Forecasting Super CME Disturbances

1
Forecasting Super CME Disturbances

• 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

2

• 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, 2007a 2007b.
  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 L1 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, 2007c
  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 for
  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 for forecasting
  the arrival time and the four ICME parameters
  that determine the interplanetary electric field
  and the dynamic pressure.

3
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
  L1, 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.

4
I

• 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 Ha 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.

5

• 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.

6

• 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 ?, 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.

7

• 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

8

• 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 Burlagas crieria the boundary using
  Marubashis 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, SHMC 21
• Single Full-Halo CME, MC
  2

9

• Fig. 3 Scatter diagrams of the duration and
  intensity of magnetic cloud Bs event versus the
  ecliptic latitude of magnetic cloud central axis
  direction. TheC 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).

10
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, HAFv2 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. Fig. 5 shows a good
  prediction of V, n, and Tp, and thus good for
  predicting arrival time.

11
Fig. 4
12
Fig. 5 Simulation of December 13th 2006 halo CME
usingXuepu Zhaos model fit to LASCO
observations From Owens presentation at CISM
All-hand Meeting, Sept. 18, 2007
13

• 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.