Robotic total station for microtopographic mappingan example from the Northern Great Plains.

Archaeological Prospection
Archaeol. Prospect. 13, 91–102 (2006)
Published online 22 August 2005 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/arp.270
RoboticTotal Station for
Microtopographic Mapping: an
Example from the Northern Great Plains
KENNETH L. KVAMME1*, EILEEN G. ERNENWEIN2
AND CHRISTINE J. MARKUSSEN1
1
Archeo-imaging Laboratory, Department of Anthropology, Old Main 330, University
of Arkansas, Fayetteville, AR 72701, USA
2
Environmental Dynamics, Ozark Hall 113, University of Arkansas, Fayetteville, AR 72701,
USA
ABSTRACT
Past human activities in cultural landscapes are often expressed by subtle variations in surface topography that reflect buried archaeological features.When seen from the air under low sunlight angles,
resultant ‘shadow marks’ form a cornerstone of site detection in aerial archaeology. Past attempts to
quantify and map such variations across large archaeological landscapes have resorted to aerial
photogrammetry, electronic totalstations, air- and ground-basedlidar, and kinematic globalpositioning systems. The most commonly used surveying instrument is the total station, but its slow rate of
data acquisition makes it poorly suited for collecting vast amounts of elevation data over large areas,
althoughit isoftenused for that task.A robotic total station, examined here, is arelativelynew technology that provides a rapid survey solution. It requires only a single person to operate the total station
by radio linkage from a control pad affixed to a wheeled reflector rod. As the rod is rolled over the
landscape it is automatically tracked, and measurements of surface topography may be acquired to
subcentimetre accuracy continuously, at a rate of one measurement per second. A case study from
the Double Ditch State Historic Site in the Great Plains of North Dakota, a fortified earthlodge village
with culturallysignificant surface expressions, exemplifiesthispotential.Thelociofprehistorichouses,
borrow pits, fortification ditches, middens and defensive mounds are clearly revealed in the topographic mapping.Copyright 2005 JohnWiley & Sons,Ltd.
Key words: DEM; robotic total station; microtopographic mapping; Great Plains USA; earthlodge
settlements
Introduction
The importance of topographic expressions of
past human activities in cultural landscapes as
signatures of archaeological sites cannot be overstated. Subtle and robust variations in the terrain
surface frequently signify the locations of buried
* Correspondence to: K. L. Kvamme, Archeo-imaging
Laboratory, Department of Anthropology, Old Main 330,
University of Arkansas, Fayetteville, AR 72701, USA.
E-mail: kkvamme@uark.edu
Copyright # 2005 John Wiley & Sons, Ltd.
structures, walls, ditches, former field boundaries, roads, trails and even entire villages. Their
presence makes possible one of the primary
means by which cultural features are made
visible to aerial archaeology. Terrain variations
caused by past human activity are responsible
for the shadow-marking phenomenon where
subtle changes in the ground surface become
readily visible from the air under low sunlight
angles (Wilson, 2000). They also cause differential frost or snow markings across the landscape
Received 18 December 2004
Accepted 7 June 2005
92
because melting tends to occur sooner on slopes
facing the sun than on shadowed slopes, and
surface depressions and leeward slopes trap
blowing snow, making such features more visible from an aerial vantage (Wilson, 2000). These
same terrain variations also have an impact on
aerial thermography when imagery is acquired
under sunny conditions (because sunlit slopes
tend to be warmer than shadowed ones; Scollar
et al., 1990), and have an impact on electrical
resistivity surveys where local summits and
depressions tend to exhibit high and low resistivities, respectively (caused primarily by correlated moisture variations; Clark, 2000).
The quantification of anthropogenic variations
in landform has therefore received large interest,
and several methods, briefly reviewed here, have
been devised to accomplish it. Coupled with this
advance is the incorporation of computer graphic techniques for mapping that allow subtle,
and even previously unknown, landscape features to be revealed. These methods include (i)
extremely dense contouring of the elevation
data (using narrow contour intervals of only a
few centimetres), (ii) pseudo-three-dimensional
views with extreme vertical exaggeration (by
several hundred per cent), and (iii) analytical
surface shading using light-source angles and
directions far surpassing those available in nature (see Newman, 1993; Doody et al., 1995;
Barratt et al., 2000; Chapman and Van de Noort,
2001; Kvamme, 2005a). This work has presented
high-resolution landscape mapping as a legitimate domain of remote sensing, because it is
non-invasive and capable of yielding new information by revealing archaeological features normally undetectable to the human eye.
