On Deriving Mass Energetics of Coronal Mass Ejections

On Deriving Mass & Energetics of Coronal Mass Ejections A Tutorial Angelos Vourlidas NRL No. RH Visit February, 2005 Angelos Vourlidas, NRL

Overview • The following questions will be addressed: – How can we derive information about CME mass/energetics? • What assumptions enter in the calculations? • What are the data analysis steps to extract quantitative CME information from white light images? – How good are the numbers? • Can we estimate the errors? How? – What can we do with this information? • What statistics tell us? • What correlations can we find? No. RH Visit February, 2005 Angelos Vourlidas, NRL

Preliminaries • Height-time plots, online movies are constructed from UNCALIBRATED LASCO images. Calibrated images are rarely shown. • All necessary calibration tools exist in the LASCO Solarsoft distribution. • This talk is relevant to CME measurements ONLY. Coronal background densities, streamers and plumes must be treated differently. • Remember, a white light CME is defined as an brightness increase relative to the background No. RH Visit February, 2005 Angelos Vourlidas, NRL

Our Objective ? Raw C 3 Image No. RH Visit February, 2005 Calibrated C 3 Image (Diff. ) Angelos Vourlidas, NRL

CME Mass/Energy Derivation Flow cme_massimg 2 total. pro C 3_massimg. pro No. RH Visit February, 2005 Angelos Vourlidas, NRL

Mass Calculations Primer Assumptions: • • • Emission is due to Thompson scattering of photospheric light from coronal electrons. All mass is on the sky plane. Plasma composition is 10% He, 90% H. Restrictions: • • • The 3 D distribution of the background and CME electrons, Ne, is unknown. The temperature of the ejected material is unknown (coronal should dominate). Emission is optically thin. No. RH Visit February, 2005 Angelos Vourlidas, NRL

Mass Calculations Primer Method: A coronagraph measures the total brightness along the line of sight. We can only measure excess brightness (ICME - IPREEVENT). Excess DN calibration Btotal Be No. of e- composition Mass Error Sources: exposure time (~0. 15%) vignetting (~1%) photon noise (<1. 4%) Phot. Calibration (0. 73%) composition (6%) stars (cancel out) Cosmic rays (few pixels) solar rotation (not important for fast events) Streamer deflections (difficult to estimate) 3 D structure (more on that later) No. RH Visit February, 2005 Angelos Vourlidas, NRL

Mass Calculation Methods SECTOR TORUS ROI • Several ways to obtain a “mass” for an event. • The choice depends on the objectives: – After the whole event? – After specific features (i. e. , core)? – Flow measurements? Best for flow calculations: Best for Most common: Position at fixed distance automated calculations: Avoid streamers, planets, Extent other & Upper boundary CMEs from CME lists/ht measurements “Typical” C 3 Mass Image No. RH Visit February, 2005 Angelos Vourlidas, NRL

Example Results — Single Event Etotal Mass mass EP Epot v. CM EK Ekin v. CM or vfront EM vesc Emag vesc More examples in Vourlidas et al (2000), Subramanian & Vourlidas (2004) No. RH Visit February, 2005 Angelos Vourlidas, NRL

How Good Are CME Mass Estimates? Real mass could be x 2 larger No. RH Visit February, 2005 Angelos Vourlidas, NRL

Effect of CME-Sky. Plane Distance on Mass Estimates? Sky-Plane PA CME mass could be 3 x less Co rre cte PA Corrected d Sky-Plane CME mass could be 5 x larger No. RH Visit February, 2005 Angelos Vourlidas, NRL

CME Mass Database (Jan 1996 – Dec 2003) Thanks to the hard work of Ed Esfandiari an up-to-date CME database has been created: • The CME information is taken from the CUA/NRL list. • The database includes full-frame mass images for every h-t data point in the CUA list (6385 events so far). • The mass is derived with the same method (sector) for all frames. • Energy and other calculations are also provided. The following information is provided for every CME frame: 1. 2. 3. 4. 5. 6. No. RH Visit February, 2005 Date/time Width Position Angle Height of CME Front Sector Area Mass 7. 8. 9. 10. 11. 12. Mass density Kinetic Energy Potential Energy Velocity (H-t) Acceleration Escape Velocity. Angelos Vourlidas, NRL

Results The analysis of the mass database is based on : • Measurements at the point of maximum mass. (Need for a single “representative” number for each event). • Does not include events with: • < 5 h-t measurements (frames). • Width > 120°. • Negative mass. • Zero pixels in sector. No. RH Visit February, 2005 Angelos Vourlidas, NRL

Results – Distributions No. RH Visit February, 2005 Parameter LASCO Solwind <Ekin> (ergs) 4. 3 1030 3. 5 1030 <Mass> (gr) 1. 7 1015 4. 1 1015 Total Mass (gr) 4. 1 1018 3. 9 1018 Mass Flux (gr/day) 3. 6 1015 7. 5 1015 Duty cycle 81. 7% 66. 5% Angelos Vourlidas, NRL

Results – Average Mass 3 1010 gr/pix or 1. 3 104 e/cm 3/Rs No. RH Visit February, 2005 The constant mass density suggests that: 1. Only the CME width is needed to derive the mass 2. The bulk of the CME material originates at high altitudes where the corona is more uniform. Angelos Vourlidas, NRL

Results – Bimodal Distribution? Do we have “failed” and “successful” CME populations? No. RH Visit February, 2005 Angelos Vourlidas, NRL

Results – Yearly Variations No. RH Visit February, 2005 Angelos Vourlidas, NRL

Review • It is easy to calculate CME mass and energetics from the LASCO images (calibration/routines available since 1996). • The accuracy of the mass values is difficult to estimate without 3 D information. Simple simulations suggest that masses could be underestimated by x 2 (on average, well-behaved (aka non-halo) events). • Thousands of measurements of several dynamical parameters for almost all CMEs are now available. • Mass images for almost all CMEs are also available (for DIYers). • Preliminary analysis of the mass/energy data yielded a couple of very interesting results: • CME mass density = constant! • There may be 2 classes of CMEs; “failed” and “successful”. • CME mass/energy distributions are power-laws (like flares!). No. RH Visit February, 2005 Angelos Vourlidas, NRL

BACKUPS No. RH Visit February, 2005 Angelos Vourlidas, NRL

Results – Mass Distribution Solwind Exponential Fit (Jackson & Howard 1993) LASCO Power-law Fit, =-1. 8 (Vourlidas & Patsourakos 2004) No. RH Visit February, 2005 Angelos Vourlidas, NRL

LASCO C 3 Photometric Performance Courtesy of A. Thiernisien No. RH Visit February, 2005 Angelos Vourlidas, NRL

Magnetic Energy Estimates • Problem: Direct measurement is not (currently) possible except • Radio gyrosynchrontron emission from energetic electrons within the CME (Bastian et al. 2001). Only a handful cases so far. • Another Approach: 1. Select fluxrope-like CMEs. 2. Assume the fluxrope feature becomes the IP Magnetic Cloud. 3. Assume magnetic flux, Φ is conserved (in the fluxrope). 4. Use in-situ measurements of Φ to normalize the magnetic energy, EM. 5. Use the coronagraph measurements of the fluxrope area, A and “length”, l to derive the evolution of EM. No. RH Visit February, 2005 Angelos Vourlidas, NRL

Magnetic Energy Estimates • Relevant Equations: Assume fluxrope is cylindrical, B & A are measured/derived from in-situ observations Φ. A is given by the no. of pixels in the LASCO images l is assumed equal to the height of the CM, l r. CM. No. RH Visit February, 2005 Angelos Vourlidas, NRL