Methodology:

We have applied the Grand Unified Incoherent Scatter Design and Analysis Package (GUISDAP) to analyse field aligned and >60 degree elevation experiment data (including scanning experiments) at 10 minute and 1 hour integration times. Different EISCAT experiments with a range of specified pulse codes have been used. GUISDAP fits theoretical incoherent scatter spectra to the log profiles derived from the autocorrelation function of the received radar signal. This produces standard Incoherent Scatter Radar parameters, such as electron density. This is carried out for each altitude step.

We require the integration time to be more than 40 minutes to be included in the 1-hour archive and more than 7 minutes to be included in the 10-minute archive (some experiments do not start/end on the hour for a multitude of reasons including scheduling so for subsequent analysis).

We have used the suggested 'magic constant'/calibration scale factor listed in the calibration tables (https://eiscat.se/scientist/data/tromso-calibration-tables/). Where these were not available, we went back through the schedule list and manually collected the relevant calibration numbers (https://portal.eiscat.se/schedule/). Where more than one magic constant were available we preferentially use

foF2, then plasma line, then foE, except for the lower altitude experiments where we used foE.

The power profile or 'raw' electron density data (Nr) comes directly from GUISDAP, these are essentially back-scatter power profiles measured by the radars at zero lag (the EISCAT data are stored as lag profiles (Virtanen et al. 2008) derived from the auto correlation function of the fitted spectra). To identify the altitude where the raw electron density is approximately equal to the fitted electron density, we found the maximum altitude where Ti/Te = 1 (this is an assumption, demonstrated in Figure 9 of Semeter & Kamalabadi (2005)). To find this altitude we smoothed the Ti/Te altitude profiles using a Savizky-Golay filter (Savitzky & Golay 1964), to identify the point where it deviated from 1 +/- 0.1. Where there were more than one power profile per integration period, we have taken a mean of Nr at each altitude below the threshold identified above.

Data collection:

The data from this archive were taken by the UHF (ultra-high frequency) and VHF (very-high frequency) radars, situated in Ramfjordmoen, near Tromso (detailed description in Rishbeth & Williams, 1985)

The UHF consists of a fully steerable, 32 metre parabolic dish and operates at frequencies close to 930 MHz, with a peak power of 2 MW.

The 3 MW VHF radar is a 40 by 120 metre parabolic trough with an elevation range of 15-90deg and an operating frequency of 224 MHz.

The data were analysed using the GUISDAP 9 package, downloaded from: https://gitlab.com/eiscat/guisdap9

Data quality:

Missing values have are represented as NaNs.