Boltek Lightning/2000 - Manual (Page 243)

R E A L - T I M E L I G H T N I N G D E T E C T I O N
243 243 243 243
will suffer. Also, at large distances the signal will degrade and
some of the information content of the signal necessary for
stroke classification will be lost in the background noise. This
will vary from stroke to stroke. At extreme range, reflections
from the ionosphere can cause an inversion of the
information used to determine the polarity of the stroke type
which will in turn cause the stroke to be plotted 180 degrees
from its true location.
The stroke rate can also cause some errors in the stroke
sorting process. In cases of very high stroke rates from
several storms (stroke count over 1000/minute) the
occasional loss of information at the beginning of a signal
may cause an error in stroke classification.
The distance that IC strokes can be detected can also be a
problem. At larger distances the total stroke counts will not
be a true indicator of storm activity due to the limited range at
which IC strokes can be detected. An example of this would
be a supercell storm over 200 miles away. At that distance a
much smaller percentage of the IC strokes would be detected
(if any) and the storm’s type could be hard to classify based
on just CG activity. An extreme case of this would be a
supercell storm that is producing only IC lightning. At long
range (200+ miles), such a storm may not be detectable or
look like a normal weak storm cell.
The detection efficiency (DE) of the hardware has been
estimated with a series of measurements with the data sets of
several storms. The DE for CG strokes is about 85% +/-
6% and the DE for IC strokes is about 20% +/- 5%. An
additional factor that effects the DE is the percentage of IC
and CG that can not be classified into stroke types. This may
result in the stroke type being classified as unknown or as
noise. Typical percentages seen (from a 108,000+ stroke
summer storm) are: