MGL1212
Plot of shot 25000 only with a 20x120 filter shows
traces 342, 349 561, 583 are bad/dead.
and traces 152, 390, 391,392, 397, 398, 398, 449, 590, 592, 636 are low amplitude
and traces 7, 8, 9, 10, 12, 13, 20, 22, 24, 29, 30, 92, 141, 449 have 1s periodic spike(s)

Text listing of the statistics of shot 25000 without any processing.

MGL1112
Plot of shot 4003 without any processing shows
traces 30, 169, 192, 257, 294, 301, 343, 349, 377, 396, 437 are bad
and traces 186, 294, 427, 428, 429, 430 are low amplitude.
Plot of shot 4003 with .2s AGC shows
traces 30, 169, 192, 257, 294, 301, 343, 349, 377, 396, 437 are bad

Text listing of the statistics of shot 4003 without any processing.
Most of the bad traces are "dead" or near 0.  These will not contribute much 
to the stack, but it'd be good to kill/zero them, especially if a prestack
gain is applied.
Traces 186, 294, 427, 428, 429 have low maximum, minimum, and average absolute value.

Plot of shot 4003 with AGC after TREDIT MINVAL 10

mintype absval minval 10 killed traces 169, 257, 294, 301, 343, 349, 377, 396, 437 
mintype kurt minval 30 killed traces 30, 169, 192, 257, 294, 301, 343, 349, 377, 396, 437

Shot R007941_1316703111.RAW killed too many of the long ranges with minval 10,
but     minval 5     worked well.
Text listing of the statistics without any processing.

Weighting by trace number is best since "good" low amplitude traces are not killed.

Dec 2011 SIO Geometrics
All work done after 20x100 filter.  Data were collect at 1ms.  
Plot of shot 1004 (day 318). Text of shot 1004 statistics.
Plot of shot 7306 (day 321). Text of shot 7306 statistics.
Plot of shot 11738 (day 324). Text of shot 11738 statistics.
Comments:
1)  Shot 1004 trace 41 and shot 7306 trace 27 are "dead" (near zero) and may be
    killed using MINTYPE ABSVAL MINVAL 1.
2)  Shot 11738 does not have any "dead" traces, but is quite noisey including
    spikes or large amplitudes that may be killed using MAXVAL.
3)  Shot 11738 trace 5 "spike" doesn't appear when plotted without filter.
    The plot without filter shows some low frequencies (<1Hz.).  Any recording filters?

2005 HLY0503
Shot 52000 is late in the cruise in fairly heavy ice and the streamer has
been heavily damaged by the ice.  Trace 24 is for the sonobuoy.
Plot with 20x100 filter and .2s agc.
Plot with 20x100 filter.  Text of trace statistics.
Plot raw shot 52000.  Text of trace statistics.
Comments:
1)  The plot with filter and agc shows that trace 1 appears to be
    leaking and the SNR is low.   Trace 18 contains many spikes.
2)  The plot with filter only shows the spike problem clearly.  
    Trace 18 likely is usable if the spikes can be removed (thus
    the trace should not be killed).
    The plot also shows that trace 1 amplitudes are extremely low.
3)  The plot without filter or agc alows show that trace 1 is bad.
    The spike problem is also evident, though spikes in the water
    column will be zeroed through muting to the water bottom.
Conclusions:
1)  The statistics of trace 1 with and without filter shows that the
    maximum value is smaller than most other traces by a factor.
    Without filter:
    Trace 1 max=19748, 2-16 max>100000, 17-19>90000, 20-23>100000
    MINVAL 30000 would kill trace 1.
2)  The standard deviation after filter of trace 1 is lower than
    another trace by a factor of 10.  This is reasonable since
    a leaky hydrophone group contains mostly 60hz.
    MINTYPE SDEV MINVAL 2500 should be used.

Ewing
Plot of shot 6410 traces 130-140.  Text for statistics for shot 6410.
Comments:
1)  No clue why trace 122 is low.  max & sdev are 1/3 of adjacent traces.
    No need to kill.
2)  Trace 130 has a classic analog (recording) IIR filter of a spike ~1.65s.
    Nothing in the statistics showing the spike.