Forecasting Super CME Disturbances 1 2 3 Super

Forecasting Super CME Disturbances 1. 2. 3. Super CMEs, such as the 2000 July 14, 2003 October 28, 2003 October 29, and 2006 December 13 full halo CMEs, generate strongest interplanetary and geomagnetospheric disturbances, that significant affect human life and modern technological systems. Forecasting super CME disturbances is the major task of space weather research. Super CMEs usually have very high propagation speed, and arrive at Earth in less than 24 hours. The extreme fast CMEs produce forward shock, compressed magnetic cloud (MC), and probably shock sheath Bs events due to the interaction of the extreme fast CME with the ambient solar wind and interplanetary magnetic field. Forecasting super CME disturbances is a challenge to space weather research, both in forecasting the arrival time and in forecasting the intensity of the ICME disturbances. The standard deviations of arrival times forecasted using Schween et al [2005], Golpawamy et al [2001] or Fry et al [2003] algorithms all are greater than 12 hours which is certainly needed to be improved. The intensity of ICME bodies-induced disturbances forecasted using Zhao and Hoeksema algorithm is valid only for slow CMEs [Schween, 2005]

4. 5. 6. We have improved the simulation of ICME propagation from near heliospheric base to the Earth by combining the time-dependent 3 -D MHD model with IPS solar wind observations and the circular cone model fit [Hayashi et al. , 2006]. We have developed the elliptic cone model to more accurately determine geometrical and kinematical parameters for halo CMEs [Zhao, 2007 a; 2007 b]. The elliptic cone model fit to the 2006 December 13 halo CME has been inputted to WSA/ENLIL model and successfully reproduced the ICME plasma structure at L 1 point [Owen, private communication, 2007]. Recent studies show that in addition to the filament orientation (SFD) [Marubashi, 1986; Bothmer and Schween, 1994; Zhao and Hoeksema, 1997] and the local inclination of HCS [Crooker, 1993; Zhao and Hoeksema, 1996], the orientation of magnetic clouds (MCs) is also associated with other manifestations of CMEs at different heliocentric distances, such as the EIT post-eruption arcades (EPA) [Zhao, 2007 c; Yurchyshyn, 2007] and the major axis of elliptic halo CMEs (MAH) [Yurchyshyn, 2007]. We plan to improve the existing empirical models summarized by Siscoe and Schween [2006] forecasting the arrival time and the polarity and magnitude of the IMF Bz generated by super CME disturbances in next 12 -24 hours. We also attempt to develop a physics-based model forecasting the arrival time and the four ICME parameters that determine the interplanetary electric field and the dynamic pressure.

1. Improvement of empirical models of CME-disturbance arrival time and intensity 1. 1 Arrival time 1. 1. 1 The existing algorithms are based on the expansion speed, Vexp, of halo CMEs [Schween et al. , 2005] or the plane-of-sky speed, Vps, of halo CMEs [Golpaswamy et al. , 2001], then use a statistical relationship between Vexp (Vps) and the radial speed, Vrad. Since cone-like CME configuration is not axial symmetric, and Vps includes effects of both expansion and projection, such statistical relationship should be questioned. 1. 1. 2 Our elliptic cone model can be used to more accurately invert the CME Vrad from white-light halo CME images than Vexp and Vps. By combining this Vrad with the prediction of solar wind speed, Vsw, from PFSS model, the observed onset of halo CMEs, and the ICME arrival time at L 1, it is expected to obtain an algorithm that takes consideration the effect of speed difference between CMEs and the ambient solar wind for predicting arrival time.

1. 2 Improvement of the empirical model of ICME bodies-induced Bs events 1. 2. 1 The existing model is based on relationship of duration and intensity of MC Bs events with the ecliptic latitude of the MC central axis field direction, and the ecliptic latitude is predicted using the relationship of the elliptic latitude with SFD orientation measured from Hα observations [Zhao and Hoeksema, 1997]. This algorithm does not include the effect of CME speed, and thus valid only slow CMEs. 1. 2. 2 Establish multiple regression of the duration and intensity of Bs events with the ecliptic latitude of MC orientation, the impact parameter, and CME propagation speed. Fig 1 shows that, in addition to the ecliptic latitude of MC orientation, the impact distance has significant correlation with both duration and intensity. And the CME speed has significant correlation with the intensity of Bs events. Using the elliptic cone model we can find out the impact parameter and CME speed. Thus we plan to establish an algorithm using multiple regression. The third row of Fig. 2 shows that the multiple regression can significantly improve the prediction of intensity of Bs events.

Fig. 1 Scatter diagrams of the duration (left column) and intensity (right column) of magnetic cloud Bs events versus various parameters that characterize magnetic clouds described as interplanetary flux ropes. The correlation coefficients are shown at the top of each diagram.