Aerial photogrammetry for terrain
characterization
Early attempts to quantify culturally induced
variations in the landscape surface relied primarily on aerial photogrammetry. Pioneering work
in the American Southwest focused on robust
standing architecture of Puebloan ruins that
were well-mapped using one-foot (30.5 cm)
intervals in photogrammetrically generated contour maps (Pouls et al., 1976). These same data
sets also revealed minor and faintly perceived
Copyright # 2005 John Wiley & Sons, Ltd.
K. L. Kvamme et al.
circular depressions in surrounding areas that
indicated the loci of buried kivas—subterranean
ceremonial structures associated with the ruins.
Recent aerial photogrammetric work offers
mappings of the ground surface of much higher
precision—Stone
(2003)
reports 10 cm—
allowing smaller terrain variations to be captured in the results. A photogrammetric example
is given below.
Electronic total stations for terrain
characterization
Laser distance measuring systems in electronic
total stations, or even antiquated optical transits,
have long been used to map anthropogenically
altered topographic surfaces to centimetre-level
accuracy. Early work of this nature is illustrated
by Frankel (1980), in Australia, and Shapiro and
Miller (1990), in the USA. Numerous similar
studies have since been conducted around the
globe, with a significant concentration of work in
Ireland. Newman (1993), for example, maps the
topographic surface of the prehistoric ritual landscape at Tara (which holds several barrows and
other ancient monuments); Doody (1993) presents similar results for a complex of Bronze
Age barrows and enclosures at Chancellorsland;
and a complete survey of the heavily vegetated
bivallate hillfort of Castle Gale, on the promontory of Carrig Henry, is described by Doody et al.
(1995).
The principal drawback of traditional total
station mappings is the tremendous amount of
labour required to survey large areas, with the
station operator commonly required to set up
and work from several datums, a second individual constantly moving and placing a reflector
rod, and the difficulty of communication
between the two (commonly alleviated by radio
headsets). Moving the reflector rod along parallel
linear traverses separated by a constant interval
(e.g. 1 m), with measurements sampled at a similar interval along each traverse, ensures a uniformly sampled matrix of elevations for creation
of a digital elevation model (DEM), but such a
field protocol can take an enormous amount of
time. Kvamme (2005a) describes the microtopographic mapping of the Fort Clark Trading Post, a
mid-nineteenth century outpost of the American
Archaeol. Prospect. 13, 91–102 (2006)
Robotic Total Station Mapping
fur trade in North Dakota. Uniform sampling of
the 4000 m2 region at 1 m intervals along x- and yaxes required approximately 4 days for the twoperson crew of station operator and rod holder.
In contrast to uniform spatial sampling, irregularly placed samples offer a number of clear
advantages because fewer measurements need to
be taken in areas of little topographic change,
and additional ones may be placed to better
characterize complex surfaces or areas of particular interest. The disadvantage of an irregular
survey protocol is that it is largely judgmental,
depending on the rod holder’s skill and perception in defining or observing significant landscape change points while meeting, at the same
time, the need to provide sufficient coverage over
large areas. In practice, a combination of uniform
and irregular sampling is most effective. This is
well demonstrated in a computer simulation by
Fletcher and Spicer (1988) that shows the benefits
of combining uniform sampling of a region supplemented with additional points judgmentally
placed over areas of significant topographic variation. Most workers have since followed this
protocol (e.g. Doody 1993; Barratt et al., 2000;
Chapman and Van de Noort 2001; Ainsworth
and Thomason, 2003; see also below).
