Title: PREDICTIVE MODEL OF POTENTIAL UNDISCOVERED

advertisement
Title: PREDICTIVE MODEL OF POTENTIAL UNDISCOVERED VILLAGE &
MOUND ARCHEOLOGY SITES FOR THE RED WING LOCALITY
Leanne Knott
Department of Resource Analysis, Saint Mary’s University of Minnesota, Minneapolis, MN
55404
Keywords: GIS, archeology, predictive modelling, spatial analysis, LiDAR, geomorphology,
mounds groups, surveying, historic mapping
Abstract
This review of literature is associated with examining the use of Geographic Information
Systems (GIS) to create a predictive model to discover previously undocumented village and
mound archeology sites for the Red Wing Locality. The Red Wing Locality represents the
largest known concentration of ancient Native American cultural sites in Minnesota,
including mounds and villages from the Oneota, Woodland, Silvernale, and Pre-Contact
archeological eras. Information reviewed is organized into the following major themes: (a)
use of GIS in archeological methodologies (b) predictive modelling (c) required data and
availability: historic maps and surveys, LiDAR, aerial photo imagery, known sites, and
geomorphology. Utilizing GIS to select areas within the locality region that may contain
undiscovered cultural resources would assist in identifying areas of cultural sensitivity and
support Phase 1 archeological resource assessments in advance of potential land
development.
Introduction
Research Problem Description
This paper reviews research associated
with the following research question:
“Where are potential undiscovered
archeological sites within the Red Wing
Locality?”
Significance of Research
The Red Wing area has long been known
as one of the richest archaeological
regions in Minnesota. (Gibbon, 2008).
Between 1050 and 1300 AD, the Red
Wing Locality supported one of the largest
populations in the Upper Midwest. The
area at the junction of the Cannon and
Mississippi Rivers was host to at least nine
large villages - perhaps more (Flemin,
2014). Silvernale is one of the earliest and
largest of at least nine large village sites
inhabited between ca. A.D. 950 and 1400
at the junction of the Cannon and
Mississippi rivers. Between ca. A.D. 1050
- 1250 (known as the Silvernale phase),
inhabitants of the Red Wing Locality
participated in intensive interaction among
several regional cultural traditions. This
phenomenon resulted in a localized
cultural development that is unique to the
sites in the Locality and that has been an
ongoing source of anthropological
research for the last 55 years (Johnson,
Schirmer, and Dobbs, 2003).
Key identifying characteristics of
Silvernale phase components include the
presence of large, fortified villages usually
surrounded by numerous conical earthen
burial mounds (Gibbon, 2008).
The distribution of Oneota and
Silvernale phase village sites across the
landscape does not display a consistently
different pattern. Bryan and Energy Park
are situated on a well-drained glacial
outwash terrace underlain by sands and
gravels. Both overlook the Cannon River
and are at least 100 feet above the
floodplain. Mero is located in a similar
setting overlooking the confluence of the
Trimbelle and Mississippi rivers. Adams is
also located on a well-drained glacial
terrace, but overlooks the back channel of
the Mississippi, rather than one of its
tributary streams. Moreover, the principal
habitation area at Adams is located several
hundred feet back from the margin of the
terrace (Gibbon, 2008).
Several conclusions can be drawn
from these site settings and provisional
site types: (1) Oneota and Silvernale
village sites occur in both high terrace and
lowland settings; (2) Silvernale and/or
Oneota villages were associated with
smaller special activity localities; and (3)
Oneota villages, although fewer in number
than Silvernale villages, fall within the
same general size categories for Silvernale
villages (Gibbon, 2008).
Undocumented Red Wing Locality
sites at are risk for destruction if their land
use is marked for development. Almost all
of the mounds and a substantial part of the
village site have been destroyed by the
construction of the Red Wing Industrial
Park and housing developments on the
uplands overlooking the industrial park
(Johnson et al., 2003).
predictive modelling as it relates to
archeological sites. Other delimitations
include availability of Red Wing Locality
data resources including historic maps and
surveys, LiDAR, aerial photo imagery,
known sites, and geomorphology. Special
attention must be given to previous GISbased archeological assessments
completed within the Red Wing Locality.
Literature Review
Use of GIS in Archeological
Methodologies
Geographical Information Systems (GIS) are
being incorporated into archaeology as a
technique to improve the understanding of
spatial organization and the relationships
among finds within specific areas (Arroyo,
2009). Although their use as a basic tool in
predicting the location of archaeological
sites or in assessing the extent of their
catchment areas is relatively common,
in general, they have less often been applied
to the study of the spatial distribution of
archaeological remains within individual
deposits (Arroyo, 2009).
