University of Southern California Dissertations and Theses

Modeling prehistoric paths in Bronze Age Northeast England

(USC Thesis Other)

Modeling prehistoric paths in Bronze Age Northeast England

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Modeling Prehistoric Paths in Bronze Age Northeast England

by

Christian Alvez

A Thesis Presented to the
Faculty of the USC Graduate School
University of Southern California
In Partial Fulfillment of the
Requirements for the Degree
Master of Science
(Geographic Information Science and Technology)

December 2016

Copyright ® 2016 by Christian Alvez. All rights reserved

To my grandmother Aurora who taught me how to read and write; to my parents Alfredo and
Amelita, and my brothers Jeff and John for showing me love; to Tatsuya for supporting and
listening; to Ariel and Yanin for giving me inspiration; and finally to Wayne Karr who taught me
how to care for and love others.

iii
Table of Contents
List of Figures ................................................................................................................................. v
List of Tables ................................................................................................................................ vii
Acknowledgements ...................................................................................................................... viii
List of Abbreviations ..................................................................................................................... ix
Abstract ........................................................................................................................................... x
Chapter 1 Introduction .................................................................................................................... 1
1.1 Background and Motivation ................................................................................................2
1.1.1. Least Cost Path analysis .............................................................................................3
1.1.2. Least Cost Path in archaeology ..................................................................................4
1.2 Study Area ...........................................................................................................................5
1.2.1. The Beaker Culture ....................................................................................................7
1.3 Research Objectives ...........................................................................................................12
1.4 Thesis Organization ...........................................................................................................12
Chapter 2 Background and Related Work .................................................................................... 13
2.1 GIS and Archaeology .........................................................................................................13
2.1 LCP Overview ...................................................................................................................14
2.1.1. LCP Studies in archaeology .....................................................................................14
2.2 Gathering Data ...................................................................................................................16
2.2.1. DEMs .......................................................................................................................16
2.2.2. Bronze Age cultural data: Settlement and burial site points ....................................19
2.2.3. Other cost variables ..................................................................................................20
2.3 Generating a Cost Surface .................................................................................................21
2.3.1. Anisotropic Modeling: Tobler’s hiking function and Naismith’s rule ....................22
2.4 LCP Algorithm ...................................................................................................................23
2.5 Validity Assessment ...........................................................................................................23
Chapter 3 Methods ........................................................................................................................ 25
iv
3.1 Data ....................................................................................................................................25
3.1.1. Acquiring and preparing the cultural dataset ...........................................................25
3.1.2. Generating the origin and destination points ...........................................................27
3.1.3. Acquiring the DEM ..................................................................................................34
3.1.4. Creating the slope raster ...........................................................................................36
3.2 Methodology ......................................................................................................................38
3.2.1. Executing the Path Distance tool .............................................................................40
3.2.2. Generating LCPs ......................................................................................................41
3.2.3. Conducting the validity assessment .........................................................................42
Chapter 4 Results .......................................................................................................................... 44
4.1 Executing LCPs .................................................................................................................44
4.2 Cluster LCPs ......................................................................................................................46
4.2.1. Cluster 1 ...................................................................................................................46
4.2.2. Cluster 2 ...................................................................................................................48
4.2.3. Cluster 3 ...................................................................................................................49
4.2.4. Cluster 4 ...................................................................................................................51
4.3 Inter-cluster LCPs ..............................................................................................................53
4.4 Aerial Imagery Comparisons .............................................................................................56
Chapter 5 Discussion and Conclusions ......................................................................................... 62
REFERENCES ............................................................................................................................. 68
Appendix A: Settlement Data Record Sources ............................................................................. 76

v
List of Figures
Figure 1 Study Area – Northeast England ...................................................................................... 6
Figure 2 Beaker artifact samples – pottery, flint tools, and arrowheads ...................................... 10
Figure 3 Overview of selected clusters of settlement and burial sites .......................................... 27
Figure 4 Cluster 1with origin and destination points (LCP nodes) .............................................. 29
Figure 5 Cluster 2 with origin and destination points (LCP nodes) ............................................. 30
Figure 6 Cluster 3 with origin and destination points (LCP nodes) ............................................. 31
Figure 7 Cluster 4 with origin and destination points (LCP nodes) ............................................. 32
Figure 8 Inter-cluster origin and destination points (LCP nodes) ................................................. 33
Figure 9 STRM 30 m resolution ................................................................................................... 35
Figure 10 ASTER GDEM V2 30 m resolution ............................................................................. 35
Figure 11 Clusters in smaller DEM (with hillshade effect) derived from
the larger SRTM DEM .................................................................................................................. 36
Figure 12 Slope raster with cultural sites ...................................................................................... 37
Figure 13 The Path Distance model, with the origin, cost surface, and backlink surface
indicated as model parameters .............................................................................................. 39
Figure 14 The Cost Path model, with the destination, cost surface, backlink surface, and the
converted LCP raster (LCP_Vector) indicated as model parameters .................................... 39
Figure 15 Path Distance tool inputs .............................................................................................. 40
Figure 16 Cost Path tool inputs ..................................................................................................... 42
Figure 17 Overview of cluster and inter-cluster primary and modified LCPs ............................. 45
Figure 18 Overview of Cluster 1 LCPs with original and modified segments ............................. 46
Figure 19 Detail of Cluster 1 – Segment 4 (Harehope Hill – West Plain Henge) ....................... 47
Figure 20 Detail of Cluster 1 – Segment 5 (West Plain Henge – Lookout Plantation) and
Modified Segment 4 (Cheviot Quarry – Lookout Plantation) ............................................... 48
Figure 21 Overview of Cluster 2 - Segment 1 (Cheviot Walk Wood – Blawearie Cairn),
Segment 2 (Blawearie Cairn – Hepburn Crag Plantation), and Segment 3 (Hepburn Crag
vi
Plantation – Rosebrough Moor Cairn 1), with Modified Segment 1 (Cheviot Walk Wood –
Millstone Hill) and Modified Segment 2 (Hepburn Crag Plantation – Rosebrough Moor
Cairn 2) .................................................................................................................................. 49
Figure 22 Overview of Cluster 3 LCPs – Segment 1 (Alwinton – Kirkhill Cemetery), with
Modified Segment 1 (Farnham – Hedley Wood); Segment 2 (Kirkhill Cemetery – Great
Tosson Quarry), with Modified Segment 2 (Holystone Common 1 – Great Tosson Quarry);
and Segment 3 (Great Tosson Quarry – Debdon Whitefield), with Modified Segment 3
(Spital Hill Cairn 1 – Debdon Whitefield) ............................................................................ 50
Figure 23 Detail of Cluster 3 – Segment 3 and Modified Segment 3, with sections of exact
coincidence of LCPs from both segments in their northeastern portions by Debdon
Whitefield.. ............................................................................................................................ 51
Figure 24 Overview of Cluster 4 LCPs – Segments 1 (NW High Carry House – Warkshaugh
Farm) and 2 (Warkshaugh Farm – Reaverhill Farm) with Modified Segment 1 (NW High
Carry House –Reaverhill Farm) ............................................................................................ 52
Figure 25 Overview of inter-cluster LCPs. ................................................................................... 53
Figure 26 Detail of inter-cluster Segment 1 (NW High Carry House – Alwinton) with Modified
Segment 1 (Wark Manor House – Farnham) ........................................................................ 54
Figure 27 Detail of inter-cluster Segment 4 (Knock Hill – Lookout Plantation) with Modified
Segment 4 (Reavely Hill – Lookout Plantation, Segment 5 (Hepburn Crag Plantation –
Lookout Plantation) with Modified Segment 5 (Millstone Hill – Lookout Plantation). ....... 55
Figure 28 Cluster 3 – Aerial of Segment 3 (Great Tosson Quarry – Debdon Whitefield) with
Modified Segment 3 (Spital Hill Cairn 1 – Debdon Whitefield) .......................................... 57
Figure 29 Aerial of the southern portions of inter-cluster Segment 1 and Modified Segment 1 .. 58
Figure 30 Figure 30 Aerial of the southern section of inter-cluster Segment 2 (Alwinton – Knock
Hill), with Modified Segment 2 (Farnham – Alnham) .......................................................... 59
Figure 31 Aerial of inter-cluster Segments 4 and 5 with Modified Segments 4 and 5; Cluster 1 –
Segment 5 with Modified Segment 4 .................................................................................... 60
Figure 32 Inter-cluster – Aerial of inter-cluster Segments 4 and 5, Modified Segments 4 and 5;
Cluster 1 – Segment 5 with Modified Segment 4, with multiple ring buffer (30%
transparency) ......................................................................................................................... 61
Figure 33 Cultural sites and LCPs in the southern portion of Cluster 1, with Inter-cluster
Segment 2 .............................................................................................................................. 64
Figure 34 Alwinton and Farnham sites, with LCPs crossing a watercourse ................................ 66
vii
List of Tables
Table 1 Sample cost surface calculations ..................................................................................... 22
Table 2 Sample cost surface calculations with weighted variables .............................................. 22
Table 3 Clusters and segments ...................................................................................................... 28
Table 4 Inter-cluster segments ...................................................................................................... 34
Table 5 Tripcevich’s (2009) vertical factor table (abbreviated) ................................................... 41
Table 6 Modified segments for sensitivity assessement ............................................................... 43

viii
Acknowledgements
I offer my gratitude to my advisors, Drs. John Wilson and Robert Vos, for their guidance and
insights, and to all my professors in the USC GIST program, including Drs. Su Jin Lee, Karen
Kemp, and Jennifer Swift. I also want to extend a very special thank you to Dr. Wendy Teeter
and Erin Eichenberg for all the lessons in field and museum archaeology.

ix
List of Abbreviations
ASTER Advanced Spaceborne Thermal Emission and Reflection Radiometer
ACS Accumulated Cost Surface
ADS Archaeological Data Service
DEM Digital Elevation Model
GIS Geographic Information System
LCP Least Cost Path
LP DAAP Land Processes Distributed Active Archive Center
LiDAR Light Detection and Ranging
SRTM Shuttle Radar Topography Mission
US United States (of America)
UK United Kingdom

x
Abstract
Numerous studies within the last 200 years have shed light on the socioeconomic patterns of the
Beaker culture during the Bronze Age, particularly in the United Kingdom. However, with the
expanding role of GIS in the field of archaeology, there is an increasing amount of spatial data
on this cultural group, allowing opportunities for analysis that can begin to describe inter- and
intrasite spatial connections. The geographic connections of pathways, for example, can illustrate
the corridors of cultural exchange that gave rise to and sustained the Beakers for over 1,000
years. Using Least Cost Path analysis, this thesis aimed to model such spatial connections in
Northeast England.
The study generated 66 anisotropic LCPs that modeled possible path connections
between sites. The first 18 LCPs served as the primary LCPs between sites – within clusters and
between clusters. Three assessment tests were conducted to validate these LCPs. First, for each
primary LCP, another LCP was generated traveling in the reverse direction. Second, new
segments that utilized pairs of nearby sites, approximating the alignment of original pairs, were
generated; the new segments also included a primary and a reverse LCP. Finally, areas with high
LCP coincidence were compared to aerial images for coincidence with paths or features. The
study found that the LCPs were mostly coincident or near coincident. However, varying degrees
of local variation in the trajectories of many LCPs were evident. Four areas with high LCP
coincidence or near coincidence were selected for aerial imagery comparison which showed
LCPs generally following watercourses. Generating LCPs can model human movement during
the Bronze Age; however, datasets that describe the environmental conditions of the period as
well cultural datasets that spatially delineate territories and taboos are needed in order to more
xi
accurately understand the efficacy of these LCPs and the costs associated with prehistoric travel
in the region.
1
Chapter 1 Introduction
The past continues to be present in our open landscapes. Prehistoric peoples created pathways
between sites and between regions which can sometimes describe the interactions of local and
regional groups, including social relationships and economic behavior (Kline 2009; Teeter,
Martinez, and Richardson 2013). However, modern developments have masked many prehistoric
trails. The northern portions of the United Kingdom (UK), for example, are rich in Bronze Age
sites but are now dominated by agricultural land that has erased the physical connections
between settlements; these connections would otherwise describe how certain objects in the era,
such as jet, were transported from one site to another. This thesis aims to model the connections
of Bronze Age settlements in northeast England according to paths of least resistance using
natural topography.
Some paths of travel in modern times are possible iterations of past trails. Paths form due
to repeated use over time, and many paths endure by conforming to new travel technologies and
demands of trade (Kline 2009; Colton 1941). In 2015, a gold route between England and Ireland
was found to have served the area since 2500 BC (Mendoza 2015). In the United States,
prehistoric trade paths connecting coastal California areas to the American Southwest became
railroad routes and highways (Colton 1941). Colton (1941, 318) has stated: “Although
individuals on foot may traverse Arizona and New Mexico from one point to another from in
almost any direction, yet the bulk of the movement would follow certain lines of geographical
least resistance”. The key to understanding the consistent choices that people make in terms of
the routes to take lies in the shape of the landscape and their associated costs to the human
traveler.

