PAPER A Layered Approach for the Discovery and Mapping of Prehistoric Sites Beneath Huron

AUTHORS ABSTRACT ’ John M. O Shea For much of modern human history (roughly the last 200,000 years), global Museum of Anthropological sea levels have been lower than present. As such, it is hardly surprising that Archaeology, University of Michigan archaeologists increasingly are looking to submarine environments to address Ashley K. Lemke some of their most pressing questions. While underwater archaeology is most Department of Sociology and commonly associated with shipwrecks, the search for submerged prehistoric Anthropology, University of Texas sites presents an entirely different set of challenges, even though many of the at Arlington same technologies are used. For Great archaeologists, the problem is how best to adapt the range of available seafloor mapping and testing techniques to the problem of identifying prehistoric sites, while operating with smaller ves- sels and the limited budgets available to “normal” archaeology. In this paper, we Introduction briefly describe the approach we have developed at the University of Michigan for t recently has been noted ( Joy, identifying 9,000-year-old caribou hunting sites beneath . The 2020) that, for 90% of modern research employs a layered research design integrating sonars, remotely operated I fi human history (roughly the last vehicles (ROVs), and scuba divers at progressively ner scales to discover and 200,000 years), global sea levels have investigate these important new archaeological sites. been lower than present. As such, it is Keywords: underwater archaeology, , sonar hardly surprising that, from Southeast Asia to the North Sea, Beringia, and South Africa, archaeologists are look- underwater to document this critical historic sites, which are often small ing to submarine environments to ad- period of climatic and cultural change. and comprise only stone tools and an- dress some of their most pressing While underwater archaeology is imal bone, the setting of the seafloor is questions. The Great Lakes in North most commonly associated with ship- critical since it represents the once dry America saw similar oscillations at the wrecks, the search for submerged land surface and environment on which beginning of the Holocene, varying prehistoric sites presents an entirely the ancient inhabitants lived. Particu- from more than 100 m higher to different set of challenges, even though larly in the context of small hunter- 100 m lower than modern lake levels. many of the same technologies are gatherer sites, the preserved environ- In the Lake Huron basin alone, it is used. Shipwreck sites represent an mental setting often provides the estimated that more than 2,500 km2 of anomaly on the seafloor that stands only clues to where the occupation modern lake bottomland was available out from their surroundings. Since sites might be found. Also in contrast for human use (Eschman & Karrow, vessels are rarely placed intentionally, to shipwreck research, the prehistoric 1985) during the Lake Stanley low- the underwater setting of the ship- archaeologist cannot know a priori stand (between roughly 10,000 and wreck is typically important only to that an occupation site even exists 8,000 cal BP) (Lewis et al., 2007). As the extent that it makes the vessel withintheareatobesearched.So archaeological and paleontological wreckage more or less visible. In while shipwreck archaeology is essen- sites from this time period are rare most cases, too, researchers are look- tially a search for a known target that on the surrounding mainland of Michi- ing for a specific vessel that was lost. will appear as an anomaly, prehistoric gan and Ontario, it is again unsurpris- Neither condition is true for sub- research requires that the ancient land- ing that archaeologists should look merged site archaeology. For pre- scape be mapped and reconstructed

May/June 2020 Volume 54 Number 3 23 before any meaningful search can hunting sites on the Alpena-Amberley separate paleolakes; Janusas et al., begin, and it requires modeling of Ridge (AAR) in central Lake Huron. 2004). The map (Figure 1) also human activities within that recon- shows extensive low-lying coastal structed environment as a basis for areas, particularly along the eastern executing a search for underwater Ancient Hunters coast of Michigan, , and archaeological sites. on the AAR southern Lake Huron that would Most research follows this general The late glacial history of the have been dry land during Lake Stan- approach, often termed the “Danish Great Lakes is complex and is charac- ley times, representing some 250,000 Model” (cf. Fischer, 1993; Benjamin, terized by a series of high and low hectares of land, which would have 2010), to discover submerged sites, water stands regulated by the inter- been exposed for settlement (Eschman particularly those