This presentation is part of the online document Acknowledging Landscapes: Presentations from the National Register Landscape Initiative.

Damita Engel: Over the past several years, Metcalf Archaeological Consultants, Inc. has investigated the Knife River Flint Quarry area in North Dakota and expanded understanding of its nature and composition. We have identified a new quarry, developed an initial predictive model of KRF natural deposits and site locational spatial distribution and are now updating that model in conjunction with the North Dakota DOT project.

Because of these investigations, Lynch Knife River Flint Quarry is now a National Historic Landmark and deep testing has expanded our knowledge of the geoarchaeology in the area. This presentation summarizes these investigations and will focus mainly on the landscape challenges and avenues of research used to construct the model and view the Lynch Quarry as a landmark in a wider context.

Knife River Flint, or KRF, is a dark brown, siliceous lithic material found throughout west-central North Dakota. The flint has a distinctive petrography that distinguishes it from other rocks in North Dakota and adjacent areas, and was a prized lithic source for native populations.

The material occurs as pebbles, cobbles and boulders and it is believed to have been eroded from a bedrock layer in the Killdeer Mountains to the northwest of the current deposits. That whole region is known as the primary source area, or PSA.

The primary source area for KRF covers approximately 2,000 square kilometers in Dunn and Mercer County. Native populations intensively exploited this material, including through intentional quarrying. Two major drainages, Spring Creek and Knife River, bisect the area. These meltwater channels served a dual purpose. They contain concentrations of KRF and also provide an attractive setting for nearby camps during procurement episodes.

KRF was possibly one of the most, if not the most, widely traded lithic material in pre-contact North America. Stan Ahler believed KRF was only surpassed in distribution by obsidian from Yellowstone Park in Wyoming. He best summed up the importance of KRF to native populations, “The lithic extraction and tool production activities which occurred at the KRF quarries had a direct and indirect effect on prehistoric peoples who lived in a 3.7 million square kilometer area in North America. The importance of KRF as a key lithic source continued through the entire post-glacial period. The quarries, workshops, camps and other sites in the KRF primary source area provided the ideal focal point for studying and monitoring many aspects of 10,000 years of cultural process and cultural evolution throughout a large part of North America.”

In 1984, Loehndorf, Ahler, and Davidson compiled the proposed National Register district in the Knife River Flint Quarries in Dunn County, North Dakota. The proposed district boundaries are still thought to encompass the core area for KRF quarrying activities by pre-contact populations, which represents the heart of the primary source area.

Unfortunately, resistance from some effected land owners has prevented the district boundary from being formally nominated. Although the primary source area has not officially been nominated for the National Register of Historic Places, two quarries are now protected. Crowley Quarry is a North Dakota State Historic Site, while the Lynch Knife River Flint Quarry became a National Historic Landmark in 2012 and we had the pleasure to prepare that nomination, which was fun. Lynch Quarry is arguably the largest and most durable quarry in the primary source area.

Native people from Paleoindians to post-contact groups intensively exploited the area with two peak extractive periods: Paleoindian, represented by the artifact on your slide right now, and, the Middle Plains Woodland. At first individual groups freely visited the quarries to acquire raw materials for personal use, but later access was apparently controlled.

Large blanks of raw material were produced here, which in turn served as trade goods later on, dispersed through trade networks among villages along the Missouri River and beyond. Tool production saw both flake and blade technologies.

In 2005, the North Dakota Department of Transportation contracted with MAC, my firm, to develop 3 semi-independent models that would assist in discussing possible impacts to cultural resources from highway construction and maintenance within the primary source area.

Predictive modeling is the process by which one or more classes of attributes are analyzed to try to best predict the probability of an outcome. The first model addressed distribution of sites and the influence of KRF in shaping prehistoric settlement patterns or site location. The second attempted to predict settings likely to contain buried archaeological materials or site burials, and the third examined distribution of quarryable KRFs and the geographic extent of the primary source area.

The study reviewed the extant literature and recorded resources, conducted fieldwork involving an intensive sampling strategy, and investigated the geoarchaeology. Geoarchaeology is a multi-discipline approach that uses the techniques and methods of the geosciences to address archaeological questions. At the time of the study, the state historical society of North Dakota had recorded approximately 3300 archaeological resources, sites and isolates both, in or adjacent to the primary source area. The literature review revealed that distribution of inventoried parcels and thus the distribution of known sites was and still is not systematic, as you can see by the great absence of survey blocks in the central part of the study area.

