Reply [to “Comments on ‘Production of polar cap electron density patches by transient magnetopause reconnections’”]

In their comment, Rodger et al. [1994] (hereafter REA), make the point that more than one mechanism is likely to be involved in the production of polar cap density patches. While we would readily agree with this, we feel REA misinterpret important aspects of the mechanism we proposed [Lockwood and Carlson, 1992], and this leads them to dismiss it as a major source of patches. REA argue that there are insufficient pre-existing gradients in the plasma density outside the polar cap for the model we propose to explain patch production. This is partly because they assume that our mechanism requires the day/night terminator to lie in a specific position relative to the equatorward edge of the convection pattern which is rarely achieved (and does not apply to the three examples we presented). We do not agree with their argument for three reasons: 1. In a true time-dependent model, such as that we proposed, the loci of flux robes can diverge, converge and even cross [Lockwood, 1993]. (Naturally, flow streamlines never cross, but the flux tube loci and the streamlines are not identical for anything other than steady-state cases or unrealistic quasi-steady approximations). For example, if we only consider a time-dependent transpolar voltage, without allowing for associated evolution of the pattern of flow, the flux robe loci are not altered - the flux robe speeds simply vary as a function of time. By contrast in our model, the pattern of flow is time-dependent and flux tube loci converge on entering the polar cap (by more than the streamlines in any one flow snapshot) which means that the gradients on the edges of the patches exceed those outside the polar cap. 2. It is, however, true that the amplitude of the difference between the densities inside and outside patches must, for our mechanism to work, be present somewhere outside the polar cap. REA argue that our model requires that this range must be in a latitudinal structure lying within a narrow band less than about 3 ø wide, at the equatorward edge of the convection pattern. They reach this conclusion because they have not appreciated a key difference between the Anderson et al. [1988] model and the one we present. That is, they (explicitly) assume that our model works by simply expanding the low-latitude edge of the convection pattern equatorward. This is not the case. Our model considers how the sequences of changes in the dayside flow pattern, due to pulsed reconnection as predicted by Cowley et al. [1991] and