associated with ear- action of early Holocene climate, the & Karrow, 1985). lier time periods that lack major flows of glacial melt waters, and the The ridge that divides the central architecture or historic records. The isostatic rebound of recently deglaci- basin deserves particular attention. search for prehistoric sites can be chal- ated land surfaces (cf. Baedke & The feature, termed the AAR, is lenging; all archaeologists must deal Thompson, 2000). Modeling these capped with Middle Devonian lime- with potential site destruction in changes is made further complex by stones and dolomites that resisted high-energy coastal settings and with short spasmodic readvances of the both fluvialandglacialerosion post-depositional sedimentation that glacial ice front (cf. Lowell et al., (Hough, 1958; Thomas et al., 1973, can bury and/or obscure traces of 1999). The most extreme of the low p. 232). Viewed in finer detail, the for- ancient sites. Archaeologists working waterstands,andthefocusofthe mation is found to be roughly 10 miles in marine environments must also current investigation, is collectively wide and exhibits a number of inter- contend with more recent coral and referred to as the Lake Stanley stage esting features including high north- organic growth. (Lewis, 2016; Lewis et al., 2007) east-facing cliffs, stretches of low Within the Great Lakes, the cold (the equivalent low water stage in coastal areas, and high plateau region- fresh water can provide extra-ordinary the basin is termed s. The northwest half of the ridge falls preservation conditions for organic ). Following Lake in American waters, and the southeast materials, such as wood and even Stanley, after 8,000 BP, lake levels half falls in Canadian waters. rooted trees. While we do not see never again drop to the lows associat- A considerable amount is known the dense accumulation of marine ed with Lake Stanley; instead, they about the regional environment dur- growth and corals as in salt water con- return to higher levels (Nipissing ing the Lake Stanley low-water phases texts, invasive species such as zebra transgression) before coming to (cf. Karrow, 2004). The general con- (D. polymorpha) and quagga mussels modern levels. This is significant sensus is that the climate of the region (D. bugensis) cover every hard surface, since it means there was a 2,000- was colder and drier than present and and the water clarity they create in year period during which the now that this changed to warmer and drier turn promotes algal growth at depths submerged portions of the Lake than present after about 7,900 BP previously unprecedented within the Huron bottomlands would have (Croley & Lewis, 2006; Lewis et al., Great Lakes. For Great Lakes archae- been available for the colonization 2007, p. 449; McCarthy et al., ologists, the problem is how best to of plant, animal, and human commu- 2015). Sarvis et al. (1999) describe exploit the potentials of this fresh- nities, but that after which it was the lake conditions during the period water research environment using the never again exposed. of 10,500–9,000 BP as similar to range of seafloor mapping and testing During the low-water Lake Stanley, those in modern Arctic lakes. This ex- techniques available. A related ques- the Huron basin contained two lakes pectation is borne out by analyses of tionishowtodosoonthelimited separated by a feature extending environmental samples collected di- budgets available to terrestrial archae- northwest to southeast across the basin rectly from the AAR, which reveal ology. In this paper, we brieflyde- from the area of Presque Isle in Michi- an open subarctic setting with numer- scribe the layered approach that we gan to around Point Clark in Ontario ous lakes and marshes with scattered have developed at the University of ( would similarly have spruce and tamarack at 9,000 BP Michigan for investigating ancient been isolated at this time, creating three (Sonnenburg & O’Shea, 2017).

24 Marine Technology Society Journal FIGURE 1 acidic forest soils rarely preserve fau- nal assemblages (see Spiess et al., Map of Lake Huron basin during the Lake Stanley lowstand. Initial three search areas are shown 1985; Storck & Spiess, 1994), hy- as labeled red boxes. pothesized shifts in technology, target prey species, and population organiza- tion are similarly inferred and untest- ed. Essentially, very little is known about the early inhabitants of the Great Lakes from the terrestrial ar- chaeological record. The lack of firm dates, intact sites, and preserved fauna has left unre- solved a number of significant long- term debates regarding Paleoindian technology, subsistence, and settle- ment systems (cf. Ellis et al., 1998). What most investigators do agree on is that this critical evidence lies be- neath the Great Lakes (Jackson, 2004; Jackson et al., 2000; Pengelly & Tinkler, 2004; Shott, 1999).