The next step was a pedestrian inventory along the Highway 200 corridor that bisects the area and cuts it in half North to South. The highway runs east-west. Also, selected parcels in the study area which are represented by the black little squares that you see. Try to fill in those gaps in between what we knew. The parcels were selected through sample regions that had not been inventoried to determine if the established primary source area boundary was accurate, and to examine those areas having land forms likely to hold significant KRF deposits, both culturally and naturally.

The parcels also included control areas not likely to hold KRF deposits. Parcels varied between 40 and 160 acres in size with most being around 80 acres. MAC recorded 57 new sites during the inventory, for a total of 700 sites in the study’s archaeological database.

At the same time, two archaeological quarrying samples were taken from these parcels from landforms with sediments with potential for KRF and/or to yield buried cultural materials. Or looking for sites, basically. The results were an archaeological, geoarchaeological database easily translated into GIS layers for model construction. GIS is an online mapping program, for those that don’t know.

The first model, the site location model, we’re going to talk about next.

Resources were grouped into 7 site types: camps, quarries, lithic procurement areas, workshops, lithic scatters, cairns, and other. The camps contained stone circles or fire cracked rocks, which suggested the presence of domestic activities. Both quarry and lithic procurement areas relate to acquisition and reduction of KRF. Both are distinguished by numerous unmodified KRF cobbles, but are differentiated from each other by the depths to those KRF deposits.

Lithic procurement areas have KRF deposits no more than 20 centimeters below the surface, the assumption being that such deposits could have easily been accessed casually and with minimal effort. Quarries, on the other hand, with KRF deposits deeper than 20 centimeters below the surface, support the assumption that obtaining this material would have required more time and energy.

As you can see here, the shaft of the pit lies to the right. I’m sorry for the grainy photo, it was from 1995. Quarry workers would have dug down until they hit the matrix containing the cobbles of KRF, then tunneled horizontally to extract as much as possible. This is definitely a lot more intensive than if it’s just spread across the top of the land form. Typically, quarries have distinct pits, waste piles, and possibly associated workshops.

This is an aerial photograph taken of the newly designated NHL Lynch Knife River Flint Quarry in the heart of the primary source area. As you can see, it has hundreds of pits, some with these distinctive waste piles.

A close-up of what it looks like on the ground. Although the workshop areas haven’t really been identified yet, they would most definitely be nearby. You wouldn’t want to haul this stuff too far. A workshop is defined as a concentrated lithic reduction area, typically located close to a quarry and with lithic material associated with early stage reduction. A lithic scatter is defined by the lack of features or evidence for quarry or lithic procurement activities. The other two categories, cairns and the catchall other category, include unusual site types found in such low numbers that the model wouldn’t really be likely to predict them. For modeling distribution of KRF and assessing its influence on settlement patterns, the quarry and lithic procurement sites were the most critical.

The site burial model is going to be next, and that was predicting the locations of buried sites based on two assumptions.  One, that sites located were attractive, prehistorically, to people. That means they were close to water, had good visibility, close proximity to resources, etc. And then two, that conditions were favorable for accumulation of sediments, something that we would be able to pick up later on in time.

The site burial model was developed to assess potential for buried cultural resources within proposed project areas. Landscape modeling tools included digital elevation models, or DEMs, and the SSURGO soil series maps.

Geoarcheological investigation showed that the position on the landscape is the most important factor governing the potential for sediment accumulation and buried archaeology, while proximity to parent material, mechanism of deposition, and the energy of transportation are additional factors.

During the project, MAC archaeologists developed a feel, or an intuitiveness, for where to find these sediments and buried archaeology, but unfortunately documenting this intuitive feel for site occurrence relative to variables on the landscape proved much more difficult. During discussions about the use of the DEM methods to model position on landscape, the archaeologists decided that capturing the full set of variables would be impractical. Instead they focused on the SSURGO maps, the soils, to determine their adequacy as a basis for modeling. Each map soil series was ranked as its potential to yield buried archaeological materials.

