When discussing EISCAT experiments, an important distinction needs to be drawn between modulation schemes and scan patterns. The scan pattern describes the pointing direction (or cycle of pointing directions) undertaken by the antenna, and the characteristic “dwell times” which the antenna spends in each. The modulation scheme describes the properties of the transmitted pulses, in terms of the pulse lengths and frequencies used and any phase coding applied. The choice of modulation scheme also largely dictates the configuration of the receiver and the subsequent signal processing strategy, which can vary widely from experiment to experiment.
Both scan pattern and modulation scheme are fundamental properties of any experiment, but the nomenclature used to describe EISCAT experiments is often confusing. Experiments with different names can, for example, have the same modulation scheme but a different scan pattern, or vice versa, or both scan patterns and modulation schemes may be different.
Let us first consider the different modulation schemes currently in use on the EISCAT radars.
EISCAT experiments conventionally use either plain pulses or pulses with an applied phase code. Plain pulses are very simple to understand, since they are simple phase-continuous pulses of RF transmission with lengths from a few microseconds to around one millisecond. These pulses are used in two types of application, namely power profiles and long pulses. Power profiles are short pulses (typically with lengths less than 10 us). The samples received from these pulses are used only for the calculation of the zero lag, so that there is no attempt to calculate lagged products from this sample series. The series of zero lags therefore simply represents the variation of power with range. This is used to derive a parameter known as the “raw electron density”, simply a range-corrected power expressed as a crude measurement of electron density because it does not take account of temperature and Debye length effects.
Long pulses tend to be of much longer duration than power profile pulses, typically in the range 200 to 400 us. These pulses therefore have a considerably larger range ambiguity, and hence long pulse measurements can only really be used in an ionosphere which varies slowly with height, at altitudes where the ionospheric scale height is considerably greater than the range ambiguity of the pulse. Lagged profiles are calculated from the time series of samples, and these are used to make up complex autocorrelation functions, the temporal extent of which is, of course, limited by the length of the sounding pulse. The length of the autocorrelation function needed for reliable measurements of plasma parameters varies with altitude, because the plasma decorrelation time varies in inverse proportion to the spectral width. In the E-region, the incoherent scatter is narrow and the autocorrelation function is correspondingly long. In the F-region, characterised by broad, double-humped spectra, the autocorrelation function displays a much more heavily damped oscillatory behaviour, with a decorrelation time of a few hundred us, in other words the same order as the length of a typical long pulse. This makes uncoded long pulses suitable for sounding the F-region, but most unsuitable (on grounds of both lag extent and range resolution) for probing the E-region.