Key points are not available for this paper at this time.
We compute the magnetic helicity injected by transient photospheric horizontal flows in six solar active regions associated with halo coronal mass ejections (CMEs) that produced major geomagnetic storms and magnetic clouds (MCs) at 1 AU. The velocities are computed using the local correlation tracking (LCT) method. Our computations cover time intervals of 110-150 hr, and in four active regions the accumulated helicities due to transient flows are factors of 8-12 larger than the accumulated helicities due to differential rotation. As was first pointed out by Démoulin and Berger, we suggest that the helicity computed with the LCT method yields not only the helicity injected from shearing motions but also the helicity coming from flux emergence. We compare the computed helicities injected into the corona with the helicities carried away by the CMEs using the MC helicity computations as proxies to the CME helicities. If we assume that the length of the MC flux tubes is l = 2 AU, then the total helicities injected into the corona are a factor of 2.9-4 lower than the total CME helicities. If we use the values of l determined by the condition for the initiation of the kink instability in the coronal flux rope or l = 0.5 AU then the total CME helicities and the total helicities injected into the corona are broadly consistent. Our study, at least partially, clears up some of the discrepancies in the helicity budget of active regions because the discrepancies appearing in our paper are much smaller than the ones reported in previous studies. However, they point out the uncertainties in the MC/CME helicity calculations and also the limitations of the LCT method, which underestimates the computed helicities.
Nindos et al. (Fri,) studied this question.