KRF distribution model, the last model. This model is definitely still being refined. The distribution of KRF, raw materials sources, and its relationship to human settlement was based on literature review and distribution of recorded resources, ground truthing, geologic mapping, and also field investigations. A major aspect was assessing geoarchaeology and dating post-Pleistocene formations to determine how these relate to human habitation. This assessment focused on Pleistocene tributaries and glacial channels, and that were be where the KRF was eroding out and flowing down through these waterways during glaciation recession. Holocene sediments in the study area mostly post-dated the Leonard Paleosols which is one of the oldest paleosols to contain cultural materials in North Dakota, which dates between 9900 years and 8800 BP.

The date ranges developed for the terraces along Spring Creek, one of those big tributaries that cut though the PSA, include 5600 BP within the terrace three soils, so it’s a little bit higher, and then going all the way down to 1800 BP at the T1 level, which is a lower terrace closer to the floodplain. Cycle of incision, which means cutting through it, began after 2900 BP and continues today.

In general, previous assessments of mechanisms of deposition and distribution of natural KRF used to define composition and the extent of the primary source area were found to be accurate. Data useful for anticipating context where natural KRF may be found outside of the study area was also gathered. Sources of unmodified or raw KRF include lag deposits on eroded uplands, glacial till, alluvial gravel deposited on the flanks of these meltwater channels, and gravel deposited as bars or fans within the beds of the stream channels themselves. These data lay out the groundwork for a future GIS layer modeling KRF distribution and its relationship to human settlement.

The preliminary end product of the modeling looks very colorful as well as seemingly robust in meaningfulness and useful data interpretation. What you’re looking at here, the red represents the heart of the Knife River Flint extraction area, so that’s pretty high and that works really well. The report summarizing the results of this study is still in draft form, but regardless this predictive model is always supposed to be dynamic, perpetually being a work in progress. As new data came in, these models will have to be revisited and refined. Implementation of the mapping program, Arc GIS program Model Builder, will make updating the models easier in the future, and the models as an active continuous information source will happen instead of a static snapshot of any data we could have at any given moment.

Our work in the PSA continues from the increased demands for surveys related to the energy boom.

In 2011, Basin Electric Power Cooperative contracted for an inventory of a proposed transmission line. Portions of this line run through the study area. The inventory identified 27 new sites, including a new quarry site, which you see in front of you here. The site is located on a hilltop, overlooking the Killdeer Mountains to the west and Lake Ilo area to the southeast. Quarry pits of varying size are evident, kind of hard to see in these photos because it was really tall grass at the time. The site also contains at least one stone ring and a possible effigy.

Other sites in the PSA located included lithic areas and camps. At least one Paleoindian site was also identified. All of this information was fed into the model. We just completed and updated file search for a one mile radius around this proposed corridor for the transmission line, and we have 500-600 new cultural resources that are on file, just to give you a sense of how much work is actually being done in the region right now. It’s kind of exciting if all of that was to be plugged into the model.

Simultaneous with the transmission line project, we are now working on the second phase of the predictor model.

NDDOT issued a contract to conduct further investigations in the study area, focusing on North Dakota Highway 200 between Beulah and Killdeer, approximately 50 miles in length. The highway is just south of and parallel Spring Creek. This inventory identified another 75 sites and counting, since we’re not done with field work yet. The original presumption was that the site density would drop off towards the eastern end as you move further away from the Killdeer Mountains, but several sites have been found in the east end as well. Parsing out why the trends aren’t fitting the expected outcome is even more exciting, as we will hopefully refine the model and look at the various reasons why the actual outcome is different.

In conjunction with the archaeological survey, ten areas were identified for deep probing, using a Giddings rig, seen here, to augment and expand upon the earlier quarrying that was done in 2009 as part of the original model.

The quarrying looked at the soils and parent material to determine how the sediment was generated. Two geomorphological reconstructions were built on the buried site model and the Spring Creek terrace sequence that I mentioned earlier.

We are happy that the radio carbon dates and the geoarchaeological interpretations compare favorably with those assessed during the initial study. To better refine the model as well as round out some of the skewed data, we developed a supplemental form that’s three pages long.

This is the first page of that supplemental form that we created and that we were hoping that could be used by archaeologists to gather more robust geoarchaeological information without overly taxing the archaeologists’ abilities to identify land forms and sediments. I have included pages 2 and 3 too, if you want to quickly run through both of those. Those next 2 slides. I like to call this geology for archaeologists, or geology for dummies, because we don’t know a lot either about that. So basically they’d be able to circle on the map where they were based on that landscape schematic. We have noted the distance to the nearest known quarry as well for easy reference by researchers, so they can use this a little bit easier.