2
To generate the model, Least Cost Path (LCP) analyses was employed. LCPs are a set of
calculations that use the slope data of an area, often with other variables such as hydrology and
land cover, to determine the possible routes with the least cost to the traveler, from a starting
point to a destination (Herzog 2014). In short, LCPs follow Colton’s (1941) logic of optimal
lines of travel.
1.1 Background and Motivation
Bronze Age England is well documented but sizeable opportunities for research remain.
Concepts of socioeconomic patterns of the era tend to be derived from southern cultures, while
some evidence points to variation throughout the region, particularly in terms of settlements
(Brück 2008). The use of multiple roundhouses in the south, for example, differ from the single
household structures in the remainder of the region. In landscape studies, spatial organization of
settlement and burial features as well as the relationship of material sources to sites lack analysis
that would show and compare regional variations (Brück 2008).
British prehistoric sites are generally spatially consistent over time. Much physical
evidence of the Bronze settlements and burials are largely the final iteration of a construction
process (Frieman 2012). Settlements were occupied, abandoned, and reused or repurposed over
several hundred years, if not thousands, as in the case of agricultural field systems in the Thames
Valley from Early to Late Bronze Age (Yates 1999). More famously, the Stonehenge monument
was built in a process lasting over 1,000 years (which encompassed the Neolithic and Early
Bronze Age), with stones transported from Wales, approximately 140 miles away (Mendoza
2015; Parker Pearson, 2013). In 2016, Mike Parker Pearson has posited that Stonehenge was
originally built in Wales and brought along through an eastward movement as descendants of the

3
monument’s dead relocated it to its current site to bring an end to inter-regional conflict
(Knapton 2016). Hence, despite changes in settlement patterns, a line of enduring paths can be
drawn between two or several sites and features that can begin to spatially describe the diffusion
or localization of socioeconomic behavior, the proximity of sites to source materials, and the
organization and cosmography of local land use – core habitation areas, farming and craft
production, and ceremonial/spiritual.
1.1.1. Least Cost Path analysis
LCPs model the spatial connections between points across a landscape (Mitchell 2012).
Examples of the multiple applications of LCP in different, disparate fields include transportation
planning – modeling roads and railways; utility planning – modeling the flow of gas and
electricity distribution networks; and ecology – modeling the behavior of animals (Mitchell
2012).
LCPs involves three algorithms to search for the least-cost path. The first is the Cost
Surface Raster. The raster provides the “cost of traveling through each cell” (Esri 2016c). It
typically involves variables such as the distance and slope values across a study area. It can
include any variable that facilitates or obstructs movement (White and Surface-Evans 2012). In
transportation, cost variables may include monetary costs of construction along a route over
another, for example (Mitchell 2012). In ecological studies, land use layers may be essential to
analyzing the cost of migration for some animals (Lee, Chon, and Ahn 2014). Whatever the case,
in GIS software such as ArcGIS, all layers will need to be reclassified to equal intervals to
generate comparable scales. Once the algorithm is run, the cost is calculated and the cost surface
is created in which each cell in the raster is assigned a cost value.

4
The second output is the Cost Distance Raster which “calculates the least accumulative
cost distance for each cell” to an origin point (Esri 2016e). To process this raster requires the cost
surface generated above and a given origin point run through the Cost Distance tool. This
generates the Cost Distance Raster and the Backlink Raster. Once the cost distance raster is
generated, an algorithm backtracks from the destination point to the origin using the lowest
values in the cost distance raster. These rasters are required to run the Cost Path tool which
provides the third and final output - a path of least cost between the two points (Esri 2016e).
1.1.2. Least Cost Path in archaeology
In archaeology, LCPs can demonstrate the movement of people across space, and often
have been used to model prehistoric pathways (Herzog 2014). While LCP calculations have been
available since 1957, cost surface analysis have been applied to archaeology since 1972 (Herzog
2014; Rahn 2005; Lee and Stucky 1998). The methods used in the field to collect spatial data
have varied and continue to vary, but the cost surface is often generated from the values of slope
and its relationship with other variable(s) that informed or shaped the decisions prehistoric
peoples made in choosing paths to get from one place to another (Rahn 2005). Indeed, Herzog
(2014) compiled archaeological LCP studies in rural areas and found that of the environmental
variables, slope tends to be the primary layer used in most research. This is not surprising
particularly since during prehistoric periods the primary mode of transport was on foot; hence, an
increase in slope would have increased the cost in time and energy for the traveler (Kondo and
Seino 2009).
LCPs have increased in complexity over time. LCPs can be executed anisotropically or
isotropically in which the calculations either consider the directionality of travel, as in the

5
former, or do not, as in the latter (White and Surface-Evans 2012). The cost of travel in isotropic
calculations is the same whether or not the traveler changes direction. Anisotropic methods
assume that directionality has implications for the cost of travel. The increase or decrease in the
slope of a route or part of a route may affect the cost of movement into and out of a point in
space (White and Surface-Evans 2012). Walking uphill has greater cost than walking downhill.
LCPs can also be implemented with partial anisotropic calculations. In a study of movement
during the Mayan Classic period, for instance, the method factored a greater cost to movement
coming from the west due to the direction of trade (Doyle, Garrison and Houston 2012).
However, despite the use of slope in many archaeological LCP studies, most LCPs are
implemented isotropically (Herzog 2014).
There are other methods to employing LCPs. Rahn (2005) observes that GIS technology
has improved in sophistication and archaeological LCAs can expand on or do away with slope
altogether. Analyses can include data on social attitudes towards specific points on a landscape
as in Llobera (2000), or improved algorithms for calculating the “visibility” relationships of
cells, which is often referred to as viewshed analysis (Rahn 2005; Lee and Stucky 1998). In the
latter case, paths are chosen according to how visible a cell is from other cells.
1.2 Study Area
Northeast England is located just south of England’s border with Scotland, and includes
Northumberland, the largest county, and Newcastle upon Tyne, its largest city (Figure 1). The
study area is roughly 280-380 miles from London. It is comprised primarily of rural, agricultural
lands interlaced with modern roads and highways, with the River Tyne and its tributaries cutting
through the southern end on a west to east orientation and emptying into the North Sea.

6

Figure 1 Study Area – Northeast England. Sources: Esri, USGS, NOAA.

7
While farm and grazing lands dominate the area, they are superimposed on a relatively
diverse topography that can impact the routes of a traveler. In the north, the Cheviot Hills flank
(and transcend) the Scottish border. Low rolling hills create some relief to the flat agricultural
topographic textures across the remainder of the region. Some sites skirt the coastal edges to the
east. The diversity in elevation and slope in the contours of the surface shape the corridors by
which pathways would have been used to connect prehistoric people from one geological
formation to the next.
1.2.1. The Beaker Culture
The Bronze Age brought new technology and cultural practices across the British Isles
(where the Bronze period lasted the longest). During the Late Neolithic and through the Early,
Middle, and Late Bronze Ages, the region was populated by people primarily associated with the
Beaker culture, so called due to the assemblage of artifacts found during this period that included
beaked drinking vessels (Crawford 1912). The earliest of these artifacts date back to
approximately 2,750 BC (Dyer 2002). In the literature, much has been discussed about whether
the Beakers were native or immigrants from continental Europe, but Parker Pearson (2013)
suggests that they were likely founded by a small number of immigrants rather than successive
groups of immigrants moving into the British Isles throughout the period as previously thought.
Dyer (2002) states that during Neolithic and into the Chalcolithic period (Late Neolithic through
Early Bronze), people in the area engaged in trading with peoples of continental Europe that
resulted in the spread of cultural ideas and exchange of material goods. In Parker Pearson’s study
of strontium and oxygen isotopes in tooth enamel, less than seven (out of 360 subjects) were
likely to have grown up outside of Britain. Thus, what defines the Beaker phenomenon is the

8
diffusion of Beaker technology and other cultural hallmarks rather than the spread of a specific
group of people.
1.2.1.1. Cultural Diffusion
The geospatial distribution of Beaker settlement sites is discontinuous across Europe as
well as in England. Research into this cultural group has included different ideas on how Beaker
culture had spread, but there is generally a lack of evidence of directionality of its diffusion
across Europe; some typological variants of certain artifacts like pottery, for example, are
sometimes contained to one small area (Linden 2007; Czebreszuk 2004; Waldren 2003; Brodie
1997). In other words, there are no known origins of Beaker technology or ideas. Since the shift
from thinking of the Beaker culture as a people to one of a cultural package, the origin of the
Beaker phenomenon has become less important; studies are now focused on the spread of
cultural material (Czebreszuk 2004).
Patterns of material exchange are well established. Jet necklaces, for example, have been
exchanged between prehistoric communities in the British Isles since the Neolithic, with Whitby
(approximately 75 miles south of Newcastle upon Tyne) as one of its major sources (Frieman
2012; Sheridan and Davis 2002). In Northern England, a form of jet was used with a source in
Scotland (Sheridan and Davis 2002). An Early Bronze Age burial in southern England near
Stonehenge, known for having the greatest amount of artifacts from the period, including copper
material with Spanish and French origins (Stevens 2008; Fitzpatrick 2002).
Beaker cultural diffusion may have been driven by trade, marriage, and war. Linden
(2007) proposes that more research is needed to examine the modes of cultural transmission such
as marital exchange, for example, in which a partner leaves their native village for their spouse’s,

9
as such arrangements can have implications for cultural diffusion across time and space. Brück
(2006) explains that the Bronze Age was predicated on an economy of gift and commodity
exchange and argues that marriage operated in a way that promoted the mechanism of the flow
of goods between groups as well as the flow of people between groups; to this extent, people and
goods are interchangeable.
Increasingly, since the 2000s, studies employ chemical analysis on human remains to
determine patterns of migration. Isotope analysis of tooth and bone in a Bavarian study have
demonstrated significant patterns of migration during the Bronze Age that have implications on
movement by marriage (Price, Grupe, and Schröter 1998). In southern England, isotope analysis
of adult subjects showed that they migrated 150-200 km, from their childhood to their final
residence (Evans, Chenery and Fitzpatrick 2006).
1.2.1.2. Burial Practices
Beaker burial practices reflect cultural change from the Neolithic to the Bronze Age
(Figure 2). The round barrows, or mounds, of single inhumations are a departure from the
multiple inhumations found in the burial mounds and monuments of the Neolithic (Bruck 2004b;
Dyer 2002). However, the number of human remains in a round barrow can sometimes vary
(Jones and Quintell 2014; Bruck 2008). In Fowler’s (2012) compilation of approximately 350
Chalcolithic and Bronze Age burial studies in Northeast England, 22 had at two to three sets of
human remains; one contained 24. It is important to note that the data also includes instances of
burials in which the number of those interred could not be determined and more in which the
number is unknown or unrecorded. It is not a surprise that these burial studies have inadequate

10
data; many British archaeological studies over the last two centuries concerning the Bronze Age
were poorly recorded excavations (Linden 2007; Brück 2008).

Figure 2 Beaker artifact samples – pottery and flint tools and arrowheads.
Source: Worcestershire Historic Environment and Archaeology Service.

Burials also feature material deposits that define the period. The first excavations of
Beakers barrows around or near Stonehenge by Richard Colt Hoare and William Cunnington in

11
1808 not only included beaker vessels but also the first bronze and copper axes and daggers
found in England (Parker Pearson 2013; Dyer 2002; Crawford 1912). In excavations since,
artifacts have also included those made of gold and jet. Fowler’s (2012) compiled dataset
includes jet necklaces and buttons, which may have served as heirlooms or relics; bronze knives
and dagger blades as well as barbed and tanged arrowheads – tools and weapons often associated
with warriors; gold earrings, and an iron spearhead; various types of beakers and other vessels
that usually dominate grave artifact assemblages; and even items made of gold and iron. While
the social positions of the dead have been defined by their associated artifact deposits in
archaeological research, there are studies that argue that artifacts may have also been offerings or
contained offerings by mourners that further expand identities and social ties (Woodward et al.
2005; Brück 2004b; Barrett 1990). This illuminates the ceremonial aspects of Beaker burials that
have particular geographic importance.
To mourn the dead includes several activities that involve community members and
various features of the local geography. Mourning revolves around preparations for burial rites
and a procession to the grave site. Later in the Bronze Age, for example, when cremation was
more widely practiced than inhumations, burials involved a procession to the pyre, deposition of
artifacts, preparation and interring of ashes, occurring along the origin and destination of the
ritual (Barrett 1990). Sometimes, artifacts might have been deposited in geologic features outside
of graves such as rivers; in some cases, human bones, or entire graves were deposited or built at
settlement boundaries (Brück 2006; Barrett 1990).

12
1.3 Research Objectives
While numerous studies of the British Bronze Age have been conducted over the last
200 years, a gap in knowledge continues, particularly in spatial analysis. Through LCP
modeling, connections and movement between sites and features of Bronze Age Northeast
England can be spatially described. The objectives of this thesis were to:
• Demonstrate the feasibility of using LCP analysis to model pathways between sites
and features in Bronze Age Northeast England.
• Identify possible pathways from settlements to burials within clusters of sites.
• Identify possible pathways between clusters of sites.
1.4 Thesis Organization
This thesis includes four remaining chapters. Chapter 2 explores the significance of GIS
and using LCP in the field of archaeology as well the processes and datasets typically used in
such endeavors. Chapter 3 describes the datasets acquired and the methods employed to generate
LCP paths in this thesis study. Chapter 4 explores the results of executing all of the LCPs
involved in the study as well as the validity assessments that were conducted. Finally, Chapter 5
discusses the significance of the results and their implications for future LCP studies in the
region.

