**4. Results**

In this section, we present the results of our analyses of LiDAR-derived imagery and geophysical surveys at the Johnston Site. We begin with a comparison of the LiDAR-derived imagery and the 1917 map of the Johnston Site. We then move to discuss the results of the large-scale magnetic gradiometer and magnetic susceptibility surveys before discussing the EMI results. Test excavations of topographic and geophysical anomalies are presented when applicable.

#### *4.1. Analysis of LiDAR-Derived Imagery Compared with the 1917 Map of the Johnston Site*

Overlaying the 1917 map based on E. G. Buck's survey work at Johnston over our LiDAR imagery allowed us to compare Buck's cartography and Myer's descriptions with the current state of topographic relief at the site (Figure 5a). It also allowed us to examine the Johnston landscape for topographic features not included on the 1917 map. Using the three clearly visible mounds at the site as anchor points for georeferencing the 1917 map to real world coordinates, we were able to confidently determine correlations and differences in topographic features between the maps. After redrawing Buck's features as polygons, we could determine where a given mound or embankment should be in relation to topographic features identifiable in LiDAR-derived imagery (Figure 5b).

**Figure 5.** Results of the 1917 Myer/Buck map of Johnston overlay with LiDAR imagery. (**a**) The 1917 Johnston map overlaid on LiDAR imagery with 60% opacity; (**b**) Features from the 1917 Johnston map as polygons on blended LiDAR imagery (Sky-view and color stretched DTM), showing fit of 1917 map overlay.

The first discrepancy we can easily identify is the 1917 bluff line that undulates along the western edge of the terrace landform. In some areas the bluff edge is depicted 60 m or more east of where the bluff is. In others, the 1917 bluff line is depicted west of the current bluff line, now dominated by incised erosional gullies. While the former can be attributed to 'sketching in' the bluff line as a feature relative to other earthen monuments, there may have been areas where the 1917 bluff line was more accurately depicted. If this were the case, we could assume that the erosional gullies present in areas that were depicted as level terrace in 1917 provide a proxy for landform evolution since the map was drawn. These two scenarios are hard to assess given the identified discrepancies between the Buck/Myer map and current topography at Pinson [59].

According to the 1917 map, the northern extent of Johnston supposedly contained Mounds 1–3 and some of the best-preserved portions of the parallel embankments. Our LiDAR-derived imagery shows that Mound 1 is still present at Johnston, conical in shape, but smaller in height and base diameter than described in 1917 (Figure 6; A to A'). Mound 2 is hard to evaluate. A topographic rise that is roughly 0.2 m in height is present in the approximate location of Mound 2 but there is no clear shape to this feature. This may be because much of Mound 2 area has eroded into the gully to the north. However, it is equally possible that this is a relic of natural topography and was interpreted as a built earthen feature in 1917. Determining whether this topographic feature is natural or cultural will require future investigations and maybe excavation.

**Figure 6.** LiDAR-derived imagery (blended image using Sky-view factor and Local dominance) and topographic profiles from the northern portion of the Johnston Site with features from the 1917 map overlaid.

Descriptions of the parallel embankments indicate the best-preserved sections extended south of Mounds 1 and 2. Our analyses of LiDAR-derived imagery and topographic profiles shows that two parallel rises exist directly south of Mounds 1 and 2 (Figure 6; B to B'). These features are both close to 15.5 m wide and 0.3 m in height. They extend approximately 80 m to the south before no longer being clearly visible in elevation profiles or other visualization methods. However, an elevation profile across the purported location of these features in a former agricultural field 300 m south of Mound 1 (Figure 6; D to D') does sugges<sup>t</sup> very subtle parallel rises are present but only 0.04 or 0.06 m tall.

The location of Mound 3 is slightly off from our georeferencing of the 1917 map. However, it is still visible as a conical mound, albeit smaller in height and diameter than the reported dimensions in 1917. Our results estimate that it is roughly 7 m in diameter and 0.3 m tall (Figure 6; C to C'). Mound 3 is currently in a precarious position, situated on a relict finger-like landform directly between two deep erosional gullies.