To sum up, energy development and the expansion of infrastructure in North Dakota threatened the rich archaeological heritage in the Knife River Flint Quarry area. But, like all theorems, archaeology projects fund research that would normally not happen, so it is definitely a love/hate relationship. Ultimately, through the refinement of the KRF model and its holistic approach, hopefully we gain the means to ensure that the significant archaeological resources in North Dakota are assessed not only on a site by site basis, but also in the wider context of land and resource use through time.

Barbara Wyatt: I was wondering, Damita, to what extent your work and your development of the predictive model led directly to the NHL or was there already sufficient information on that important area to know that this was exceptionally special?

Damita: Okay thanks. It did definitely help in support. Obviously the quarry site spoke for itself, just looking on the landscape.  But it also shored up our deep testing that we conducted to the geoarcheology. It really augmented some of the known factors and could get some really firm dates on the deposition of those sediments in a direct area, so that was very nice.

Barbara: The above ground landscape is incredible and it looks like it has a lot of integrity. I don’t know if that’s ranch land or what. Could you describe a little bit what the uses of the area you were [inaudible] in?  I’m sure it’s quite varied because you were looking at a large area, but just generally speaking the images you showed of wide open spaces.

Damita: I think probably from East Coast standards it definitely is. But even for the plains it actually is really wide open and a lot of places haven’t been developed agriculturally. You get the prairie and the ranchland that haven’t been affected a lot. The oil and energy boom that’s happening is really significantly impacting that view – the viewshed on a lot of these places. And that’s just going to keep going as our demand increases for energy. The only other real impact besides roads and railroads in this area and rural settlements is some lignite mining that had occurred at the quarry itself, and that’s represented by the larger pits in that picture that are actually collapsed mines underneath these resources, so that kind of is a little bit of a threat, but overall it’s very pristine out there in that immediate area. As you go across the landscape in the whole entire primary source area, the further west you get the more of an impact with the development of oil energy. It’s definitely an effect. But again,  like I said, without those projects a lot of the archeology wouldn’t have been documented either. So definitely lucky.

Barbara: So referring to the fact that this was a huge compliance case. Did it also involve evaluation of the historic landscape or beyond archeology? Were there some assessments, not by your company perhaps, but were others also looking at the integrity of that cultural landscape?

Damita: Well we document everything that’s older than fifty years old, regardless if it’s pre-historic or historic or [inaudible] architecture to try and get that all on and when we were developing this model, we were looking at it definitely for the lithics for KRF. We were also looking in other areas away from this flint resource, if you could try and model by putting in other attributes into the model and getting a different layer to come up like a historic settlement of Euro-Americans or pretty much anything. So the ranching lands, or something like that, will definitely be tied into the quality of the soils out there to ranch or to farm, so it’s definitely applicable in different places, but you need the money and research to really delve into other components to it. But it is possible. To the best of my knowledge there hasn’t been a lot of over-arching studies, but the SHPO could speak more to that on file.

Nancy Brown: Nancy Brown. I was curious about who owns the land. Perhaps you said at the beginning but I missed it, and I have a second question.

Damita: Okay. The land for the NHL that was just designated is held privately. I think the main land owners are the Lynches and they have done their hardest work to preserve as much as they can of that area. And they worked in conjunction with seven other land owners, I believe. And the [inaudible]mostly again, the one that was proposed but wasn’t nominated, is also privately held. Going across the landscape for both counties the majority of it is held in private land. There is a little bit of Forest Service land. I believe it picks up clear out west, south of the Killdeer Mountains.

Nancy: And who’s the lead federal agency for the transmission line?

Damita: The lead is Western and RUS are the two. We support that one.

Nancy: And have you had a historic landscape architect write up a description of this more in cultural landscape terms at this point? You know, connecting the dots.

Damita: We have not.

Nancy Brown: The research is fascinating, and I think that would be a terrific addition.

National Center for Preservation Technology and Training
645 University Parkway
Natchitoches, LA 71457

Email: ncptt[at]nps.gov
Phone: (318) 356-7444
Fax: (318) 356-9119