2. Flare energy and fast electrons via Alfvén waves H. S. Hudson & L. Fletcher SSL/Berkeley and Glasgow U. Predictions for Hinode/SOT flare observations

3. Kosugi memorial, April 27 2007 GONG SOHO/MDI B LOS Maps  B LOS Maps Sudol & Harvey (2005), X10 flare of 2003 Oct. 29

4. Kosugi memorial, April 27 2007 Flare of 2001 Aug. 25 GONG + TRACE 1600A Other examples with GOES times The changes are stepwise, of order 10% of the line-of-sight field, and primarily occur at the impulsive phase of the flare

5. Kosugi memorial, April 27 2007 Replacing the thick-target model Large-scale restructuring => flare energy Energy transport by Alfvén-mode waves, not electrons The “jerk” as an alternative source of seismic waves Fletcher & Hudson 2007

6. Kosugi memorial, April 27 2007 Unknowns in the new scenario The partition of flare energy release into Alfvén mode and fast mode is unknown There are clear possibilities for electron acceleration not involving beams, but little development in the literature The role of fast-mode waves is not so evident in 3D, but mode conversion guarantees their presence

7. Kosugi memorial, April 27 2007 High Alfvén speeds Solar astronomers (those not familiar with radio astronomy?) commonly underestimate active-region field strengths and Alfvén speeds v A = 3 x 10 4 km/s is certainly possible (500 G, 10 9 cm -3 ) The condition  < m e /m p is also certainly possible, hence inertial Alfvén waves

8. Kosugi memorial, April 27 2007 The McClymont “jerk” Yohkoh and Mees observations of rapid sunspot motion during the flare of 15 Nov. 1991 Estimate of jerk penetration depth of 1.2 Mm in 4 min Energy coupling estimated at 3x10 29 to 2x10 30 ergs Anwar et al., Solar Phys. 147, 287 (1993)

9. Kosugi memorial, April 27 2007 Predictions of Hinode B variation (B => B+  b during flare) Coronal implosion (Hudson 2000) I z = constant (Melrose) - Difference field  b is potential - Ampere’s law integral is an easy test Conjugate McClymont jerks

10. Kosugi memorial, April 27 2007 Electrodynamics of the collapsing trap Betatron acceleration Fermi acceleration Inductive energy release via chromospheric “electrojet”

11. Kosugi memorial, April 27 2007 Whose ideas to follow? Obayashi 1975 Hirayama 1974

12. Kosugi memorial, April 27 2007 A more physically correct flare theory Obayashi (1975) and Hudson (2000) got it right: A flare represents a collapse of the stressed coronal field Flare theories must incorporate the full tensor conductivity in the chromosphere The role of reconnection has been greatly overstated We do not need a “black box” electron accelerator in the corona; the energy transport is via Alfvén waves instead

13. Kosugi memorial, April 27 2007 Conclusions The pattern of field changes may make it possible to identify the physics of flare causation and energy supply The “McClymont Jerk” could play a role in launching flare seismic waves We should encourage research related to the Hinode observations of before/after vector fields, both theoretical and observational

14. Kosugi memorial, April 27 2007 End Thanks for discussion and input: George Fisher, Bernhard Kliem, Brian Welsch…

15. Kosugi memorial, April 27 2007 Death to the thick target Spatial scales have become very small and make extreme requirements - extreme particle distribution functions - strong self-field - “number problem” The “dentist’s mirror” of albedo shows no significant X-ray directivity The occurrence distribution of >100 keV events shows no significant directivity

16. Kosugi memorial, April 27 2007 H. Wang, 1993 (cf. Anwar et al., 1993) First clear evidence for flare-associated field changes?

17. Kosugi memorial, April 27 2007 Giovanelli (1948) Gold & Hoyle (1961) Longcope & Noonan (2000) Anzer & Pneuman (1982) A Cartoon Sampler http://solarmuri.ssl.berkeley.edu/~hhudson/cartoons/ Priest & Milne (1980) Hudson (2000) Liu et al. (2005)

18. Kosugi memorial, April 27 2007 Courtesy Török & Kliem Notes: (1) This simulation shows a kink instability. (2) The simulation has no realistic chromosphere, so the current patterns are merely illustrative at this time. (3) The simulation does not connect one equilibrium state with another.