In SJT93 it was asserted on the basis of an examination of time-series plots that most of the channels were spending a large fraction of the time at background. This was then used to justify a bulk statistical analysis of the count-rate histograms to determine the background levels. This method is, however, valid only when 2 conditions are met:

1. The count rate is indeed truly at background for most of the time.
2. The background level doesn’t change during the analysis interval.

Unfortunately, with an extra 8 months of data it now appears that neither of these conditions is met. Therefore, in this analysis I have reverted to a method more closely allied to that used by GMS91 in the pre-Jupiter estimations, of selecting times of low count rate by eye and fitting a level to them.

General Trends. The first stage of the analysis was to examine the entire mission to date (1990 day 318 to 1991 day 120) at 6-hour time averaging. This allows the identification of the general trends and of intervals of similar behavior.

On the basis of this, the mission to date may be divided into 6 intervals with distinct properties.

1. Switch On - End Jan. 1991 - All channels show occasional excursions to low count rates. The interval is too short to be able to reliably determine trends. This was the only interval analysed in the GMS91 study.
2. End Jan. 1991 - August 1991 (Conjunction) - During this interval there was continuous heliospheric activity, and none of the channels approached background levels at any time.
3. August 1991 - Feb. 1991 (Encounter) - Most channels show some times at low count rates (although E1 and E1' drop less often than other channels). The higher-energy DE channels seem to show an increasing trend, but we must beware of the low point right at conjunction. Most other channels appear more or less constant.
4. Feb. 1991 - April 1993 (Exit from current sheet and disappearance of slow solar wind) - Most channels (other than low energy electrons) come to low count rates essentially every rotation. Lowest channels show nearly constant levels, but the higher energy channels show a steady rise.
5. April 1993 - Jan. 1994 (Reappearance of the electron streams) - All channels come to low rates regularly. The levels are shifted down relative to the previous interval but the same trends continue.
6. Jan. 1994 - April 1994 (End of data) - Increasing trend in high-energy channels slows or stops. Low energy channels increase back towards the interval 4 levels.

Minimum Levels. The determination of the lowest levels displayed by each channel was made by selecting one (or more) lowest minima from each of the intervals outlined above (apart from #2) for examination in more detail. The case of interval 3 is special because the spacecraft conjunction occurred early in this interval; the count rates of many channels were at their lowest at this time. However, because of the low telemetry rate and resulting intermittent data at this time we cannot trust these values and have not used them. (We have, however, flagged cases where the conjunction rate appears lower than the quoted value in the data tabulations.)

These “lowest minima” were then plotted in 1-hour QCA filtered averages with error bars to allow the lowest counting rate to be estimated using a “cross-hair” cursor; error limits were also estimated in the same way. This method is preferable to calculating averages as it allows much better elimination of low-level events and of data dropouts. The levels thus estimated are summarized in Tables 5.1 through 5.5. For each channel the date given is the approximate day at which the rate was determined (the actual interval used may be more or less than one day, depending on variations). The rates are given in counts/second followed by the estimated error in the final digit in parentheses (thus 0.340(7) means 0.340±0.007)

In the case of the Wart channels (W1-8 and Z2-4, Z2A), the rates are too low to allow this method to work and so there levels were estimated using 2, 4, or 10 day averages and simply selecting the lowest level seen. In cases where zeroes were seen, the level is given as a limit at the 1 count/averaging period level. These levels are given in Table 5.6.

Table 5.1 Minimum count rates in each channel for interval number 1

Table 5.2 Minimum count rates in each channel for interval number 3

Table 5.3 Minimum count rates in each channel for interval number 4

Table 5.4 Minimum count rates in each channel for interval number 5

Table 5.5 Minimum count rates in each channel for interval number 6

Table 5.6 WART channels lowest levels

Next: 5.12 Method (continued)

Return to Chapter 5 Table of Contents