Ashkelon, Israel: 2007 and 2009

We had two field seasons of GPR collection at the complex tel site of Ashkelon, on the southern coast of Israel.  Over 2 years we collected many grids of data, and I am showing just a few of the images here that have helped us understand portions of this complex site. 

This project was funded by the Leon Levy Foundation and Harvard University. Federica Boschi from University of Bologna helped in data collection and processing in 2007 and Yossi Salmon (University of Haifa) and James Conyers in 2009.

In 2007 we collected some long lines on the rampart to the east of the site to look for the Islamic rampart, which is well exposed along the north end of the site.  These are the profiles we collected:

Profiles 8 and 9 showed a nice buried surface parallel to the ground, which we hypothesized would be the Islamic rampart surface.

In 2008 Daniel Master (Wheaton College) and his crew excavated these features with a front-end loader.  Here is what he said about those results:  "As we scraped the section, we noticed that instead of hitting the stone
Fatamid Ramparts, the excavator had cut through at least three ramparts which we could date ceramically to the Middle Bronze Age.  This was utterly baffling to our late period archaeologists.   Did they only build a stone
moat around the gate?  Is there a second wall system farther out?   No other site has a half stone- half earthen rampart like this.  (in contrast, the Bronze Age archaeologists on site were happy to have their theories about
the age and scope of the rampart confirmed.As with the earlier work, it appears that the GPR was able to pick up sensitive differences in the soils.  As before, it was unable to distinguish these soil differences from a stone surface."

With the information from the excavations, it now appears that the GPR profile could see two of these three buried rampart surfaces, which were quite old, and generated from earthen features and not the stone, like the Islamic features to the north.

We also collected two grids within one of the large excavations (Excavation 38) to produce images from the different layers in the tel.  These are the two grids we collected:


 

The reflection profiles are quite impressive and a number of features can be seen in the late Bronze Age layers. The amplitude slice maps show a number if features on the floor, as do the profiles, shown below:

The profiles helped us interpret what was seen in the slice maps. This "small object" in the profile below turned out to be a ceramic pot when excavated:

Other more subtle features such as this hard-packed floor showed up nicely also in profile, as well as in the slice-maps:

In almost all our grids at the site the reflections were so good we could see some features immediately during collection. Sometimes reflection features were visible in "real time" on the screen during collection, such as this profile:

Here I am collecting these profiles at the Basilica.  We found that the columns were all in the wrong place, and had been placed there just to make this "look" like a cool ancient site:

Using the GPR Viewer project we always analyzed our reflection profiles during data analysis:

For instance, when we zoomed on on this buried feature (a column base from the Roman age Basilica) we can see the base for the column that has been partially disturbed when the column fell over, leaving the small hole just below the hyperbola.

We also collected profiles on the beach, looking for the Bronze age harbor, which has never been found.  Here is Federica collecting reflection data with the 400 MHz antennas:

A number of interesting features were discovered on the beach, but it was difficult to interpret them, as we didn't know any information about the stratigraphy.  Here is one of the profiles that appeared to show a truncation into bedrock below the beach sand:

After an analysis of field notes from years past Daniel Master was able to tell us that this bedrock layer was an early Holocene age clay layer, deposited with sea level was much lower and this area was a coastal marsh.  The bedrock "cut" in the profile above turned out to be a backhoe trench that was dug in the late 1980s to test the stratigraphy.

We also collected some large grids in un-explored areas that were VERY complex, filled with modern and ancient materials.  We were able to annotate each of the slices based on an interpretation of the reflection profiles to produce maps of what is below the surface:

We returned to Ashkelon in 2009 to follow up on previous work, and collect more data in areas of interest.  A good deal of time was spent on the beach trying to get a handle on the bedrock stratigraphy, and still look for harbor features. There have been some very interesting changes along the beach in the last 2 years, and it appears that there has been significant salt water intrusion into the sand, perhaps due to beach erosion or the withdrawal of ground water from on-shore.  Also a cut off of sand from the south due to the breakwater constructed down by Gaza may be to blame.  The areas where we collected good data in 2007 were now totally unsuitable for GPR, with attenuation of energy just a few centimeters below the ground.  That attenuation can be seen in this profile:

In one area of the beach where good reflections could still collected in 2009 fresh water filled the sand void spaces and not salt water.  In those fresh water areas super reflections were collected that show a number of bedrock features including the "Kukar" bedrock (cemented aeolian sandstone of Pleistocene age...shown in red in the profile above and below), and the clay layers above it (in blue and green):

Other areas appear to show rubble under the beach sand, likely due to past beach erosion, or perhaps blocks of rubble that have fallen from the tel on the beach.