As archaeology is a field inherently
concerned with locating sites within a
landscape, GIS has offered a perfect
opportunity to allow the rapid mapping
and dissemination of site and settlement
pattern information (McCool, 2014).
Digital geodatabases allow the
accumulation of vast amounts of
information which can be readily accessed
with simple tools, such as overlay, or
utilized in spatial analysis. One of the
most daunting tasks in approaching any
archaeological project is in understanding
how various elements relate, as well as the
extent of work done by prior investigators.
Such problems are greatly ameliorated by
the compilation of information within a
single
Delimitations of the Problem
Delimitations of this literature review
consist of assessing the scope of
established analytical techniques for
2
resource that allows users to compare
known relationships and develop future
directions for research in the area
(McCool, 2014).
Archaeological predictive models
have been built to predict where people in
the past chose to settle, to hunt, to bury the
dead, to create or discard objects in
particular locations to the exclusion of
others (McEwan, 2012). The statistical
foundation is the first step; the next should
be to use model output data to better
understand behavior in the archaeological
past. The combination of GIS, quantitative
techniques and quantitative, experiential,
and landscape theory to interpret model
output could be one way forward
(McEwan, 2012).
The application of GIS
technologies in archaeology over the past
recent years has yielded important
expertise that can be successfully
exploited by both archaeological research
and Cultural Resource Management
(CRM)( Balla, Pavlogeorgatos, Tsiafakis,
Pavlidis, 2013).
GIS may be one of the most
important technological innovations in
archaeology in the past twenty years
(Ebert, 2004). It has made various types of
spatial analysis, especially over large
regions, possible, practical and potentially
sophisticated. GIS has had an impact in
the way that archaeology is done, both in
the academe and in the Cultural Resource
Management (CRM) industry (Ebert,
2004). We can further delineate GIS by
recognizing a hierarchy of three levels
of application in archaeology: 1) visual
ization; 2) management; and 3) analysis
(Ebert, 2004).
social, cultural or emotional variables that
might have influenced past human
behavior and new ground has been broken
recently. This work remains vital if
predictive models are to aid archaeological
research (McEwan, 2012). For although
predictive models can help archaeologists
better understand the environmental
patterns of site distributions, or predict the
locations of sites yet to be discovered, it is
more important to transform the statistical
output of a model into a better
understanding of the behavior of past
peoples (McEwan, 2012).
Balla et al., (2014) stated that the
application that has set GIS as a
mainstream tool in the field of
archaeology is predictive modelling. The
basic principle upon which this scientific
field was based on is that the locational
selection of human activity in the ancient
times was related to the current period
environmental and geographical
conditions. Based on these conditions that
characterize a location, repeating patterns
can be identified. These, compared to
patterns of other areas with similar
geographic features found at the same
period, may result in the identifying of
new locations that may also have been
occupied by similar human activities.
Predictive modelling aims at establishing a
causal relationship between certain
environmental parameters and known
archeological site locations (Balla et al.,
2014).
The data used to create an
archeological predictive model always
arise from the relationship of archeological
sites with the natural and cultural
environment. “It is clear, however, that the
input parameters of an archeological
predictive model should be associated with
the study area and the subject of study.”
(Balla et al., 2014). In reference, for
example, it is reported that many studies,
Predictive Modelling
McEwan (2012) stated that attempts have
been made to create models based on
3
which examine the locational processes of
ancient settlements (both before and after
the introduction of GIS techniques to
archeology), suggest that, apart from
socioeconomic factors, features such as
topographic relief, distance from water
bodies or soil cover type, had an important
role. Therefore, in any case, it is necessary
to study thoroughly the particular type of
archeological site and extract the criteria
that led to the specific human decision
rules. “It is clear that those factors-criteria
can vary even for the same type of
archeological site, as they may be related
to a specific time period, region or specific
cultures (Balla et al., 2014).
work took place at the site for another 62
years until Lloyd Wilford (University of
Minnesota) excavated there in 1947 and
1950. In 1947, Wilford excavated two
mounds south of the tracks and in the
“eastern part of the field” north of the
tracks. This work in the village area
revealed a number of subsurface features
that Wilford interpreted as storage pits and
fire hearths. Later, in 1950, Wilford
excavated in the “western part of the
field”. All in all, Wilford excavated a total
of 700 square feet (ca.65 square meters) of
the village (Johnson et al., 2003).
