/on : TSI vs MUV (240 nm)

Sun Climate Symposium Savannah, GA, 11/11/15


TIM/TSI SIM SOLSTICE

- TSI and VIS are in-‐phase during whole analysis period (2003 – 2011). -‐ Only 5 outliers from SIM are excluded for the EEMD analysis to remove the outstanding cycles appeared in purple doied curve. -‐ Outliers from SIM are chosen from high frequency mode comparison with TSI.

−10

0

10

−10

0

10rotational variation : TSI vs Hα 656.3nm

2003

−10

0

10

−10

0

102004

−10

0

10

−10

0

102005

−10

0

10

−10

0

102006

−10

0

10

−10

0

102007

10

Rota/onal Varia/on : TSI vs H-‐α (656 nm)



W/m

2 /nm

TIM/TSI SIM unscreened SIM

Slope = 0.0011

TSI and SIM H-‐α (656 nm)

−5

0

5

−5

0

5rotational variation : TSI vs UV2008

−5

0

5

−5

0

52009

−5

0

5

−5

0

52010

−5

0

5

−5

0

52011


11

Rota/onal Varia/on : TSI vs. H-‐α (656 nm)


TIM/TSI SIM unscreened SIM

rotational mode : 2005

wavelength (nm)120 140 160 180 200 220 240

20

05

: d

ay

of

yea

r

50

100

150

200

250

300

350

400 600 800 1000 1200 1400 1600-3

-2

-1

0

1

2

3

-5 0 5

50

100

150

200

250

300

350

Phase of the solar rota/onal modes : 2005



mode of TSI: 2005 mode of SSI: 2005

days : 2005

days : 2005

-‐  Rota/onal varia/ons are well defined in UV and 600-‐ 900 nm, but not at photospheric wavelength (250 – 600 nm) and IR above 900 nm.

SOLSTICE SIM

-‐  Correla/ons begin to grow at 250 nm, correla/ons from SOLSTICE, SIM and SATIRE track each other.

-‐  Correla/ons from SIM is >0.8 in 600 – 900 nm.

-‐  Correla/ons from SATIRE are nearly 1 above 400 nm, but shih at opaque minimum at 1650 nm.

13

wavelength (nm)400 600 800 1000 1200 1400 1600 1800 2000 2200 2400

r

-0.2

0

0.2

0.4

0.6

0.8

1Correlation of the 27-day mode between SSI and TSI (400-2400 nm)

SIMSATIRE

wavelength (nm)100 150 200 250 300 350 400

r

-0.2

0

0.2

0.4

0.6

0.8

1Correlation of the 27-day mode between SSI and TSI

SIM

SOLSTICE BSOLSTICE A

SATIRE


Linear correla/ons of the 27-‐day mode between TSI and SSI

SATIRE-S

SIM

SATIRE-S

SOL-A SOL-B SIM

Summary

�  Solar rotational modulations of spectral irradiance are wavelength dependent.

�  Rotational variations of VIS and NIR (~600-900 nm) are in-phase with that of TSI, but variations of FUV and MUV are not always in-phase.

�  Close agreement of the rotational variations from SIM and SOLSTICE is encouraging. We can address rotational variations from SORCE SSI measurements with high confidence.

�  Rotational variations from SATIRE-S matches with observed SSI modes below 280 nm, and exactly matches with TSI mode above 400nm.


Lee, Cahalan, and Wu 14

Acknowledgment

�  SORCE Team

�  TIMED and SATIRE Teams

�  Tom Woods, Jerry Harder, and Marty Snow


Back ups



Low frequency and High frequency varia/ons of TSI and Lyman-‐α

01/04 01/05 01/06 01/07 01/08 01/09 01/10 01/11 01/12 01/131360

1360.2

1360.4

1360.6

1360.8

1361

1361.2

1361.4TIM/TSI

W/m

2

#10-35

5.5

6

6.5

7

7.5

8(a) low frequency variation : TSI vs Lyman-,

SOLSTICE/Lyman-,

01/04 01/05 01/06 01/07 01/08 01/09 01/10 01/11 01/12 01/13-4

-3

-2

-1

0

1

2

3

4TIM / TSI

W/m

2

#10-3-1.5

-1

-0.5

0

0.5

1

1.5(b) high frequency variation : TSI vs Lyman-,

SOLSTICE/Lyman-,

W/m

2

W/m

2

•  The phase of low and high frequency varia/ons of TSI and Lyman-‐α line are compared.

•  Amplitudes of high frequency varia/on

•  Low frequency varia/ons are in-‐phase with each other.

•  High frequency varia/ons?

Low frequency varia/on

HIgh frequency varia/on

Distribu/on of low and high frequency modes


•  TSI HF mode is not normally distributed.

•  SIM SSI difference between high and low TSI regime has a dip at 1600nm.

W/m2-4 -3 -2 -1 0 1

frequ

ency

0

50

100

150

200

250

300

350 TSI HF mode : Histogram and fitted Normal curve

W/m2-4 -3 -2 -1 0 1

frequency

0

20

40

60

80

100

120

140

160

180 TSI HF mode : Histogram in high and low regime


W/m2/nm #10-3-3 -2 -1 0 1 2 3

frequency

0

5

10

15

20

25

30

35

40

45Lyman , HF mode : Histogram

( ) ( )OxIOxII −≈−= 1exp)( 00λ1. Relation between irradiance I and Opacity à here O is opacity and x is distance times density. Note O depends on the wave length lambda

2. Assume there are two main source of absorption such as O1: H- and O2:H2 à Then

( )22110 1)( xOxOII −−=λ

Eric Richard’s TSI vs SSI



3.  Let us see if the following fig makes sense

à the y axis is defined as

( ) ( )( )

( ) ( )( )

1

111

)(

2211

2,221,11

22110

2211022110

ref

refref

ref

ref

ref

ref

xOxOxxOxxO

xOxOIxOxOIxOxOI

III

d

−−

−+−=

−−

−−−−−=

−=λ



4. We can compare two wave length 1000 and 1600.

( ) ( )( )

)1000()1000(1)1000()1000(

)1000(2211

2,221,11

ref

refref

xOxOxxOxxO

d−−

−+−=

( ) ( )( )

)1600()1600(1)1600()1600(

)1600(2211

2,221,11

ref

refref

xOxOxxOxxO

d−−

−+−=

5. Note in general lager fluctuation for lambda 1000 than 16000 because

( ) ( )refref xOxOxOxO 22112211 )1600()1600(11

)1000()1000(11

−−>

−−

6. Also, depending on relative abundance (density ) of H- and H2, that is x1 and x2, the sign of d(1000) and d(1600) can be different. That is while one grows , the other could decrease.



SSI composite: TSI(high) – TSI(low)



SSI composite: TSI(asc) – TSI(des)


solar rotational modulations of ssi and correlations with...

Documents