The central portion of Johnston has the two largest mounds (4 and 5) recorded at the site, in addition to four small conical mounds situated near them (Figure 7). Mounds 4 and 5 are still easily visible on the ground and in our data visualizations. Mound 5 was listed as a polygon in the 1917 map but appears more like a deflated low-lying rectangular mound in our topographic visualizations. The impact of agricultural plowing may have caused, or at least contributed, to this slight change in shape. Currently, Mound 5 is roughly 3.6 m tall and is 39.3 × 45.3 m in size at its base, while the surface spans 22.9 × 25 m (Figure 7; E to E'). Mound 4, the largest mound at Johnston, is a classic rectangular platform mound measuring 5.8 m tall, 57 × 59.5 m at the base, and exhibiting a surface extent of 32.8 × 34.6 m (Figure 7; F to F'). Multiple circular depressions on the surface of Mound 4 are

visible, suggesting either looter attempts to recover artifacts from the mound, or potentially concavities from tree falls.

**Figure 7.** LiDAR-derived imagery (blended image using Sky-view factor and Local dominance) and topographic profiles from the central portion of the Johnston Site with features from the 1917 map overlaid.

On a narrow finger-like residual landform that extends west toward the SFFDR floodplain between Mounds 4 and 5 is the location of two small conical mounds. These are labeled Mounds 6 and 7 in the 1917 map. Mound 6 is the more visible of the two, potentially because it is situated at the center of this landform. Mound 6 is about 5.9 m in diameter at the base, and roughly 0.4 to 0.5 m in height. Mound 7 is harder to evaluate. It is positioned at the edge of the landform, surrounded by erosional gullies. The 1917 map describes it as a 'half oval' or crescent. We can distinguish a roughly 0.3 m rise at the edge of this landform but if this is the remnants of an earthen mound, very little remains intact. The small rise extends 4 m from the edge of the landform and measures roughly 7 m north to south (Figure 7; G to G').

Mounds 8 and 9 on the 1917 map are situated south of Mound 4 and connected via purported earthen 'walls'. We can detect no topographic evidence of earthen walls that connect these three mounds to one another. Moreover, it is not clear that the subtle topographic relief associated with the location of these two mounds confirm human construction. If they are built earthen architecture, then of the two, Mound 8 is the most visible in our imagery analyses. It measures 20 m in diameter and is roughly 0.6 m in height. Mound 9 is much harder to discern in LiDAR-derived visualizations and topographic profiles. Still, a topographic rise is present in the area where the 1917 map denotes the location of this feature. This rise is 20 m in diameter and 0.2 m in height (Figure 7; H to H').

According to the 1917 map of Johnston, the southern portion of the site contains an earthen wall that extends to a conical mound, Mound 10. Like the walls connecting Mounds 4, 8, and 9, no surficial evidence of this earthen wall exists. However, there is a clear topographic rise associated with the purported location of Mound 10 (Figure 8a; J to J'). In this case, our LiDAR-derived visualizations sugges<sup>t</sup> that this rise associated with this area is 0.4 m tall and roughly 25 m in diameter. Like Mounds 8 and 9, it is not clear from topographic analyses alone that this is a human-made feature. We discuss below how we confirmed it as such. An exciting and unexpected discovery in our examination of LiDAR-derived imagery of the southern portions of the Johnston Site comes in the form of a non-mound feature that was not in the historic survey of the site. West of Mound 8 and 9 we discovered a depression encircled with an embankment. This feature is morphologically like the feature designated the "Duck's Nest" at PMSAP (Figure 8a; I to I'; Figure 8b; K to K') [49] (pp. 155–158).

**Figure 8.** LiDAR-derived imagery (blended image using Sky-view factor and Local dominance) and topographic profiles from the southern portion of the Johnston Site with features from the 1917 map overlaid, with comparison of the Duck's Nest at PMSAP. (**a**) LiDAR-derived imagery from the Johnston Site. The profile line from I to I' is absent to better show the circular embankment feature; (**b**) LiDAR-derived imagery from the Duck's Nest at PMSAP set to the same scale, with profile line K to K' removed to better show the feature.

