2003 SCAPE GEOARCHAEOLOGY FIELD SEASON
by Garry Running

Personnel: Karen Havholm (Department of Geology, UWEC) Garry Running (Department of Geography and Anthropology, UWEC) Dion Wiseman (Department of Geography, Brandon U)

Students: Sarah Buss and Adam Lange (Department of Geography and Anthropology, UWEC) Laura Roskowski (Department of Archaeology, UCalgary)

After attending the field conference the team worked mainly in the Forks, Glacial Lake Hind Basin and Tiger Hills localities.

Tiger Hills Locality
The geoarchaeology team consulted with Scott Hamilton regarding his interpretations of soils, buried soils, and stratigraphy observed in the archaeological sites under investigation. Most of these sites are shallowly buried or near surface sites associated with the modern surface soil profiles. A few of these sites may be stratified or represent palimpsest occupations or activity areas. However, even where these sites are within over-thickened (cumulic) A-horizons, processes of pedoturbation may have sufficiently disturbed archaeological material to make difficult distinguishing one period of occupation from another.

In addition, a Geoprobe core was recovered, and two profiles were described at the Craig Bessant site. The Geoprobe core was recovered from the terrace within which the site is located. One of the profiles described was located within an excavation unit at the site. The other profile was located in a cutbank that exposed sediments within the youngest fluvial terrace observed adjacent to the Craig Bessant site.

The Craig Bessant site is located along the Souris River just downstream from the "elbow". The elbow is the local name for the point of stream capture, where the Souris River abandons the Pembina meltwater spillway and turns north to flow through the Tiger Hills and toward the Assiniboine River.

Although geoarchaeological investigations of the Craig Bessant site are in a preliminary stage, of significance are two Holocene terraces and at least one terminal late-Holocene cut terrace which can be observed along this reach of the Souris River. Moreover, buried soil profiles within vertical accretion facies deposits are subhorizontally oriented, 10-30 cm thick overall, with 2-4 cm thick A-horizons, similar to those observed elsewhere in SCAPE investigations in similar settings. Perhaps these "wimpisols" as we informally refer to them in the field are more widespread in occurrence geographically than we once thought. Hopefully, what we learn from Laura Roskowskis work in the Forks locality, Judy Klassen and Liz Robertson's work in the Cypress Hills locality regarding their genesis and the paleoenvironmental signal they represent will be applicable to this area.

The Forks Locality
Efforts this year focused on two objectives: 1) refining the terrace sequence model developed last year for the South Saskatchewan River from the St. Louis site downstream to the confluence with the North Saskatchewan River, and 2) correlating interpretations of site stratigraphy at the St. Louis site amongst the investigators who have been working there.

1. Our preliminary terrace sequence model, developed after the 2003 field season, is based on subsurface investigations at two points, the Harper Valley site, and the St. Louis site, coupled with limited topographic data (Harper Valley site), and a map of terraces, their extent and locations based on aerial photographic interpretation. This field season we collected subsurface data at three more locations within the study reach. Geoprobe cores were collected, at each of these new locations by the investigators and students identified above, from 5 terraces. We confirm our preliminary model regarding the Holocene age of T1 and T2, the presence of "wimpisols" in lateral accretion facies deposits within T1 and T2, and the general stratigraphic and elevational relationships between terraces. However, we now have data that suggests T3 as mapped last year is really T3-4-5. T3-5 appears to be 3 terminal late-Pleistocene cut terraces graded to, presumably, levels of glacial lakes Agassiz and/ or Saskatchewan. Efforts to map the extent, locations, elevations, and trends of these terraces across the study reach are ongoing and will depend on accurate aerial photographic interpretation. To aid in mapping T1-5 in the study reach. Dion Wiseman, and Sarah Buss and Adam Lange collected elevational data using dGPS along a number of transects where the 5 terrace treads were observable. These data will be incorporated into our GIS database and used to refine the terrace maps. We found last year that aerial photographic interpretation alone, without supporting elevational data based on dGPS transects, was insufficient to generate an accurate map of terraces within the study reach.
   
