| 3 minute read |

Details are sketchy, as the saying goes, but the rumor is that the US bombing data from SE Asia was accidentally discovered in the basement of a data clerk, when it was being cleared out after he died. Whatever the real story is, this data is still very topical, some 40 years after the fact.

For example in 2006, two researchers showed that this database revealed that bombing in Cambodia was much heavier than previously thought. Another recent study by Cornell researchers highlights the ineffectiveness of this kind of bombardment. This is all the more pertinent when one considers that many types of un-exploded ordnance that remain in these areas are still highly dangerous and pose an ongoing threat to people living in there.

Despite the probable incompleteness of the data, the US National Archives and Records Administration (NARA) in Washington DC compiled the data from “US Department of Defense contemporaneous records of combat activities” and have made it available to Governments of Lao PDR. Cambodia, and Socialist Republic of Vietnam by DSCA-HAMA “in support of humanitarian demining.”

In total, the database contains information on some 16 million sorties flown between October 1st, 1965, and May 15, 1975 – a month after the Khmer Rouge took Phnom Penh, and just two weeks after the fall of Saigon to the Việt Cộng. Each sortie comprise a number of aircraft carrying a number of ordnance (quite a number in the case of the B52). The data itself is quite detailed and includes the aircraft type, number of ordnance, type of ordnance and BDA (Bomb Damage Assessment).

Here is a typical record from the database:

Visualizing the data poses several challenges including the sheer number of bombs dropped and also changes over time. A typical plot of all points looks something like this:

Therefore, it seems logical that bomb density (weight per unit area) would be the way to go. However, the metadata accompanying the database that the load, in pounds (Lbs), “should not be considered accurate.” This is clear in some records where the recorded weight is impossibly high. So time for some data cleanup.

First, I summarized the data by ordnance class e.g. “MK-82 500 LB Bomb”, and determined the lowest weight recorded for that ordnance. Then I cross checked the ordnance description with the weight and made corrections. e.g. a bomb with the description “MK-82 500 LB Bomb” weighs 500Lbs (give or take). For more cryptic names, e.g. “CBU-28/BLU438”, I searched the internet to find the weight of that ordnance.

Once the correct weights were assigned, I then joined the data back to the original table, multiplied number of bombs by weight and voila! Lbs per sortie. This is then converted to Lbs per square mile. You can view the density map here.

There are some quite interesting patterns in the data, for example, bombing in Laos is often concentrated along roads and trails, especially the Ho Chi Minh Trail(s).

You can download the shapefile of this data here. There is a metadata description in the enclosed text file.