This major event in Great Lakes history [the draining of Lake Al- gonquian]…exposed about half of the present lake floor areas as dry land. This drainage must have had a significant effect on local climates and flora and fauna, opening large tracts to col- onization. Since such drainage was rapid, one can only speculate on its effect on the economy, reli- gion, and society of the human In terms of the regional archaeology, deposits (cf. Storck, 1997). The population (Karrow & Warner, the earliest generally accepted human Paleoindian occupation is followed 1990, p. 17). occupation in the upper Great Lakes by the Archaic, which covers the span is associated with the regional fluted- of time from nearly 10,000 until This is why the AAR has such point Paleoindian tradition. The about 2,000 BP (e.g., Ellis et al., great significance. It was a unique Paleoindian period is typically divided 1990; Monaghan & Lovis, 2005, dry land area during the critical time into three phases (Ellis, 2004; Julig, pp. 72–78; Lovis, 2009; Shott, 1999). period between 10,000 and 7,500 2002; Lothrop et al., 2016) defined While this progression of phases is BP, and unlike coastal areas that primarily on point styles and raw ma- widely accepted, it is dating and char- have been subject to extensive re- terial utilization. With the possible acter is largely inferred from archaeo- working and burial due to coastal pro- exception of the Gainey phase, none logical sequences in other parts of cesses and subsequent lake level rises of these complexes is well dated, and . Since there are very (cf. Barnett, 2015), sites on the few of the sites in either Michigan or few intact sites and few carbon-dated AAR occupy a midwater location Ontario are represented by intact site contexts, and since the region’s with little sediment cover and have

May/June 2020 Volume 54 Number 3 25 sat untouched since the time of three research areas that presented using multibeam (MB) sonar. The human occupation—an ideal setting contrastive landforms that were eth- MB survey used an R2 Sonic 2024 for submerged archaeology. nographically known to be desirable multibeam echosounder with an for caribou hunting: a water crossing, F180 vessel attitude and positioning a naturally occurring “choke point” (a system. The extensive survey was very narrow section of the AAR), and not expected to reveal actual hunting Mapping Ancient Hunting open ground along a paleolake margin structures—although in one case, it Sites Under Water (see Figure 1). These three areas did (O’Shea & Meadows, 2009)— The AAR provides a number of ranged in size from 17 to 56 km2 but rather to provide a detailed view advantages for an archaeological with depths ranging from 20 to 50 of the lake bottom. This more de- investigation. On the one hand, it m (with depth here being a potential tailed view highlighted the three- represents a relatively narrow and cir- surrogate for age, as lake levels gradu- dimensionality of the landforms, cumscribed area in which the cultural ally rose from the early Lake Stanley specifically the water crossing, choke activity would have been confined levels; i.e., the deeper the site, the point, and paleolake areas originally and thus where archaeological sites older it could be as water levels were selected. In addition, the sharp varia- may be found. On the other hand, at their lowest at the beginning of tion in acoustic reflectivity on the given what is already known about Lake Stanley). In addition, and in col- AAR revealed in the backscatter the paleo environment, the kinds laboration with Dr. Robert Reynolds from these surveys also provided a of cultural activities expected and of the Department of Computer Sci- very useful contrast between sandy what they might look like can also ence at Wayne State University, we and harder bottom conditions. This be predicted. Given the subarctic began an effort directed at both contrast was particularly important environment, it is expected that peo- modeling the ancient submerged en- given the offshore location of the sur- ple living on the AAR would have vironment and then populating it vey areas (i.e., 55–85 km) and the low pursued cold-adapted animals, such with migrating caribou, using artifi- likelihood that any substantial quanti- as caribou. These factors simplified cial intelligence (AI) methods to ty of sand would be transported from the problem from a search perspective enable the computer-simulated cari- land. As such, areas of sand observed andallowedustodomoreserious bou to learn and transit the ancient on the AAR had to have been depos- cultural modeling to predict the spe- environment (Fogarty et al., 2015). ited during Lake Stanley and there- cific kinds of settings in which sites Together, the physical methods of ex- fore provided an important indicator would be found. Specifically, we hy- amination, both remote and direct of ancient waterways (lakes, rivers) pothesized that the AAR would have (see below), and the AI combined to and marshes that existed when the provided a corridor for the seasonal form a layered approach to the inves- AARwasdryland.Thishypothesis migration of caribou herds and that tigation. Physical techniques, such as was confirmed by the discovery of the hunters pursuing the animals side-scan sonar (SSS) mapping, pro- marsh testate amoebae in these sand would have exploited similar tech- gressed from extensive to intensive, deposits (Sonnenburg & O’Shea, nologies, in the form of stone lines while the computer simulation pro- 2017). and hunting blinds known ethno- vided a virtual database of all the Following an initial examination graphically among precontact caribou information being generated by the of promising locations using a small hunters in the Arctic (Brink, 2005; physical examination and testing under- remotely operated vehicle (ROV; see Stewart et al., 2004). water (Figure 2). description below) and teams of The starting point for the investi- The first level in the layered search SCUBA-trained archaeologists, the gation was existing high-resolution strategy saw the complete coverage of next layer of search exploration was bathymetry for the Lake Huron the three test areas using SSS at an in- initiated. It was clear that the prelim- basin. Using these data, we were termediate frequency of 330 kHz over inary survey blocks were still very able to create a generalized model of depths between 20 and 40 m and at a large areas to investigate archaeologi- the ancient land surface that would range of 100 m. A 115-km2 portion cally; therefore, the next step involved have been dry land during Lake Stan- of the central AAR, which partially defining smaller areas of the lake ley, and on this basis, we identified overlapped Area 1, was also surveyed floor that could be examined more

