Automation is essential because there are ~70 satellites yielding ~4000 crossings of the auroral zone every day.

Birkeland current signatures are derived from the data using the cross- track component, in which the current signatures are largest and uncertainties in the background field are smallest.

Preliminary data processing occurs in three steps:

1. Subtract model of Earth's magnetic field.

2. Cross-correlate residuals with model field to identify orientation, gain and offsets. Apply corrections for these effects.

3. Filter to allow periods shorter than 26 minutes

These steps are displayed in Figure 4.

Figure 4. Steps in processing the Iridium magnetic field data. Top panel shows the original cross track component for one day. In this case, Feb. 7-8, 1999. Horizontal lines show the full scale limits of the panel below 'Model Subtraction'. Second panel shows the data after subtracting the model magnetic field. Spike features are now apparent, especially from 1400 to 2200 UT. These are the auroral current signatures. Horizontal lines show full scale of the following panel. Third panel shows the data after applying post processing calibration corrections to the data (a calibration matrix is determined by comparing the residuals with the model field). Although the auroral signatures are clearer, orbit period and half orbit period variations are still apparent. These are probably due to a combination of tiny attitude variations and errors in the model field. These residual variations are removed by applying a high pass filter, cuttoff at 26 minute period. The high pass filtered data are shown in the bottom panel. This result is the basis for subsequent analysis. Auroral zone candidates are identified by performing statistics on these data and assigning a threshold sufficiently above the gaussian noise distribution to reliably identify outliers and real physical signals.

Figure 4