Coronal Mass Ejections and Geomagnetic Storms

Transcription

1 Coronal Mass Ejections and Geomagnetic Storms Marilena Mierla 1,2 1. Royal Observatory of Belgium, Brussels, Belgium 2. Institute of Geodynamics of Romanian Academy, Bucharest, Romania

2

3 with Zi = b0 + b1 xi bj xij (1) CMEs Big gap 1 AU ε = 107 V B2 l02 sin4 (θ/2) [J/s] (2) gapthis 0.01 AUas W Geomagnetic = εdt [J], where we calculated energy ICMEs storms storm mai to t is time interval of geomagnetic tf ε t0 ability of occurrence of intense 0 given by i-th observation ofzuccarello et al phase. The variables in ε formula are V sola our case we used 25 observations wind velocity [m/s], B intensity of th V; AW; mass gnetic storms), bj are model V interplanetary magnetic field [T], l0 a constant equa e derived (known as regression to 7 Earth radii (RE) 11/2 : [m], and θ angle between th egeosphere, are ; values in our case Np two components of interplanetary magnetic field independent variables listed By and Bz. The factor l0 was empirically determined an planetary magnetic field, and one dependent By applying regression model expl to I and j=0 to J, I being 25 and J represents effective cross-sectional area,el) Dst index. The choice of variables is we obtained a 100% probabilitydst that April are independent variables (Akasofu, of have solartriggered wind-magnetospher Tpet al. by The regression previous coefficients works. Caneare (2000) 1981) CME should an intense g ve. interaction and was added to fit energy input t hatiterative southward field component storm. θ an method.magnetic Z is estimated total output. authors argue Fig. 25. COR1 images (upper left panel), COR2 images (upper right panel), LASCO-C2 (lower left panel),some HI1 (lower middle panel), and HI2-tha ded in odds ICME is most important factor in estimated hm of of occurrence Energy transfer from w A (lower right panel) recorded on April 3, 2010, 09:50 UT this (COR1), 11:54 UTis3.2 (COR2), 11:06 UT (LASCO-C2), 16:49 UT (HI1), andsolar April factor rar low (Lu et al., 1998; Knipp et5, al ationship between CMEs and geomagnetic 00:09 UT (HI2-A). agnetic storm. phi Koskinen magnetosphere 1998; and Tanskanen, 2002; Østgaard an From Sun to Wu and Lundstedt (1997) stated that two basic e regression coefficients and from Using parameter introduced by th A Tanskanen, 2003; Tanskanen et al., 2002) because ations of solar wind parameters would give y Earth of occurrence CMEs may: of intense scaling factor we (lcomputed energy transferred computed assuming that thfr 0) was Transfer of energy from e predictions of geomagnetic storms: [B, n, V] s B set of 25 ICMEs which produced into estimated magnetosphere during g energy inputwind equals energy dissipatio -V]Deflect n, (B S = lbzl for Bz < 0 and BS = 0 for Bz > 0). tic storms (Dst < -150 nt) in solar ICME to magnetosphere: storm of and April 11, ( joule heating auroral precipitation in th ameters include se considerations. Rotate trained logistic model with 21 From parameter ionosphere) and ring current dissipation. - Akasofu Akasofu (1983) function: In th ormula introduced by Srivastava (2005) is e remaining four for validation. The β - 1Get deformed study, we chose εinitial value of 4 l(θ/2) 0=7R[J/s] E of th = 107 V B2 l02 sin d in this model is shown in Table 1. = with Zi = b0 + b1 xi bparameter. j xij (1) tf Zi d1in validation are shown in bold. +e W = calculated this by energy aset al. t 0 εdt ε (2014) Using we function obtained Wang - Wang et al. 2014: when ofcme was first of intense probability occurrence i istime Manchester et al t0 is 1.47 interval geomagnet EIN = 3.78of 10to nsw0.24 VSWtime BT0.86 (sin2.7of(θ/2) ) [J/s] (3 CO taken i-thcme gnetic(asstorm givenfrom by observation phase. The variables in ε formula are n 2 and 3: start and end of

4 Filling gap?

5 Rouillard 2011 Filling gap?

6 One needs to know: CME propagation 1) CME characteristics (area, mass, speed, direction of propagation, magnetic configuration..) 2) properties of medium through which CMEs propagate (density, speed, magnetic configuration) 3) interaction between CME and medium (physical processes: deflection, rotation, deformation, reconnection, erosion see e.g. Wang et al. 2004, Dasso et al. 2006, Lynch et al. 2009, Lugaz et al. 2011, Manchester et al. 2014)