#### *4.2. Gradiometer Results from the Johnston Site*

Our gradiometer survey of Johnston was limited to the central and southern portions of the Johnston Site, where large open spaces could easily accommodate our cart-based instrument. There are also areas with large amounts of metal in the northern portions of the survey area that would make a gradiometer survey unsuccessful. Viewing the full extent of our magnetic coverage brings into focus the magnitude of agricultural plowing and drainage modifications in this portion of the site in the historic era (Figure 9). Intensive plow scars are seen in the data trending north-south, with the exception of the extreme southern portion of the survey area, where scars trend east-west—an indication of a former field boundary. Some plow scars are intense, potentially reflecting early deep-chisel plowing. The long history of agricultural practices seen in the gradiometer data undoubtedly impacted the topography of small features like the purported 1917 parallel embankments and smaller mounds. We

can also identify the remnants of several lightning strikes across our survey area, something more archaeologists are discussing in recent research (see [84]).

**Figure 9.** Gradiometer coverage at the Johnston Site with features from the 1917 map overlaid (in yellow). Numbers denote extant mounds. A clear example of a lightning strike is present in the southern extent of the image.

We focus our discussion on results from selected magnetic features examined in test excavations that offer insight for our research questions outlined here. To begin, in the northern half of our gradiometer survey we were able to delineate numerous clusters of pit features denoted by rounded and spatially distinct magnetic highs (Figure 10a). Limited test excavations of a sample of these features have shown that they are either hickory nut (*Carya* sp.) roasting pits or storage pits with little to no cultural material but are filled with dark organically enriched material. Several pits are aligned in a row to the north of Mound 4, suggesting not only the spatial contemporaneity, but potentially association with Mound 4 as well. Test excavations of one pit in this group revealed a dark fill but no diagnostic cultural artifacts to associate the feature with a specific time of use (Figure 11a). We have several charred botanical samples that have been submitted for AMS 14C dating, but those dates are the focus of another article we are preparing.

**Figure 10.** Gradiometer coverage north of Mound 4 at the Johnston Site with features from the 1917 map overlaid (in yellow). (**a**) Birds-eye view of gradiometer data north of Mound 4 exhibiting magnetic features discussed in the text; (**b**) Close-up of gradiometer data from areas around Mound 5 depicting the summit of Mound 5 and the buried ditch enclosure associated with the mound.

A large irregular feature measuring nearly 11 m in diameter is visible in the northeastern portion of our gradiometer data. Test excavations revealed a deep (>1 m) pit refilled with comparatively dark organic sediment (Figure 11b). Unfortunately, this feature contained no diagnostic artifacts to denote when the feature was used or refilled. In between the area where parallel embankments were mapped in 1917, and where the potential 'structure A' was situated (evidenced by burned daub present in plowed soil), we discerned the magnetic remnants of an isolated lightning strike roughly 12 m in diameter. Soils below plow zone were examined through a test excavation in this area and were documented as a reddish clayey sand. Plow scars cutting into this reddish horizon were visible below the plow zone. Therefore, it is possible that the structural feature 'A' mapped in this locale in 1917 was the result of a lightning strike. However, further research must be conducted to confirm this hypothesis. It is interesting to note that a noisy area of magnetic gradient borders the lightning strike to the west and southwest. This dense area of subsurface magnetic features likely represents a concentrated area of past human activity. Additional areas of seemingly 'noisy' magnetic phenomena are represented by non-patterned high and low magnetic features visible to the northwest of the locale inside the parallel embankments mapped in 1917. These can be characterized by high magnetic susceptibility due to the intensity of the response to plowing in this area. No magnetic response correlates with the location of the parallel embankments, leaving us wondering whether subsurface remnants of these features still exist.