Required Data and Availability
Werbrouk, Antrop, and Zwertvaegher
(2010) state that LiDAR is an active
remote sensing technique which provides
sub-randomly distributed threedimensional terrain point data. It uses a
laser pulse to measure the terrain elevation
derived from the time between emission
and reception of reflected pulses. Through
laser altimetry, the x, y, and z-coordinates
are registered. The x and y coordinates
represent the planimetric position and the
z-coordinate the altimetric position. The
final data is generally provided in a
comma-delimited ASCII format.
Several approaches exist to obtain
the bare earth surface out of raw airborne
laser scanning point cloud. After removing
points that were reflected by trees or
shrubs, the remaining points are assumed
to represent the ground surface. In this
way it is possible to look for archeological
features under the trees (Humme,
Lindenberg, and Suer, 2010). The filtering
is done by means of the geostatistical
Kriging method, a special case of BLUP,
Best Linear Unbiased Prediction. In this
method information on correlation
between observations and relative quality
of individual observations is used to attach
LiDAR
Historic Maps and Surveys
During the mid-1880s, T. H. Lewis
recorded and mapped more than 2000
mounds clustered in this locality (Winchell
1911). At the turn of the century, J. V.
Brower and W. M. Sweeney also recorded
mounds in the region, particularly on
Prairie Island, and documented the
presence of several village sites (Gibbon,
2008). Modern archaeological
investigations began in the late 1940s with
Moreau Maxwell’s excavations for Beloit
College at the Mero site across the river in
Wisconsin and Lloyd Wilford’s
excavations at the Bryan and Silvernale
sites. Wilford recognized that the clearest
expression of Middle Mississippian traits
in the state occurred within this complex
of sites, which he referred to as the
Silvernale focus, a focus whose ceramics
were “clearly related to Aztalan, to Apple
River, and to the Monks Mound aspect.”
(Gibbon, 2008).
Lewis mapped most of the mounds
at the site and noted the presence of a
village site as well. However, no further
4
weights to observations for a prediction at
a certain location, e.g. near a Celtic field.
If a long correlation distance is used, more
weight is attached to more far
observations, which will result in a smooth
representation of the ground surface
(Humme et al., 2010).
Excavations by Lloyd Wilford of the
University of Minnesota in the 1950s and
by the Institute for Minnesota
Archaeology (IMA) in the 1980s revealed
the presence of square to rectangular
houses, a log palisade that may have been
10-12 feet high, thousands of artifacts, and
maize (corn) and maize-processing tools.
While some pottery vessels are similar to
Middle Mississippian vessels to the south,
the site’s bone, antler, and tooth tool
assemblage is Plains-related. A high
precision radiocarbon dating study by
Clark Dobbs (1993:2-3) suggests “that
Bryan was occupied for no more than one
or two generations and that the most
expansive period of occupation at Bryan
occurred between A.D. 1190 and 1223,”
the likely peak of Middle Mississippian
influence in the Red Wing locality
(Gibbon, 2008).
The Silvernale site (21GD3,
21GD17) is a large Silvernale phase
village and associated mound group on a
large, low terrace beside the Cannon River
floodplain. Excavated by Lloyd Wilford of
the University of Minnesota in the 1940s
and 1950s, Silvernale became the type-site
of the phase, which dates between A.D.
1050 and 1200. The site is especially
important because of its apparent mixture
of Middle Mississippian-related and
Oneota artifact traits and ceramic styles
and forms. At least 400 small earthen
burial mounds are thought to be associated
with the village site (Gibbon, 2008).
The Energy Park site (21GD158,
21GD52) is similar to other Silvernale
phase village sites in the Red Wing
locality in that it consists of a village area
surrounded by a group of earthen burial
mounds. However, it differs from major
village complexes like Bryan, Silvernale,
Mero, and Adams in its smaller size, less
extensive artifact deposits, and shorter
period of use, and by the presence of a
Aerial Photo Imagery
Balla et al., (2011) stated that the use of
aerial photographs for locating buried
antiquities first appeared in the early 20th
century. In the first years of satellite
images high market costs, low territorial
analysis and the lack of user-friendly
software for their processing and for
extracting data resulted in its not being
included by archaeologists in the group of
the main archaeological provisions.