26 Marine Technology Society Journal FIGURE 2 in acoustic mapping of the micro- regions (as opposed to the traditional Schematic model of layered search strategy. towed side scan, which was used to map the much larger preliminary areas). The project utilized a base-level Iver3 AUV, and over the course of the research, we have experimented with differing configurations of acoustic and photography packages (Figure 3). The bulk of the research to date has been conducted using SSS mounted on the AUV (Edgetech 2205 operated at 600 and 1,600 KHz simultaneously). Bottom mapping with an AUV was attractive for several reasons: (1) the AUV can flymuchclosertothebot- tom, which accentuates the three- dimensionality of the seafloor; (2) it obviates the need for complex layback calculations when mosaicking the re- sults of the survey; and (3) it does not experience image distortion produced by heave or other motions transmitted from the surface towing vessel (al- though it was not entirely immune to weather and sea conditions; see below). For initial survey coverage, a relative- ly long 50-m range was utilized with a grid providing 100% overlap coverage of the survey area. Unlike the extensive towed SSS survey, the AUV surveys were expected to reveal potential hunt- ing structures and other built features and alignments, particularly given its closeness to the lake bottom and high- resolution imagery. The more detailed intensively. Initially, four “micro- vestigate other potential activities not bottom imaging produced a clearer regions” were selected for a more related to caribou hunting. view of the immediate environment concentrated investigation. These lo- Each micro-locality has an area of and the configuration of natural fea- calities were selected based on the 0.5 km2 with the typical area surveyed tures, which would have been relevant character of the immediate landscape being a 1,000 × 500 m rectangle. The to ancient hunting and habitation. Fi- and also represented areas within orientation of the micro-region survey nally, the mosaicked AUV imagery which stone hunting features had block was determined by the shape represented a detailed and georefer- already been identified during pre- of the underlying landform in order enced map to guide more intensive lo- liminary ROV and SCUBA observa- to produce the most effective sonar calized site investigation and mapping. tions. The total number of micro- mosaic. This layer of investigation Using these detailed mosaic maps regions has been expanded to nine, employed a small autonomous under- of the micro-regions, a thorough ex- as the research has broadened to in- water vehicle (AUV) to conduct close- amination of potential targets was

May/June 2020 Volume 54 Number 3 27 FIGURE 3 has progressed over time and loca- tions have been revisited, a cumulative AUV. (A) Iver3 being launched from research vessel. (B) SSS image of high ground area in central Lake Huron from AUV survey; swath is 100 m and clearly shows cliffs, sand ripples, (12+ year) record of environmental and individual boulders. Photograph taken by A. Lemke. conditions on the Lake Huron bot- tomlands has been created, which permits changing conditions and the progression of invasive species to be directly observed. For targets that continue to look promising after ROV inspection, the research shifts from discovery mode to documentation and sampling, which entails both remote and direct access to the site. Since our primary targets were constructed stone features, the creation of an accurate map of the structure is of prime importance. Tra- ditionally, this would involve divers’ stretching tapes, a process that is both time consuming and inaccurate. Such maps are also inherently 2-D. To over- come these drawbacks, we adopted two approaches for mapping large fea- tures, one remote and one direct. The remote approach involved the use of scanning sonar. In essence, the sonar head is held in place on a tripod sitting on the lake floor, and the sonar head then rotates through 360° to produce a scaled acoustic image. We utilized a Kongsberg MS1000 unit along with its proprietary software. The scanning unit is deployed from the research vessel and rapidly pro- duces a high-resolution acoustic image of the feature in real time. The operator can vary the range or performed using an ROV (Figure 4). gets and for recording the location of scale of the image on the fly, and mul- An Outland 1000, a hand-deployable recovered video imagery. Potential tiple collected images can be mo- ROV, was used for this work. In ad- sites were examined in real time on saicked. As the system can produce dition to the standard video cameras the research vessel, and later in the either high-resolution acoustic images and forward scanning sonar, paired laboratory, the video footage was or point cloud data, it can provide an forward-pointing parallel red lasers parsed by location and entered into accurate 3-D image of the target, were attached to provide a scale on a linked Geographic Information Sys- which can then be manipulated for visual images, and a USBL transpon- tem (GIS) to permit a fuller ex- viewing from multiple angles, accurately der was added (Tritech Micronav amination of areas during the off measured, georeferenced, and exported 100) to permit locational control for season. An unanticipated benefitof into other image analysis or computer navigating the ROV to designated tar- the video database is that, as research simulation systems (see Figure 5).