13
Chapter 2 Background and Related Work
This chapter begins with an overview of the current role of GIS in archaeology and the
applications of LCP analysis. A discussion on the types of data necessary to run archaeological
LCPs, including variables and the cost surface raster follows. The chapter concludes with a look
into the different validity assessments implemented in previous studies.
2.1 GIS and Archaeology
GIS has become increasingly central to archaeological work. There is no one exacting
definition of GIS, but it often involves computerized hardware and software in which digital
spatial data can be collected using GPS units, stored and managed in a geodatabase, and
processed, analyzed, and visualized using software with automated statistical formulas (Arias
2013; Kline 2009; Kvamme 1999). These different components can be connected with
compatible software formats to facilitate the flow of data from one digital architecture to another.
For example, digital data can be collected in the field using Trimble or Garmin units (common
brand names of GPS devices in North America) and uploaded as shapefiles into ArcGIS, a major
spatial data software developed by Esri. In ArcGIS, the spatial data can be housed in a
geodatabase and then processed and/or analyzed using ArcMap, a mapping interface within
which a few hundred geoprocessing and spatial analysis tools can be employed (Esri, 2016b).
The implications for archaeology are vast. For instance, processing and storing data from the
field to a geodatabase can be done in a matter of hours whereas manually processing the data can
take several days, or even months. Through GIS, the capacity and ability of archaeological
projects in managing and analyzing data over the last three decades has been transformed from

14
slow and manual to quick and simultaneous processes with greater flexibility for scale in map
visualizations (McCoy and Ladefoged 2009; Kvamme 1999).
The increasing use of GIS has led to prospecting, predictive modeling, and behavioral
modeling since the 1980s (McCoy and Ladefoged 2009). Prospecting aims to locate
archaeological sites, particularly buried features, typically by geophysical survey and remote-
sensing methods; predictive modeling utilizes the association of environmental elements to
known archaeological sites to estimate the probability for certain human activities to occur; and
behavioral modeling, which typically uses spatial statistics to analyze artifacts, features, and
associated environmental variables and to extrapolate human social behavior. It is also in
behavioral modeling under which archaeological LCP analysis falls (McCoy and Ladefoged
2009).
2.1 LCP Overview
On its own, LCP analysis simply calculates movement and generates a path with the
lowest cost from one point in space to another. It is in this simplicity and in choosing the
appropriate variables to solve a particular problem that it owes its applicability to the number of
fields that employ it. In the prehistoric archaeological context, it can model the way people
interacted with their environment, or more specifically, how people consistently chose the paths
to traverse an area.
2.1.1. LCP Studies in archaeology
As mentioned, LCP analysis can generate possible paths between two points across an
area. It considers the cost of time and energy for human travel, which makes it valuable to
archaeological studies for a number of reasons, including the potential for: (1) identifying

15
prehistoric trails; (2) identifying additional sites along a path; and (3) understanding the cultural
significance and meaning of the spaces associated with the path, and of the path itself. There
have been numerous archaeological studies that have employed LCPs to these ends.
One of the primary goals of archaeological LCPs is to identify prehistoric trails. Melmed
and Apple (2009) modeled seven possible continuing paths from existing prehistoric trail
segments. The study suggests with greater environmental constraints, the accuracy of the LCP
analysis increases, which is the case with Schild’s (2016) thesis which executed LCPs to
determine the least cost Bronze Age trading route in an area of Southern Turkey while
demonstrating the appropriateness of using LCPs to find possible prehistoric routes. Teeter,
Martinez, and Kennedy Richardson (2013) conducted a study of historic and prehistoric trails in
Santa Catalina Island, California that employed ethnographic data from which possible potential
trails were identified and the area on and around the trails surveyed. The survey results were then
later tested by LCP analysis. Surveys have found additional sites and artifacts. Kline (2009)
conducted LCP analysis to map out the network of footpaths in Lost Valley, California, using
several sets of start and end points. The study found various foothpaths crossing and fanning out
in the valleys with many following drainages.
Other studies are discussed or referenced in the following sections and subsections as
they relate to a particular subject or topic. Some of these have been compiled in notable works
such as in White and Surface-Evans (2012) which features various approaches to LCP analysis;
Herzog (2014) which explores the shortcomings of executing LCPs, including the use of cost
layers for interpreting cultural variables in multiple studies.

16
2.2 Gathering Data
Prehistoric routes are not only cultural but topographic in nature. To model prehistoric
routes, spatial datasets that characterize the local landscape and cultural points of reference are
key. The model must also necessarily present the origin and destination of a traveler and the
consistent choices travelers have made over what routes to take over time. These choices,
theoretically involves minimizing time and energy (White and Surface-Evans 2012). The
topographic dataset(s) must be able to provide values from which the variables of cost can be
derived. Since these variables typically involve elevation and slope, LCPs of prehistoric routes
relies on two primary types of datasets: (1) Digital Elevation Models (DEMs) that describe the
shape of the terrain; and (2) cultural site points that represent archaeological sites. When
applicable (and available), datasets of other variables that influence travel within a study area
may also be included in or factored into the analysis.
2.2.1. DEMs
DEMs represent the surface of the Earth and are essential to LCP analysis. They can be
used as the primary base raster upon which all geospatial calculations are performed. In addition,
the elevation attributes of DEMs can be used to generate slope data, which in turn can be used as
a variable to generate a cost surface raster. To select the appropriate DEM for use in a project,
the advantages and disadvantages of data sources and structures of a DEM must be weighed.
Data sources for DEM surfaces include ground survey, existing topographic maps, and
remote-sensing. Ground surveys can be expensive and time consuming but can yield data with
high accuracy; and digitization of existing topographic maps can be less costly but yield
considerably less accuracy than other methods (Nelson, Reuter, and Gessler 2009). Remote

17
sensing methods such as photogrammetric land-surface models (aerial imagery) can produce
high resolution photographs of the surface and high levels of accuracy but sometimes require
greater data storage and processing time; and LiDAR, which can be costly and require high post-
processing time, but, when used in ideal conditions (fair weather), can yield data with sub-meter
accuracy (Wilson 2012; Nelson, Reuter, and Gessler 2009).
DEM datasets are structured in three ways: triangulated, contour, and grid (Hutchinson
2008; Wilson and Gallant 2000). Triangulated irregular networks (TINs) are derived from
surface specific point elevation data through ground survey and photogrammetric stereo models
(Hutchinson 2008). TINs represent the surface with triangulated elements joined at the vertices
and configured in varying sizes and density to approximate terrain variation; it enables the DEM
to efficiently handle breaks in the terrain and other irregularities (Wilson and Gallant 2000).
TINs can be applied to LCP analysis, especially in relatively small areas as using TINs is
computationally intensive (Antikainen 2013).
Contour data structures are usually derived from existing topographic maps, and
sometimes from aerial images (Hutchinson 2008). The structure employs polygons that reflect
the contours of the topography, and essentially renders the spaces between the contour lines as
smooth spaces (Hutchinson 2008; Wilson and Gallant 2000). These spaces and lines can produce
errors in the DEMs, resulting in impressions of a terraced surface and other related anomalies or
errors, also referred to as “artifacts” (not to be confused with the archaeological definition of the
same term) (Lock and Pouncett 2010).
The gridded DEM is a surface raster comprised of square cells and are typically derived
from aerial imagery (Hutchinson 2008). Each cell is attributed with an elevation value that
altogether creates the relief or shape of the Earth’s surface. Gridded DEMs are simple and easy

18
to use; however, unlike TINs, these structures sometimes cannot efficiently incorporate
discontinuities in the terrain (Wilson and Gallant 2000). Nonetheless, their structural simplicity
makes for relatively easier computation across relatively larger areas, and thus more suited to the
scale of this thesis’ analysis (Hutchinson 2008).
Various gridded DEMs at different resolutions can be acquired through numerous
sources. Since the 1990s, elevation data collection has grown in sophistication, resulting in finer
and more accurate datasets (Wilson 2012; Nelson, Reuter, and Gessler 2009). The U.S.
Geological Survey, for example, offers two commonly used global DEMs in archaeology (as
well as other fields) to analyze large areas: The Advanced Spaceborne Thermal Emission and
Reflection Radiometer (ASTER) Global Digital Elevation Model (GDEM) and the Shuttle Radar
Topography Mission (SRTM). Millions of images were processed to comprise the ASTER
GDEM’s 22,702 1° x 1° tiles, with the newest version (ASTER GDEM V2) available at a 30 m
resolution (U.S. Geological Survey 2014). ASTER GDEM V2, released in 2011, contains
significantly less discontinuities and artifacts than the first iteration (U.S. Geological Survey
2011b; Krieger, Curtis, and Haase 2010). Data collected from the space shuttle Endeavor
comprises the SRTM DEM, for which a 30 m resolution released in 2014 (U.S. Geological
Survey 2015b). SRTM differs from ASTER’s optical-based data collection: ASTER must collect
cloud-free images or mask clouds in post-processing in contrast to SRTM’s radar system which
can penetrate through cloud masses (U.S. Geological Survey 2015a). However, SRTM’s relative
shortcomings lie in discontinuities in higher elevation rugged terrains due to the angles that it
takes; these discontinuities present themselves as gaps or voids in the DEM which in SRTM’s
latest (and perhaps final) 30 m iteration, were primarily filled by ASTER data as ASTER’s
stereoscopic data collection features a nadir view (U.S. Geological Survey 2015a).

19
Since ASTER’s Version 2 (V2) release, a few published archaeological studies feature
comparisons between ASTER and SRTM. Aside from the obvious resolution differences
between the 30 m ASTER and the initial 90 m SRTM, some discrepancies occur in their vertical
and horizontal accuracies (Wilson 2012; Nelson, Reuter, and Gessler 2009). Rademaker, Reid,
and Bromley (2012) observe a significant vertical bias in the ASTER DEM over SRTM in their
study of Paleoindian sites within the mountainous area of Nevado Coropuna in Peru. In fact, in
the U.S. Geological Survey’s (2011a) ASTER validation study over the continental U.S.
(CONUS), ASTER GDEM V2 provided higher or lower elevations than SRTM depending on
land cover type. In Doyle, Garrison, and Houston’s (2012) Mayan study, three LCPs were
executed using the 30 m ASTER, 90 m SRTM, and a 5 m AIRSAR DEM. The study identified
only the areas within the differences between LCPs generated by the ASTER and AIRSAR as
the most probable prehistoric pathways. The ASTER and AIRSAR LCPs appear more consistent
with each other than with the coarser SRTM. As the 30 m resolution SRTM was published in
2014 (approximately two years prior to this writing), very few comparisons with ASTER GDEM
V2 have been published. Preferences may lie in the field in which analysis is conducted, the size
and location of a study area, and the context/framework of a particular study.
2.2.2. Bronze Age cultural data: Settlement and burial site points
The objective of gathering archaeological data for this thesis is to acquire points
representing the origins and destinations of human movement. There are various sources of
spatial data from the U.K. that can be utilized, as there have been numerous archaeological
studies of the Bronze Age since the early 1800s. Choosing cultural datasets for analysis to model
a relatively large area requires the most comprehensive collection of spatial data of known sites

20
available from a reliable source. Fortunately, there is an open source website that provides
archaeological data from the U.K., the Archaeological Data Service (ADS).
ADS includes studies of the U.K. and the region by British archaeologists. Beginning in
1996, the website acts as a clearinghouse for research data as well as an archival site for data that
are not associated or housed in any other web-based repository. A quick search for Bronze Age
data can yield several tens of thousands of studies, both with spatial data and without. This
extensive collection is largely due to the larger number of site-specific archaeological studies
that have been conducted over the years. This means that to analyze an appropriate set of sites
over a large region is to conduct an exhaustive search of all sites with spatial data, organized
according to research subject, and determine their applicability for analysis. A dataset of
settlement point data was compiled from these varied sources.
Additionally, one study available on ADS involves a sweeping collection of Bronze Age
sites in a relative large region. Fowler (2012) provides a compilation of data from mortuary
studies in Northeast England area that includes 137 burial sites (cairns and barrows) with a few
hundred inhumation and cremation deposits along with their associated artifacts.
2.2.3. Other cost variables
While slope has been the primary factor used in most archaeological LCP studies,
numerous studies have used other variables in determining travel costs. Variables such as
vegetation, hydrology, and land cover can play a role in the choices people make in terms of
paths for travel. A study on communication routes in the Mediterranean region, for example,
factored in avoidance of swamps and other areas prone to flooding (Fiz and Orengo 2008; Bell,

21
Wilson, and Wickham 2002). In their multi-criteria study of prehistoric movement in Michigan,
Howey (2007) utilized both vegetation land cover and waterways cost layers.
Using different variables and methods can make a difference in the outcome of an LCP
analysis. Zakšek et al. (2008) conducted a viewshed analysis in which a layer was created by
recalculating the slope with directionality and aspect to account for path visibility. The authors
ran their model and an isotropic LCP analysis (with only slope as cost) for comparison. The
study found significant differences – the isotropic path was the shortest and the lowest route,
while their model generated lines with better visibility along routes with some gentle slopes.
Ultimately, reconstruction of human path selection continues to present challenges.
2.3 Generating a Cost Surface
The cost surface is a raster with each cell containing the total cost value of all variables
involved in an LCP analysis (Esri 2016c). As mentioned in Chapter 1, to qualify all variables for
analysis, they must be reclassified to the same range of values and at the same intervals in order
to perform meaningful calculations. If a given range is 1-8, then the range of values in each of
the rasters must only contain values within that range wherein the higher the classification the
greater the cost value. In Bell, Wilson, and Wickham (2002), for example, marshes were given
the highest classification of 12, whereas grasslands, which are usually flat or offer very little
slope, were given the lowest rating.
The values of corresponding cells are then added to estimate the total cost for each cell in
the cost surface raster. For example, if the value of a cell in a slope raster is 5 and the
corresponding cell in another cost variable raster is 3, then that same cell in the cost surface
raster is 8 (Esri 2016c) (Table 1).

22
Table 1. Sample cost surface calculations.

Variable 1 Variable 2 Total Cost
Set 1 5 3 5 + 3 = 8
Set 2 2 7 2 + 7 = 9

When applicable, variables can also be weighted according to their estimated influence,
such as in Doyle, Garrison, and Houston (2012), as mentioned in Chapter 1. In ArcGIS, the
weights are expressed in percentages (Table 2):
Table 2. Sample cost surface calculations with weighted variables.