**Figure 11.** Test excavation of gradiometer features identified in the northern portions of the survey area. (**a**) Profile of a pit in the row situated north of Mound 4. Profile shows dark feature fill absent of cultural material beneath plowzone; (**b**) Photograph of nearly black feature fill from the large deep pit in the northeastern portion of the survey area; (**c**) Planview of the enclosure ditch beneath Mound 5 prior to full excavation.

Gradiometer results associated with Mound 5 are particularly intriguing. Our results identify a rectangular platform shape to this mound that contradicts the 1917 descriptions of the mound as a polygon. This discrepancy is likely the result of the intensive plowing at the site. The summit of Mound 5 exhibits numerous magnetic features potentially associated with platform activities. A lightning strike is visible on the eastern flank of the mound. However, most interesting to this area is a small ditched enclosure that we identified under the northern flank of Mound 5. This feature is morphologically like those identified in the Middle Ohio Valley and associated with Hopewellian ritual activity. The ditch encloses approximately 185 m<sup>2</sup> and exhibits a magnetically enhanced feature 3 m in diameter at its center. A test unit cross-cutting a small section of the ditch confirmed this interpretation and revealed the ditch contained burned earth and dense amounts of charcoal, suggesting the termination of the enclosure was intentional and likely occurred prior to the construction of Mound 5 (Figure 11c).

Gradiometer results from the fields south of Mound 4 raise additional questions about the indigenous landscape. Gradiometer coverage over Mounds 8 and 9 are noisy and reflect isolated higher magnetic susceptibility in these areas, visible in the non-uniform magnetic characteristics of the plow scars between the mounds (Figure 12a). However, no clear association is visible between the mound locations and our data, leaving questions to whether the mounds mapped in 1917 are now less visible due to plowing, or if they are relicts of natural topography. There are no clear clusters of pits south of Mound 4 like we see north of the mound. However, some individual pits can be discerned southeast of Mound 4. An area of low magnetic gradient is present in the southern field with four potential pits situated at the center of the area (Figure 12a; Low nT Area). We mapped numerous lightning strikes across this open area south of Mound 4.

**Figure 12.** Gradiometer coverage south of Mound 4 at the Johnston Site with features from the 1917 map overlaid (in yellow). (**a**) Birds-eye view of gradiometer data south of Mound 4 exhibiting magnetic features discussed in the text; (**b**) Close-up of gradiometer data from areas around Mound 10 depicting the shape of the monument and internal subsurface features; (**c**) Close-up of gradiometer data from the circular feature southwest of Mound 10 depicting associated high magnetic features.

The area mapped historically as Mound 10, which has no topographic relief, was among the significant outcomes of our assessment of the magnetic data. Here we identified a rectangular feature and a round pit, both tested in a 1 × 10 m long trench (Figure 12a,b). These results sugges<sup>t</sup> there was a mound present, but it was likely rectangular and not round or conical. The magnetic response to Mound 10's base dimensions may reflect magnetically enhanced topsoil that eroded over time to the bottom of the mound, revealing a 27 × 20.5 m base dimension.

Test excavations confirmed the presence of at least two different mound fills related to Mound 10's construction and identified a small basin shaped hearth under the mound fill. Analyses continue to assess how these features relate to one another. (Figure 13). A small clay cooking ball was recovered from the mound fill. Artifacts like these imply a Late Archaic (ca. 4000–1000 BC) presence on the site [85]. Additional excavations and a robust radiocarbon dating program is needed to confirm the construction age of Mound 10.

**Figure 13.** Profile of test excavations at Mound 10 indicating intact construction fills and sub-mound hearth.

Southwest of Mound 10 we identified a potential circular post structure 10 m in diameter (Figure 12c). This feature appears similar to other post-enclosures or paired-post structures in the Middle Ohio Valley [2,22,36,86–91]. We have ye<sup>t</sup> to test this feature with excavations; however, the exterior of the circular feature appears to have numerous possible pits associated with it. A central magnetic feature is also present within this circular structure. This may be a central hearth, a depositional or refuse feature, or a central post. Excavations should clarify the nature of this structure.

