Spatial Distribution of the First 4,800 Eastern Woodlands Radiocarbon Dates

4/13/2016

I've been working on compiling a database of radiocarbon dates from the Eastern Woodlands. While my interest in doing this is mainly driven by my own research goals (the driving force right now is my desire to be able to discuss the possible abandonment of portions of the Southeast at the end of the Early Archaic period in a symposium I'm pulling together for the 2016 SEAC meetings this fall), I know these data will be useful to others as well.

Here is a GIS map showing the counts by county of the first 4,870 dates that I've gotten plugged in:

I started with a spreadsheet sent to me by Shane Miller and combined it with data available online from PIDBA (both of those sources were focused on dates from early sites across the east), the Louisiana Division of Archaeology, the Virginia Department of Historic Resources, the Kentucky, Ohio and West Virginia Radiocarbon Database (Cultural Resource Analysts, Inc.), a list of Tennessee radiocarbon dates (Tennessee Archaeology Network), and A Comprehensive Radiocarbon Date Database from Archaeological Contexts on the Coastal Plain of Georgia by John A. Turck, and Victor Thompson. After combining these datasets into a single database (which took some effort), I did a sweep to eliminate redundancies and flag obvious errors. I added a column for county and linked that to a separate table of county names attached to UTM coordinates of the approximate center of the county. That lets me query the database to spit out a table containing a listing of dates and associated UTMs. I imported that into GIS and then did a "join" to count the number of points per county. Voila.

There are still numerous errors and omissions in the database as it currently stands, which is why I'm not prepared to supply the raw data at this point. I've got many dates that are missing key pieces of information (error, site number, county, etc.), and the columns for references are a total mess at this point. As I work through the process of cleaning all that up and trying to fill in blanks, I'll be adding new data. I know of some print publications that will help me fill in some of the large blank areas, and I suspect there are other online or electronic sources of data out there. I've got the UTM coordinates for the counties in most of the Midwest and Southeast (I still haven't done Mississippi and Florida), but I haven't yet started on the tier of states immediately west of the Mississippi River (Arkansas, Missouri, Iowa, and Minnesota) or the Northeast and New England.

John

4/13/2016 01:11:00 pm

I have a general question about the sampling work that leads to plots like this, please take this with a grain of salt I come from a physics background so I may be missing the point.

When I look at this map or some of your Kirk information I find myself wondering if there is just a lot of missing information. For instance with the Kirk's weren't you sourcing the data from known collections? What if the material was still in/on the ground in the places you were seeing them as "missing".

I guess the point is to fill in the information as completely as possible and then see what information is really missing but then how would you know whether something was truly missing or just hadn't been found yet...inference I guess.

Andy White

4/13/2016 01:29:50 pm

The maps that I'm making at this point (both the Kirk maps and the radiocarbon map in this post) are simply to show where the data are coming from - they're not analytical. In both cases, the "holes" in the data are indeed because I just don't have data from those areas yet.

The distribution of Kirk points (and any other kind of point) will have real geographic limits. Until you collect data from an area, though (e.g., by looking at a lot of collections), you don't know if what you're looking for it there is or not. If I want to find the "edges" of the distribution of Kirk points (which I do), I'll need to collect data in areas where they are as well as areas where they're not. If I look at hundreds and hundreds of points from Bobo county in Somewhere state, for example, but see no Kirk points rather than the 5% I've seen in other areas looking at similar total numbers of points, I can make a reasoned interpretation that they're probably just not there. That can always be proven wrong later, however.

John

4/13/2016 01:43:54 pm

Very interesting, thanks!

Joe Fitzgerald

4/15/2016 09:55:34 am

It's too bad that the giant skeletons found by Smithsonian archaeologists subsequently disappeared and cannot be carbon dated and included in the database!

PFERK

4/13/2016 02:54:06 pm

Thanks.... these maps can only be made by dedicated work-a-holics....

Matt Boulanger

4/13/2016 05:38:12 pm

Just emailed you my data for the Northeast and Middle Atlantic. It should help fill in that area.

Andy White

4/14/2016 11:10:57 am

Thanks. I'm working on it now. Delaware is my favorite state, since there are only three counties.

Sara link

4/17/2016 06:10:16 am

Can I make a suggestion?

The orange color scheme looks great until you're trying to differentiate between counties next door. You should make the difference in gradients more obvious.

Other than that - I'm actually kinda fascinated by this - speaking from layman's terms.

Andy White

4/17/2016 06:21:21 am

I'm going to use a wider color gradient (i.e., more divisions) for the next "in progress" map, which should have about 8000 dates. I'll also experiment with multi-chromatic gradients to see if they convey density information well (since the data are not really continuous, like elevation, I'm not sure it will be readable).

Comments are closed.

Spatial Distribution of the First 4,800 Eastern Woodlands Radiocarbon Dates

Archives

Categories