LOOMING up in front of them was the Temple of the Moon, an ancient,
flat-topped pyramid near Trujillo. As they inched across the dry Peruvian earth
with their high-tech sensors, John Hildebrand and Larry Conyers hoped that
somewhere beneath their feet were the mysterious remains of the Moche鈥攖he
long-lost civilisation that built the temple in pre-Inca times. With any luck,
their instruments would tell them exactly where those remains were buried.
At the end of the afternoon, they examined their data. With growing
excitement, they identified the wispy outlines of dozens of rooms and canals,
and turned the site over to Peruvian archaeologist Santiago Useda for
excavation. But as Useda dug, his delight cooled. Though he found the canals,
almost half the rooms located by Hildebrand and Conyers were filled-in pits left
by pot-hunting looters. How could the archaeologists have been so misled by
their state-of-the-art scanners?
Since the 1950s, archaeologists have tried out all sorts of promising gadgets
designed for geophysical prospecting, including magnetometers, resistance meters
and ground penetrating radar. The problems really begin when they try to
interpret what these scanners reveal. 鈥淎n infinity of things could give you the
same signal,鈥� says Ralph von Frese, an archaeo-geophysicist at Ohio State
University in Columbus. 鈥淵ou don鈥檛 get something back that says, `This is an
arrowhead鈥�. You get a bunch of distortions and disturbances and you have to sort
through them to determine if you鈥檙e looking at a real target that鈥檚 worthwhile
digging for.鈥�
Advertisement
It has to be worthwhile. Digging takes time and time is money. Even a single
trench 1 metre wide and 3 metres long can take days to excavate. And all the
while, says von Frese, 鈥測ou don鈥檛 know if what you鈥檙e after is a sacred object
or just a tin can鈥�. That ambiguity has left many archaeologists reluctant to
replace their spades with high-tech scanners.
But many people are anxious to make these geophysical surveys more than just
an educated guess. If the ruins of settlements and towns could be mapped in
enough detail, then digging might not be necessary at all. Or archaeologists
might be able to answer a question about an ancient civilisation simply by
excavating, say, one room in a particular house, leaving the rest of the site
undisturbed.
John Isaacson, an archaeologist at the US Army鈥檚 Construction Engineering
Research Labs (CERL) in Urbana, Illinois, aims to resolve the problems with
geophysical surveys once and for all, but it鈥檚 a tough task he has set himself.
Isaacson and colleagues from CERL and the University of Illinois in
Urbana-Champaign have just completed the world鈥檚 first archaeological
test-bed.
Dubbed the Controlled Archaeological Test Site (CATS), it is seeded with
features and artefacts typical of archaeological sites in the American Midwest.
If researchers can know in advance what sort of thing they are looking at and
precisely where everything is buried, they should be able to sort out the
meaningful signals from mere background noise. They will be able to evaluate new
sensors under controlled conditions and eventually, CATS may help
archaeogeophysicists find buried treasure without turning over a single clod of
earth.
Up against the law
Isaacson got the idea for CATS from his job as cultural resource manager with
the US Army. He has to make sure that military construction projects comply with
laws passed to protect archaeological sites. But he faces a daunting task.
鈥淎nything over 50 years old has to be evaluated,鈥� Isaacson says. 鈥淭here is lots
of land. There is just a huge problem.鈥� The Army controls nearly 12 million
acres, and surveying land almost always means digging test pits. Excavation can
cost up to $30 000 a site.
Archaeologists in Europe have had great success with geophysical techniques
for detecting things like walls, roads and hearths. 鈥淵ou can hardly miss them,鈥�
says Armin Schmidt, a lecturer in archaeological geophysics at the University of
Bradford. They have a bewildering variety of techniques at their disposal, from
ground-penetrating radar, which beams radio waves into the soil and picks up the
reflections from buried features, to magnetometers which can detect fired
pottery and bricks as well as hearths and kilns. This is because heat from the
cooking fires make the atoms in these materials weakly magnetic. House
foundations resist the flow of electrical current more than the surrounding
soil, so these can be detected by measuring the resistance between two
electrodes stuck into the soil. But things haven鈥檛 gone as smoothly in the
Americas, where the remnants of towns, villages and camps left by older, more
indigenous civilisations contain ephemeral features鈥攎ud walls, burnt wood
or the stamped-down soil of house floors. Unlike the remnants of Greek temples
or Roman aqueducts, these features are much harder to distinguish from the
surrounding soil. As a result, the notion of geophysical prospecting hasn鈥檛 won
over many hearts and minds in North America. 鈥淧eople say: `Joe at the site next
door tried it and it didn鈥檛 work鈥�,鈥� says Hildebrand, who is based at the Scripps
Institution of Oceanography in La Jolla.
