Ground Penetrating Radar at the Molalla Encampment Site

During our six week stay in Grand Ronde for the FMIA field school, we each had to pick something that interested us for our leadership project. Students focused on the magnetometer, GPS, keeping the public informed of the project through the FMIA Facebook page, and more.  When I first started thinking about what my leadership project would be, I couldn’t decide because there were just too many topics I was interested in.  Then one of the FMIA TAs Ian suggested I focus on ground penetrating radar (GPR) for my leadership project and it seemed like the perfect topic.

First, I started to learn what the does and how it is used in archaeological survey.  The GPR consists of a radio antenna that, while it is dragged across the ground surface, sends radio waves into the ground.  The GPR’s data logger records the time it takes for a signal to be reflected back to the ground surface, measuring the depth and location of each reflected signal (Conyers 2012:28-31). The data logger shows this data in real time on the screen and also records the data to be downloaded and processed later. The Grand Ronde Tribal Historic Preservation Office (THPO) owns a GSSI data logger and carriage with a 400 MHz antenna, which FMIA used as part of our low-impact archaeological methodology.  We were interested in using it to get an idea of what might be underground at the Molalla Encampment, one of the initial settlements on the Grand Ronde Reservation.

GPR grids molalla

FMIA 2016 Molalla Encampment site GPR grids

The Molalla Encampment site filmed from the north looking south. (Video Credit: Katy Leonard-Doll)

In order to do this, we split the Molalla Encampment site into 12 grids and operated the GPR over about five days.  We started by operating the GPR over 20 x 20 m grids but had to modify the dimensions of each grid based on the unique shape of the Molalla Encampment site.  Within each grid we ran the GPR in 0.5 m transects in order to get overlap of the antenna readings.  We also operated the GPR in a zig zag fashion moving north for one transect and then south for another.  When operating the GPR we needed a team of at least three people, one person as the operator and two people moving the tapes every 0.5 meters.  Our team rotated through these positions as we collected data for the 12 grids.  In the data logger we set the GPR to take readings one meter below the ground surface and set the scan unit (the number of reads per meter) to 50 so it would take a reading every two centimeters.  Each team member recorded information such as the file numbers, dimensions, and operational pattern for each grid in our notebooks to make sure we had that data backed up in case any was lost on the data logger.

FMIA 2016 student Katy Leonard-Doll operates the GPR at the Molalla Encampment site.(Video credit: Tiuana Cabillan)

Operating the GPR wasn’t too difficult but when it came to processing the data I had no idea where to begin. Briece Edwards, Senior Archaeologist for the Confederated Tribes of Grand Ronde, walked me through it.  He told me that when viewing the data it was a good idea to come with an idea of what you’re looking for but at the same time not making your eyes see something that isn’t there.  We started by looking at the raw (unprocessed data). Briece said he likes to look at the raw data because you get a picture of all the data before anything useful gets filtered out. With the unprocessed data we played around with different gains which shifts where the zero is up the scale, and the color transforms which change the colors of the grids to make certain data stand out. Also when viewing the raw data, we created a Super 3-D file which makes it possible to view multiple grids next to each other.  For example we created a Super 3-D file which included grids 02-06 and could see them all lined up on top of each other.  To see these grids transitions from the ground surface to one meter down we hit the animation button which slowly moves down the one meter depth centimeter by centimeter.  The bright colored data are referred to as anomalies which indicate where the radio waves hit something in the ground that had a different reading than the radio waves right before it.

Here is an example of an animated super 3-D file of GPR grids 2-13:

In the video, the cursor highlights some interesting anomalies.  In grid 5 (NW corner of grids) there is a straight line running east to west which Briece identified as a drain pipe. (Video Credit: Briece Edwards)

Next, Briece showed me how to do the basic GPR data processing called “easy processing.”  The first step is to move the zero point of the data up and get rid of the surface grass area (which was about the top 18 cm of the data we collected) under the tab “time zero.”  The second step was under the “background removal” tab which filtered out data that wasn’t relevant to what we were looking for.  The next step was under the tab “test and apply filters” but we left the automatic settings alone because they worked for our purposes.  The fourth step or type of processing we used was called “migration” which allowed us to narrow a target in the data.  Once we did these “easy processing” steps we went back and looked at the grids in their 3-D mode again.  Once in this mode, we went to the properties menu and changed the transparency of the grids. This allows you to choose between showing all the depths stacked on top of one another or not which can make certain pieces of data stand out.  In this menu we also changed the background color of the grids from white to black to see if that made the data stand out more on the grids (Conyers 2012:40-43).  What I gathered from doing these easy processing steps is that it is mostly about making certain parts of the data stand out in their 3-D mode to better pinpoint interesting anomalies in the data to investigate further.