28 Marine Technology Society Journal FIGURE 4 of imagery is often collected as HD videos and then frame grabbed to ROV deployed at the Drop45 site in central Lake Huron. Sample marker from archaeological test unit 15 is visible in the foreground. obtain the still images needed, we col- lected still images at regular intervals of 1–3 s/shot. This approach obviated the intermediate post-processing step, and it ensured that each image con- tained its unique meta-data, which en- hanced the stitching together of the imagesbytheprogram.Weutilized Agisoft Metashape (formerly PhotoScan Pro) software for processing the images. The resulting 3-D images can be accu- rately measured and, with appropriate texturing, provide a compelling image of the feature that can be scaled and viewed from multiple angles. The im- ages can also be exported to a variety of generally available 3-D viewing for- mats,suchas.pdf,andcanbedirectly placed into an existing computer simu- lation environment. The second approach to docu- et al., 2019). Divers equipped with With the mapping and documen- menting and imaging features was GoPro cameras would slowly swim tation steps complete, the work moves photogrammetry, a technique common around the feature to produce the initial on into more traditional archaeological in maritime archaeology (see McCarthy overlapped imagery set. While this type activities of excavation and sample

FIGURE 5

Scanning sonar. (A) Scanning sonar unit being prepared for deployment. Note sonar head at base. (B) Sector scan image of Target 5 in Lake Huron. Image radius is 20 m with unit in the center. Red circles represent 4-m intervals.

May/June 2020 Volume 54 Number 3 29 recovery. This work is conducted by we hope to have all data streams avail- AAR has been conducted using a Par- SCUBA-trained archaeologists, and able for online access. ker 2530 as the research platform) the goal of technology at this stage is The second and more experimen- and less expensive, hand-deployable to reduce the burdens of navigation tal approach focuses on the dynamic survey equipment. In this regard, the and documentation on the divers VW model of the research area creat- Outland 1000 ROV has been partic- so that their limited bottom time is ed by Robert Reynolds and his team ularly valuable due to its smaller size devoted to sample recovery. For this, (cf.Fogartyetal.,2015;Stanley, and versatility. Versatility and flexibil- the ROV is again the work horse. The 2019). The idea is by continually up- ity are essential if submerged site re- ROV is deployed ahead of the divers dating the VW model as new data on search is to be conducted within the and provides a critical visual link be- the environment or site locations and budgets of (non-glamor) archaeology. tween the surface and the divers. It structures are accumulated, the VW Of course, the use of less expensive confirms the location of the target, will come to encompass essentially gear or options does come with its creates a visual pre-disturbance record everything we know about the AAR own price. For example, while the of the site, and directs the divers to and will present this information in lower-end AUV utilized produced all their work areas (as well as providing a form that can be visually inspected of the advantages we expected, it also a redundant guide back to the boat). and experienced as virtual reality. came with a series of limitations. Per- On-site, the ROV records the sam- While finer grained data, such as re- haps, the biggest disappointment was pling activities on the lake bottom, covered stone artifacts and wood, are the limited navigational accuracy of and once sampling is completed, it not yet incorporated in the system, the base model Iver3. The absence records the precise location of each initial trials of the VW model conduct- of the more sophisticated inertial sample unit collected. The ROV can ed with traditional caribou hunters in guidance package meant that, even also remain on-site during dive sur- have produced encouraging over the relatively short distances of face intervals and produce a final, results and suggested ways to both the micro-localities (1,000 m), the post-testing image of the site. improve and generalize the VW model. AUV could not maintain an accurate track. We were also reminded that, while AUVs are unaffected by the After Discovery