Variable 1 *Influence
(40%)
Variable 2 *Influence
(60%)
Total Cost
Set 1 5 5 x .4 = 2 3 3 x .6 = 1.8 2 + 1.8 = 3.8
Set 2 2 2 x .4 = .8 7 7 x .6 = 4.2 .8 + 4.2 = 5

2.3.1. Anisotropic Modeling: Tobler’s hiking function and Naismith’s rule
In most LCP analyses, the cost is framed in terms of time and/or energy. In isotropic
modeling, the accumulated cost surface (ACS) is produced by averaging the cost of movement
from one cell to the next as an LCP algorithm finds its way to each cell from a source/origin
point and an LCP is then generated by selection of a string of cells with the lowest values
between the source and the destination (Herzog 2014). However, the cost of moving into a cell
may differ depending on slope and the direction of travel, as in anisotropic modeling (Tobler
1993). For example, a hiker may pass a point along a mountain peak trail faster on their return
from the top than from the bottom. Tobler (1993) produced a function, referred to as Tobler’s
Hiking Function, that added 0.05 to slope in an exponential formula to describe the directional
discrepancy in walking velocity. Another commonly used rule is Naismith’s Rule which suggests
4 km/hour and 600 m/hour of ascent on slopes greater than 12°, subtracting 10 minutes from a
300 m downhill trek on 5°-12° slopes, and an additional 10 minutes on downhill slopes greater

23
than 12° (Yang et al. 2014; Magyari-Sáska and Dombay 2012). In Magyari-Sáska and Dombay
(2012) which compared the time rates for each formula, Tobler’s Hiking Function estimated
shorter times than the refined Naismith’s Rule.
Using these rules may depend on the software by which an LCP analysis is conducted.
Naismith’s Rule can be run using GRASS software while Tobler’s Hiking Function can be used
with ArcGIS.
2.4 LCP Algorithm
After the generation of an ACS, an algorithm searches for a string of lowest values that in
turn produces the LCP. The most enduring LCP algorithm is Dijkstra’s (1959), which conducts a
sweeping search for the least cost path in all directions, regardless of accumulating distance, until
it reaches the destination point (White and Surface-Evans 2012; Husdal 2000; Dijkstra 1959).
Due to such an exhaustive search, LCPs produced by the algorithm tend to be longer than the
true optimal path for it does not search for alternative shorter routes that would cost less in time
to travel (Herzog 2014). Another algorithm used in archaeological LCP analyses is the A* which
includes a similar algorithm to Dijkstra but the search parameters must be set by the user, and
thus limiting LCP returns to within those parameters (White and Surface-Evans 2012).
2.5 Validity Assessment
Some methods for validating archaeological LCPs exists, but none are necessarily
considered standard practice (Schild 2016). Ground truthing, visual comparison with aerial
images, introducing minor errors, and changes in variables have been implemented (Schild
2016).

24
Ground truthing is a practice by which archaeological LCPs can be tested by surveys and
existing archaeological data (White 2012). A survey would involve a search for artifacts and/or
features along the LCP. Melmed and Apple (2009) used ground truthing to validate LCPs with
varying results due to the relatively uniform topography of the study area. On the other hand,
Teeter, Martinez, and Kennedy-Richardson (2013), implemented the inverse process and used
LCPs to validate archaeological data.
Kline (2009), who looked at probable prehistoric trails by running LCPs, superimposed
the network of LCP paths over satellite imagery and visually confirmed a relationship between
LCP paths and sites. White (2012) performed ground truthing (although not from running LCPs
but from identifying anomalies in aerial imagery) and observed that the efficacy of this method
depended on the degree of landscape modification since prehistoric times. However, Kline
(2009) posits that many of the modern trails, roads, and rails evolved from prehistoric trails.
Validating an LCP model can also include changing the parameters of an analysis such as
the cost variables and the source and destination points. Schild (2016) used a sensitivity analysis
introducing random points, relocating the source and destination, and re-running the analysis.
Rademaker, Reid, and Bromley (2012) validated their study by changing the walking speed
parameters of foragers. In short, the scope of validity assessments is diverse and depends on the
needs of the study and its archaeological contexts.

25
Chapter 3 Methods
This chapter describes of the methods employed in generating LCPs between cultural
sites in Northeast England. The method involved: (1) acquiring and preparing DEMs; (2)
identifying clusters within the study area and selecting sites for analysis; (3) generating the cost
surface and the LCPs; and (4) validating the analysis. All processes described were performed in
ArcGIS 10.4, referred to, henceforth, simply as ArcGIS.
3.1 Data
3.1.1. Acquiring and preparing the cultural dataset
To model a matrix of prehistoric routes, employing a cultural dataset with site points is
ideal. The cultural datasets were acquired from ADS. The first dataset was downloaded in
Microsoft Excel .csv format (Fowler 2012). It is a large collection of data from different
archaeological mortuary studies in the study area from the last 200 years (Fowler 2012). The data
was comprised of individual points spanning 2,100 years from 3360 BC. to 1260 BC. The
objects in the dataset are instances of inhumations and cremations tabulated along with their
associated burial features (e.g. cists, cairns, pits), artifact deposits, location type, estimated age
range, elevation, and source of data. Sites excluded from this collection are those that are too
difficult to date. Since the dataset is a record of each inhumation and/or cremation, there was
redundancy in the number of burial sites – some sites include more than one set of human
remains. A new dataset was then derived from the original to only contain the unique locations
of these burial features (n=137).
This thesis primarily utilized the location attribute of each site. The British National Grid
was used for spatial reference in the original dataset, so the new burial sites dataset was first

26
transformed to the geographic coordinate system (GCS) OSGB 1936 (i.e. Ordnance Survey
Great Britain 1936) as it was exported into ArcGIS, and then projected to the British National
Grid datum. Note that projecting coordinates from OSBG 1936 to another datum can cause
significant errors to global projected coordinated systems (PCS) and other projections because
the datum is confined to Britain only (Ordnance Survey 2015). The data was incorporated into
the geodatabase and added to ArcMap.
To provide greater accuracy, the dataset was further reduced to reflect a more relevant
spatiotemporal range. In the raw data, Bronze Age artifacts appear with increasing frequency and
diversity beginning in 2200 BC. The material types, for example, include bronze, jet, and lead
ore that were not found in the burial sites prior to 2200 BC. Therefore, sites were selected in the
burial site layer that fall within the 2200-1260 BC range, with 1260 BC as the most recent of the
burials. After the sites were selected, they were imported into the geodatabase as a new feature
class. The final iteration of the dataset totals 46 burial sites.
The second dataset is comprised of settlement points that was manually compiled from
ADS (Appendix A). The ADS website includes a feature called “Archsearch” that serves as a
metadata catalog with over 1 million archaeological studies and resources. A search for Bronze
Age settlements in Archsearch yielded numerous site points. Each resource page supplies a small
amount of information on the particular site, including a brief description, the coordinates, and
historical period. Sites described or deemed as Bronze Age settlements in the Northeast England
region were selected for inclusion in the dataset. Each site point was entered into a .csv file in
Microsoft Excel with a simple set of attributes, consisting of Site name, National Grid reference
(NGR), and Easting and Northing coordinates (Appendix A). Sites without names in the catalog

27
were given one after their civil parish. A total of 59 settlement sites was collected and imported
into ArcGIS as a feature class with the same projection as the burial dataset.
3.1.2. Generating the origin and destination points
Generating the origin and destination points first required a process of selection. Two
steps were employed to generate the origin and destination points. First, clusters of sites were
identified from a visual evaluation of site points in ArcMap (Figure 3). Clusters are sets of sites
with each one within three kilometers of their nearest neighbor. Second, adjacent segments of
origin and destination points were then assigned for LCP analysis from these clusters.

Figure 3 Overview of selected clusters of settlement and burial sites. Sources (basemap): Esri,
USGS, NOAA.

28

Four clusters were selected, with each cluster broken down into segments (Table 3). In
the north central portion of the study is Cluster 1 (Figure 4). Cluster 1 includes 35 settlement
sites and 20 burial sites with a north-south orientation spanning approximately 30 km (Figure 4).
Five pairs of origin and destination points, or segments, were chosen for LCP analysis.
Hazeltonrig and Knock Hill comprise Segment 1; Knock Hill and Brands Hill comprise Segment
2; Brands Hill and Harehope Hill comprise Segment 3; Harehope Hill and West Plain Henge
comprise Segment 4; and West Plain Henge and Lookout Plantation comprise Segment 5.
Table 3. Clusters and Segments.

Origin Site Destination Site
Segment Cluster 1
1 Hazeltonrig Knock Hill
2 Knock Hill Brands Hill
3 Brands Hill Harehope Hill
4 Harehope Hill West Plain Henge
5 West Plain Henge Lookout Plantation
Cluster 2
1 Cheviot Walk Wood Blawearie cairn
2 Blawearie cairn Hepburn Crag Plantation
3 Hepburn Crag Plantation Rosemoor Cairn 1
Cluster 3
1 Alwinton Kirkhill Cremation Cemetery
2 Kirkhill Cremation Cemetery Great Tosson Quarry
3 Great Tosson Quarry Debdon Whitefield
Cluster 4
1 NW High Carry House Warkshaugh Farm
2 Warkshaugh Farm Reaverhill Farm

29

Figure 4 Cluster 1 with origin and destination points (LCP nodes). Sources (basemap): Esri,
USGS, NOAA.

30
Cluster 2 included one settlement and twelve burial sites with an approximately SE-NW
orientation, from the Cheviot Walk Wood site to the Rosebrough Moor site (Figure 5). Three
origin-destination pairs were selected for analysis.

Figure 5 Cluster 2 with origin and destination points (LCP Nodes). Sources (basemap): Esri,
USGS, NOAA.

31
Cluster 3 included two settlement sites and 21 burial sites, with a semi-circular alignment
and a slight E-W orientation, from Alwinton to Debdon Whitefield (Figure 6). Three origin-
destination pairs were selected for analysis.

Figure 6 Cluster 3 with origin and destination points (LCP Nodes). Sources (basemap): Esri,
USGS, NOAA.
Cluster 4 included two settlement and six burial sites with a NE-SW orientation from the
High Carry House site to Reaverhill Farm (Figure 7). Three origin-destination pairs were

32
selected for analysis. All of the origin-destination pairs in the four clusters were converted into a
feature class in ArcGIS.

Figure 7 Cluster 4 with labeled origin and destination points (LCP Nodes). Sources (basemap):
Esri, USGS, NOAA.

33
Finally, to model connections between clusters, origin and destination points for paths
between clusters were also selected (Figure 8). The five pairs listed in Table 4 were chosen for
analysis.

Figure 8 Inter-cluster origin and destination points (LCP Nodes). Sources (basemap): Esri,
USGS, NOAA.

34

Table 4. Inter-cluster segments.

Origin Site Destination Site
Segment Inter-cluster 1
1 High Carry House Alwinton
2 Alwinton Knock Hill
3 Knock Hill Hepburn Crag Plantation
4 Knock Hill Lookout Plantation
5 Hepburn Crag Plantation Lookout Plantation

3.1.3. Acquiring the DEM
The DEMs were acquired from the U.S. Geological Survey Land Processes Distributed
Active Archive Center website (LP DAAP; http://gdex.cr.usgs.gov/gdex/) that provides free
spatial data for registered users. After selecting the study area in an interactive map, the 30 m
SRTM and ASTER GDEM V2 datasets were chosen for download in a compressed folder
(Figures 9 and 10). The folders were then imported into ArcGIS by unzipping the file and saved
in a local drive that stores all relevant raw spatial data for the project. The DEMs were then
imported as a raster feature class into the project geodatabase and visualized in ArcMap. The 30
m SRTM with voids filled was chosen for analysis as “artifacts” were demonstrably less than in
the 30 m ASTER GDEM. Recall also the vertical bias in ASTER, as discussed in Section 2.2.1.
Once in ArcGIS, the DEMs were clipped and reprojected for the study area at hand. This step
meant that each DEM was projected from WGS84 to the British National Grid to match the
datum of the cultural dataset, OSGB 1936 (as was discussed in Section 3.1.1).

35

Figure 9 SRTM 30 m resolution, with elevation values in meters. Source: U.S. Geological
Survey Land Processes Distributed Active Archive Center (LP DAAP;
http://gdex.cr.usgs.gov/gdex/).

Figure 10 ASTER GDEM V2 30 m resolution, with elevation values in meters. Source: U.S.
Geological Survey Land Processes Distributed Active Archive Center (LP DAAP;
http://gdex.cr.usgs.gov/gdex/).

36
3.1.4. Creating the slope raster
Prior to creating the slope raster, the DEM was resized (Figure 11). The original
SRTM30 was much larger than necessary to cover the study area. This would result in slower
processing times. Thus, a smaller version was derived from the larger DEM.

Figure 11 Clusters in smaller DEM (with hillshade effect; values in meters) derived from
the larger SRTM DEM. Sources (basemap): Esri, USGS, NOAA.

37
In ArcGIS, a temporary polygon feature class was created to reflect an area extent large
enough to include all four clusters. Using the “Clip” tool, the new area polygon was utilized as
the output extent. The Clip tool creates a new raster by extracting a smaller area from a source
raster using specific inputs of extent (X and Y maxima and minima) (Esri 2016a). In this case,
the tool drew the extent from the area polygon.

Figure 12 Slope raster (with values in degrees) with cultural sites. Sources (basemap):
Esri, USGS, NOAA.

The slope surface was one of the components required for the production of a cost layer,
which in turn, was key to generating LCPs. Once the new DEM was created, the slope raster was
generated using the “Slope” tool in the ArcGIS Spatial Analyst toolbox (Figure 12). The Slope

38
tool creates a continuous grid of cells, with each cell given a slope value calculated from changes
in elevation from one cell to the next (Esri 2011). In this case, the output measurement of the
raster is expressed in degrees.
3.2 Methodology
Due to the high number of LCPs (n=66) produced, the study employed two models for
the execution of the Path Distance and Cost Path tools using the ModelBuilder application in
ArcGIS. A model in ModelBuilder includes the entire package of tools, inputs, and outputs to
solve a particular spatial problem (Allen 2011). A model can be built as a workflow using a
visually friendly interface that does not require any knowledge of coding; rather, the architecture
of the model only needs the specific inputs, tools, and the desired output.
Prior to building the models, the environmental settings for the model properties were
customized. Customization minimizes errors and offers greater efficiency in processing data. In
the Processing Extent environment, the SRTM30 DEM (Figure 12) was chosen as the Snap
Raster. This meant that the model only processed data which fell within the extent of the DEM.
Any additional data that extended beyond this extent were out-of-bounds and not included in the
analysis. Next, the raster cell size in the Raster Analysis Environment was set to 30 to match the
cell size of the DEM. And, finally, the current workspace was set to the project geodatabase so
the workspace became the default output location for any new data layers.
Once the DEM, the derived slope, and the cultural dataset were prepared, the generation
of a cost surface, and finally, generation of the least cost path – was assembled in ModelBuilder
(Figures 16 and 17). The process is sequential, with each step a requirement for the next.

