Army Scientist Cut UXO False Alarms

[Lenny Siegel <lsiegel@cpeo.org>]

ARMY SCIENTISTS ASSERT UXO "FALSE ALARM" BREAKTHROUGH
(May, 1998)

My summary of the Defense Science Board (DSB) Unexploded Ordnance (UXO)
Remediation report included the DSB's suggestion that within 3 to 5
years the number of false detections in UXO surveys could be reduced
tenfold, cutting the cost of range remediation in half. This was based
upon technical briefings that took place more than a year ago. To some,
it seemed that DSB was promising too much.

This week, however, at the UXO Conference in Anaheim, scientists from
the U.S. Army Engineering and Support Center in Huntsville claimed that
they have already reached that goal, using "digital geophysical mapping"
to "post-process" data from existing sensors. Before I explain the new
technique, let me describe the conventional way of sweeping a range for
subsurface UXO.

Typically, after clearing surface ordnance and sometimes vegetation or
other obstacles, UXO technicians rope off narrow alleys. They walk down
those alleys, carrying handheld sensors, such as passive magnetometers
or electromagnetic induction sensors (coin detectors). Wherever the
device emits a squeal, the technician plants a small bright flag in the
ground. Later technicians return to dig for the magnetic anomaly that
cause the squeal. This method, known as "mag and flag," often generates
hundreds of "false anomalies, metallic rocks, munitions fragments, or
other human-made metallic objects for each piece of live ordnance.
Consequently three quarters or more of the cost of clearing a range can
be devoted to digging holes to remove those non-explosive objects.

The trick, therefore, is to determine which anomalies are likely to be
explosive before digging any holes. There are a number of possible
approaches, including sensors that determine the shape of anomalies and
sensors that "smell" for explosive compounds. The new Huntsville
approach is based upon the fact that existing sensors are capable of
collecting much more information than human operators recognize when
they decide whether the squeal of the device deserves a flag.

Digital geophysical mapping utilizes that data, effectively mapping the
electromagnetic response in a computer. High precision navigational
equipment allows technicians to return to the site of suspected ordnance
later without implanting flags. Once the data is in the computer,
operators can analyze and re-analyze the data in their effort to
discriminate between ordnance and other anomalies and also to estimate
the depth of ordnance penetration.

The Huntsville scientists start with a simple concept, comparing the
electromagnetic signature of an anomaly against signatures of known
ordnance, at a variety of depths and angles. To strengthen the
comparison, they focus on ordnance types known to have been deposited in
the area. With each piece of UXO they unearth, they add information on
the signature of that specific piece - which is influenced by the soil,
moisture, and other local conditions - to their library of comparison
signatures.

In early trials of this method, they report that they have already
dropped the false alarm rate by 90% - that is, there are one tenth the
"dry holes" - from conventional mag and flag operations. They also say
that they can detect ordnance of a given size deeper than with other
techniques.

Their success has not yet been widely recognized. I don't know whether
it has yet been subject to independent review. The approach makes a lot
of sense, however, and I am confident that they will soon be able to
prove their claim.

The Army has already hired researchers to work on a more sophisticated
post-processing system, and they are likely to achieve even better
results. But I fear that improvement will be incremental. Thus far, the
Army scientists have been improving their capability through better
field work and processing algorithms. However, they haven't made much of
an effort to understand why different types of ordnance, at different
depths, in different locations and in different soil, respond to the
sensors is a specific way. Digital geophysical mapping will work best
once scientists understand why each type of ordnance, in each specific
medium, generates a specific electromagnetic signature or other signal.

Still, UXO detection and discrimination is an area where there is plenty
of room for improvement, largely because until recently no one was
trying to figure out how to solve the problem.

Lenny Siegel
Director, Center for Public Environmental Oversight (AKA SFSU
CAREER/PRO)
c/o PSC, 222B View St., Mountain View, CA 94041
Voice: 650/961-8918 or 650/969-1545
Fax: 650/968-1126
lsiegel@cpeo.org