We use the first accurate measurements of current densities in the plasma sheet to calculate the half‐thickness and position of the current sheet as a function of time. Our technique assumes a Harris current sheet model, which is parameterized by lobe magnetic field B0, current sheet half‐thickness h, and current sheet position z0. Cluster measurements of magnetic field, current density, and plasma pressure are used to infer the three parameters as a function of time. We find that most long timescale (6–12 hours) current sheet crossings observed by Cluster cannot be described by a static Harris current sheet with a single set of parameters B0, h, and z0. Noting the presence of high‐frequency fluctuations that appear to be superimposed on lower frequency variations, we average over running 6‐min intervals and use the smoothed data to infer the parameters h(t) and z0(t), constrained by the pressure balance lobe magnetic field B0(t). Whereas this approach has been used in previous studies, the spatial gradients now provided by the Cluster magnetometers were unavailable or not well constrained in earlier studies. We place the calculated half‐thicknesses in a magnetospheric context by examining the change in thickness with substorm phase for three case study events and 21 events in a superposed epoch analysis. We find that the inferred half‐thickness in many cases reflects the nominal changes experienced by the plasma sheet during substorms (i.e., thinning during growth phase, thickening following substorm onset). We conclude with an analysis of the relative contribution of ∂BZ/∂X to the cross‐tail current density during substorms. We find that ∂BZ/∂X can contribute a significant portion of the cross‐tail current around substorm onset.