Fig. 2 Multiple correlation coefficients (“c” in panels) and multiple regressions of MC Bs event, (Duration, D, in left column and intensity B in right column), with MC parameters, ecliptic latitude of central axis λ, impact distance p, CME speed U, and central axial field strength Bax. The open (filled) circles denote the observed (predicted) duration and intensity of MC Bs events.

1. 2. 3 Prediction of the ecliptic latitude of MC orientation The correlation coefficients between orientations of various manifestations are as follows CME manifestations Correlation coefficient SFD and MCL 76% [Zhao and Hoeksema, 1997] HMA and MCL 77% [Yurchyshyn, 2007] EPA and HMA 95% [Yurchyshyn, 2007] HMA and HCS 68% [Yurchyshyn, 2007] HCS and HMC 94% [Yurchyshyn, 2007] The results strongly suggests that SFD, EPA, and HMA are located in corona and HCS, and MCL are in heliosphere. The latter is consistent with the conclusion that the field orientation of MCs is well conserved through the heliosphere [Kang et al. , 2006]. Statistical results also show that about 15% of MC orientations have more than 70 degrees rotation with respect to the SFD orientation [Zhao and Hoeksema, 1998; Thernisien et al. , 2006]. The MC orientation rotation is suggested to be associated with torus instability [e. g. , Torok and Kliem, 2005], that is hard to predict from solar observations. Post-Eruption Arcade (PEA) is better than SFD as cadidate of low-hight rope CMEs: * Both in quiet and active regions * in “super-AR” there may be more than one PRL and PEA orientation that cause different orientation of CMEs originating in a same AR [Liu, Webb and Zhao, 2006]. * Determining helicity and central field strength (? ) using HMI vector field measurements and NFFF model

It has shown that 60% magnetic clouds observed between Aug. 1978 and Feb. 1982 were encountered at sector boundaries [Crooker et al. , 1998]. The other 40% are expected to be encountered at UBL [Zhao & Webb, 2003]. We plan to establish the relationship of the ecliptic latitude of MC orientation with the local inclination of UBL and HCS as well as the relationship of the ecliptic latitude of MC orientation with the orientation of EPS and HMA. 1. 2. 4 Determination of the MC orientation (1) Identifying MC from in situ observations using Burlaga’s crieria & the boundary using Marubashi’s method (see above data set) (2) Determine MC orientation using Expanding MC model [Marubashi, 1997], i. e. by fitting flux rope model to the data to improve the orientation from MV technique. (3) Fig. 3 shows the improvement obtained using the Expanding MC model. 1. 2. 5 Data set Zhang et al. [2006] summarize observed CMEs and associated ICMEs and storms between 1995 and 2006. We plan to analyze single full halo CMEs and the associated events in heliosphere. Here are number of Events: Single Full-Halo CME, SH+MC 21 Single Full-Halo CME, MC 2

Fig. 3 Scatter diagrams of the duration and intensity of magnetic cloud Bs event versus the ecliptic latitude of magnetic cloud central axis direction. The”C” in diagrams denotes the correlation coefficients. The line is the least square fit to the scatter diagram. The coefficients obtained using the data computed with the expanding flux rope model (left column) are significantly greater than (more than 10%) those obtained using the combined data set (right column).

2. Development of physics-based forecasting model 2. 1 Arrival time 2. 1. 1 Existing Physics-Based Models [Siscoe & Schwenn, 2006] STOA (Dryer, 1974), ISPM (Smith & Dryer, 1990], HAFv 2 [Fry et al. , 2003]. All using kinematic models and the smallest standard deviation is ± 12 hours 2. 1. 2 Our New Physics-Based Model Hayashi et al [2006] has shown the improvement in predicting CME arrival time by using circular cone model [Zhao et al. , 2002] and the MHD model (see Fig. 4). By combining elliptic cone model fit with MAS/ENLIL model further improves the prediction of the arrival time, as shown in Fig. 5 shows a good prediction of V, n, and Tp, and thus good for predicting arrival time.

Fig. 4

Fig. 5 Simulation of December 13 th 2006 halo CME using Xuepu Zhao’s model fit to LASCO observations [From Owen’s presentation at CISM All-hand Meeting, Sept. 18, 2007]

2. 2 forecasting Speed, Density, Duration & Intensities of MC Bs Events 2. 1 The interplanetary electric field and the dynamic pressue are parameters to ditermine the response of geomagnetic sphere to ICME disturbances. Therefore, MC Parameters that need to be predicted are CME speed Vcme (km/s), plasma density, n, and the duration and intensity of MC Bs event, D and Bs [Siscoe & Schween, 2006] 2. 2 The physics-based model for predicting Vcme, n, D and Bs. The MHD simulations mentioned above is carried out by putting a plasma structure at inner boundary as initial and boundary conditions of the model. The plasma structure is constructed on the basis of cone model fitting results. To better model the interaction between interplanetary magnetic field and CME internal field, we plan to add an magnetic flux rope –like structure to examine the effect of interplanetary flux rope. This model is expected to simulate magnetic disturbance as well as plasma disturbances.