New technologies for microtopographic
characterization
Several new technologies have emerged in recent
years that are capable of rapidly producing high
resolution DEM over large areas. One is airborne
lidar (LIght Detection And Ranging) that offers
horizontal sampling densities well below 1 m
with absolute vertical accuracy as good as
15 cm (although relative accuracy between
nearby measurements is considerably higher
(Flood, 2001). Barnes (2003) illustrates an
impressive archaeological application with a
lidar produced DEM of a 2.4 km2 region surrounding the Iron Age Sidbury Hillfort, England,
that clearly reveals earthworks, barrows, trails
and tracks. Ground-based lidar systems, more
commonly referred to as ‘three-dimensional laser
scanners’, have also been used to create high
resolution DEMs (Neubauer, 2004). Designed
primarily for creating three-dimensional models
of nearby architecture, numerous stations must
Copyright # 2005 John Wiley & Sons, Ltd.
93
be used to produce large-area DEMs and the low
oblique scanning angle, even when placed on a
raised platform, means that surface vegetation—
including blades of grass—must be filtered out
from the scans using complex algorithms (Limp,
2004).
Surveyor-grade real time kinematic global
positioning systems (GPS) have also been
explored for mapping topographic surfaces at
archaeological sites and producing DEMs holding great detail. Barratt et al. (2000) describe a
35 ha survey at the Roman town of Viroconium
(one of the largest urban centers in Roman
Britania). Using a series of previously established
datums, a direct radio link between a base station
and a roving unit allowed differential corrections
on the fly, with subcentimetre accuracy. A 5-m
traverse separation was generally used, although
2- or 1-m traverse separations were used where
more information was desired. Measurements
were typically sampled every 2 m along traverses, although minimum intervals as small as
0.1 m were occasionally adopted over important
features. Chapman and Van de Noort (2001)
used similar GPS methods in two wetlands
surveys in England by collecting data points
every 8 m in ‘background’ regions and at
less than 1 m intervals in areas of high archaeological potential. A recent publication by English Heritage (Ainsworth and Thomason, 2003)
illustrates similar case studies.
Use and availability
All of the foregoing approaches are capable of
producing high-resolution DEMs of great accuracy (or relative accuracy in the case of aerial
lidar) that reveal culturally generated terrain
variations, including those of great subtlety.
Yet, depending on location and budget, some
methods may not be readily available. Aerial
lidar is practically inaccessible to most archaeological projects in the USA, for example, and
surveyor-grade kinematic GPS, although growing in popularity, is expensive and therefore not
widely used. For this reason, the total station, a
common archaeological mapping tool, remains
of necessity the primary method for landscape
characterization in many regions. The purpose
of this paper is to present a relatively new
Archaeol. Prospect. 13, 91–102 (2006)
94
development in total station technology that
offers very rapid survey coverage at half the
labour costs of a conventional unit. The robotic
total station rivals or surpasses kinematic GPS in
accuracy and speed of survey using far less
expensive instrumentation.
The robotic total station
The robotic total station is a relatively new electronic distance measuring (EDM) system with
many automated capabilities. The Trimble 5600
is one such instrument that contains a highly
accurate EDM (allowing subcentimetre work), a
four-speed servo for rapid and accurate aiming,
AutolockTM technology for automatic tracking of
the target as it is moved, and robotic capabilities
allowing one-person survey through radio linkage (halving crew requirements). The servodriven aiming, Autolock and robotic capabilities
mean that all operations may be conducted by a
single person through a control pad affixed to the
reflector rod that is radio-linked to the total
station (Figure 1). A further innovation that
K. L. Kvamme et al.
enables a tremendous leap in survey speed is a
wheel attached to the base of the reflector rod.
This allows the operator to simply roll the
rod over the landscape, controlling the survey
by radio from the control pad, while the total
station automatically tracks the roving rod and
records data. For expedient surveys data may be
acquired in a continuous mode at a rate of
one measurement per second. The x, y (positional) and z (elevation) coordinates are transmitted back to the control pad on the roving
unit and displayed in map form in real time
(Figure 1). Data are easily downloaded as
comma delimited, x, y, z data triplets, in an
ASCII format for computer processing and map
creation.