7 CME propagation One can use: Observations near Sun (SOHO, STEREO, PROBA2, SDO) interplanetary space (STEREO/HI) in situ (ACE, WIND, DISCOVR, SOHO, STEREO) Models empirical: drag based models (Cargill+ 1996, Vrsnak ) constant or cessation of acceleration before 1 AU (Gopalswamy et al. 2000, 2001) MHD: ENLIL (Odstrcil 2003) EUHFORIA (Pomoell et al. 2017)

8 CME propagation The Astrophysical Journal, 743:101 (12pp), 2011 December 20 Solar wind background: Wang-Sheeley-Arge + ENLIL Forces acting on CME: - propelling Lorentz force - drag force. Temmer et al Figure 10. Same as Figure 9 but for WSA+ENLIL.

9 CME propagation ( ) Ø CME 3D speeds give G2000 proj. V G2000 3D V 12-h aver. Upstream SW slightly better predictions than projected CME speeds Ø The observed CME transit times from Sun to G2001 proj. V G2001 3D V LE ICME V 1 AU show a particularly good correlation with upstream solar-wind speed. Kilpua, Mierla et al G2000 Gopalswamy et al G2001 Gopalswamy et al. 2001

10 CME propagation deep minimum Ø wide-angle view point from STEREO is crucial to detect solar counterparts for weak ICMEs Ø narrow CMEs (angular widths 20 ) can arrive at Earth and an unstructured CME may result in a flux rope-type ICME. Ø Ten out of 16 (63 %) of associated CMEs were stealth CMEs. Kilpua, Mierla et al. 2014

11 Geomagnetic storms - major storms are produced by CMEs (Goplaswamy et al. 2007, Zhang et al. 2007) - Geomagnetic storms are highly correlated with Bz, V!Bz, CMEs speeds as well as with ram pressure (Gonzalez et al. 1994, Srivastava and Venkatakrishnan, 2004, Gopalswamy et al. 2008, Echer et al. 2013). - The major storms and superstorms meet GT criteria (Echer et al. 2005, Gonzalez and Tsurutani 1987). Zhang et al (GT criteria: a necessary IP condition for an intense geomagnetic storm to occur is presence of an intense m.f. (Bs > 10 nt) for long durations (t > 3 h) of time.)

12 Geomagnetic storms Period1: Period2: sunspot number Period 1 Period 2 moderate Dst - weak southward IMF and lack of strong ICMEs led to weak Dst activity in Period2. - Low solar wind densities may have furr weakened ring current response and solar wind magnetosphere coupling efficiency. - No difference in solar wind speed between 2 periods hours moderate AE moderate E Y intense Dst intense AE intense E Y Kilpua et al F

13 Energy transfer - Coupling functions Dst(nT) Dst Bz Bz (nt) Dst(nT) Dst Kp Kp(x10) Time (hours) Time (hours) 20 0 x x Dst(nT) Dst Akasofu Akasofu (J/s) Kp(x10) Kp Akasofu Akasofu (J/s) Time (hours) Time (hours) 0 Oprea, Mierla et al CMEs in SC23, Dst < -150 nt r(bz, Dst) -2 = 0.76 r(bs V, Dst) -2 = r(v, Dst) -3 = r(ρ, Dst) -1 =

14 Energy transfer - Coupling functions Ein ε PC Ey Dst Geomagnetic storm on April 11, 2001 Besliu-Ionescu, Mierla and Maris, 2016

15 Summary Still needed: To understand CME propagation into IP space: Ø To improve background solar wind Ø To understand interaction between CMEs and SW Ø To improve forecast of CME arrival at spacecraft and forecast of Bz To understand coupling between ICME and magnetosphere

16 Or recent studies not shown in this talk CME propagation: - Using STEREO/HI and/or drag approaches: Vrsnak+ 2010, 2013, Maloney+ 2010, Moestl+ 2014, Colaninno+ 2013, Mishra+ 2013, Lugaz+ 2013, Hess+ 2014, Wang+ 2014, Temmer+ 2015, Zic+ 2015, Shi+ 2015, Zhao+ 2016, Rollett+ 2016, etc. - MHD: Lugaz+ 2011, Vrsnak+ 2014, Isavnin+ 2014, Webb+ 2014, Manchester+ 2014, Shen+ 2014, Pizzo+ 2015, Wang+ 2016, etc. Geomagnetic storms: - extreme geomagnetic storms: Cid+ 2014, 2015, etc. - SC23 storms: Andriyas+ 2017, Hema+ 2017, etc. - SC24 storms: Kataoka+ 2015, Wu+ 2016, Bisht+ 2017, Selvakumaran+ 2016, etc. - prediction: Kataoka+ 2016, Kubicka+ 2016, etc.