Phase-1: First we assume that every 30-sec average ( = Tmin, the highest resolution from the PM) is a shock candidate. Then a "10-1-10" (form a 10*Tmin = 5 minute average, skip 30 seconds, form another 10*Tmin = 5 minute average) forward-shock check is applied. This test checks the two 5-minute averages to determine if the changes in speed, flux, Mach number AND thermal speed exceed pre-set values. (These pre-set values were selected by attempting to maximize the number of actual shocks that are accepted while simultaneously minimizing the number of non-shocks, and are specific to the PM; they will undoubtedly be different for other solar wind instruments). An additional test requires that the maximum rate in any of the 6 energy-per-charge steps be at least 1.25 * the minimum rate (this avoids spurious results in the presence of very high fluxes of energetic particles). If the event passes the test it is then subjected to a 2-1-2 check {1 min averages}. The same 4 solar wind parameters are checked, but with different pre-set values. If it fails the forward 10-1-10 checks, the event is subjected to a 10-1-10 reverse-shock test, and, if that test is passed, to a a 2-1-2 reverse-shock test.
If the event passes both the 10-1-10 AND 2-1-2 checks, the 10-1-10 parameters are used to calculate a "Quality". The Quality for a series of shock events (each with a different 30-second center-time) occurring in a 5 minute window after the initial shock are compared to determine which center-time has the best Quality. That event (i.e., that 30-sec data point) is selected as the candidate shock time.
Phase-2: The shock candidate is next located in a 2-dimensional space formed by Quality and the parameter DeltaVsw/StdVsw (jump speed divided by the cumulative standard deviations from the 5-min averages). This 2-d space is divided into a number of "zones". Candidates in Zone 0 are rejected. Candidates remaining are then classified as Zones 1 - 4 (if a forward shock candidate) or Zone 1 (reverse shock candidates).
Phase-3: Forward events in zones 1 to 3 and reverse events in zone 1 are subjected to a third screening in another 2-d space formed by the Quality and DeltaNp/StdNp parameters. Events in zones 1 to 3 that pass the third screening and all events in zone 4 are reported as shocks.
Even more details are availableHere are summary results from shockspotter v1.7-5 for the interval 21 Jan 1996 to 31 Dec 2001. Note: zone10, 20 and 30 below refer to shocks that fell into zone1, 2 or 3 in Phase-2, but were rejected in Phase-3. fwd* events are those for which the average max/min PM rate was 1.11
All Events:
zone zone zone zone zone zone zone zone zone zone
0 10 20 30 1 2 3 4 >0 1-4
fwd 476 9 22 3 38 28 36 67 203 169
fwd* 13 0 0 0 0 0 0 1 1 1
REV - 3 - - 13 - - - 16 13
489 12 22 3 51 28 36 68 204 183
Shock Events (from the "Figs" web page)
zone zone zone zone zone zone zone zone zone zone
0 10 20 30 1 2 3 4 >0 1-4
fwd 6 0 0 1 17 20 34 67 139 138
fwd* 0 0 0 0 0 0 0 0 0 0
REV - 0 - - 6 - - - 6 6
6 0 0 1 23 20 34 64 145 144
Of the 153 forward shocks on the 'Figs' web page, 138 were accepted by Shockspotter.
Of the 15 that were not found, 8 were rejected in Phase-1, 6 in Phase-2, and 1 in Phase-3.
Of the 6 reverse shocks on the 'Figs' web page, all 6 were found by Shockspotter.
Of the 169 FWD events found by Shockspotter in combined zones 1-4, 138 were shocks.
Of the 13 zone 1 REV events found by Shockspotter, 6 were shocks.
A number of the false shock identifications (i.e., false-positives) are in fact
unusual time periods, representing pressure changes, non-shock discontinuities,
and possibly 'steepened' structures that are either decaying or may later evolve
into shocks. The missed shocks (false-negatives) may have been caused by
changes in the jump parameters that were too small to trigger the various
thresholds; by exceptionally turbulent pre-existing solar wind conditions; or
conceivably even by instrumental anomalies.