The Trimble 5600 has been used by the University of Arkansas on a number of archaeological sites for the specific purpose of creating dense
DEMs over large areas with 1 m sample separations or less and subcentimetre accuracy (e.g. see
Herrmann, 2004). Several difficulties have been
resolved during this work. The original 15 cm
diameter rod wheel that came with the unit
was far too small for efficient movement over
Figure 1. The Trimble 5600 robotic total station (left), rod-mounted control pad with real-time display (middle), and wheeled
reflector rod in use (right).
Copyright # 2005 John Wiley & Sons, Ltd.
Archaeol. Prospect. 13, 91–102 (2006)
Robotic Total Station Mapping
terrain—it frequently became stuck between
clumps of grass, in surface detritus, or rodent
holes. A larger 35 cm diameter wheel was easily
installed and remedied this situation. Maintaining the rod nearly vertical while it is in motion
remains a constant problem, but one that experience helps to mitigate. With the height of the
reflector typically 2 m above the ground surface,
a simple calculation reveals that if the rod is
angled by only 15 off-vertical, horizontal accuracy can shift by more than a 0.5 m and vertical
accuracy by 0.07 m. The Windows CE operating
system occasionally ‘crashes’ during survey,
causing delays because the instrument must be
rebooted and the station set-up procedure rerun.
Initially, battery life appeared to be an issue with
a ‘low battery’ signal appearing after only 1–2 h
of use. It has been established that the gauge is
inaccurate, and that batteries last for an entire
day. The instrument’s range also poses difficulties. The maximum range using the more accurate miniprism under Autolock tracking is only
about 150 m in clear weather, with that distance
rapidly diminishing under light rain or even
high humidity. A minimum effective range of
about 10 m has also been determined, dictated by
the inability of the Autolock tracking to keep up
with the relatively fast moving rod at close distances. Finally, obstructions to the laser beam,
whether caused by signage, vegetation, rising
ground, or people, are a regular difficulty that
must be closely monitored, because the Autolock
tracking may be lost and must be reacquired
before survey can recommence (the operator,
too, may cause beam blockage, but this circumstance is easily avoided by ensuring that the
reflector is above the operator).
Case study: mapping microtopography
at the Double Ditch State Historic Site
The Double Ditch State Historic Site (32BL8) is a
large, fortified, earthlodge village located on a
high terrace above the Missouri River, near
Bismarck, North Dakota (historically, the earthlodge was a dome-shaped dwelling composed of
timbers, 15–18 m in diameter, and covered with
0.5 m of earth). When Lewis and Clark of the
Voyage of Discovery first reported the site in
Copyright # 2005 John Wiley & Sons, Ltd.
95
1804, it had been abandoned by the Mandan
about 20 years previously, most probably owing
to a smallpox epidemic that hit the region in the
early 1780s (Ahler, 2004). As it exists today,
Double Ditch is one of the largest and best
preserved of the remaining earthlodge village
sites. It is distinctive in its many clear terrain
expressions of cultural features and by its large
size, with an approximately 6 ha area showing
significant topographic expressions. Two robust
fortification ditches, in places more than 1 m
deep, are readily seen in the ground surface
(prehistorically, they were lined on the interior
side with a palisade of large posts). Nearly 100
shallow depressions 12–50 m in diameter signify
the locations of former earthlodges and soil
borrowing areas. Above them lie numerous large
mounds (probably part of defensive works)
and mounded middens, some approaching
3 m in height. Hailey (2005) illustrates many of
these features from an aerial vantage (see also
Figure 2).
Double Ditch has been the focus of an intensive 4-year remote sensing project that has been
combined with archaeological excavations to
better understand and document this important
site to Northern Great Plains prehistory (Ahler,
2004; Kvamme, 2005b). The remote sensing has
included aerial imaging and thermography of
the entire park, complete geophysical surveys
by magnetic gradiometry and electrical resistivity methods, and large areas of groundpenetrating radar and electromagnetic induction
surveys. Significant findings (many verified by
excavation) include the identification of thousands of subterranean corn storage pits (about
1.5 m deep by 1.5 m in diameter) that testify to
high levels of maize horticultural production.
Extensive areas of earth borrowing for earthlodge coverings and fortifications were also
defined. Most significantly, the discovery of
two previously unknown fortification ditches,
invisible on the surface, expanded the village’s
size to about 10 ha. These and other findings
have begun to alter perceptions of the complexity
of the village way of life on the northern Great
Plains (Ahler, 2004).