Surgical precision
Isaacson realised that to get more out of geophysical surveys, archaeologists
needed a way of knowing how buried objects were likely to appear to their
sensors. This would let them swiftly interpret the information. Only then could
they rely on these techniques to tell them where and what the objects were. And
it would help if the test didn鈥檛 involve burrowing into real archaeological
sites. 鈥淭here needs to be some sort of scientific bench test, and that just
didn鈥檛 exist,鈥� Isaacson says. 鈥淭here was just no place you could go to try new
补辫辫谤辞补肠丑别蝉.鈥�
His solution was to create his own archaeological test site. By starting with
familiar objects such as bones and pots in precisely mapped locations,
researchers would already know exactly what they were looking for before they
turned their instruments on. With funding from the Center for Preservation
Technology and Training, an arm of the National Park Service in Natchitoches,
Louisiana, Isaacson and University of Illinois geophysicist Tom Riley launched
the project as a one-year course in experimental archaeology. After some
background research, the team settled on a few basic items and set about
鈥減lanting鈥� them.
Today, CATS is an unremarkable acre of land northwest of the University of
Illinois campus. A metre or so beneath the surface, the team has made its own
history. They built packed-earth house floors with storage and fire pits. They
built a burial mound with log-lined crypts containing the remains of two pigs.
There鈥檚 a grid of clay bricks buried at different depths. There鈥檚 an earth oven,
鈥渞ipened鈥� by daily meals of chicken and yams cooked in it by a graduate student.
There are also remnants of palisades, trenches and embankments. The location and
depth of each feature is known to within a millimetre.
Researchers at CERL are now planning experiments at CATS that they hope will
lead to new and better ways of 鈥渟eeing鈥� underground. Right now,
archaeogeophysical surveying is not really about seeing anything鈥攄irectly,
at least. It鈥檚 about locating 鈥渁nomalies,鈥� places in the ground where things
change. 鈥淭hat doesn鈥檛 tell you it鈥檚 an archaeological feature,鈥� says Mark
Gilberg, research co-ordinator at the National Park Service鈥檚 Center for
Preservation Research and Training. 鈥淚n the end, you鈥檙e going to have to dig to
really know what it is.鈥�
But Charles Marsh, a materials engineer at CERL, isn鈥檛 satisfied. He is
developing a new way to see with sound waves. Working at the University of
Illinois with experts in acoustics, signal processing and soils, Marsh is
adapting sonar technology developed by the US Navy for torpedoes to create the
archaeologist鈥檚 equivalent of a medical ultrasound scanner. And one of the first
places he鈥檒l test it will be at CATS.
In Marsh鈥檚 laboratory, the metre-wide nose cone from a sonar-guided torpedo
is mounted on a movable platform which runs above a large bin of sand. Inside
the cone is a phased array antenna鈥攁 flat grid of 52 small transducers.
These convert electricity into pulses of sound and listen for the echoes when
these pulses bounce back from a target. The strength of the returning signals
can be converted into a map of 鈥渞eflectors.鈥� In the ocean, the reflectors are
enemy ships; in Marsh鈥檚 lab, they are targets buried in the sand.
William O鈥橞rien, an expert in medical ultrasound at the University of
Illinois, is working with Marsh to turn the sonar into a ground imaging system.
It hasn鈥檛 been easy. The array was designed to operate in the ocean and the
large change in density between the sand and the air above it means that most of
the sound pulses bounce straight back from the surface of the sand. So O鈥橞rien
and Marsh set a children鈥檚 paddling pool on top of the sand box and filled it
with water. The water, they hoped, would reduce the abrupt density change that
the pulses see. And when they dipped the tip of the torpedo in and turned on the
array, the sound passed beyond the surface of the sand to the objects buried
below.
