Comment on “Modeling the magnetosphere for northward interplanetary magnetic field: Effects of electrical resistivity” by Joachim Raeder

In a recent article Raeder [1999] presented a series of MHD simulations using a resistive MHD model to describe the interaction of the terrestrial magnetosphere with the solar wind for due northward interplanetary magnetic field (IMF) conditions. Depending on the initial conditions, the simulated magnetotail remained open for at least 6 hours after the IMF turned from south to north, or remained closed at all times when the simulation started with northward IMF. When an increasing amount of uniform resistivity was added, the magnetotail became closed, and its length decreased with increasing resistivity. At the same time, increasing uniform resistivity diffused the bow shock heavily. Raeder [1999, p. 17,357] speculated that models that predict the rapid closure of the magnetosphere and the formation of a steady, finite-length tail are “possibly in error due to numerical resistivity.” The variation of explicit resistivity of Raeder [1999], intended to mimic the artificial resistivity present in MHD codes, is an interesting exercise. However, we disagree with Raeder’s statements about the numerical resistivity being the cause of closed magnetotails for northward IMF. In fact, as it will be described in this comment, there are several reasons to believe that numerical resistivity does not play a role in the formation of closed magnetotails. In this Comment we point out the following: 1. A number of investigators, using a wide variety of MHD codes, have reported closed configurations for the northward IMF condition. Raeder’s [this issue] implication that all these codes are only as accurate as the first-order Rusanov solution, provided by him for comparison, is contradicted both by the marked similarities of some codes (particularly that of Raeder [1999] and that of Fedder and Lyon [1995]) and by published convergence studies [e.g. Powell et al., 1999]. 2. Closed configurations have been obtained with MHD codes with similar underlying schemes as the approach employed by Raeder [1999] and on meshes more highly resolved than those employed by him [cf. Song et al., 1999]. 3. In the simulations presented by Raeder [1999], added uniform resistivity simultaneously resulted in closed magnetotails and substantial diffusion of the bow shock (the upstream plasma pressure in the equatorial plane significantly exceeded the thermal pressure of the ambient solar wind). By interpreting the difference between his results and those of other investigators in terms of numerical resistivity, Raeder [1999] reaches unjustified and incorrect conclusions. In particular, in his simulations, short, closed tails come together with the disappearance of well-defined bow shocks. Since other simulations obtaining closed magnetotails have sharp and well-resolved bow shocks, Raeder’s [1999] simulations imply that these results cannot be due to uniform numerical resistivity.