39

Figure 13 The Path Distance model, with the origin, cost surface, and backlink surface indicated
as a model parameter (P).

Figure 14 The Cost Path model, with the destination, cost surface raster, backlink surface raster,
and the converted LCP raster (LCP_Vector) indicated as a model parameter (P).

40
3.2.1. Executing the Path Distance tool
The “Path Distance” tool is executed after creating the slope surface. The tool required
origin points for the input feature source data (Tables 3 and 4) and the DEM as the input surface
raster which the tool used to measure surface distance (Figure 15; Esri 2016f). The SRTM 30 m
DEM was also used as the input for the vertical raster. This functions to generate the slope. The
tool produced the cost surface raster and the output backlink which were required to run the Cost
Path tool. Each time a segment is run, a new origin point for that segment is used as input.

Figure 15 Path Distance tool inputs.

41
Due to the diversity of the terrain and range of slope, an anisotropic analysis was used.
This was calculated using Tripcevich’s (2009) vertical factor table, which takes the reciprocal of
Tobler’s (1993) formula: TIME (HOURS) TO CROSS 1 METER or the reciprocal of Meters per
Hour =0.000166666*(EXP(3.5*(ABS(TAN(RADIANS(slope_deg))+0.05)))) (Table 5). The
table can be interpreted as a vertical factor parameter. The resulting surface, or cost surface, was
comprised of values reflecting the cost of moving from one cell to the next throughout the extent
of the DEM from the origin. When paired with a destination point in the Cost Path Tool as the
input cost raster, each resulting LCP was in the direction towards the destination from the origin.
Table 5. Tripcevich’s (2009) vertical factor table (abbreviated).

Slope (deg) Vertical Factor
-90 -1
-80 -1
-70 2.099409721
-50 0.009064613
-30 0.001055449
-10 0.00025934
-5 0.000190035
-4 0.000178706
-3 0.000168077
-2 0.000175699
-1 0.000186775
0 0.000198541
5 0.000269672
10 0.000368021
30 0.001497754
50 0.012863298
70 2.979204206
80 -1
90 -1

3.2.2. Generating LCPs
To generate the LCP, the “Cost Path” tool was employed. The tool required three data
layers as input. These were the destination points (see Tables 3 and 4) as well as the cost surface

42
and the output backlink produced from executing the path distance tool (Figure 16). Like the
origin points in the execution of the Path Distance, the input for the destination data changes
with each new segment tested in the model. Once the Cost Path tool is run, it creates a raster of
LCP as the output. To better visualize the LCPs, the tool “Raster to Polyline” was employed to
convert the LCP raster to a vector format.

Figure 16 Cost Path tool inputs.
3.2.3. Conducting the validity assessment
Three forms of validity analysis were performed. The first validity assessment generated
reverse LCPs for the original cluster and inter-cluster segments (see Tables 4 and 5). Since each
LCP is anisotropic, a path distance analysis would result in a cost surface raster that accounts for
the costs associated with traveling from particular origins. This means that the values in a cost
surface raster may be different when an LCP is executed from the opposite direction. The second

43
validity assessment included generating new segments, referred to as modified segments, for
each of the clusters using new sets of origin and destination points that approximated the
alignment of the original segments. For comparison, each new segment was selected using sites
close to the original nodes (Table 6). In addition, reverse LCPs were also generated for the new
segments. Henceforth, the original LCP and its reverse counterpart are referred to as the original
set, and the modified LCP and its reverse counterpart are referred to as the modified set. The
third assessment compared all LCPs to aerial images of paths for coincidence or proximity. A
multiple ring buffer consisting of 0.5, 1, and 1.5 km bands was generated to estimate proximity.

Table 6. Modified segments used for sensitivity assessment.

Origin Site Destination Site
Segment Cluster 1
1 Alnham 3 Reavely Hill
2 Turf Knowe Humbleton Hill
3 Humbleton Hill Cheviot Quarry
4 Cheviot Quarry Lookout Plantation
Cluster 2
1 Cheviot Walk Wood Millstone Hill
2 Hepburn Crag Plantation Rosebrough Moor Cairn 2
Cluster 3
1 Farnham Hedley Wood
2 Holystone Common Great Tosson Quarry
3 Spital Hill Debdon Whitefield
Cluster 4
1 NW High Carry House Reaverhill Farm
Inter-Cluster
1 Wark Manor House Farnham
2 Farnham Alnham
3 Alnham Millstone Hill
4 Millstone Hill Lookout Plantation
5 Reavely Hill Lookout Plantation

44
Chapter 4 Results
This chapter describes the results from performing the tasks specified in the preceding chapter.
The first two sections detail the LCPs executed within clusters and between clusters. Each
section also includes comparisons with reverse LCPs and LCPs generated from modified
segments as part of the validity assessments. It is important to note that LCPs in the original
configuration and reverse LCPs were generated from the same set of nodes and are thus treated
as a set in a segment, referred to here as original sets. LCPs from modified segments have no
more than one common node with the original sets, if at all, and are thus discussed as a separate
measure of validity. Moreover, modified LCPs come as a set in a segment, with another path
generated from the opposite direction, referred to here as modified sets.
For map visualizations, LCPs are superimposed on the SRTM30 DEM in hillshade effect,
with the exception of the aerial imagery in the last section that compares selected segments with
original and modified LCPs to check the coincidence with visible geologic features.
4.1 Executing LCPs
The study generated a total of 66 LCPs in 18 original and 15 modified segments for all 33
source/destination nodes, approximating a network of paths cutting through clusters and
connecting selected sites between clusters (Figure 17).
Original sets showed some divergence from each other but overall they indicated
consistency/coincidence or near coincidence in all four clusters. Clusters 1, 2 and 4 showed some
divergence. In Cluster 4, for example, Segment 1 diverged by 0.5 km or more. Cluster 3
presented the least amount of divergence, with < 0.3 km divergence throughout.

45

Figure 17 Overview of cluster and inter-cluster primary and modified LCPs.
LCPs from modified sets, however, sometimes diverged to a greater extent from the
original LCPs, and from each other within a segment. Modified Segments 2 and 3 in Cluster 3
for example, diverged by more than 1.5 km. On the other hand, the modified LCPs produced just
one exact path of coincidence in Modified Segment 1. By running the Intersect tool in ArcGIS,
sections of exact coincidence within segments were identified. (In maps throughout this chapter,
these are referred to as intersections, rather than coincidence, after the tool).

46
4.2 Cluster LCPs
4.2.1. Cluster 1
LCPs were generated for the five segments of Cluster 1 (Figure 18). Overall, the original
sets were consistent with each other, exactly coinciding along several sections, particularly the
northern portion of Segment 1 and the southern portion of Segment 4 (Figure 19). Segment 1
also produced the most divergence, > 0.6 km, in its southern portion.

Figure 18 Overview of Cluster 1 LCPs with original and modified segments.

47

Figure 19 Detail of Cluster 1 – Segment 4 (Harehope Hill – West Plain Henge).
Modified LCPs in Cluster 1 were also generally consistent within each segment but
sometimes diverged from each other, although to a lesser degree than the original sets. Modified
Segments 1 and 3 showed the greatest divergence but only by < 0.4 km. However, the primary
and the reverse LCPs in Modified Segment 4 exactly coincided with their counterparts in
Segment 5 in the northern portions of the segments (Figure 20). Additionally, the modified sets
varied in proximity and coincidence with the original sets. Modified Segment 1 followed a low
cost corridor approximately 3 km from the original set and never coincided. Modified Segment 2
steered away from the alignment of the original Segment 2, following a more direct route until it

48
nearly coincided with Segment 3 along a narrow corridor. Modified Segment 3 roughly
paralleled the original Segment 4 and was never coincident or near coincident.

Figure 20 Detail of Cluster 1 – Segment 5 (West Plain Henge – Lookout Plantation) and
Modified Segment 4 (Cheviot Quarry – Lookout Plantation).
4.2.2. Cluster 2
Three original and two modified LCP sets were generated in Cluster 2 (Figure 21).
Segments 1 and 2 were mostly consistent and showed the most intersections, while Segment 3,
presented the greatest divergence. However, the divergences along Segment 3 were < 0.5 km at
their widest. In addition, the original primary LCP of the segment exactly coincided with its
counterpart in Modified Segment 2 all along its eastern section. Divergence also characterized

49
Modified Segment 1, by approximately 1 km at its widest, particularly in the southern portion
where there is greater topological variation. While each modified set shared a common node with
the original sets, it was only at these points where intersections occurred.

Figure 21 Overview of Cluster 2 - Segment 1 (Cheviot Walk Wood – Blawearie Cairn), Segment
2 (Blawearie Cairn – Hepburn Crag Plantation), and Segment 3 (Hepburn Crag Plantation –
Rosebrough Moor Cairn 1), with Modified Segment 1 (Cheviot Walk Wood – Millstone Hill)
and Modified Segment 2 (Hepburn Crag Plantation – Rosebrough Moor Cairn 2).
4.2.3. Cluster 3
Three original and three modified segments comprised Cluster 3 (Figure 22). All
segments were consistent, with LCPs coinciding in several sections throughout the cluster.
Segment 1 and Modified Segment 1 ran parallel (< 0.3 km apart) from each other along a low
cost corridor, coinciding or near coinciding near the northwestern nodes of Alwinton and

50
Farnham. Segment 2 and Modified Segment 2 also ran parallel but to a wider degree (nearly 1
km apart) until all LCPs connect at the Great Tosson Quarry site.

Figure 22 Overview of Cluster 3 LCPs – Segment 1 (Alwinton – Kirkhill Cemetery), with
Modified Segment 1 (Farnham – Hedley Wood); Segment 2 (Kirkhill Cemetery – Great Tosson
Quarry), with Modified Segment 2 (Holystone Common 1 – Great Tosson Quarry); and Segment
3 (Great Tosson Quarry – Debdon Whitefield), with Modified Segment 3 (Spital Hill Cairn 1 –
Debdon Whitefield).
Segment 3 and Modified Segment 3 showed most LCPs in exact coincidence (Figure 23).
LCPs of Segment 3 were divergent but ran closely parallel as they traverse a valley. They
became coincident or near coincident in the higher elevation, northeastern portion of the
segment. LCPs from Segment 3 and Modified Segment 3 also coincided in this portion, with the

51
primary and reverse LCPs exactly coinciding with their counterparts all the way to Debdon
Whitefield (Figure 23).

Figure 23 Detail of Cluster 3 – Segment 3 and Modified Segment 3, with sections of exact
coincidence of LCPs from both segments in their northeastern portions by Debdon Whitefield.
4.2.4. Cluster 4
Cluster 4 was comprised of two original segments and one modified segment (Figure 24).
All segments were consistent with each other, with Segment 2 exactly coincident throughout
most of its run. Segment 2 showed the most divergence but < 0.2 km at its widest.

52

Figure 24 Overview of Cluster 4 LCPs – Segments 1 (NW High Carry House – Warkshaugh
Farm) and 2 (Warkshaugh Farm – Reaverhill Farm) with Modified Segment 1 (NW High Carry
House –Reaverhill Farm).
A long, narrow topographic depression that runs along all nodes likely contributed to the
largely consistent stretch of Segments 1 and 2. This consistency began approximately at the
middle of Segment 1 and continued all the way to Reaverhill Farm. However, Modified Segment
2 took a more direct route between NW High Carry House and Reaverhill Farm.

53
4.3 Inter-cluster LCPs
LCPs were generated for the five original and five modified segments connecting the four
clusters (Figure 25). The original and modified segments were mostly consistent, with many
LCPs exactly coinciding. However, some divergence occurred.

Figure 25 Overview of Inter-cluster LCPs.

54
Consistency characterized the original segments of the inter-cluster LCPs. For example,
in its nearly approximately 30 km run between the nodes of NW High Carry House and
Alwinton, Segment 1 was mostly consistent, particularly in the middle of the segment, along a
long narrow corridor (Figure 26). The segment was also consistent with Modified Segment 1
(Wark Manor House – Farnham), with the primary and reverse LCPs from each segment exactly
coinciding with their counterparts in some sections. Marked divergence occurred in both
segments by their nodes.

Figure 26 Detail of Inter-cluster Segment 1 (NW High Carry House – Alwinton) with Modified
Segment 1 (Wark Manor House – Farnham).

55
The remainder of the original and modified segments were coincident or near coincident,
with some divergence. Segment 4 and Modified Segment 4 exhibited the widest divergence by as
much as 1.3 km. However, the primary and reverse LCPs from the two segments exactly
coincided with their counterparts in a valley between their nodes (Figure 27). In contrast,
Segment 5 and Modified Segment 5 ran along the same corridor and included primary and
reverse LCPs that exactly coincided with their counterparts; however, all LCPs were consistent
with the exception of the southeastern portion of the segments where they ran parallel and < 0.7
km apart.

Figure 27 Detail of Inter-cluster Segment 4 (Knock Hill – Lookout Plantation) with Modified
Segment 4 (Reavely Hill – Lookout Plantation, Segment 5 (Hepburn Crag Plantation – Lookout
Plantation) with Modified Segment 5 (Millstone Hill – Lookout Plantation).

56
4.4 Aerial Imagery Comparisons
Due to their high coincidence of LCPs, four areas were selected for comparison with
aerial imagery: (1) Cluster 3 – Segment 3 with Modified Segment 3; (2) Inter-cluster Segment 1
with Modified Segment 1; (3) Inter-cluster Segment 2 and Modified Segment 2; and (4) Inter-
cluster Segment 4 and 5 with Modified Segment 4 and 5, and Cluster 1 – Segment 5 with
Modified 4. Watercourses were found in three of the areas. One area that included a river system
flowing through from which the distances to the LCPs was measured using the multiple ring
buffer tool in ArcGIS.
Cluster 3 – Segment 3’s original set and the corresponding modified set are largely
consistent, particularly in the northeastern sections of the segments where the primary and
reverse LCPs exactly coincided with their counterparts (Figure 28). The LCPs intersect the river
just once as they cross the bottom of the valley, with the original primary LCP nearly coinciding
with a modern bridge. The segments cut through the northwestern edge of a small village and
exploited a low cost corridor on the northern side of a small hill between the river and Debdon
Whitefield. While the southeastern section of the primary segment diverged, the LCPs were no
more than 0.3 km apart. The divergence occurred at the bottom of the valley, between the Great
Tosson Quarry site and the edge of the village. This divergence was perhaps due to what appears
to be a relative uniformity of elevation in this particular area. However, it is possible that subtle
changes in the topography may have caused the LCPs to favor one corridor over the other. Note
that despite the divergence in the original segment in this area, the modified segment remained
consistent with only a marked divergence by the village edge where it diverged along with its
counterpart.