Near the close of the project in 2004, it was
decided to further document Double Ditch’s core
village area through creation of a high resolution
Archaeol. Prospect. 13, 91–102 (2006)
96
K. L. Kvamme et al.
Figure 2. Aerialview of Double Ditch under low sunlight conditions showing anthropogenic variationsin microtopography (photograph credit: Tommy Ike Hailey).
DEM. Such a data set would document the state
of the site’s surface as it existed in 2004, and
could serve as a baseline for subsequent change
studies resulting from normal erosion or intensive tourist traffic. The data would also be useful
for examining certain aspects of this cultural
landscape—for instance, size distributions of
‘house’ (earthlodge) or earth borrowing depressions, middens and mounds. When merged with
evidence of the two hidden fortification ditches
revealed by the geophysics, a DEM might also
allow better understanding of the distribution of
the many large and built-up mounds as components of these fortifications (see below). Finally, it
is well known that terrain form holds significant
relationships with various types of remote sensing, such as electrical resistivity and aerial thermography. A high-resolution DEM would
potentially allow study of terrain effects on these
data sets.
Copyright # 2005 John Wiley & Sons, Ltd.
Parkwide ‘background’ DEM through
photogrammetry
A topographic map with a 25 cm contour interval
was available for the entire 700 800 m park
area. It was created by photogrammetry in 1994
by a local surveying company, KBM Inc., at the
request of the State Historical Society of North
Dakota. Although the focus of the robotic total
station survey was to lie only within the confines
of the approximate 260 280 m area of the village proper, it was decided to first create a parkwide DEM using the relatively crude KBM data
as a background into which the results of the
more detailed survey could be inserted. The
KBM map reveals many of the Double Ditch’s
principal topographic features in the elevation
contours (Figure 3). Hundreds of isolated elevations are also included on the map at significant
change points (e.g. mound summits, low points
Archaeol. Prospect. 13, 91–102 (2006)
Robotic Total Station Mapping
97
Figure 3. The photogrammetrically produced 25-cm contour map of the entire park by KBM, Inc. (left) was used to create aTINgenerated DEM (upper right), but only after significant editing to close line labelling gaps (indicated by arrows, inset, lower right).
in depressions). Fortunately, the spatial component of the KBM data existed in digital form that
allowed extraction of elevation contours and
points into a geographical information system
(GIS) as line and point ‘vectors.’ Unfortunately,
the attribute information for each vector was
unavailable except in paper copies, necessitating
each vector to be manually tagged with its elevation. Significant editing was also required to
connect labelling gaps left in contour lines (inset,
Figure 3). Finally, hundreds of surplus lines that
denoted other park features (fence lines, roads,
monuments, trails) had to be removed. These
elevation vectors were then registered to the
project’s arbitrary plane coordinate system utilizing stone monuments built in the 1930s that
surround the park periphery as ground control
points. Overall RMS error for the 700 800 m
park area was 0.292 m.
A DEM was produced by first pre-processing
each line vector to add vertices to ensure no data
gaps larger than 2 m. This enabled a more
detailed triangular irregular network (TIN)
describing the park’s surface to be generated
with fewer large gaps in the Delauney triangulation, resulting in a smoother surface (Burrough
and McDonnell, 1998). Using this TIN surface, a
Copyright # 2005 John Wiley & Sons, Ltd.
raster DEM was created with a horizontal spatial
resolution of a 0.5 m, compatible with several of
the geophysical data sets. The result, although
illustrating very well Double Ditch’s prominent
characteristics (Figure 3), tends to lack many
details, inadequately portraying sizes and shapes
of many of the site’s features owing to the limited
information content offered by 25-cm contour
lines. These findings verified the need for the
more accurate representation by robotic total
station survey.