The chronological and social
dimensions of burials are often clarified
through programs of excavation; however,
a full understanding requires knowledge of
the spatial geography that comes from
systematic, large-scale programs of
archaeological survey and mapping over
wider regions (Oltean, I., 2013). In this
respect, aerial and satellite imagery can be
of particular value. Recent research in the
Ponto-Danubian region has employed for
the first time, with exceptional results, an
integrated program of aerial photography
and satellite remote sensing to identify and
map barrow cemeteries and settlements,
enabling the appreciation of ancient
landscapes as an unprecedented level
(Oltean, 2013).
Known Sites
Bryan (21GD4, 21GD45) is a large
Silvernale phase village and earthen burial
mound complex on a high terrace
overlooking the Cannon River not far from
its juncture with the Mississippi River.
5
Middle Mississippian-like flat-topped
rectangular “temple” mound. This unusual
combination of traits suggests that the site
was a ceremonial center or at least played
some other special role in the locality. The
Energy Park site is particularly important
today, for the village area remains largely
undisturbed. The site was probably
occupied in the twelfth century (Gibbon,
2008).
predictive model can be useful to both
researchers and land use planners in
identifying potential undiscovered sites
and protect these cultural resources into
the future.
References
Arroya, A. 2009. Assessing What Lies
Beneath the Spatial Distribution of a
Zooarcheological Record: The Use of
GIS and Spatial Correlations at El Miron
Cave (Spain). Archaeometry, 51(3), 506524.
Balla, A., Pavlogeorgatos, G., Tsiafakis,
D., and Pavlidis, G. 2013. Locating
Macedonian tombs using predictive
modelling. Journal Of Cultural Heritage,
14(5), 403-410.
Ebert, D. 2004. Applications of
Archaeological GIS. Canadian Journal
Of Archaeology, 28(2), 319-341.
Fleming, E. 2014. Red Wing Archeology.
Science Museum of Minnesota web link.
Gibbon, G. 2008. The Silvernale Phase in
Southeastern Minnesota. University of
Minnesota web link.
Gouma, M., van Wijngaarden, G., and
Soetens, S. 2011. Assessing the effects of
geomorphological processes on
archaeological densities: a GIS case
study on Zakynthos Island, Greece.
Journal Of Archaeological Science,
38(10), 2714-2725.
Humme, A., Lindenberg, R., Suer, C.
2010. Revealing Celtic Fields From
LiDAR Data Using Kriging Based
Filtering. Delft University of
Technology, Delft Institute of Earth
Observation and Space Systems, 1-5.
Johnson, D., Schirmer, R., and Dobbs, C.
2003. Geophysics and Archaeology at
the Silvernale Site (21GD03),
Minnesota. Midwest Archaeological
Conference, 49th Annual Meeting,
Milwaukee, Wisconsin October 16-19.
Geomorphology
Gouma, van Wijngaarden, Soetens, (2011)
stated that geomorphological processes are
considered a key factor in the preservation
and visibility of archaeological surface
finds across the landscape. The
methodology of the survey includes the
pick-up survey and the detailed
geomorphological mapping of the study
areas and has been adapted to suit the
landscape characteristics, the topography
and the artefact distributions.
The assessment of the survey
results is based on detailed
geomorphological mapping of the study
area. The results of the models are
comparatively applied on the
archaeological artefact distribution maps,
in order to test the hypothesis that the
numbers of artefacts in the survey tracts
are related to topographic factors and
geomorphic processes (Gouma et al.,
2011).
Summary/Conclusion
Archeological resources in the Red Wing
Locality have been mapped for many
years but the full extent of mound and
village sites remains unknown. Leveraging
the power of GIS methodologies
(including historic maps and surveys,
LiDAR, aerial photo imagery, known
sites, and geomorphology) to create a
6
McCool, J. 2014. PRAGIS: a test case for
a web-based archaeological GIS. Journal
Of Archaeological Science, 41133-139.
McEwan, D. 2012. Qualitative Landscape
Theories and Archaeological Predictive
Modelling-A Journey Through No Man's
Land?. Journal Of Archaeological
Method & Theory, 19(4), 526-547.
Werbrouck, I. I., Antrop, M. M., Van
Eetvelde, V. V., Stal, C. C., De Maeyer,
P. h., Bats, M. M., & ... Zwertvaegher,
A. A. 2011. Digital Elevation Model
generation for historical landscape
analysis based on LiDAR data, a case
study in Flanders (Belgium). Expert
Systems With Applications, 38(7), 81788185.
7
Download