Discussion motions of a tow vessel, they are not As should be clear from the pre- Submerged site archaeology, as a immune to surface conditions. For ceding narrative, the layered approach field, is still in its infancy as various example, during rough surface condi- to site discovery and documentation approaches, some borrowed from tions, the unit’s ability to acquire an produces a diversity of digital data maritime and shipwreck research and adequate GPS fix was notably degrad- reflecting the differing scales at others from terrestrial archaeology, ed, even as wind and waves tended to which data are collected. The effective are being drawn upon to develop push the small unit off course. We integration of these differing data ways forward. The layering of tech- remain convinced that AUVs have a streams is essential. For this, we have niques described here was developed pivotally important role to play in adopted two approaches, a traditional specifically for the research in Lake the layered search system and in un- GIS database and the creation of a Huron, and it has proven an effective derwater archaeology more generally. de facto database within a virtual world method for conducting a multi-year The technology exists today, but we (VW) computer simulation. program of archaeological research. must await future innovations that Since every piece of data collected The techniques applied were particu- will allow more accurate navigation by the project, regardless of scale, has larly well suited for identifying past at lower costs. a spatial component, it can all be ac- environmental conditions and the pres- The same caveat applies to the use commodated within a single GIS for ence of stone structures—advantages of small vessels. Small research vessels, central Lake Huron. Everything not always shared in other marine ar- such as the Parker 2530 we use, bring from the largest 56-km2 search area chaeology efforts. The efforts have versatility and flexibility to research to the location of a single 100-ml en- also demonstrated that sustained re- efforts. They are relatively inexpensive vironmental sample can be plotted. search can be conducted using small to operate and, through their dedi- By the completion of the research, vessels (nearly all the research on the cated use, allow researchers to take

30 Marine Technology Society Journal advantage of short-term windows of evaluating the ‘Danish Model’ for international of caribou behavior across Lake Huron using favorable weather and sea conditions. practice. J Isl Coastal Archaeol. 5:253-70. cultural algorithms and influence maps. In: These benefits are achieved, however, https://doi.org/10.1080/15564894.2010. Caribou Hunting in the Upper Great Lakes: at the cost of smaller working spaces, 506623. Archaeological, Ethnographic, and Paleoen- vironmental Perspectives, eds. Sonnenburg, limited crewing capabilities, and the Brink, J. 2005. Inukshuk: Caribou drive lanes E.P.,Lemke,A.K.,&O’Shea, J.M., pp. 31-52. inability to remain continuously on on Southern Victoria Island, Nunavut, Canada. Ann Arbor, MI: Museum of Anthropology, the water for multiple days. Arctic Anthropol. 42(1):1-28. https://doi. University of Michigan. Looking beyond the AAR, it is org/10.1353/arc.2011.0084. easy to envision how other common Hough, J. 1958. Geology of the Great Lakes. Croley, T., II, & Lewis, C. 2006. Warmer Urbana, IL: University of Illinois Press. techniques, such as sub-bottom pro- and drier climates that make terminal Great fi fi ling, lidar, and coring, can be t Lakes. J Great Lakes Res. 32:852-69. https:// Jackson, L. 2004. Changing views of late into a layered search strategy to per- doi.org/10.3394/0380-1330(2006)32[852: Palaeo-Indian in Southern Ontario. In: The mit research in areas with differing WADCTM]2.0.CO;2. Lake Palaeo-Indian Great Lakes: Geological conditions of bottom visibility and and Archaeological Investigations of Late Ellis, C. 2004. Hi-Lo: An early lithic complex sedimentation. Overall, the twin and Early Holocene Environments, in the Great Lakes. In: The Late Palaeo-Indian trends of miniaturization and increas- eds. Jackson, L., & Hinshelwood, A., pp. Great Lakes: Geological and Archaeological ing capability will ensure that impor- 25-56. Ottawa, Canada: Archaeology Papers, Investigations of Late Pleistocene and Early Canadian Museum of Civilization: Mercury tant archaeological sites on the bottom Holocene Environments, eds. Jackson, L., & ’ Series. https://doi.org/10.2307/j.ctv16x58.9. of the Great Lakes, or the world s Hinshelwood, A., pp. 57-84. Ottawa, Canada: oceans, will become increasingly acces- Archaeology Papers, Canadian Museum of Jackson,L.,Ellis,C.,Morgan,A.,& sible to archaeological research and dis- Civilization: Mercury Series. https://doi.org/ McAndrews, J. 2000. Glacial lake levels and covery. 