To summarize, the results of our gradiometer survey at Johnston revealed numerous subsurface magnetic features that exhibit no topographic relief. The identification of such features allows us to discern activity areas across space in ways LiDAR-derived imagery do not. The gradiometer data have little correlation with some of the built features outlined on the 1917 map, such as the parallel embankments north of Mound 4 and the 'streets' or 'walls' that connect Mounds 8 and 9 to Mound 4. Moreover, there is little evidence that Mounds 8 and 9 exhibit clear magnetic signatures, leading us to question the nature of their existence. The gradiometer data over Mound 5 reveals the original shape of the mound to be rectilinear and shows that a uniquely different monument, a small ditched enclosure, preceded its construction.

#### *4.3. Results from Large-Scale Surface Magnetic Susceptibility Surveys at the Johnston Site*

Our large-scale surface magnetic susceptibility survey at Johnston was intended to complement the results of our gradiometer and LiDAR imagery survey and analyses. By examining the magnetic nature of the near-surface across the site, we hoped to better understand the accuracy of the 1917 map, interrogate the noisier portions of the gradiometer survey, and examine areas of the Johnston Site that we were not able to survey with our cart-based gradiometer or that contained large amounts of metal.

The results of our magnetic susceptibility survey exposed multiple areas of high susceptibility likely related to past human activities. We consider these 'activity areas' in the broadest sense of the term. North of Mound 4, the highest readings and largest activity area spatially coincides with the western opening in the parallel embankments from the 1917 map (Figure 14b). Probably not coincidentally, this entrance area leads to one of the springs denoted on the 1917 map. This suggests either a long history of pre-Contact indigenous people accessing this natural feature, or a short and intensive period of activity around the spring. Other large activity areas north of Mound 4 identified by spatially-distinct areas of high magnetic susceptibility correlated with the 'A' and 'B' purported structures inside the parallel embankments on the 1917 map, as well as an area situated along the tree line to the west of these structures. While we do not have gradiometer data over the activity area associated with the northern-most structure 'B' on the 1917 map, we do have comparative gradiometer data to assess correlations between the two magnetic datasets and the activity area associated with structure 'A' on the 1917 map and the area to the west at the tree line. The activity area associated with structure 'A' exhibits a large lightning strike right over the mapped structure, in addition to an area of elevated background gradiometer readings and isolated magnetic highs likely related to archaeological features (e.g., posts and pits) (see Figure 10a). The activity area at the tree line exhibited what we characterize as high levels of 'noise' in the gradiometer data, stemming from intense plow scars and numerous isolated high and low magnetic features (see Figure 10a). Therefore, we consider both areas as important locales exhibiting strong evidence for past human use of the Johnston Site. Moreover, the correlation with highly magnetic plow scars in this area indicates that highly magnetic plow scars elsewhere at the Johnston Site, as well as at other sites in the eastern U.S., may be an indicator for areas that have enhanced magnetic susceptibility from pre-Contact human occupation [29].

**Figure 14.** Surface magnetic susceptibility data from the Johnston Site. (**a**) Location and intensity of magnetic susceptibility readings at Johnston; (**b**) Gridded raster (kriging, 2.2 m cell size) of magnetic susceptibility data classified into 16 natural classes.

Magnetic susceptibility data over Mounds 4 and 5 exhibit high readings associated with the mounds themselves and with areas directly adjacent to them. For Mound 5, high levels of magnetic susceptibility extend to the west of the current mound boundaries. This may be the result of historic plowing that has displaced mound fills westward, or these data may be indicating o ff-mound activity areas. Magnetic susceptibility data around Mound 4 show high values extending to the northwest and southeast of the mound. These values may represent erosion of mound fill in these directions. We do not have the gradiometer coverage over the northwestern areas of Mound 4 for comparison, but data to the southeast exhibits highly magnetic plow scars similar to the activity area at the tree line in the northwest portion of our magnetic survey. However, in this locale there are not as many isolated magnetic features in our gradiometer data, lending support to the hypothesis that high magnetic susceptibility around Mound 4 may be related to eroded mound fills, and the spreading of those sediments during historic plowing.