57

Figure 28 Cluster 3 – Aerial of Segment 3 (Great Tosson Quarry – Debdon Whitefield) with
Modified Segment 3 (Spital Hill Cairn 1 – Debdon Whitefield). Sources (basemap): Esri,
DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AEX,
Getmapping, Aerogrid, IGN, IGP, swisstopo, and the GIS User Community.

The LCPs of inter-cluster Segment 1 with its corresponding Modified Segment 1 were
mostly consistent and nearly coincident, particularly in the southern portion of each segment
(Figure 29). The segments cut through agricultural lands and closely followed a river. A few
modern structures were found along the river, but no recognizable village existed. The LCPs
intersected the river several times along their routes.

58

Figure 29 Aerial of the southern portions of Inter-cluster Segment 1 and Modified Segment 1.
Source (basemap): Esri, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS,
USDA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP, swisstopo, and the GIS User
Community.
The southern section of Inter-cluster Segment 2 was mostly consistent with Modified
Segment 2 (Figure 30). Specifically, the primary LCP of Segment 2 was exactly coincident with

59
its modified counterpart, as were the reverse LCPs. Some divergence occurred within each
segment, but for < 0.4 km. The LCPs traversed a corridor where a hilly formation met flat
agricultural lands.

Figure 30 Aerial of the southern section of Inter-cluster Segment 2 (Alwinton – Knock Hill),
with Modified Segment 2 (Farnham – Alnham). Source (basemap): Esri, DigitalGlobe, GeoEye,
Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AEX, Getmapping, Aerogrid, IGN,
IGP, swisstopo, and the GIS User Community.
Inter-cluster Segments 4 and 5 with Modified Segments 4 and 5 join Cluster 1 – Segment
5 with its corresponding Segment 4 along a stretch of a wide corridor cut through by a river

60
(Figure 31). The corridor is also agricultural but modern buildings and small villages dot the
landscape close to the river. Multiple LCPs exactly coincide within this corridor, including two
pairs of primary LCPs and two pairs of reverse LCPs. A multiple ring buffer was generated from
the primary LCP of Cluster 1 – Segment 5. The buffer showed that all LCPs were within 1.5 km
of the intersected feature (Figure 32). However, Inter-cluster Segment 5 and Modified Segment
5, which took a direct route between their nodes, extend outside of the buffer near the Hepburn
Crag Plantation and Millstone Hill sites.

Figure 31. Aerial of Inter-cluster Segments 4 and 5 with Modified Segments 4 and 5; Cluster 1 –
Segment 5 with Modified Segment 4. Sources (basemap): Esri, DigitalGlobe, GeoEye, Earthstar
Geographics, CNES/Airbus DS, USDA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP,
swisstopo, and the GIS User Community.

61

Figure 32. Inter-cluster – Aerial of Inter-cluster Segments 4 and 5, Modified Segments 4 and 5;
Cluster 1 – Segment 5 with Modified Segment 4, with multiple ring buffer (30% transparency).
Sources (basemap): Esri, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS,
USDA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP, swisstopo, and the GIS User
Community.

62
Chapter 5 Discussion and Conclusions
Degrees of local topographic variation shaped the generated LCPs, and the results of this thesis
have only begun to describe a network of paths during the Bronze Age in Northeast England.
The study generated relatively consistent original sets of LCPs that, on their own, already point
to possible corridors of prehistoric travel. However, the addition of the modified sets as another
kind of validity test provided another dimension to this analysis, particularly when existing
nearby sites were used as nodes. They demonstrated the contiguity or lack of contiguity of the
paths in connecting with other sites in a cluster as well as between clusters, and hence,
challenges to or confirmation of the validity of the primary LCPs. Through this level of
validation, four areas were identified with a high likelihood of having supported prehistoric
paths. This chapter explores the significance of the patterns that emerged from generating LCPs,
the significance of using only one variable (slope) in the study, and finally some concluding
comments with directions for future research.
The modified sets demonstrated that topographic variations at the local scale can impact
the route of an LCP. In Cluster 2, for example, Modified Segment 1 was highly divergent (1 km
apart at its widest) in its southern portion by the Cheviot Walk Wood node (see Figure 21).
Immediate north of the node is a small depression flanked by varying degrees of slope as the area
transitioned to higher elevation. It appears that the divergence of the modified set was due to this
variation, with the primary LCP favoring a route along a corridor to the west than the reverse
which followed a corridor to east. A narrow strip of small hills between these corridors preserved
the divergence until both LCPs met at the Millstone Hill node.

63
Variation can lead to divergence even in wide, open spaces that appear to have largely
uniform elevations and/or slopes. The northern portion of Inter-cluster Segment 4 and its
modified counterpart follow a wide open corridor with some relatively minor undulations (see
Figure 27). Recall that Segment 5 and Modified Segment 5 ran parallel with the Segment 4 and
Modified Segment 4 along the same corridor and were coincident or nearly coincident all along
the area. However, a cursory look at the hillshade DEM showed that there is a cluster of small
formations in between the divergent LCPs that appear to be of higher elevation, effectively
creating shallow corridors of low cost values on either side. The fact that the trajectory changes
as a result of the slightest variations in relief is consistent with the effects of local variations
discussed above; even in wide open spaces that offered few restrictions, significant divergence
can occur.
While modified sets in the study served to test the validity of the original primary LCPs,
they also generated low cost paths that can serve as possible prehistoric routes. Indeed, multiple
routes between sites are a possibility in this region. For example, the northeastern part of
Segment 1 in Cluster 1 was coincident or nearly coincident by Knock Hill, and hence, this is a
route with a high likelihood for serving as a location of prehistoric paths (Figure 33). However,
its counterpart, Modified Segment 1, exploited a deeper, more restrictive corridor to the east,
following a more or less parallel path to the other segments. Settlement sites dominate this
portion of the dataset, with the nearest ones such as Cat Crag in the west and Chesters Burn in
the east > 0.5 km away from the closest parts of the LCP segments. More travel would have
occurred along paths connecting settlements, and these sites could possibly be served by these
segments, which means that either one is viable and warrants further study or ground-truthing.

64

Figure 33 Cultural sites and LCPs in the southern portion of Cluster 1, with Inter-cluster
Segment 2.
Rivers and streams offer persistently low cost corridors throughout their courses and
emerged as areas with the highest prevalence of LCP coincidence in the study. Logically, LCPs,
when given the opportunity, exploit the low cost values in topography associated with
watercourses and the adjacent valley bottoms. This is evident in the examples of aerial imagery
comparisons discussed at the end of Chapter 4. Indeed, High Carry House, the southern node of
Inter-cluster Segment 1, is also a node in Cluster 4 (see Figure 29), in which all three segment
nodes lie adjacent to the stream. All original segments ran parallel with the general flow of the

65
water. This is consistent with Kline (2009) who found many generated LCPs following drainage
systems.
Since the only variable utilized in the study is slope, then only the values of slope guide
the LCPs. Indeed, many LCPs crossed streams and rivers due to the absence of the added costs
that may accompany the portrayal of rivers in a hydrological dataset. But the study chose to
forgo the use of hydrology and other variables for several reasons. In terms of hydrology,
modern datasets do not necessarily describe the conditions of the Bronze Age. Early Bronze Age
was also the onset of a wetter climate that, in turn, also transitioned into a warmer period at the
start of the Iron Age (Brown 2008). Hence, different hydrological datasets at finer temporal
lenses need to be combined with cultural datasets within the same temporal ranges. Additionally,
the watercourses within the study area are minor rivers and streams that may have been welcome
features for prehistoric travelers; they may have even been a necessity (Colton 1941).
Smaller rivers and streams did not hinder movement in the Bronze Age. The Beakers
built small bridges to allow for crossings (Bruck 2004a). The settlement site of Alwinton, for
instance, is located just across a river from the burial site of Farnham; a hydrological dataset
would have likely triggered LCP algorithms to find other low cost avenues and the LCPs would
not have been as consistent as the first segments of Cluster 4 (Figure 34). In terms of land cover,
some increase in deforestation relative to the Neolithic also characterized the period with the
continued dependence on agriculture, effectively creating patches all around what was then a
largely wooded terrain (Dyer 2002). If it was possible to derive such hydrological and land cover
datasets from existing climate studies during the Bronze Age (see Brown 2008), future LCP
studies in the region would greatly benefit from the inclusion of these data layers.

66

Figure 34 Alwinton and Farnham sites, with LCPs crossing a watercourse. Sources (basemap):
Esri, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AEX,
Getmapping, Aerogrid, IGN, IGP, swisstopo, and the GIS User Community.
This thesis generated 66 LCPs that demonstrated how sites may have been connected
through a network of paths. While numerous, the number of sites involved is small amount
relative to the number of sites remaining in the dataset. Nonetheless, the thesis showed that
possible prehistoric paths can be modeled using LCPs when three types of validation are
employed to test the primary LCPs; these assessments produced areas with varying levels of
likelihood for having hosted prehistoric trails. These methods are particularly applicable to
analysis of areas with multiple sites.
While slope was the only variable used as a proxy for cost, some error may have been
introduced as a result of the calculation method. ArcGIS calculates slope using a 3 x 3 cell
moving window and assigns the value of a center cell to cells without considering the elevations

67
of cells just outside of these 3 by 3 cell moving windows or neighborhoods (Esri 2016d). This
can result in local anomalies and using a larger window may yield more accurate slope values
and least cost paths across the land surface.
Moreover, the additional processing to produce the SRTM 30 m DEM may have also
introduced residual errors that affect slope calculations, and ultimately the computed paths
between nodes. Data from the ASTER and the lower resolution GMTED were used to fill some
of the voids in the previous SRTM version (USGS 2015a). Filling the voids can result in a
smoother terrain that can also affect LCP trajectories, given the LCP algorithm’s sensitivity to
small and subtle changes in the topography.
More studies should be implemented to deepen our understanding of prehistoric
movement in Northeast England. In addition to finding appropriate and finer temporal datasets,
other methods can be employed to further validate LCPs. In the archaeological context,
interfacing LCPs with the presence of cultural material would substantiate computed paths –
either with recovered artifacts and features with spatial data or through ground truthing with
surveys and excavations. Furthermore, human social behavior is not always determined by the
environment. Future studies must also employ ethnographic and similar data to incorporate
territorial boundaries and taboos that would inform the path choices of a traveler.

68
REFERENCES
Allen, David W. 2011. Getting to Know ArcGIS ModelBuilder. Redlands: Esri Press.
Antikainen, Harri. 2013. “Comparison of Different Strategies for Determining Raster-Based
Least –Cost Paths with a Minimum Amount of Distortion.” Transactions in GIS 17, 1:96-
108. Accessed August 24, 2016. http://dx.doi.org/10.1111/j.1467-9671.2012.01355.x.

Arias, Veronica Maria. 2013. “Application of GIS and Spatial Data Modeling to Archaeology: A
Case Study in the American Southwest.” PhD diss., The University of New Mexico.
Accessed February 7, 2016. http://hdl.handle.net/1928/23385.

Barrett, John C. 1990. “The Monumentality of Death: The Character of Early Bronze Age
Mortuary Mounds in Southern Britain.” World Archaeology 22, 2:179-189. Accessed
July 7, 2016. http://www.jstor.org.libproxy2.usc.edu/stable/124875.

Bell, Tyler, Andrew Wilson, and Andrew Wickham. 2002. “Tracking the Samnites: Landscape
and Communications Routes in the Sangro Valley, Italy. American Journal of
Archaeology 106, 2:169-86. Accessed July 11, 2016. http://dx.doi.org/10.2307/4126242.

Brodie, Neil. 1997. New Perspectives on the Bell-Beaker Culture. Oxford Journal of
Archaeology, 3:297-314. Accessed June 9, 2016. http://dx.doi.org/10.1111/1468-
0092.00042.

Brown, Tony. 2008. “The Bronze Age Climate and Environment of Britain.” Bronze Age Review
1: 23-33. Accessed August 11, 2016.
https://www.britishmuseum.org/pdf/BAR1_2008_2_Brown_d.pdf.

Brück, Joanna. 2004a. “Bronze Age Britain and Ireland. In Ancient Europe, 8000 B.C. to A.D.
1000: Encyclopedia of the Barbarian World, edited by Peter Bogucki and Pam J.
Crabtree. Vol 2: Bronze Age to Early Middle Ages (c. 3000 – A.D. 1000). New York:
Charles Scribner’s Sons. 54-61. Accessed July 7, 2016.
http://go.galegroup.com.libproxy1.usc.edu/ps/i.do?id=GALE%7CCX3400400124&v=2.1
&u=usocal_main&it=r&p=GVRL&sw=w&asid=4b6e8e31762d1abd193ed0cfe9520e06.

Brück, Joanna. 2004b. “The Relational Construction of Identity in Early Bronze Age Burials in
Ireland and Britain.” Journal of Social Archaeology 4, 3:307-333. Accessed July 7, 2016.
http://dx.doi.org/10.1177/1469605304046417.

Brück, Joanna. 2006. “Death, Exchange and Reproduction in the British Bronze Age.” European
Journal of Archaeology 9, 1:73-101. Accessed July 7, 2016.
http://dx.doi.org/10.1177/1461957107077707.