When looking at the GPR data, we were looking for right angles in the anomalies that could indicate a possible structure, and we were looking for anomalies that formed a circular feature that probably doesn’t form naturally.  Here are a few anomalies from the Molalla Encampment that Briece found interesting:

This data set is made up of grids 3 and 4.  In this video clip, the cursor highlights interesting anomalies in the NW corner of the grids as well as on the west side of the grids between the 50 and 80 meter markers.  (Video Credit: Briece Edwards)

This second data set is of Molalla Encampment grids 9-11.  The cursor highlights an interesting curved line anomaly in the east side of grids 9 and 10.  The striping seen in grid 9 is probably from plowing or recent mowing of the area. (Video Credit: Briece Edwards)

In the clips above there are some striping areas that upon first glance, I thought were an indication of strong anomalies but Briece told me that these are actually operator error probably from switching operators in the middle of a grid. Everyone operates the machine slightly differently and this can manifest in the data.

In addition to locating anomalies in the GPR data that are of interest, we used the GPR data in conjunction with other methodologies as part FMIA’s low impact archaeological approach, which emphasizes doing the least amount of harm while still gaining the most amount of knowledge.  By using other intensive geophysical survey methods such as gradiometry along with archival research and mapping we can narrow points of archaeological interest before disturbing the ground surface.  We can then find specific places of interest to eventually do surface collection or set up an excavation unit.  For instance, Briece and I noticed an interesting rectangular anomaly in the southwest corner of grid 4 and then looked to other data sets (e.g., gradiometry data, surface collection data, surface topography) to see if they too showed a possible point of interest as well.  We looked at the gradiometer data and saw that there was a large dipole in the same location as the GPR anomaly.  We also looked at an aerial photograph of the area from 1955 and zoomed in on the Molalla Encampment. We saw a few structures located on the site that are not present today.  We noticed that one of the structures is in the same vicinity as a rectangular anomaly in the GPR and dipole in the gradiometer data.

This GPR data set is from grids 2-4.  The first anomaly the cursor highlights in the SW corner of grid 4 and NW corner of grid 3 is the anomaly that corresponds with the gradiometer data as the aerial photograph.  (Video Credit: Briece Edwards)

This is a zoomed-in video of the GPR data set grids 3 and 4 which better highlights the anomaly discussed above. (Video Credit: Briece Edwards)

Data_7.19.2016

This picture is of the gradiometer data set in the same location as the anomaly discussed above which can be seen here located along the west wall of the data set as a highlighted red anomaly.  (Photo Credit: Dr. Sara Gonzalez)

1311422806658

This is a zoomed-in aerial photograph of the Grand Ronde area from 1955. You can see the location of the previous Molalla Encampment site circled in red.  (Photo Credit: Briece Edwards and the Grand Ronde THPO archives)

1955

Zooming in further, the red circle is the area of the anomaly we are interested in. A possible structure the GPR and gradiometer are picking up on is shown in blue.  (Photo Credit: Briece Edwards and Grand Ronde THPO archives).

There is no way to know for sure if the geophysical survey data points to the anomaly and dipole as this structure but this demonstrates how using different methodologies can help to locate points of interest before breaking the ground.  Now Briece, Dr. Gonzalez, University of Washington (UW) graduate student Ian Kretzler, and the tribe can decide whether this anomaly warrants further archaeological investigation.  Using this approach of comparing data from different methodologies while in the FMIA field school in Grand Ronde emphasized to me the value of using intensive geophysical survey methods along with other complementary methods to find specific points of interest not only to archaeologists but to the community who can then decide if excavation is desired or necessary.  Working with Dr. Gonzalez, Briece Edwards, and Ian Kretzler showed me how minimizing harm while maximizing the amount of information gained can be accomplished in a respectful and beneficial way and that something as technology-based as GPR can fit right in to that approach and its accompanying methodology.

Works Cited

Conyers, Lawrence                                                                                                                 2012 Interpreting Ground-penetrating Radar for Archaeology. Walnut Creek: Left Coast       Press.

Leave a Reply

Your email address will not be published. Required fields are marked *