Seen from Fig. 3 a and Fig. 4 a, at the onset of the flare, three bright regions were lying in the SXR emission streak along L3, among which, the feet regions A and B were further enhanced in the impulsive phase (Fig. 3 b and Fig. 4 b) and actually became the nonthermal electron depositing sites giving strong HXR and chromospheric features (Paper I), while the region D, or the top of loop L3, faded in the impulsive phase. The morphological comparison carried out in the previous sections seem to suggest that the interaction between L1 and L3 may occur at the site of D. Indeed, in Paper I, we have spotted a hot component of the coronal plasma and found that both thermal and nonthermal processes were taking place at the very onset of the flare; the HXR images synthesized from Yohkoh observation also reveal a loop top source at this time (Takakura et al. 1994). Combining the above facts, we are able to establish the following scenario of the initial phase of the flare: the flare was triggered by the interaction between two current-carrying loops at the top of loop L3 where the in situ energy release occurred through the rapidly heated plasma and energized electrons.
Fig. 5 illustrates the SXR flare at the onset of the event superposed on the map of free energy distribution. The concept of free energy used here is defined from the magnetic shear (Hagyard, Low, & Tandberg-Hanssen 1981), which is calculated as the following (Wang et al. 1996):
where Bs is the source field (Hagyard, Low, & Tandberg-Hanssen 1981), or the departure of the observed vector field Bo from the potential field Bp deduced from the observed longitudinal field, Bs=Bo-Bp, which contains information on the path of atmospheric currents whose footpoints on the photosphere are mapped by vertical currents (Moreten & Severny 1968; Hagyard 1988; Wang et al. 1996)
It is seen that the flaring loop L3 is where the free energy gets most concentrated and the regions A, B and D are coaligned with the energy density maxima. To be more careful, since the loop L1 and L3 are fairly overlapped along the inversion line M1, we cannot tell definitely whether the free energy concentration is actually occurring along the interface between L1 and L3 or along the strongly sheared loop L3. Nevertheless, in either case, energy release in this flare should be taking place at the sites of free energy concentration (Machado et al. 1988; Low & Wolfson 1988). Particularly, one such site is the foot A, which is also the locus of the strong concentration of magnetic longitudinal flux, vertical current and free energy, and the other site is B, located over the reverse line of magnetic longitudinal field and vertical current, where strong non-potentiality is dominant and the separatrix layer may be involved in the energy release (Mandrini et al. 1995; Wang et al. 1997; Wang 1997).
A distinctly different case is seen for the pattern L2, which fairly deviate from the free energy concentration around M2, implying that the flaring of L2, unlike L3, could be a secondary effect of energy transfer from the initial flaring site. We are encouraged to hold that a certain energy flux from the flaring loop L3, was input along the loop L2 and impacted on region C, noting that the gradual radiative behavior and special spectral feature in the region C support this idea (Paper I).
From above, a schematic interpretation of the trigger of energy release in the flare on October 27 is shown in the cartoon in Fig. 6.
© European Southern Observatory (ESO) 1998
Online publication: July 7, 1998