How Extreme Can Solar Events Be ? 1st Varsiti Meeting June 7, 2016 Bulgaria　9:30-10:00 invited How Extreme Can Solar Events Be ? Kazunari Shibata Kwasan and Hida Observatories, Kyoto University, Kyoto, Japan Collaborators : Hiroyuki Maehara, Takuya Shibayama, Yuta Notsu, Shota Notsu, Satoshi Honda, Daisaku Nogami, Takuya Takahashi, Hiroaki Isobe

Carrington flare (1859, Sep 1, am 11:18 ） 　The first flare that human beings observed by Richard Carrington (England) 　white flare for 5 minutes 　very bright aurora appeared next day morning at many places on Earth, e.g. Cuba, the Bahamas, Jamaica, El Salvador, and Hawaii. Largest magnetic storm (> 1000 nT) in recent 200 yrs. Telegraph systems all over Europe and North America failed. Telegraph pylons threw sparks and telegraph paper spontaneously caught Fire　（Loomis　1861） http://en.wikipedia.org/wiki/Solar_storm_of_1859

Will the Carrington-class flare occur again ? Can much bigger flares, superflares (>10^33 erg), occur on the Sun at present ? If yes, what is frequency of superflares ? Why and how can superflares occur on the Sun ? To answer these questions is the subject of my talk.

？ statistics of occurrence frequency of solar flares, microflares, nanoflares nanoflare microflare 1000 in 1 year 100 in 1 year 10 in 1 year 1 in 1 year 1 in 10 year 1 in 100 year 1 in 1000 year 1 in 10000 year solar flare ？ Largest solar flare superflare superflares C M X X10 X100 C M X X10 X1000 X100000

How can we observe superflares on the Sun ? If empirical statistics rule of solar flares is applied to much larger flares (superflares), then the frequency of superflares with energy 1000 times larger than the largest solar flares might occur once in 10000 years. However, the period of modern observations of the Sun with telescope is only 400 years. How can we observe the Sun for 10000 years ? If we observe 10000 solar type stars (similar to our Sun) for 1 year, we can get the data similar to the data obtained from 10000 years observtions of the Sun ! Prof Sekiguchi kindly told me that the Kepler satellite is taking such data !

Kepler satellite (NASA) Space mission to detect exoplanets by observing transit of exoplanets 0.95 m telescope Observing 160,000 stars continuously (from 2009 to 2013). Among them, 80000 are solar type stars. ~30 min time cadence (public data)

Superflares on Solar Type Stars : Our study (Maehara et al. 2012) Hence we searched for superflares on solar type stars using Kepler satellite data, which include data of 80000 solar type stars Since the data are so large, we asked 1st year undergraduate students to help analyzing these stars, 　　because students have a lot of free time (2010 fall) Surprisingly, we (they) found 365 superflares on 148 solar type stars (G-type main sequence stars)

Superflares on Solar type stars Published in Nature (2012, May) Superflares on Solar type stars Undergraduate students H. Maehara, T. Shibayama, S. Notsu, Y. Notsu, T. Nagao, S. Kusaba, S. Honda, D. Nogami, K. Shibata  

typical superflare observed by Kepler Brightness of a star and a flare Total energy ~ 10^35 erg Time (day) Maehara et al. (2012)

typical superflare observed by Kepler What is the cause of stellar brightness variation ? typical superflare observed by Kepler Brightness of a star and a flare Total energy ~ 10^36 erg It is likely due to rotation of a star with a big star spot Time (day) Maehara et al. (2012)

time Model calculation of stellar brightness variation KIC6034120 model(green) inclination = 45° Starspot radius 0.16 R* Stellar brightness 2％(平均基準) 5 days time Notsu et al. 2013

time Model calculation of stellar brightness variation KIC6034120 model(green) inclination = 45° Starspot radius 0.16 R* Stellar brightness 2％(平均基準) 5 days time Notsu et al. 2013

Hudson argued that Light Curve of superflare stars is different from that of the Sun. Is this True ? Solanki Solar Physics (2004) 224: 197–208 60 days (Aug– Sep 2000) Superflare Star days

