Substorm growth and expansion onset as observed with ideal ground-spacecraft THEMIS coverage

We present a fortuitous case of an isolated substorm on 29 March 2009, observed by the Time History of Events and Macroscale Interaction during Substorms (THEMIS) probes clustered at similar to 11-14 R-E with simultaneous coverage by the THEMIS ground network. The four probes are at roughly the same radial and azimuthal location, with one probe staying near the neutral sheet during entire growth phase and during the ensuing transition to the substorm expansion phase. Prior to substorm onset, THEMIS observed the damping of earthward convection and development of an embedded tail current sheet with half thickness <= 0.15 R-E and current density similar to 20 nA/m(2), while the total magnetic field closest to the neutral sheet was below 2 nT. Tail activity was observed to start prior to substorm onset tailward of the THEMIS probes (<-14 R-E) with the gradual increase of earthward flow and total pressure in front of an earthward moving bursty bulk flow event (BBF). There was no evidence of an outward propagating rarefaction wave prior to BBF onset. Peak flow was accompanied by two short dipolarization pulses, followed by a sharp reduction of the total pressure (up to 30-50%) and a decrease in the lobe magnetic field. By analyzing the lobe field waveforms, we infer their reconnection origin and argue that different onset-related magnetotail phenomena discussed in past literature (BBFs, bubbles, current disruption, nightside flux transfer events, earthward traveling convection regions) are views of the same dissipative structure formed by reconnection, when it is observed from different vantage points. Although the first ionospheric signatures of the substorm were observed near the equatorward edge of the auroral oval, the adaptive magnetospheric model maps the breakup to the tail current sheet at 15-20 R-E, consistent with other estimates of the substorm onset location available in our case. The event provides definite evidence in favor of tail reconnection within a thin current sheet as the primary substorm initiation process. It also demonstrates that actual changes of tail current magnitude in the magnetotail can be an order of magnitude larger, compared to the magnitude of 3-D current inferred from ground observations.