Solar wind input between substorm onsets during and after the October 18-20, 1995, magnetic cloud

We study an aspect of substorm occurrence during Earth passage of the Octob er 1995 magnetic cloud and the corotating stream overtaking it. The substor ms were identified primarily by five latitudinal arrays of ground-based mag netometers around the globe and the interplanetary observations were from t he Wind spacecraft. The aspect we study is the energy and magnetic flux acc umulated in the geomagnetic tail between successive substorm onsets, which we calculate by integrating different theoretical and empirical expressions for the solar wind-magnetosphere coupling rates. These coupling functions, which depend strongly on the interplanetary magnetic field (IMF) north-sou th component, B-z, have all been proposed previously as appropriate measure s of solar wind input into the magnetosphere and indeed most correlate well with the cross-polar cap potential. We contrast the effect on magnetospher ic substorms of three solar wind states: A 14 hour long period of continuou s and strong IMF B-z < 0 and a 16 hour long period of continuous and strong IMF B-z > 0, both states due to the magnetic cloud; and a 22 hour long per iod of intermittent IMF B-z polarity of large, but weaker amplitude, due to Alfven waves in the faster stream. We find that (1) the average rate of su bstorm onset occurrence during the B-z < 0 phase in the cloud is higher tha n during the intermittent B-z polarity in the wake, by about a factor of 1. 4. (2) The long B-z > 0 phase of the cloud elicited no substorm onsets. (3) The total time for which B-z is negative between substorm onsets is less v ariable than the interonset interval itself, but still varies between simil ar to 0.5 hour and similar to 4 hours. (4) The integral of any individual s olar wind-magnetosphere coupling function between successive substorm onset s is also variable. (5) The interonset integral varies differently for diff erent coupling functions, even for those which measure the same quantity (e nergy or magnetic flux). (6) For all energy and magnetic flux coupling func tions we find a positive correlation between the interonset integral and th e value of the coupling function around the time of the preceding substorm onset. The correlation is best for the epsilon and nu B-s functions. This r esult is consistent with those substorm models in which substorm onset occu rs at some fixed energy or magnetic flux threshold and where the amount of energy or magnetic flux lost following each substorm onset is proportional to the solar wind input rate at the time of onset.