DEM by robotic total station
A rapid data collection protocol was established
at Double Ditch for the Trimble 5600 survey. It
made use of extant 20 m square data collection
units, previously located by total station for the
geophysical surveys, that were marked across
the site by 1 m long PVC pipes placed vertically
in the ground at their corners (see background
in Figure 1). By utilizing the 20 m units, the
topographic survey could be conducted piecemeal, unit-by-unit. In each 20 m unit, survey
was accomplished by placing pin flags at 5-m
intervals at opposite ends of each unit to serve as
guides for the survey transects. With an initial
Archaeol. Prospect. 13, 91–102 (2006)
98
sampling density goal of one measurement per
metre in both x- and y-axes, traverses were
made approximately every metre by zigzagging
through each 20 m unit and visually estimating
approximate traverse locations vis-à-vis the pin
flags, PVC corner markers and the real-time
display on the rod’s control unit (Figure 1).
Within each traverse, the wheeled reflector rod
was moved at a rate of somewhat less than
1 m s 1 to achieve better than one measurement
per metre based on the instrument’s continuous
mode sampling rate. It is emphasized that additional elevations were also acquired over particularly variable, small and culturally significant
landscape features in each 20 m unit.
Approximately 15 min were normally required
to survey a 20 m unit, including pin flag set-up
and the walking of 20 traverses. With fieldwork
K. L. Kvamme et al.
coming to a close and the entire village core
surveyed with 1 m traverses (the area within
the innermost ditch; Figures 2 and 3), it was
decided to attempt complete survey of the larger
village area encompassed by the outer topographically indicated ditch with 4 days remaining to
the project (Figures 2 and 3). This was accomplished by changing traverse separation from
1 m to 2 m (although samples within each traverse were maintained at approximately 1 m
intervals). This decreased survey time to as little
as 10 min per 20 m unit, but resulted in somewhat diminished detail in the resultant DEM that
occasionally becomes apparent in certain forms
of display.
A mapping of a portion of the data postings
reveals much about the nature of our surveys
(Figure 4). With data collected continuously, the
Figure 4. Data postings from the robotic total station showing 1m versus 2 m traverse separations and clustering of data points
near 20 m collection unit boundaries, indicated by the backgroundgrid.Holesin the point distributionrepresent the lociofon-going
excavations (E) and certain data clusters indicate judgmentally placed measurements (J).
Copyright # 2005 John Wiley & Sons, Ltd.
Archaeol. Prospect. 13, 91–102 (2006)
Robotic Total Station Mapping
99
Figure 5. Digital elevation model (DEM) produced by the robotic total station (left), overlaid on the larger photogrammetrically
generated DEM (right).
zigzag survey in 20 m collection units caused
data points to accumulate at the end of each
north–south traverse, or when the operator happened to stop in mid-transect (perhaps to speak
with tourists). The 1 m versus 2 m traverse
separations are apparent, and the fact that traverses are generally parallel and uniformly separated is encouraging. The few holes visible in the
data were caused by the need to avoid open
excavations, and some data clustering or additional survey lines point to areas of extraordinary
topographic change or special interest (Figure 4).
Geographical information system methods
were used to thin dense point clusters by removing measurements separated by less than 0.5 m.
A TIN surface was then generated between the
remaining points, and a 0.5 m horizontal-resolution raster DEM was created from this surface.
The result holds remarkable detail, showing not
only house and borrow pit depressions, ditches,
mounds and middens, but very small cavities
throughout the landscape and a generally richer
texture. Many of the small cavities reflect looter’s
pits from early in the past century but, significantly, some may represent collapsed prehistoric
subterranean storage pits (Figure 5). The DEM
Copyright # 2005 John Wiley & Sons, Ltd.
generated by robotic total station was ultimately
mated with the larger photogrammetrically produced DEM using GIS overlay methods. This
allowed creation of an ultimate park-wide DEM
for display and analysis purposes (Figure 5).