   
2. Karen Havholm described a series of profiles exposed in T2 at the St. Louis site. She and Andrea Freeman correlated their observations with those of Butch Amundson who is conducting the detailed geoarchaeological analysis at the St. Louis site. Sedimentological and other data generated over the past two years by Havholm and Running from within the study reach will be turned over to Andrea Freeman. Andrea will collate our data, and Butch's data from the St. Louis site, with hers to further refine our preliminary terrace sequence model with regard to: age ranges for terrace fills, terrace chronology, and perhaps, if we're lucky, a testable hypothesis regarding the conditions under which terraces within the study reach were abandoned and new floodplains were established.

The Glacial Lake Hind Basin (GLHB)
Efforts this year focused on providing a better understanding of the sediments and buried soils associated with archaeological sites in the Crepeele dune field and at the Atkinson site. The Crepeele dune field is one of at least 18 dune fields within the GLHB. Archaeological surveys conducted this year by James Graham and his crews reveal a strong correlation between parabolic dunes within this dune field and archaeological sites. James Graham and his survey crews can provide more detail on their archaeological investigations in the Crepeele dune field. A similar relationship is observed within the Makotchi Ded Dontipi (MDD) area located a few miles away.

Several observations were made in the Crepeele dune field investigations this year. First, the relationship between parabolic dunes, and interdunal wetlands, and preferential use (or at least comparatively greater density of archaeological sites than surrounding non-dune landscapes) of these eolian landscapes by precontact groups first observed at MDD is repeated here. This relationship between parabolic dune-dominated landscapes and human use is not restricted to the MDD area. Second, the model we (Garry Running, Karen Havholm, Matt Boyd, and Woody Wallace) have developed for parabolic dune age (late-Holocene - Unit D), and processes of formation (dune formation when sandsheet transport is slowed upwind of wetlands), for the GLHB in general apply to the Crepeele dune field. Third, the uppermost buried soil profiles associated with archaeological material in the Crepeele dune field appear to represent approximately 1500-2000 years of soil development (pending confirmation by 14C analysis of archaeological material recovered). These buried soil profiles exhibit an A-E-B horizonation typical of soil formation under forest vegetation. This suggests that forest vegetation observed in dune fields today is not a recent vegetation response to EuroCandadian fire suppression. Rather, forest vegetation existed in association with dune fields within the GLHB during at least the late-Holocene. Therefore, the mosaic of microhabitats associated with dune fields included forest vegetation, and perhaps was an important component of the dune field landscape that was attractive to human groups over at least the late-Holocene. Next field season we intend to develop our model of eolian landform development, vegetation response, and human land-use by investigating more archaeological sites, and describing more sedimentary sequences and buried soil profiles in the Crepeele dune field.

The Atkinson site itself is a multi-component site located along the banks of the Souris River a few miles from the MDD, the Crepeele dune field and Flintstone Hill. The oldest archaeological component is dated to ~6,100 BP and includes diagnostic material consistent with the Mummy Cave complex. Our efforts focused on correlating the fluvial and eolian deposits at the Atkinson site with the stratigraphic sequence observed at Flintstone Hill. We can confirm that the geomorphic units observed at Flintstone Hill are consistent with those observed at the Atkinson site. In addition, Unit B, which is a mid-Holocene eolian dune unit, is observed at the Atkinson site ~ 75 meters upstream from the archaeological excavations. Organic materials from below Unit B were recovered this year by Matt Boyd and will be submitted for radiocarbon analysis. Our interpretation based on investigations conducted this year suggest that basal Unit B deposits at the Atkinson site may be older than those observed at Flintstone Hill. If so, they will be the oldest known eolian dune deposits yet observed in the Canadian Prairies and adjacent portions of the northern Great Plains in the U.S. This is very exciting and fieldwork next year will focus on refining our understanding of the stratigraphic sequence observed at the Atkinson site, with particular emphasis on Unit B. Sarah Buss, now a graduate student at Appalachian State University will conduct fieldwork at the Atkinson site next year as part of her Master's research. With assistance from Karen Havholm, Dion Wiseman, Matt Boyd and Garry Running, she intends to provide a detailed 3-D map, and chronostratigraphic model, of the eolian and fluvial geomorphic units observed at the Atkinson site. In addition, efforts to recover organic material suitable for paleoenvironmental analysis will focus on sediments within an abandoned meander observed along the northern margins of the site.