Sierra Samaritans - National Ski Patrol

avalanche & mountaineering
table 2
Probabilities of detection calculated for a
first pass of various probe lines calculated
by PROBE.
Probe grid (cm) PROBE POD on first pass
30x30 99.9%
40x40 97%
50x50 88%
60x60 75%
70x70 63%
80x80 51%
75x70 59%
❚ CONTINUED FROM PAGE 45
possible locations, we did our best to
model a body as it might rest in all variations
in the wake of an avalanche. The
PROBE program generates a set of overlapping
spheres to model the torso and
limbs of a buried body in random orientation
with consideration for the natural
range of the movement of the limbs.
Using averages, the victim modeled in the
program is 175 cm in height and consists
of about 950 spheres.
After developing the model of the victim’s
body, we had the computer generate
a model of a snowfield and provide coordinates
and dimensions for a body buried
in that field. The computer was then
instructed to choose a random starting
point from which to begin a search for the
“victim.” With the grid predetermined
and held consistent at 75 by 70 cm and the
victim’s location established, the program
conducted a probe search across the grid
until the victim was found or until the
probing had clearly missed finding the
body on the first pass.
To test the validity of several variations
of the shape and size of the target, the programmers
ran 10,000 probing trials on the
computer. That is, the program ran 10
random starting points for probe searches
for 1,000 separate victims. As shown in
table 1, this model gives an even lower
POD for the 75 by 70 cm grid (except in
cases of vertical body orientation).
We were not just interested in understanding
past results and unexpected
failures. We wanted to find a probing
table 3
table 4
Expected time to discovery for various probing grids
GRID DESCRIPTION (CM), HPS = HOLES PER STEP
30 by 30 3HPS
40 by 40 3HPS
50 by 50 3HPS
70 by 70 2HPS
70 by 75 2HPS
80 by 80 2HPS
60 by 60 3HPS
technique we can recommend and teach
that will yield higher POD percentages
for various target areas, so we started by
collecting more statistics.
EXPECTED TIME UNTIL
DISCOVERY
If first-pass POD percentages were the sole
criteria for devising a probing strategy,
searchers would obviously always use a 30-
by 30-cm fine probe (table 2). The 99.9
percent POD offers some level of assurance
that the victim will be found. But at
what cost? Such a time-consuming grid
search would never be warranted if
searchers hoped to rescue a live avalanche
victim (table 3).
But even when probe lines are used for
recovery rather than rescue operations
some thought must be given to the speed
of the search in terms of the morale and
well-being of the search team. Some past
studies determined probing efficiency by
calculating probes per area per second, but
we chose instead to calculate the time it
would take to find a victim by simulating
how probe lines actually work.
0 50 100 150 200 250 300
EXPECTED TIME TO DISCOVERY (MINUTES)
Comparison of several statistics for several grid sizes. (Percentages are plus or minus 1 percent.)
% FOUND ON VARIOUS PASSES
Technique 1st 2nd 3 or more Time per pass Average passes ETD (min.)
30 cm sq. 99.9% 0.1% 0% 487 min. 0.50 244
40 cm sq. 97% 2.8% 0.04% 274 min. 0.53 144
50 cm sq. 88% 12.1% 0.8% 175 min. 0.64 112
60 cm sq. 74% 20.8% 4.4% 122 min. 0.83 100
Trad. 70x75 60% 26.5% 13.6% 96 min. 1.15 111
To determine the expected time to discovery
(ETD) for each size grid, our calculations
called for a 20-person probe line
inside a space of 10,000 square meters
(1 hectare). We also used an estimate of 3.7
seconds per probe, 4.4 seconds per step, a
descent speed (i.e., the rate at which rescuers
return to the bottom of the debris
field for the second pass) of 1 second per
meter, and 60 seconds for reorganization.
We did our best to keep the numbers reasonable
and consistent across the various
trials (again, see table 3).
After the first pass, we had to define
what we needed to do in order to keep
probing until the body was found. To
maximize the probability of finding the
body, the probe coordinates for the second
pass were offset from the first by half
of the grid dimensions. That is, after covering
the standard 70 cm by 75 cm grid,
the second pass was offset by 35 cm up the
hill and 37.5 cm laterally for each probe
step. Thus the probe holes of the second
pass were exactly halfway between the
holes left by the first pass.
CONTINUED ON PAGE 72
46 Ski Patrol Magazine | Winter 2005