69
Brück, Joanna. 2008. “Prospects and Potential in the Archaeology of Bronze Age Britain.”
Bronze Age Review 1: 23-33. Accessed April 23, 2016.
https://www.britishmuseum.org/pdf/BAR1_2008_3_Bruck_d.pdf.

Colton, Harold Sellers. 1941. “Prehistoric Trade in the Southwest.” The Scientific Monthly 52,
no. 4: 308-319.

Crawford, O.G.S. 1912. “The Distribution of Early Bronze Age Settlements in Britain.” The
Geographical Journal, vol. 40, no. 2:184-197. Accessed June 11, 2016.
http://www.jstor.org.libproxy2.usc.edu/stable/pdf/1778466

Czebreszuk, Janusz. 2004. “Bell Beakers from West to East.” In Ancient Europe, 8000 B.C. to
A.D. 1000: Encyclopedia of the Barbarian World, Vol. 1, edited by Peter Bogucki and
Pam J. Crabtree, 476-485. New York: Charles Scribner’s Sons.

Doyle, James, Thomas G. Garrison, and Stephen D. Houston. 2012. “Watchful Realms:
Integrating GIS Analysis and Political History in the Southern Maya Lowlands.”
Antiquity 86, 792-807.

Dijkstra, E.W. 1959. “A Note on Two Problems in Connexion with Graphs.” Numerische
Mathematik 1, 1:269-271. Accessed July 16, 2016.
http://dx.doi.org/10.1007/BF01386390.

Dyer, James. 2002. Ancient Britain. London, UK: Routledge.

Esri. 2016a. Clip. Accessed April 23, 2016. http://pro.arcgis.com/en/pro-app/tool-reference/data-
management/clip.htm.

Esri. 2016b. Create and Share Maps, Analytics, and Data. Accessed April 3, 2016.
http://desktop.arcgis.com/en/.

Esri. 2016c. Creating a Cost Surface Raster. Accessed June 12, 2016.
http://pro.arcgis.com/en/pro-app/tool-reference/spatial-analyst/creating-a-cost-surface-
raster.htm

Esri. 2016d. How Slope Works. Accessed September 8, 2016.
http://desktop.arcgis.com/en/arcmap/10.3/tools/spatial-analyst-toolbox/how-slope-
works.htm

Esri. 2016e. Understanding Cost Distance Analysis. Accessed June 12, 2016.
http://desktop.arcgis.com/en/arcmap/10.3/tools/spatial-analyst-toolbox/understanding-
cost-distance-analysis.htm.

70
Esri. 2016f. Understanding Path Distance Analysis. Accessed September 8, 2016.
http://pro.arcgis.com/en/pro-app/tool-reference/spatial-analyst/understanding-path-
distance-analysis.htm

Esri. 2011. Slope (Spatial Analyst). Accessed April 23, 2016.
http://help.arcgis.com/En/Arcgisdesktop/10.0/Help/index.html#//009z000000v2000000.h
tm.

Evans, J.A., C.A. Chenery, and A.P. Fitzpatrick. 2006. “Bronze Age Childhood Migrations of
Individuals Near Stonehenge, Revealed by Strontium and Oxygen Isotope Enamel
Analysis.” Archaeometry 48, 2:309-321. Accessed July 7, 2016.
http://dx.doi.org/10.1111/j.1475-4754.2006.00258.x.

Fiz, Ignacio and Hector A. Orengo. 2008. “Simulating Communication Routes in Mediterranean
Alluvial Plains.” in Layers of Perception. Proceedings of the 35th International
Conference on Computer Applications and Quantitative Methods in Archaeology (CAA),
Berlin, Germany, April 2-6, 2007, Kolloquien zur Vor- und Frühgeschichte 10. 316–21,
edited by Axel Posluschny, Karsten Lambers and Irmela Herzog. Accessed June 10,
2016. http://proceedings.caaconference.org/paper/102_fiz_orengo_caa2007/

Fitzpatrick, A.P. 2002. “’The Amesbury Archer’: A Well-Furnished Early Bronze Age Burial in
Southern England.” Antiquity 76, 293:629-630. Accessed July 8, 2016.
http://search.proquest.com.libproxy1.usc.edu/docview/217561765?accountid=14749

Fowler, Chris. 2012. Chalcolithic and Early Bronze Age Burials in Northeast England.
Archaeological Data Service: York. Accessed June 6, 2012.
http://dx.doi.org/10.5284/1017128.

Frieman, Catherine. 2012. “Going to Pieces at the Funeral: Completeness and Complexity in
Early Bronze Age Jet “Necklace” Assemblages.” Journal of Social Archaeology, 12:
334-355. Accessed June 12, 2016. http://dx.doi.org/10.1177/1469605311431400.

Herzog, Irmela. 2014. “Least-cost Paths – Some Methodological Issues.” Internet Archaeology
36. Accessed April 3, 2016. http://dx.doi.org/10.11141/ia.36.5.

Howey, Meghan C.L. 2007. “Using Multi-Criteria Cost Surface Analysis to Explore Past
Regional Landscapes: A Case Study of Ritual Activity and Social Interaction in
Michigan, AD 1200-1600.” Journal of Archaeological Science 34, 11: 1830-1846.
Accessed September 9, 2016. http://www.dx.doi.org/10.1016/j.jas.2007.01.002

Hutchinson, Michael F. 2008. “Adding the Z Dimension.” In The Handbook of Geographic
Information Science, edited by John P. Wilson and A. Stewart Fotheringham. Oxford,
UK: Blackwell Publishing Ltd.

71
Husdal, Jan. 2000. “How to Make a Straight Line Square. Network Analysis in Raster GIS with
Time-dependent Cost Variables.” Thesis, University of Leicester. Accessed July 16,
2016. http://www.husdal.com/2000/09/01/how-to-make-a-straight-line-square/

Jones, Andy M. and Henrietta Quintell. 2014. “Saucer Barrows: Places for Ritual within Wessex
Early Bronze Age Chalkland Barrow Cemeteries.” Oxford Journal of Archaeology 33,
4:339-359. Accessed July 3, 2016. http://dx.doi.org/10.1111/ojoa.12041.

Kline, Victoria. 2009. “Regional Network Analysis Situating Lost Valley in the Inter-Site
Landscape.” Thesis, San Diego State University. Accessed February 21, 2016.
http://130.191.3.5/~anthro/docs/theses/V_Kline_2009.pdf.

Knapton, Sarah. 2016. “Original Stonehenge was Dismantled in Wales and Moved to Wiltshire,
Archaeologists believe.” The Telegraph. June 6, 2016. Accessed June 12, 2016.
http://www.telegraph.co.uk/science/2016/06/05/original-stonehenge-was-dismantled-in-
wales-and-moved-to-wiltshi/

Kondo, Yasuhisa and Yoichi Seino. 2009. “GPS-aided Walking Experiments and Data-driven
Travel Cost Modeling on the Historical Road of Nakasendo-Kisoji (Central Highland
Japan),” In Making History Interactive. Computer Applications and Quantitative
Methods in Archaeology (CAA). Proceedings of the 37
th
International Conference,
edited by Bernhard Frischer, Jane Webb Crawford and David Koller. Willamsburg,
Virginia. Accessed July 4, 2016.
http://proceedings.caaconference.org/paper/21_kondo_seino_caa2009/

Krieger, Tabatha, Bill Curtis, and Jeff Haase. 2010. ASTER V2 Evaluation.
https://lpdaacaster.cr.usgs.gov/GDEM/Appendix_C_NGA_GDEM2_validation_report.pd
f.

Kvamme, Kenneth L. 1999. “Recent Directions and Developments in Geographic Information
Systems.” Journal of Archaeological Research 7, no. 2: 153-201. Accessed April 2,
2016. http://dx.doi.org/10.1007/BF02446276.

Lee, Jay and Dan Stucky. 1998. “On Applying Viewshed Analysis for Determining Least-Cost
Paths on Digital Elevation Models.” International Journal of Geographic Information
Science 12, 8:891-905. Accessed July 4, 2016.
http://dx.doi.org/10.1080/136588198241554.

Lee, Jung A., Jinhyung Chon, and Changwoo Ahn. 2014. “Planning Landscape Corridors in
Ecological Infrastructure Using Least-Cost Path Methods Based on the Value of
Ecosystem Services.” Sustainability 6, 11:7564-7585. Accessed July 4, 2016.
http:dx.doi.org.librproxy2.usc.edu/10.3390/su6117564.

Linden, Marc Vander. 2007. “What linked the Bell Beakers in Third Millennium BC Europe?”
Antiquity 81, 312:343-352. Accessed June 16, 2016.

72
http://libproxy.usc.edu/login?url=http://search.proquest.com.libproxy1.usc.edu/docview/2
17570278?accountid=14749

Llobera, Marcos. 2000. “Understanding Movement: A Pilot Model Towards the Sociology of
Movement,” in Beyond the Map. Archaeology and Spatial Technologies, 65–84, edited
by Gary Lock. Amsterdam: IOS Press.

Lock, Gary and John Pouncett 2010. “Walking the Ridgeway Revisited: The Methodological and
Theoretical Implications of Scale Dependency for the Derivation of Slope and the
Calculation of Least-Cost Pathways.” In Making History Interactive. Computer
Applications and Quantitative Methods in Archaeology (CAA). Proceedings of the 37th
International Conference, Williamsburg, Virginia, United States of America, March 22-
26 (BAR International Series S2079), pp. 192-203, edited by B. Frischer, J. Webb
Crawford and D. Koller. Oxford: Archaeopress.

Magyari-Sáska, Zsolt and Ştefan Dombay. 2012. “Determining Minimum Hiking Time Using
DEM.” Geographia Napocensis 4, 2:124-129. Accessed July 16, 2016.
http://geographianapocensis.acad-
cluj.ro/Revista/volume/nr_2_2012/pdf/Magyari_Dombay.pdf

McCoy, Mark D. and Thegn N. Ladefoged. 2009. “New Developments in the Use of Spatial
Technology in Archaeology.” Journal of Archaeological Research 17, no. 3: 263-295.
http://www.dx.doi.org/10.1007/s10814-009-9030-1.

Melmed, Anamay and Rebecca Apple. 2009. “Trails Through the Landscape of the Colorado
Desert.” Proceedings of the Society for California Archaeology, 21: 226-230. Accessed
February 21, 2016.
https://www.scahome.org/publications/proceedings/Proceedings.21Melmed.pdf

Mendoza, Jessica. 2015. “Ancient Trade Route Shows Gold’s Enduring Power.” The Christian
Science Monitor. Accessed June 8, 2016.
http://search.proquest.com.libproxy2.usc.edu/docview/1686536911?pq-
origsite=summon&accountid=14749

Mitchell, Andy. 2012. The Esri Guide to GIS Analysis, Modeling Suitability, Movement, and
Interaction. Redlands: Esri Press.

Nelson, A., H.I. Reuter, and P. Gessler. 2009. “DEM Production Methods and Sources.”
Developments in Soil Science 33: 65-85. Accessed September 9, 2016.
http://www.dx.doi.org/10.1016/S0166-2481(08)0003-2.

Ordnance Survey. 2015. A Guide to Coordinate Systems in Great Britain: An Introduction to
Mapping Coordinate Systems and the Use of GPS Datasets with Ordnance Survey
Mapping. Accessed July 20, 2016.

73
https://www.ordnancesurvey.co.uk/docs/support/guide-coordinate-systems-great-
britain.pdf

Parker Pearson, Mike. 2013. Stonehenge – A New Understanding: Solving the Mysteries of the
Greatest Stone Age Monument. The Experiment: New York.

Price, Douglas, Gisela Grupe, and Peter Schröter. 1998. “Migration in the Bell Beaker Period of
Central Europe.” Antiquity 72, 276:405-411. Accessed July 7, 2016.
http://libproxy.usc.edu/login?url=http://search.proquest.com.libproxy2.usc.edu/docview/2
17557201?accountid=14749

Rademaker, Kurt, David A. Reid, and Gordon R.M. Bromley. 2012. “Connecting the Dots: Least
Cost Analysis, Paleogeography, and the Search for Paleoindian Sites in Southern
Highland Peru.” In An Introduction to the Least Cost Analysis of Social
Landscapes,edited by Sarah L Surface-Evans and Devin A. White, 32-45.. Salt Lake
City: University of Utah Press.

Rahn, Brian. 2005. “Cost Surface-Derived Least-Cost Paths: A Case Study from Iron Age
Orkney.” Internet Archaeology 19. Accessed July 4, 2016.
http://dx.doi.org/10.11141/ia.19.2.

Schild, Alex. 2016. “Archaeological Least Cost Path Modeling: A Behavioral Study of Middle
Bronze Age Merchant Travel Routes Across the Amanus Mountains, Turkey.” Thesis,
University of Southern California.

Sheridan, Alison and Mary Davis. 2002. “Investigating Jet and Jet-like Artefacts from
Prehistoric Scotland: The National Museums of Scotland Project.” Antiquity 76, 293:812-
825. Accessed July 17, 2016. http://dx.doi.org/10.1017/S0003598X00091298.

Stevens, Fay. 2008. “Elemental Interplay: The Production, Circulation and Deposition of Bronze
Age Metalwork in Britain and Ireland.” World Archaeology 40, 2:238-252. Accessed July
8, 2016. http://www.jstor.org/stable/40388209?seq=1&cid=pdf-
reference#references_tab_contents.

Teeter, Wendy G., Desiree R. Martinez, and Karimah O. Kennedy Richardson. 2013. “Cultural
Landscapes of Santa Catalina Island”. In California’s Channel Island: The Archaeology
of Human-Environment Interactions, edited by Christopher S. Jazwa and Jennifer E.
Perry, 156-171. Salt Lake City: The University of Utah Press.

Tobler, Waldo. 1993. “Three Presentations on Geographical Analysis and Modeling: Non-
Isotropic Geographic Modeling; Speculations on the Geometry of Geography; and Global
Spatial Analysis.” UC Santa Barbara: National Center for Geographic Information and
Analysis. Technical Report 93-1. Accessed July 8, 2016.
http://escholarship.org/uc/item/05r820mz#.