Discussion
A 60 80 m segment of the photogrammetrically
produced DEM shows its inaccuracies and simplification of the surface when compared against
the more detailed results obtained by the robotic
total station (Figure 6). It is clear that 25-cm
contours are inadequate for revealing many features in a cultural landscape, even at a site with
such robust terrain expressions as Double Ditch
(given vertical precision differences between the
two data sets of 1 cm versus 25 cm, further comparisons are pointless). A cross-section across
several of Double Ditch’s house, ditch and
mound features suggests the potential of the
robotic total station’s data for exploring such
issues as house size, deposit formation around
houses, ditch depth and mound formation
(Figure 6). The full analytical worth of these
Archaeol. Prospect. 13, 91–102 (2006)
100
K. L. Kvamme et al.
Figure 6. Comparison of details and feature content for the same region between DEMs generated by photogrammetry (left) and
the robotic total station (right). Note absence of features in the left profile.
data has not yet been explored, but probably will
include GIS-based studies of viewsheds and visibility from key defensive mounds relative to the
site’s fortifications, probable corridors of travel
within the village, volumetric studies of mounds
and borrow pits, and house size studies. Terrain
effects on geophysical data, particularly thermal
and electrical resistivity, also may be explored.
Some of the DEM’s potential for better understanding Double Ditch is suggested in Figure 7,
where the magnetic gradiometry data have been
draped over the elevation surface. In this view,
Figure 7. A segment of the magnetic gradiometry data overlaid on the park-wide DEM showing relationships between large
mounds and fortification ditches (labelled1^4). Ditches 3 and 4 are visible only to magnetometry.
Copyright # 2005 John Wiley & Sons, Ltd.
Archaeol. Prospect. 13, 91–102 (2006)
Robotic Total Station Mapping
topographically expressed fortification Ditches 1
and 2 may be seen in the fore- and middleground, while magnetically revealed Ditches 3
and 4 are seen along the background (a series of
regularly spaced, U-shaped bastions occur at
regular intervals along outer Ditch 4; Kvamme,
2005b). The underlying topographic data allow
simultaneous visualization of the magnetic
anomalies and the loci of the large mounds,
many of which probably served a defensive
function such as viewing platforms, bastions or
perhaps ramparts. Analysis of imagery such as
Figure 7 may ultimately better clarify the role
that mounds played in Double Ditch’s defenses.
Acknowledgements
Eileen Ernenwein, doctoral student at the University of Arkansas, was in charge of directing
the Trimble 5600 survey at Double Ditch,
although most of it was physically performed
by Andrew Gold (after training), a local resident
of Bismarck. Christine Markussen, M.A. student
at the University of Arkansas, did most of the
work to create the photogrammetric DEM. This
research was made possible by a grant from the
PaleoCultural Research Group, Stanley A. Ahler,
Director, of Flagstaff, Arizona, in league with the
Historic Preservation Division, Fern Swenson,
Director, of the State Historical Society of North
Dakota. Stan Ahler is thanked for engineering
the transition to a larger survey wheel for the
reflector rod. Use of the Trimble 5600 was generously granted by the Center for Advanced
Spatial Technologies, University of Arkansas,
directed by W. Fredrick Limp. The Trimble
5600 was obtained through a Major Research
Instrumentation grant awarded to the University
of Arkansas by the U.S. National Science Foundation (grant number BCS-0321286), W. Fredrick
Limp, Kenneth L. Kvamme, Sreekak Bajwa,
Stephen Beaupre, and Steve Burian, Principal
Investigators.
References
Ahler SA. 2004. Introduction and background for
2003–2004 studies. In Archaeological Investigations
During 2003 at Double Ditch State Historic Site,
North Dakota, Ahler, SA (ed.). Unpublished report
submitted to the State Historical Society of North
Dakota, Bismarck; 1–16.
Copyright # 2005 John Wiley & Sons, Ltd.
101
Ainsworth S, Thomason B. 2003. Where on Earth are
We? The Global Positioning System (GPS) in Archaeological Field Survey. English Heritage: Swindon.
Barnes I. 2003. Aerial remote-sensing techniques
used in the management of archaeological monuments on the British Army’s Salisbury Plain
training area, Wiltshire, UK. Archaeological
Prospection 10: 83–90.
Barratt G, Gaffney V, Goodchild H, Wilkes S. 2000.
Survey at Wroxeter using carrier phase, differential GPS surveying techniques. Archaeological
Prospection 7: 133–143.
Burrough PA, McDonnell RA. 1998. Principles of
Geographical Information Systems. Oxford University Press: Oxford.