10.2307/j.ctv16x58.10. eastern Great Lakes Palaeo-Indians. Geoarch- aeology. 15:415-40. https://doi.org/10.1002/ Ellis,C.,Goodyear,A.,Morse,D.,& (SICI)1520-6548(200006)15:5<415::AID- Tankersley, K. 1998. Archaeology of the GEA2>3.0.CO;2-2. Corresponding Author: Pleistocene-Holocene transition in Eastern ’ John M. O Shea North America. Quatern Int. 49/50:151-66. Janusas, S., Blasco, S., McClellan, S., & Museum of Anthropological https://doi.org/10.1016/S1040-6182(97) Lusted, J. 2004. Prehistoric drainage and ar- Archaeology, University of Michigan 00060-8. chaeological implications across the submerged 310 East University Avenue north of Tobermory, Ellis, C., Kenyon, I., & Spence, M. 1990. Ann Arbor, MI 48109 Ontario. In: The Lake Palaeo-Indian Great The archaic. In: The Archaeology of Southern Email: [email protected] Lakes: Geological and Archaeological Inves- Ontario to A.D. 1650, eds. Ellis, C., & Ferris, N., tigations of Late Pleistocene and Early pp. 65-124. London, UK: Occasional Publi- Holocene Environments, eds. Jackson, L., & cation of the London Chapter: OAS. References Hinshelwood, A., pp. 303-14. Ottawa, Canada: Baedke, S., & Thompson, T. 2000. A 4,700 Eschman, D., & Karrow, P. 1985. Huron Archaeology Papers, Canadian Museum year record of lake level and isostasy for Lake basin Glacial Lakes: A review (Spec. Paper 30). of Civilization: Mercury Series. https://doi. Michigan. J Great Lakes Res. 26(4):416-26. In: Evolution of the Great Lakes, org/10.2307/j.ctv16x58.18. https://doi.org/10.1016/S0380-1330(00)70705-2. eds. Karrow, P., & Calkin, P., pp. 79-93. Joy, S. 2020. Coastally adapted: A model for University of Waterloo, Waterloo, Ontario: Barnett, P. 2015. Potential for deeply buried eastern coastal Paleoindian sites. Paper presented Geological Association of Canada. archaeological sites in Ontario based on glacial at 53rd Annual Conference on Historical and history. In: Caribou Hunting in the Upper Fischer, A. 1993. Stenalderbopladser på Underwater Archaeology. Boston, MA. Great Lakes: Archaeological, Ethnographic, bunden af Smålandsfarvandet. En teori afprøvet Julig, P. (Ed.) 2002. The Sheguiandah Site: and Paleoenvironmental Perspectives, eds. ved dykkerbesigtigelse [Stone age settlements Archaeological, Geological, and Paleobotanical Sonnenburg,E.P.,Lemke,A.K.,&O’Shea, in the Småland Boght. A theory tested by Studies at a Paleo-Indian Site on Manitoulin J.M., pp. 5-12. Ann Arbor, MI: Museum of diving]. In: Man and Sea in the Mesolithic, ed. Island, Ontario. Ottawa, Canada: Archaeology Anthropology, University of Michigan. Fischer, A., pp. 371-84. Oxford, UK: Oxbow. Papers, Canadian Museum of Civilization: Benjamin, J. 2010. Submerged prehistoric Fogarty, J., Reynolds, R., Salaymeh, A., & Mercury Series. https://doi.org/10.2307/j. landscapes and underwater site discovery: Re- Palazzolo, T. 2015. Serious game modeling ctt22zmf6d.

May/June 2020 Volume 54 Number 3 31 Karrow, P. 2004. Ontario geological events McCarthy,F.,McAndrews,J.,&Papangelakis, Sonnenburg,E.,&O’Shea, J. 2017. Archae- and environmental change in the time of the E. 2015. Paleoenvironmental context for early ological landscapes during the 10-8 ka Lake late Palaeo-Indian and early archaic cultures Holocene caribou migration on the Alpena- StanleylowstandontheAlpena-AmberleyRidge, (10,500 to 8,500 B.P.). In: The Lake Palaeo- Amberley Ridge. In: Caribou Hunting in Lake Huron. Geoarchaeology. 32(2):230-47. Indian Great Lakes: Geological and Archaeo- the Upper Great Lakes: Archaeological, https://doi.org/10.1002/gea.21590. logical Investigations of Late Pleistocene and Ethnographic, and Paleoenvironmental Per- Spiess, A.E., Curran, M.L., & Grimes, J.R. Early Holocene Environments, eds. Jackson, spectives, eds. Sonnenburg, E.P., Lemke, A.K., 1985. Caribou (Rangifer tarandus L.) bones L., & Hinshelwood, A., pp. 1-23. Ottawa, &O’Shea, J.M., pp. 13-30. Ann Arbor, MI: from New England Paleoindian sites. Canada: Archaeology Papers, Canadian Mu- Museum of Anthropology, University of North American Archaeologist. 