A series of magnetic susceptibility readings in tree cover over Mounds 6 and 7 show good correlations with the topographic remnants of these features and high magnetic susceptibility values. However, we also identified an oval area of high magnetic susceptibility in between the mounds that sugges<sup>t</sup> a subsurface activity area or cluster of archaeological features.

South of Mound 4 our magnetic susceptibility survey shows a large area of moderately high readings. This includes the area over Mounds 8 and 9; however, a more isolated area of moderately high magnetic susceptibility is associated with Mound 8. Two isolated areas of high magnetic susceptibility lie south of where Mounds 8 and 9 were mapped in 1917. Like our survey results north of Mound 4, these two areas correlate with plow scars in the gradiometer data that appear more magnetic than those elsewhere. Therefore, we consider this as good evidence for activity areas.

Magnetic susceptibility data associated with Mound 10 was moderately high only north of the mound remnants identified through the gradiometer data, despite an increased sampling density in this area. Highly magnetic plow scars are present in this area as well, suggestive of an o ff-mound activity area associated with Mound 10. Low to moderately low magnetic susceptibility readings are correlated with the circular feature west of Mound 10 in the gradiometer data. This may indicate that this feature was not used for a long period of time or peripheral to more intensive activity areas in the core of the site.

The results of our large-scale surface magnetic susceptibility survey identify areas at Johnston that were likely used by humans but were absent in other remote sensing imagery (surface or geophysical). As such, it provides an important supplement to understand the spatial relationships between human activities and monumental architecture at the site.

#### *4.4. Results from an Electromagnetic Induction Survey of Mound 8*

After assessing the LiDAR-derived imagery, gradiometer data, and surface magnetic susceptibility data for the Johnston Site, we were still unsure of whether a few of the small conical mounds at the site were indeed pre-Contact indigenous mounds rather than relict topography. However, we only had the opportunity to survey one with a slingram EMI meter, so we decided to focus on Mound 8.

Mound 8 has four datasets. The magnetic susceptibility data from the 0.5 and 1 m coil separations (ca. 0.3 and 0.6 m depth) show parallels with the gradiometer and surface magnetic susceptibility data. In data from both depths, this includes higher background magnetic susceptibility around the locale where Mound 8 was mapped in 1917 and the identification of isolated magnetic features that likely represent intact subsurface archaeological features associated with the topographic rise we can identify in the LiDAR-derived visualizations (Figure 15a,b). The data from the 0.5 m coil shows a transitional arc of high and low magnetic susceptibility in the southeastern portion of our survey block that mirrors the bend of Mound 8 as it was mapped in 1917.

**Figure 15.** EMI data from the Mound 8 locale with the 1917 features overlaid in red. (**a**) Magnetic susceptibility data from the 0.5 m coil separation; (**b**) Magnetic susceptibility data from the 1 m coil separation; (**c**) Conductivity data from the 0.5 m coil separation; (**d**) Conductivity data from the 1 m coil separation.

The conductivity results of our EMI survey over Mound 8 depict no clear high values correlating with the complete coverage of the supposed Mound 8 location (Figure 15c,d). A high conductivity pattern is common among plowed down earthen mounds in the eastern U.S. because the remnants of clay-rich sediments used to build mounds often retain more moisture than surrounding soils (see [78,79,87]). However, in both the 0.5 and 1 m coil datasets (ca. 0.75 and 1.5 m depth), a high conductivity feature is present and spatially centralized over the topographic rise denoted as Mound 8. This may sugges<sup>t</sup> that intact mound remnants are located here and are retaining more moisture, or there is a centralized non-mound feature under this 'rise'. In either case, the sum of data in this locale supports an interpretation that Mound 8 is anthropogenic, although this needs to be confirmed through excavation.