Chapman HP, Van de Noort R. 2001. Highresolution wetland prospection, using GPS
and GIS: landscape studies at Sutton Common
(South Yorkshire), and Meare Village East
(Somerset). Journal of Archaeological Science 28:
365–375.
Clark A. 2000. Seeing Beneath the Soil: Prospection
Methods in Archaeology. Reprinted. Routledge:
London. Originally Published 1990, B.T. Batsford
Ltd: London.
Doody M. 1993. Ballyhoura Hills Project interim
report. In Discovery Programme Reports: 1, Project
Results 1992. Royal Irish Academy: Dublin; 20–30.
Doody M, Synnott P, Tobin R, Masterson B. 1995. A
topographic survey of the inland promontory fort
at Castle Gale, Carrig Henry, Co. Limerick. In
Discovery Programme Reports: 2, Project Results
1993. Royal Irish Academy: Dublin; 39–44.
Fletcher M, Spicer, D. 1988. Clonehenge: an experiment with gridded and non-gridded survey data.
In Computer and Quantitative Methods in Archaeology 1988, Vol. 2, Rahtz SPQ (ed.). British Archaeological Reports, International Series 446(ii): 309–
325.
Flood M. 2001. Laser altimetry: from science to
commercial lidar mapping. Photogrammetric Engineering and Remote Sensing 67: 1209–1217.
Frankel D. 1980. Contour-plans and surface plotting: aids for the field archaeologist. Journal of
Field Archaeology 7: 367–372.
Hailey TI. 2005. The powered parachute as an
archaeological aerial reconnaissance vehicle. Archaeological Prospection 12: 69–78.
Herrmann JT. 2004. Interpreting Leetown through the
integration of aerial and ground-based remote sensing.
Unpublished M.A. thesis, Department of Anthropology, University of Arkansas, Fayetteville.
Kvamme KL. 2005a. Terrestrial remote sensing
in archaeology. In Handbook of Archaeological
Methods, Maschner HDG, Chippendale C (eds).
AltaMira: Lanham, MD; in press.
Kvamme KL. 2005b. Geophysical Findings at Double
Ditch State Historic Site (32BL8), North Dakota,
2001–2004. Unpublished report submitted to the
Archaeol. Prospect. 13, 91–102 (2006)
102
State Historical Society of North Dakota:
Bismarck, ND.
Limp WF. 2004. Application of ground-based LIDAR
and other innovative photogrammetric methods to the
documentation and interpretation of historic structures and archaeological sites. Paper presented at
the Partners in Environmental Technology Symposium, sponsored by the Strategic Environmental Research and Development Program,
Washington, DC.
Neubauer W. 2004. GIS in archaeology—the interface between prospection and excavation. Archaeological Prospection 11: 159–166.
Newman C. 1993. The Tara survey interim report.
In Discovery Programme Reports: 1, Project Results
1992. Royal Irish Academy: Dublin; 70–93.
Pouls BG, Lyons TR, Ebert JI. 1976. Photogrammetric mapping and digitization of prehistoric
architecture: techniques and applications in
Chaco Canyon National Monument, New Mex-
Copyright # 2005 John Wiley & Sons, Ltd.
K. L. Kvamme et al.
ico. In Remote Sensing Experiments in Cultural
Resource Studies: Non-destructive Methods of
Archaeological Exploration, Survey, and Analysis,
Lyons TR (ed.). National Park Service, U.S.
Department of the Interior and University of
New Mexico: Albuquerque; 103–114.
Scollar I, Tabbagh A, Hesse A, Herzog I. 1990.
Archaeological Prospection and Remote Sensing.
Cambridge University Press: Cambridge.
Shapiro G, Miller JJ. 1990. The seventeenth-century
landscape of San Luis de Talimali: three scales
of analysis. In Earth Patterns: Essays in
Landscape Archaeology, Kelso WM, Most R (eds).
University of Virginia Press: Charlottesville;
89–101.
Stone J. 2003. Cawthorn Camps, North Yorkshire—
a photogrammetric approach. Archaeological Prospection 10: 153–157.
Wilson DR. 2000. Air Photo Interpretation for Archaeologists. Arcadia Publishing: Charleston, SC.
Archaeol. Prospect. 13, 91–102 (2006)