Flare energy vs rotational period Solar Rotation Fast rotation (young) Slow rotation (old) Stars with period longer than 10 days cf solar rot period ~ 25days There is no hot Jupiter in these superflare stars against previous prediction (Schaefer+ 2000) 25days Maehara+(2012),　Notsu+ (2013)

Flare energy vs sunspot area Superflares on solar type stars Once in 1000 years 10^35 erg X10000 Once in 100 years 10^34 erg X1000 Once in 10 years 10^33 erg X100 Once in 1 year 10^32 erg X10 0.01 su Sunspot area (in unit of solar surface area) 10 in 1 year 10^31 erg 　 X 100 in 1 year 10^30 erg M Solar flare 1000 in 1 year 10^29 erg 　 C Sammis et al. 2000 0.0001 0.001

Flare energy vs sunspot area Superflares on solar type stars Once in 1000 years 10^35 erg X10000 Once in 100 years 10^34 erg X1000 Once in 10 years 10^33 erg X100 Once in 1 year 10^32 erg X10 0.01 Solar flares su Sunspot area (in unit of solar Surface area) 10 in 1 year 10^31 erg 　 X 100 in 1 year 10^30 erg M Solar flare 1000 in 1 year 10^29 erg 　 C Sammis et al. 2000 0.0001 0.001

Flare energy vs sunspot area (magnetic flux) ? Stellar flares Solar flares Shibata et al. (2013)

Indirect Measurement of Rotational Period and Spot Area are True Indirect Measurement of Rotational Period and Spot Area are True ? => Spectroscopic Observations of Superflare Stars Notsu et al. (2015) PASJ

Projected rotation velocity (v sin i) We can estimate projected rotation velocity (v sin i) from the Doppler broadening of absorption lines. Fast rotators ⇒ wide line profile Slow rotators ⇒ narrow line profile Line of sight Rotation Axis We explain this in a bit detail in this slide. We can estimate projected rotation velocity (v sin i) from the broadening of absorption lines in our spectroscopic data. This figure is an example of our spectroscopic data of superflare stars. As shown here, fast rotating stars have wide line profiles, while slowly rotaing stars like the Sun have narrow line profiles. We estimated projected rotation velocity (v sin i) of the target superflare stars in this way, but we must note here the velocity estimated here is only “projected” rotation velocity since we cannot know the inclination angle between rotaion axis and line of sight. ※Measurement methods Takeda et al.(2008etc)

Rotation Period ⇔Brightness variation period ? Most of the data points locate below the line of i=90° ⇒“Brightness variation≒Rotation” is OK!! Edge-on view(sin i =1) Velocity estimated from brightness variation period (Vlc[km s-1]) v sin i [km s-1] Pole-on view Line of Sight Rotation Axis Then, in this figure, we compared the projected rotation velocity with the brightness variation of Kepler data. This figure is a bit complex, so here I try to explain in detail. The vertical axis of this figure is stellar projected rotation velocity (v sin i), estimated from the line broadening of our spectroscopic data. The horizontal axis is velocity estimated from the period of the brightness variation by using this equation. If the brightness variation completely corresponds to the rotation, the values of horizontal axis and vertical axis become equal. But, some data points in this figure have differences between these two values. We consider such differences are explained by the inclination angle effect, in other words, the effect of sin I in the vertical axis. In this figure we plot four lines corresponding to other inclination angle values.   As a result, this figure shows two important results. First, most of the data points locate below the line of inclination angle I is 90 degree. In other words, for almost all the stars, v sin i in the vertical axis is less than the velocity in the horizontal axis. This is consistent with our assumption that the brightness variation is caused by the rotation since the inclination angle effect, in other words, the effect of sin I in the vertical axis can cause the data points located below .the line of i = 90 degree..  In addition, stars distributed in the lower right side of this figure are expected to have small inclination angles such as I (inclination)=10 degree, and to be nearly pole-on stars. As a result of this figure, the rotation velocity from spectroscopic data is consistent with the rotation velocity from the brightness variation period. This fact supports the assumption that the brightness variation of superflare stars is caused by the rotation. pole-on view ※Sun: Vrot～2 [km s-1]