So far they have found the system can detect targets but it can鈥檛 yet produce
clear images of them. They鈥檝e experimented with stove pipes, metal plates and
glass bottles, but even after computer enhancement they still look more like
blobs than real artefacts. Eventually, Marsh says, they hope to build a ground
sonar that can see targets as small as 2 centimetres at a depth of 20
centimetres. And they鈥檒l test it at a location at the CATS site which is already
equipped with a test-bed of acoustic targets鈥攊ncluding paint cans buried
at different depths and orientations and a series of metal rods placed various
distances from each other. 鈥淚t鈥檚 just like an eye chart,鈥� Marsh says. 鈥淚t tells
you how well the array can see.鈥�
Some day it may also be possible to determine the composition of a feature
directly鈥攚hether it鈥檚 made of wood, stone or bone鈥攚ithout disturbing
any soil. Joyce Baird, an environmental engineer with CERL, will test a new
approach at CATS that attempts to do just this. Data from one kind of
sensor鈥攚hether it鈥檚 a radar system, a magnetic field detector or a
conductivity meter鈥攄oesn鈥檛 necessarily narrow down the choices enough to
identify an ancient artefact with any degree of certainty.
So Baird will be combining the signals from different sorts of sensors. Then,
she hopes, the data could reveal more than any single sensor could. 鈥淢ultiple
sensors give you the ability to zero in on what the actual object is,鈥� Isaacson
says. For instance, combining magnetic field detection with ground conductivity
measurements could help archaeologists identify wooden objects. Wood is
invisible to magnetic field detectors but conductivity measurements often
highlight the changes in electrical properties around and within a wooden
object. 鈥淚f you use a couple of techniques, you can eliminate certain kinds of
objects,鈥� she says.
CERL鈥檚 Mark Ginsberg wants to try something similar, this time using nuclear
quadrupole resonance. This technique uses radio waves to excite nuclear
quadrupoles鈥攏uclei with a nonspherical distribution of charges. Once
excited, they re-emit the energy as electromagnetic radiation at a
characteristic frequency. Measure this frequency and you can map the identity
and location of the atoms emitting it.
This technique works well for nitrogen. That鈥檚 why the technology is being
evaluated for locating landmines, and explosives such as RDX and TNT, which
contain a lot of nitrogen. But Ginsberg is more interested in the technique鈥檚
potential for locating calcium, the mineral that builds bones. And if the system
were tuned so its field tweaks calcium atoms, it could detect the remains of
skeletons.
Decaying pigs
CATS has the perfect target for Ginsberg鈥檚 bone detector. The pigs in
Isaacson鈥檚 burial mound have decayed, leaving behind bones and various organic
chemicals in the soil that Ginsberg could use to test his idea. A bone detector
would be very handy in the US right now. Following the Native American Graves
Protection and Repatriation Act, archaeologists have to consult closely with
local Native American groups when disinterring human remains, and they are not
eager to allow what some consider to be desecration of their ancestors鈥� final
resting ground. 鈥淲e don鈥檛 want to inadvertently excavate a human interment,鈥�
Isaacson says. 鈥淲e would much rather know where they are and leave them
补濒辞苍别.鈥�
The CATS project is the first of its kind, but not the only archaeological
test-bed in the works. Researchers and Native American groups are teaming up in
the state of Washington to build one part of a larger test site based at the US
Department of Energy鈥檚 waste storage site at Hanford in Richland. Most of it is
earmarked for training and research into environmental cleanup. But part is also
set aside for pit houses, fires and other features.
These new sites could be critical to making geophysical surveys more than
just an educated guess. 鈥淚n the end, people are hoping to not dig at all when
the information received is already sufficient to answer the question,鈥� says
Schmidt. And if a closer look becomes necessary, archaeologists could excavate a
feature or artefact with surgical precision, leaving the rest of the site
undisturbed.
It might be a while before archaeologists agree to trade in their spades for
sonar scanners. But with one test site completed and another on the way,
researchers and instrument manufacturers will finally have the test-beds they
need to advance the craft. 鈥淚t鈥檚 been a long time coming,鈥� Isaacson says. 鈥淭he
main battle the geophysicists have been fighting for the last 10 or 15 years is
to try and impress upon people that these techniques actually work.鈥�
As for Ginsberg, he鈥檚 eager to take a crack at identifying those pigs. And he
imagines greater possibilities than simply detecting bones. In principle, he
says, he could turn multiple nuclear quadrupole resonance scans of the ground
into three-dimensional images of buried skeletons. 鈥淭here are a couple of
annoying technical details,鈥� he admits, 鈥渂ut it would be damned interesting to
迟谤测.鈥