6:145-56. seum of Civilization: Mercury Series. https:// Michigan. https://doi.org/10.2190/JP8K-0V8F-HLPV- doi.org/10.2307/j.ctv16x58.8. McCarthy, J.K., Benjamin, J., Winton, T., & XWGN. Karrow, P., & Warner, B. 1990. The geo- Van Duivenvoorde, W. (Eds.) 2019. 3D Stanley, S.D. 2019. Capso: A multi-objective logical and biological environment for human Recording and Interpretation for Maritime cultural algorithm system to predict locations of occupation in southern Ontario. In: The Ar- Archaeology. Cham, Switzerland: Springer. ancient sites. Ph.D. dissertation, Department chaeology of Southern Ontario to A.D. 1650, https://doi.org/10.1007/978-3-030-03635-5. of Computer Science, Wayne State University. eds. Ellis, C., & Ferris, N., pp. 5-36. London, Monaghan, G., & Lovis, W. 2005. Modeling UK: Occasional Publication of the London Stewart, A., Keith, D., & Scottie, J. 2004. Archaeological Site BurialinSouthernMichigan. Chapter: OAS. Caribou crossings and cultural meanings: East Lansing, MI: Michigan State University Placing traditional knowledge and archaeology Lewis, C. 2016. Understanding the Holocene Press. in context in an Inuit landscape. J Archaeol closed-basin phases (lowstands) of the Lau- O’Shea, J.M., & Meadows, G.A. 2009. Method Th. 11:183-211. https://doi.org/10. rentian Great Lakes and their significance. Evidence for early hunters beneath the 1023/B:JARM.0000038066.09898.cd. Geosci Can. 43:179-97. https://doi.org/ Great Lakes. P Natl Acad Sci USA. 106 10.12789/geocanj.2016.43.102. Storck, P.L. 1997. The Fisher Site: Archaeo- (25):10120-3. logical, Geological and Paleobotanical Studies Lewis, C., Heil, C., Hubeny, J., King, J., Pengelly, J., & Tinkler, K. 2004. Lake level at an Early Paleo-Indian Site in Southern Moore, T., & Rea, D. 2007. The Stanley changes and aboriginal cultural manifestations Ontario, Canada. Ann Arbor, MI: Museum of unconformity in Lake Huron Basin: evidence in areas adjacent to and including Niagara Anthropology, University of Michigan https:// for a climate-driven closed lowstand about Peninsula. In: The Lake Palaeo-Indian Great doi.org/10.3998/mpub.11395063. 7900 14C BP, with similar implications for Lakes: Geological and Archaeological Investi- the Chippewa lowstand in Lake Michigan Storck, P.L., & Spiess, A.E. 1994. The sig- gations of Late Pleistocene and Early Holocene Basin. J Paleolimnol. 37:435-52. https://doi. nificance of new faunal identifications attrib- Environments, eds. Jackson, L., & Hinshelwood, org/10.1007/s10933-006-9049-y. uted to an early Paleoindian (Gainey complex) A., pp. 201-24. Ottawa, Canada: Archaeology occupation at the Udora site, Ontario, Lothrop Papers, Canadian Museum of Civilization: , J., Lowery, D., Spiess, A., & Ellis, Canada. Am Antiquity. 59(1):121-42. https:// Mercury Series. https://doi.org/10.2307/ C. 2016. Early human settlement in north- doi.org/10.2307/3085506. eastern North America. PaleoAmerica. 2(3): j.ctv16x58.14. Thomas, R., Kemp, A., & Lewis, C. 1973. 192-251. https://doi.org/10.1080/20555563. Sarvis , A.P., McCarthy, F., & Blasco, S. 1999. fi 2016.1212178. The sur cial sediments of Lake Huron. Can J Explaining the lowstand in Georgian Bay ap- Earth Sci. 10:226-71. https://doi.org/10.1139/ Lovis proximately 7,200 years ago: A paleolimno- , W. 2009. Hunter-gatherer adaptations e73-023. and alternative perspectives on the Michigan logical approach using microfossil evidence. archaic: Research problems in context. In: In: Leading Edge ’99. Niagara Escarpment Archaic Societies, Diversity and Complexity Commission, Burlington, Ontario. Available at: Across the Midcontinent, eds. Emerson, T., http://citeseerx.ist.psu.edu/viewdoc/download? McElrath, D., & Fortier, A., pp. 725-54. doi=10.1.1.728.9390&rep=rep1&type=pdf. Albany, NY: SUNY Press. Shott, M. 1999. The early archaic: Life after Lowell, T., Larson, G., Hughes, J., & Denton, the glaciers. In: Retrieving Michigan’s Buried G. 1999. Age verification of the Lake Gibben Past: The Archaeology of the Great Lakes forest bed and the advance State, ed. Halsey, J., pp. 71-82. Bloomfield of the . Can J Earth Sci. Hills, MI: Cranbrook Institute of Science 36:383-93. https://doi.org/10.1139/e98-095. Press.

32 Marine Technology Society Journal