74
Tripcevich, Nicolas. 2009. Cost Distance Analysis. Accessed July 14, 2016.
http://mapaspects.org/node/3744.

U.S. Geological Survey. 2011a. Validation of the ASTER Global Digital Elevation Model
(GDEM Version over the Conterminous United States, by Dean Gesch, Michael Oimoen,
Zheng Zhang, Jeffrey Danielson, and David Meyer. Accessed July 16, 2016.
http://www.jspacesystems.or.jp/ersdac/GDEM/ver2Validation/Appendix_B_CONUS%2
0_GDEMv2_validation_report.pdf

U.S. Geological Survey. 2011b. ASTER Global Digital Elevation Model Version 2 – Summary of
Validation Results, by Tetsushi Tachikawa, Manabu Kaku, Akira Iwasaki, Dean Gesch,
Michael Oimoen, Zheng Zhang, Jeffrey Danielson, Tabatha Krieger, Bill Curtis, Jeff
Haase, Michael Abrams, Robert Crippen, and Claudia Carabajal. Accessed July 16, 2016.
https://lpdaacaster.cr.usgs.gov/GDEM/Summary_GDEM2_validation_report_final.pdf

U.S. Geological Survey. 2014. Routine ASTER Global Digital Elevation Model. Accessed July
16, 2016. http://dx.doi.org/10.5067/ASTER/ASTGTM.002.

U.S. Geological Survey. 2015a. The Shuttle Radar Topography Mission (SRTM) Collection User
Guide. Accessed July 16, 2016.
https://lpdaac.usgs.gov/sites/default/files/public/measures/docs/NASA_SRTM_V3.pdf.

U.S. Geological Survey. 2015b. Shuttle Radar Topography Mission (SRTM) Void Filled.
Accessed July 16, 2016. https://lta.cr.usgs.gov/SRTMVF.

Waldren, W. 2003. “Evidence of Iberian Bronze Age ‘Boquique” Pottery in the Balearic Islands:
Trade, Marriage or Culture?” Oxford Journal of Archaeology 22, 4:357-374. Accessed
July 7, 2016. http://dx.doi.org/10.1046/j.1468-0092.2003.00193.x.

White, Devin A. 2012. “Prehistoric Trail Networks of the Western Papaqueria: A Multifaceted
Least Cost Graph Theory Analysis.” In Least Cost Analysis of Social Landscapes, edited
by White, Devin A. and Sarah L. Surface-Evans. (eds). Salt Lake City: University of
Utah Press.

White, Devin A. and Sarah L. Surface-Evans. (eds). 2012. Least Cost Analysis of Social
Landscapes. Salt Lake City: University of Utah Press.

Wilson, John P. 2012. “Digital Terrain Modeling.” Geomorphology 137, 1:107-121. Accessed
July 8, 2016. http://dx.doi.org/10.1016/j.geomorph.2011.03.012.

Wilson, John P. and John C. Gallant. (eds). 2000. Terrain Analysis: Principles and Applications.
New York: Wiley.

Woodward, Ann, John Hunter, Rob Ixer, Mark Maltby, Philip J. Potts, Peter C. Webb, John S.
Watson, and Michael C. Jones. 2005. “Ritual in Some Early Bronze Age Gravegoods.”

75
Archaeological Journal 162, 1:31-64. Accessed July 7, 2016.
http://dx.doi.org/10.1080/00665983.2005.11020620.

Yang, Mingyu, Frieke van Coillie, Min Liu, Robert de Wulf, Luc Hens, and Xiaokun Ou. 2014.
“A GIS Approach to Estimating Tourists’ Off-road Use in a Mountainous Protected Area
of Northwest Yunnan, China.” Mountain Research and Development 34, 2:107-117.
Accessed July 16, 2016. http://dx.doi.org/10.1659/MRD-JOURNAL-D-13-00041.1.

Yates, David Thomas. 1999. “Bronze Age Field Systems in the Thames Valley.” Oxford Journal
of Archaeology, 2: 157-170. Accessed June 10, 2016. http://dx.doi.org/10.1111/1468-
0092.00077.

Zakšek, Klemen, Elise Fovet, Laure Nuninger, and Tomaž Podobnikar. 2008. “Path Modeling
and Settlement Pattern.” In Layers of Perception, Proceedings of the 35th International
Conference on Computer Applications and Quantitative Methods in Archaeology (CAA),
April 2-6, 2007, (Kolloquien zur Vor- und Frühgeschichte, 10). Accessed July 14, 2016.
http://proceedings.caaconference.org/paper/101_zaksek_et_al_caa2007/

76
Appendix A: Settlement Data Record Sources

Historic England. 1993. Tynedale. York: Archaeology Data Service. Accessed July 16, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=1862160.

Historic England. University of London Institute of Archaeology. 1971. Levens Park (Site B).
York: Archaeology Data Service. Accessed July 16, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=1839351.

Historic England. n.d. Eston Nab. York: Archaeology Data Service. Accessed July 16, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=978803.

Historic England. n.d. No Title. York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=969489.

Historic England. n.d. No Title. York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=968760.

Historic England. n.d. No Title. York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=1291307.

Historic England. n.d. No Title. York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=1291290.

Historic England. n.d. No Title. York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=1234664.

Historic England. n.d. No Title. York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=975744.

Historic England. n.d. No Title. York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=976573.

Historic England. n.d. No Title. York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=975138.

Historic England. n.d. No Title. York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=974681.

History of Berwickshire Naturalists Club. 1861. Greaves Ash/Linhope Camp. York: Archaeology
Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=1866491.

77

Newcastle University. 1980. Standrop Rigg, Linhope. York: Archaeology Data Services.
Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=1865481.

Northumberland Archaeological Group. 2000. Wether Hill. York: Archaeology Data Service.
Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=1885610.

Northumberland County Council. 1979. Chesters Burn Unenclosed Round House Platform,
Field Plots and Clearance Cairns (dataset). York: Archaeology Data Service. Accessed
July 24, 2016. http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=955506.

Northumberland County Council. n.d. Apperley Dene Fort. York: Archaeology Data Service.
Accessed July 16, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=963932.

Northumberland County Council. n.d. Bronze Age Unenclosed Settlement and Field System with
Burial Cairns South West of Brands Hill, 350m ENE of Langlee. York: Archaeology Data
Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=955994.

Northumberland County Council. n.d. Cairnfield, Unenclosed Hut Cirle Settlement and Area of
Cord Rig 190m North East of Pigdon’s Leap. York: Archaeology Data Service. Accessed
July 24, 2016. http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=955491.

Northumberland County Council. n.d. Cat Crag Settlement of Unenclosed Round Houses. York:
Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=955445.

Northumberland County Council. n.d. Cochrane Pike, Unenclosed Settlement. York:
Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=957412.

Northumberland County Council. n.d. Defended Settlement, 700m South East of Glen Aln. York:
Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=958557.

Northumberland County Council. n.d. Defended Settlement, Field System, Cairnfield, Round
Cairn and Medieval Settlement, 300m South East of Barrow. York: Archaeology Data
Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=955193.

78
Northumberland County Council. n.d. Hazeltonrig Hill 1, Bronze Age Unenclosed Settlement.
York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=955521.

Northumberland County Council. n.d. Hut Circle 320m West of Threestoneburn House. York:
Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=955949.

Northumberland County Council. n.d. Knock Hill North Settlement of Unenclosed Round
Houses. York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=955446.

Northumberland County Council. n.d. The Law, Belling. York: Archaeology Data Service.
Accessed July 16, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=960360.

Northumberland County Council. n.d. No Title. York: Archaeology Data Service. Accessed July
24, 2016. http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=969278.

Northumberland County Council. n.d. Prehistoric Settlement at Scaud Knowe. York:
Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=955505.

Northumberland County Council. n.d. Prehistoric Unenclosed Hut Circle Settlement South East
of Long Crags. York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=955987.

Northumberland County Council. n.d. Reavely Hill, Unenclosed Round House. York:
Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=957321.

Northumberland County Council. n.d. Turf Knowe, Settlement of Unenclosed Round Houses.
York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=957311.

Northumberland County Council. n.d. Unenclosed Hut Circle Settlement 1060m North West of
Ilderton Moor. York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=957690.

Northumberland County Council. n.d. Unenclosed Hut Circle Settlement 655m South West of
White Gate. York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=966657.

Northumberland County Council. n.d. Unenclosed Hut Circle Settlement, Associated Field
System and Cairnfield on Standrop Rigg, 820m North West of Linhope Spout. York

79
Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=955444.

Northumberland County Council. n.d. Unenclosed Hut Circle Settlement, Cairnfields and a
Rectangular Enclosure 1km South-West of Debdon Whitefield. York: Archaeology Data
Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=957137.

Northumberland County Council. n.d. Unenclosed Hut Circle Settlement, Field System,
Cairnfield and Cord Rig Cultivation North West of Linhope Spout. York: Archaeology
Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=955441.

Northumberland County Council. n.d. Unenclosed Round House, 200m North of Barrow Cleugh.
York: Archaeology Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=954481.

Northumberland County Council. n.d. Unenclosed Scooped Settlement on East Slope of
Harehope Hill, 750m South East of High Akeld Cottages. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=955654.

Northumberland County Council. n.d. Unenclosed Settlement 220m South-West of Manor House.
Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=961894.

Northumberland County Council. n.d. (West Plain) Site of Henge 1000yds SE of Milfield. York:
Archaeology Data Service. Accessed July 16, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=956225.

University of Durham Archaeological Services. 2005. Catterick Army Training Estate. York:
Archaeological Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=1834657.

University of Durham Archaeological Services. 1998. ‘Little Haystacks,’ Haystack Hill. York:
Archaeology Data Service. Accessed July 16, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=1939702.

University of Durham Archaeological Services. 1997. Little Maltby Farm. York: Archaeology
Data Service. Accessed July 24, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=1829508.

Waddington, Clive. 2007. Cheviot Quarry, Northumberland: Excavation and Interpretation
[dataset]. Archaeological Research Services. York: Archaeology Data Service. Accessed
July 24, 2016. http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=1612025.

80
Yorkshire Dales National Park Authority. n.d. Bronze Age Unenclosed Settlement, High Eskeleth
[dataset]. York: Archaeology Data Service. Accessed July 16, 2016.
http://archaeologydataservice.ac.uk/archsearch/record.jsf?titleId=2122768

Abstract (if available)

Abstract Numerous studies within the last 200 years have shed light on the socioeconomic patterns of the Beaker culture during the Bronze Age, particularly in the United Kingdom. However, with the expanding role of GIS in the field of archaeology, there is an increasing amount of spatial data on this cultural group, allowing opportunities for analysis that can begin to describe inter- and intrasite spatial connections. The geographic connections of pathways, for example, can illustrate the corridors of cultural exchange that gave rise to and sustained the Beakers for over 1,000 years. Using Least Cost Path analysis, this thesis aimed to model such spatial connections in Northeast England. ❧ The study generated 66 anisotropic LCPs that modeled possible path connections between sites. The first 18 LCPs served as the primary LCPs between sites—within clusters and between clusters. Three assessment tests were conducted to validate these LCPs. First, for each primary LCP, another LCP was generated traveling in the reverse direction. Second, new segments that utilized pairs of nearby sites, approximating the alignment of original pairs, were generated

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Archaeological least cost path modeling: a behavioral study of Middle Bronze Age merchant travel routes across the Amanus Mountains, Turkey

PDF

Modeling geopolitics in Tikal through least cost paths

PDF

Applying least cost path analysis to search and rescue data: a case study in Yosemite National Park

PDF

Predicting the presence of historic and prehistoric campsites in Virginia’s Chesapeake Bay counties

PDF

Modeling nitrate contamination of groundwater in Mountain Home, Idaho using the DRASTIC method

PDF

Residential housing code violation prediction: a study in Victorville, CA using geographically weighted logistic regression

PDF

Using Maxent to model the distribution of prehistoric agricultural features in a portion of the Hōkūli‘a subdivision in Kona, Hawai‘i

PDF

A threat-based least-cost path decision support model for national security resource allocation along the US-Mexico border

PDF

Deriving traverse paths for scientific fieldwork with multicriteria evaluation and path modeling in a geographic information system

PDF

Does the Bay Area have a social center? Delimiting the postmodern urban center of the San Francisco Bay Area

PDF

Precision agriculture and GIS: evaluating the use of yield maps combined with LiDAR data

PDF

Validation of volunteered geographic information quality components for incidents of law enforcement use of force

PDF

Predicting archaeological site locations in northeastern California’s High Desert using the Maxent model

PDF

A Maxent-based model for identifying local-scale tree species richness patch boundaries in the Lake Tahoe Basin of California and Nevada

PDF

Modeling historic structure preservation candidacy on Fort Ord

PDF

Evaluating the utility of a geographic information systems-based mobility model in search and rescue operations

PDF

Identifying areas of high risk for avian mortality by performing a least accumulated-cost analysis

PDF

A spatial analysis of violent crime cold spots: testing the capable guardian component of the routine activity theory

PDF

Light rail expansion in Houston using viable path corridors and least cost path: alternatives for the failed University and Uptown lines

PDF

A comparison of GLM, GAM, and GWR modeling of fish distribution and abundance in Lake Ontario

Asset Metadata

Creator Alvez, Christian (author)

Core Title Modeling prehistoric paths in Bronze Age Northeast England

School College of Letters, Arts and Sciences

Degree Master of Science

Degree Program Geographic Information Science and Technology

Publication Date 09/28/2016

Defense Date 09/06/2016

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag archaeology,GIS,LCP,least cost path,OAI-PMH Harvest,prehistoric paths

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Wilson, John (committee chair), Kemp, Karen (committee member), Lee, Su Jin (committee member)

Creator Email alvezchristian@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-308580

Unique identifier UC11280652

Identifier etd-AlvezChris-4827.pdf (filename),usctheses-c40-308580 (legacy record id)

Legacy Identifier etd-AlvezChris-4827.pdf

Dmrecord 308580

Document Type Thesis

Format application/pdf (imt)

Rights Alvez, Christian

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA