Separation of spatial and temporal structure of auroral particle precipitation

[1] Knowledge of the dominant temporal and spatial scales of auroral features is instrumental in understanding the various mechanisms responsible for auroral particle precipitation. Single spacecraft data always suffer from temporal/spatial ambiguity. In an effort to separate the temporal and spatial variations of the aurora, we use electron and ion precipitation data from two co-orbiting satellites, F6 and F8 of the Defense Meteorological Satellite Program (DMSP). The two spacecraft have almost identical polar orbits with a small difference in period. As a result the time difference between the two measurements varies with time. We use two statistical tools in order to determine the most probable lifetimes and spatial dimensions of the prevalent auroral features. The first tool is cross-correlation analysis between the magnetic latitude series of electron and ion, number and energy fluxes measured by the two DMSP spacecraft. As one spacecraft overtakes the other, the variable time lag between the two measurements results in different cross-correlation of the two series. We explore the dependence of this variation on the time lag between the satellites. We find that the electron precipitation exhibits a decreasing correlation between the two spacecraft with increasing time lag, whereas there is only a small similar effect for the ion precipitation data. The second statistical tool is cross-spectral analysis, for which we compute the so-called coherence function as a function of frequency (or inverse wavelength) and hence size of the auroral features. The coherence function is a measure of the stability of auroral features of different sizes. We investigate its variation as a function of the time separation between the two measurements. We show that the coherence function of both electrons and ions remains high for up to 1.5 min spacecraft separations for all features larger than about 100 km in width. For smaller features the coherence is lower even for time lags of a few seconds. The results are discussed in the context of characteristic temporal and spatial auroral scales deduced from complementary studies and expected from theory