Flare energy vs. area of starspots Spectroscopic rotational velocity　(Subaru obs) Stellar flares low inclination angle Photometric rotational velocity Solar flares Notsu, Y. et al. (2015)

(BigBear Solar Observatory data) Strong magnetic field area around starspots show strong Ca II emission!! This figure in the left side is an example of photometric image of the Sun, and this figure in the right side is a map of Ca II line intensity of the same date. This figure shows strong magnetic field area around starspots show strong Ca II emission and reflects Ca II line is a well-known indicator of stellar magnetic activity of solar-type stars.   Then we can indirectly investigate the existence of large starspots by using the intensity of Ca II lines. The Sun with Ca II K line The Sun with visible light (BigBear Solar Observatory data) We can indirectly investigate the existence of large starspots by using the intensity of Ca II lines.

Indirect estimation of starspot coverage with Ca II lines As the magnetic activity enhanced, the core depth become shallow because of the greater amount of the emission from the chromosphere. Chromospheric activity ⇒These stars have large starspots ! The core depth becomes shallow. Large starspots Superflare stars Superflare stars Superflare stars 18Sco (Solar-twin) This figure is an example of Ca II 8542 lines of superflare stars and the Sun. As the magnetic activity enhanced, the core depth become shallow because of the greater amount of the emission from the chromosphere. The core depth of Ca II lines of superflare stars shown here is shallower than that of the Sun, and as a result, these stars have large starspots compared to the Sun.

Starspot coverage vs Ca II 8542 intensity ↑ Sun r0 (8542) r0: Normalized intensity of Ca II 8542 Next, in this figure, we compared Ca II intensity with the brightness variation amplitude of Kepler data.   The vertical axis of this figure is normalized intensity of the line center of Ca II 8542 line. The upper side value correspond to the existence of large starspots on the basis of spectroscopic data. The horizontal axis is the amplitude of the brightness variation of Kepler data. We can see a rough correlation in this figure between Ca II intensity and the amplitude of brightness variation. If we assume that the brightness variation is caused by the rotation, the brightness variation amplitude in the horizontal axis corresponds to the starspot coverage of these stars. Then, we can say that there is a rough positive correlation between the starspot coverage on the basis of Kepler data and Ca II intensity in the vertical axis. (下部表示) This rough correlation shows us all the target stars that are expected to have large starspots on the basis of their large amplitude of the brightness variation in the horizontal axis also show the existence of large starspots on the basis of spectroscopic data in the vertical axis. In other words, as for the viewpoint of starspot sizes, results from Kepler data are consistent with the spectroscopic results. Brightness Variation Amplitude ≒ starspot coverage All the stars that are expected to have large starspots (from large brightness variation amplitude) show high (Ca II) magnetic activities.

Superflare stars have large spots Hypothetical image In visible light (photosphere) Hypothetical image In CaII line (chromosphere)

Comparison of statistics between solar flares/microflares and superflares Shibata et al. 2013 nanoflare microflare solar flare ？ Largest solar flare superflare C M X X10 X100

Comparison of statistics between solar flares/microflares and superflares X10000 flare Shibata et al. 2013 Superflares of 1000 times more Energetic than the largest solar flares occur once in 5000 years ! nanoflare microflare 1000 in 1 year 100 in 1 year 10 in 1 year 1 in 1 year 1 in 10 year 1 in 100 year 1 in 1000 year 1 in 10000 year solar flare Largest solar flare superflare C M X X10 X100 C M X X10 X1000 X100000

Comparison of statistics between solar flares/microflares and superflares X1000 flare Shibata et al. 2013 Superflares of 100 times more Energetic than the largest solar flares occur once in 800 years ! nanoflare microflare 1000 in 1 year 100 in 1 year 10 in 1 year 1 in 1 year 1 in 10 year 1 in 100 year 1 in 1000 year 1 in 10000 year solar flare Largest solar flare superflare C M X X10 X100 C M X X10 X1000 X100000

