Name the missing

Forensic anthropologist John E. Byrd envisions a data warehouse that would help identify decades-old skeletal remains of missing U.S. service personnel.

Olivier Douliery

Army researchers use a data warehouse to uncover identities of lost soldiers

Old bones speak to John E. Byrd, and he is building a data warehouse to help them divulge their names.

Byrd, one of 31 forensic anthropologists at the Army Central Identification Laboratory, wants to close as many cases as possible of U.S. service members missing in action over the last six decades. Despite its Army name, the lab is at Hickam Air Force Base, Hawaii, and draws its staff from all branches of the military.

Some of the lab's toughest cases involve commingled skeletal remains, which are difficult to sort after years of burial or weathering. A data warehouse could help sort and match the remains to their individual identities.

Founded 30 years ago to identify remains from the Vietnam War, the lab has since extended its mission to MIAs from World War II, the Korean conflict and the Cold War.

Site recovery teams of 12 to 14 researchers go to crash and battle sites in Southeast Asia and the Pacific islands to excavate evidence as carefully as archaeologists dig up ancient ruins. A typical mission takes 30 to 45 days.

Byrd said they especially look for military personnel who were declared dead without further details. Even nearly six decades after World War II, there are still family members who want to know what happened to their brothers, fathers or uncles.

The lab has occasionally identified civilians who were missing and presumed killed, such as CIA operatives behind the Iron Curtain and a few journalists lost in Vietnam.

'Every day is Memorial Day' at the lab, Byrd said.

The process of identifying remains of people dead for several decades is more complicated than, for example, finding the name of a car-crash victim. Byrd and his colleagues prefer to examine complete skeletal remains, but sometimes they have only fragments of bones and teeth.

Several pieces

They never rely on a single piece of evidence for a positive ID, Byrd said. There must be several pieces of evidence such as mitochondrial DNA, bone size, teeth that match dental records and artifacts found at the scene.

'We do digs the way we used to when I was a professor,' said Byrd, who taught at East Carolina University in Greenville, N.C., before he joined the Army lab as a civilian scientist in 1998.

With crime-scene tape and string, recovery teams divide up an aircraft crash site into a grid and study each segment. Positions of bones and personal effects are pieces of the complex puzzle.

Sex, age, race and stature'four key anthropological markers'are all documented for U.S. military personnel, Byrd said. For soldiers of the Vietnam era, forensic odontologists can compare teeth with antemortem X-rays. World War II and Korean War dead, however, rarely have X-rays in their dental records, so researchers must rely on dental charts, which are less precise.

Testing mitochondrial DNA, which comes from a different part of each human cell than nuclear DNA, is the latest clue to the puzzle. Mitochondrial DNA is less fragile than the nuclear type over years of weathering or water immersion.

The Air Force DNA Identification Laboratory in Rockville, Md., analyzes the DNA samples.
Researchers extract mitochondrial DNA from weathered skeletal remains and hair and compare the sequences with those from maternal-line relatives of missing service members. But Byrd said mitochondrial DNA matching is expensive, time-consuming and not unique to the individual, as nuclear DNA is.
From 1990 to 1994, North Korea gave the U.S. government 208 boxes of human remains alleged to be American personnel missing in action during the Korean War, lab spokeswoman Ginger L. Couden said. The North Koreans claimed that the remains came from several battlefields, prison camps and aircraft crash sites.

The 208 boxes, however, do not necessarily represent 208 individuals, Couden said. Remains are commingled, and the North Koreans failed to pass along forensic information from the recovery scenes.

When researchers have a small group of remains'say, from a downed helicopter with three crew members'they spread the remains on a table and separate the individuals. Bigger groups make the table method impractical.

Data entry

Byrd said his data warehouse would collect shape and size data from bone and tooth specimens, DNA test results, crash-scene observations such as rust stains and dog tags, personnel data and dental records of missing service members. Then the search software could find matches and patterns in the huge collection beyond what human researchers can see on their own.

The Army lab now uses an Oracle8i database, running under Unix, to house the center's Automated Recovery and Identification System, said Gary Stephens, the networking project director. CARIS contains antemortem dental records of many missing personnel. Images of scanned documents related to the CARIS data go into Convera RetrievalWare FileRoom from Convera Corp. of Vienna, Va.

Some of the researchers maintain small Microsoft Access databases linked to CARIS and use Microsoft Excel spreadsheets to prepare data for uploading to CARIS, Stephens said.

In perspective

The warehousing approach lends itself to step-by-step documentation and minimizes reliance on subjective opinions from forensic anthropologists, Byrd said.

It also can reduce the time that remains spend in the lab before they are sent to family members for burial.

The lab wants access to other organizations' relevant data, but that is still in the discussion stage, Stephens said.

Byrd has been consulting with Steve Ousley, a Smithsonian Institution anthropologist who is also a computer programmer and co-author of a widely used software application, FORDISC 2.0, for classifying adult remains according to body measurements.

Anthropologists have studied collections of skeletons to devise ways of inferring overall height from the lengths of arm and leg bones.

The lab researchers use a number of statistical applications'such as regression, variance and cluster analysis'from SAS Institute Inc. of Cary, N.C., Byrd said.

Some scientists also use FORDISC, available from the University of Tennessee, and statistical programs from SPSS Inc. of Chicago, Stephens said.

The Army lab recently received the 2003 Enterprise Intelligence Award for Government from SAS for the work it has done.

'We hope these methods for sorting commingled remains can send more missing people home,' Byrd said.

From the remains that the North Koreans released in the early 1990s, the lab has identified 10 individuals, Couden said. It continues to analyze the other remains in hopes of identifying others.

Reader Comments

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Please type the letters/numbers you see above