"Small Scatters of Generally Unobtrusive Artifacts"

When searching for an image for this post, I came across several works by E.B. Banning applying search theory to archaeology:

  • Sweep widths and the detection of artifacts in archaeological survey. (2011) [Science Direct]
  • Detection functions for archaeological survey (2006). [JSTOR]
  • Archaeological Survey (2002 book). [Google books]

Now what would archaeologists be doing with sweep widths?  Looking for nails, shards, and other small objects in the soil.  What they nicely call "small scatters of generally unobtrusive artifacts on the surface".

In WiSAR we call them clues.

The abstract sounds a familiar lament about measuring detectability:

Yet archaeologists have expended relatively little research on the factors that affect detection of such artifacts or the impacts they have on the reliability of surveys.

Therefore:

We employ mock survey of plowed fields ‘seeded’ with a variety of artifacts in order to evaluate the effectiveness of pedestrian survey (fieldwalking) with respect to search time and transect spacing. ....

That is, they did a sweep width experiment.  They placed small nails, glass shards, ceramic shards, and tokens into three small plowed and gridded fields (which had no prior objects!).  Then they measured many things including:

  • Detection probability with time
  • Detection as a function of lateral range
  • Detection by time of day and sun angle
  • Detection by artifact type and search speed

I thought it might be possible to extend these results into cases where searchers are looking for small clues.  But there are two obstacles.

First, the results show that sweep width for such items varies greatly.  Consider this graph:

Banning et al 2011. Half lateral range curve for cluelike objects.

For half lateral-range curves, the sweep width for each object is twice the area under that object's curve. There's a very big difference from small nails up to pottery shards. Here's a graph differentiating small and large nails (I don't recall why the small-nail curve is slightly different from above):

Banning et al, Figure 13. Lateral range curves for nails.

Second, surveying archaeologists travel very slowly. These results were obtained with an average search speed of about 8 meters per minute". That's right, about half a kilometer per hour. They might spend 20 minutes in each 25m x 50m plot. Speeding up a bit reduced the sweep widths noticeably.

OK, so... searching for such small clues is really hard. At usual search speeds, we should expect not to see them.

But it's also interesting to see how broadly applicable search theory is, and how it's being used for different but related problems.

Author: ctwardy

Charles Twardy started the SARBayes project at Monash University in 2000. Work at Monash included SORAL, the Australian Lost Person Behavior Study, AGM-SAR, and Probability Mapper. At George Mason University, he added the MapScore project and related work. More generally, he works on evidence and inference with a special interest in causal models, Bayesian networks, and Bayesian search theory, especially the analysis and prediction of lost person behavior. From 2011-2015, Charles led the DAGGRE & SciCast combinatorial prediction market projects at George Mason University, and has recently joined NTVI Federal as a data scientist supporting the Defense Suicide Prevention Office. Charles received a Dual Ph.D. in History & Philosophy of Science and Cognitive Science from Indiana University, followed by a postdoc in machine learning at Monash.