Comparison of statistics between solar flares/microflares and superflares X100 flare Maehara+ 2015 Superflares of 10 times more Energetic than the largest solar flares occur once in 500 years ! nanoflare microflare 1000 in 1 year 100 in 1 year 10 in 1 year 1 in 1 year 1 in 10 year 1 in 100 year 1 in 1000 year 1 in 10000 year solar flare Largest solar flare superflare Maehara et al. 2015 C M X X10 X100 C M X X10 X1000 X100000

Fundamental Question Why and how can superflares occur on Sun-like stars (i.e., present Sun) ? Superflares occur because of the presence of large spots. => Why and how can large spots be generated on Sun-like stars (i.e., present Sun) ?

What is the spot area statistics ? Teff=5500-6000 K Many stars without superflares Show evidence of large spots ! What is the spot area statistics ? Teff=5500-6000 K Solar Rotation Amplitude (in unit of total area of a star) Spot Area Maehara+ (2016) Candelaresi+ (2014) Period (day)

Statistics of Spot Area on the Sun 0.001 0.01 0.1 Courtesy of Ishii (2015)

Statistics of Spot Area on the Sun and Sun-like Stars 0.001 0.01 0.1 Large spots exist in many Sun-like stars though frequency Is small Maehara+ (2016)

Necessary time to generate magnetic flux producing superflares (Shibata et al. 2013) Why and how can large spots be generated on the present Sun ? (Shibata et al. 2013) The necessary time to generate magnetic flux of 1024 Mx that can produce superflares of 1035 erg are 40 years　（<< 5000 years) (but > 11 years) only 8 years (< 11 years) to generate 2x1023 Mx producing superflares of 1034 erg => easily occur !? Is it possible to store such huge magnetic flux below the base of convection zone ? => big challenge to dynamo theorist !

Most active Sun-like star 4 superflares in 500 days ! Shibayama et al 2013 Stellar Brightness variation day KIC10471412 (rot period 15 days)

What is Solar/Stellar Cycle dependence of Flare frequency ? Shibayama et al 2013 1024 Mx 1025 Mx 1023 Mx nanoflare microflare Most active Sun-like stars solar flare Solar maximum 2001 Largest solar flare superflare Solar minimum 2008

Evidence of a superflare ? Corresponding to 10^34 erg superflare If this is due to a solar flare (Miyake et al. Nature , 2012, June, 486, 240)

Okayama 3.8m New Technology Telescope of Kyoto Univ Spectroscopic Observations of Solar type stars causing superflares will be extremely important Okayama 3.8m New Technology Telescope of Kyoto Univ (under construction) New Technology 1. Making Mirrors with Grinding 2. Segmented mirror 3. Ultra Light mounting High speed photometric and spectroscopic observation of Transient objects Gamma ray bursts Exoplanets Stellar flares (superflares) Budget for operation Is still lacking. Please support us ! Will be completed ~ 2017 courtesy of Prof. Nagata (Department of Astronomy , Kyoto University)

Summary Using Kepler data, we found 365 superflares (10^33-10^36 erg) on 148 solar type stars (among 80000 stars) during 120 days (Maehara+ 2012). => 1547 superflares from 279 solar type stars during 500 days (Shibayama+ 2013). Superflares occur on Sun-like stars (5600-6000K and slow rotation) with frequency such that superflares with energy 10^33-10^35 erg (X100-X10000 solar flare) occur once in 500-5000 years These stars have large star spot (Notsu+ 2013). Rotational velocity and large star spot of 50 superflare stars has been confirmed by spectroscopic observations (Notsu+ 2015) Hence we cannot reject the possibility that superflares of 10^33 – 10^35 erg would occur once in 500 – 5000 years on the present Sun (Shibata+ 2013, Nogami+ 2014, Maehara+ 2015) Evidence of superflares on the Sun 1000 years ago ? (Miyake+ 2012, 13)

Thank you for your attention