GIS LECTURE NOTES

WEEK 1: KEY CONCEPTS AND DEFINITIONS

Course literature:

 All chapters except 7.1,7.2,7.4, 7.5, 8.8

 Formula’s: only those mentioned in lecture sheets

GIS a computerised set of tools for collecting, storing, retrieving at will, transforming, and
displaying spatial data from the real world for a particular set of purposes. It links spatial and
tabular information enabling the visualisation of patterns, relationships and trends.

GIS as a toolbox vs. science (steers research – e.g. analysis of pathways)

5 components:

1. Hardware– input and output (plotters) devices. Processing devices (PC/datacentres)

2. Software (ArcGIS/QGIS)
3. Data
4. People – system & application managers/specialists
5. Methods – algorithms, guidelines, specification standards

Main applications in archaeology:

 Management of archaeological resources

o Cultural resource management
o Storing/maintaining data, predictive models
 Excavations
o Collecting/analysing/presenting data
 Landscape archaeology
o Surveys
 Modelling human behaviour
o Pathways, agent based modelling, visibility

Key concepts:

1) Data models  spatial data structures (vector/raster models)

2) Layers  topographic/thematic
3) Map-database integration  combining map data with attribute data

Raster model:

 Continuous data (e.g. elevation, groundwater, artefact densities)

 Simple – based on single numerical value

Pro’s and cons: SEE PPTX. (raster model)

Vector model:

 Consists of geographical primitives and topology (spatial data)

 More complex
 Works better with clearly bounded entities
 Flexible database linkages make it ideally suited for database intensive application (e.g.
archaeological features, sites, monuments, administrative boundaries, network analysis etc.)
 GIS LECTURE NOTES

4 aspects of spatial GIS data:

1) Locational

A record of the position, in geographic space that determines where something is and what shape it
takes (geographical object, projections, coordinates)

 Earth is divided in meridians

 Projections required to convert 3D sphenoid (earth) to a flat, 2D surface.
o Cylindrical: preserves distances or arcs
o Conical: preserves angles
o Azimuthal/planar: preserves distances

2) Attribute

A record of attribute components which determines what the geographic objects represent and their
properties

 Vector GIS
 Database fields should be determined beforehand

3) Topological

A record of the logical relationships between different geographic objects (e.g. road connections,
flow direction)

 Can be calculated afterwards and addressed from the start

 Calculated: is within. Intersects, contains
 Topological properties should be stored prior to attempting analysis.

4) Metadata

Thorough documentation (explanation) of the contents of the layer in the geodatabase (metadata
component)

 What information do we need to attach to our map to enable others to find it in the future
and use it the right way?
 Necessary for good data exchange.

WEEK 2: SOURCES OF SPATIAL DATA

 Data can be expensive  takes time to generate
 Defining data needs:
o Conceptualisation
o Data specification (proxy):
 Geographic aggregation (squares)
 Temporal aggregation (mm/hr)
 Units of measurement (i.e. sieving)
 Purpose of spatial data determines level of generalisation (scale, error)
 Primary vs. secondary data
 GIS LECTURE NOTES

Sources for spatial data:

1) Paper maps

 Hand drawn excavation maps

 Digitalisation/Vectorisation of a paper map to digital map
o Heads-up (monitor) vs. Heads-down (XY tablet)
o Ground control points  Georeferencing!
o Within acceptable error margins
o Avoid overlaps!

4 stages of vectorisation:

1. capturing spatial data

2. Entering attribute data
3. Error Checking
4. Drawing Clean-up

Georeferencing: Defining the spatial coordinates for the data so it can be positioned and scaled in
relation to other maps. Used when digitalising paper maps.

Georectification: Combined process of correcting distortion and placing in a coordinate system (also
rubber sheeting, often used with historic maps).

2) Survey measurements (Total stations, GPS devices)

 GPS has an accuracy of 2-3cm (not very high)

o Replaces robotic total station
o (differential) GPS allows for the correction of errors

3) Images (Remote sensing)

Remote sensing: the collection and interpretation of phenomena without physical contact.

 Photographic
 Digital:
o Panchromatic (visible wavelengths)
o Multispectral
o False colour
 Always find the data and sensor specifications (raster data)
 Aerial photos require georeferencing/georectification

4) Lidar

Pulses from airplane to surface

5) Internet

Pros:

 Up to date
 Interactive
 From the source
 Free (INSPIRE)
 GIS LECTURE NOTES

Cons:

 Need access to the internet

 Reliability is put into question

WEEK 3: EXPLORATORY DATA ANALYSIS

 Visual inspection
o GIS enables the visualisation of (archaeological) patterns

Queries: formal request for a subset of data

 Attribute query (Phenomenal)

o Relational queries
 Topological (Geographical)
o Spatial queries
o Distance: Buffers  involves the selection of a subset of a dataset based on its
distance to a defined primitive.

SQL (Structured Query Language):

SELECT <variable>
FORM <table>
WHERE <condition>
ORDER BY <variable>
INTO <(new_table)>

Combined use og different Boolean operators requires brackets

OR = expanding selection
AND = limiting selection

Data Summary:

 Group by – will create a single record for each unique value of the given attribute
 Count

Creating new data with queries:

 Sum, average, min, max, count

 Statistics (univariate)
 Spatial functions

Thematic Maps – Data classification

 Finding patterns in the spatial data

 Presenting attribute data on the map (i.e. a choropleth)
 Classification methods
o Equal interval/equal step
o Standard deviation
o Quantile/quintile (equal count)
o Natural Break
o Equal area
 GIS LECTURE NOTES

o Geometric intervals

CASE STUDY – TERNEUZEN

 Determine location, nature size of sites (MA/Mesolithic)

 No excavation! Non-invasive research
o Using coring data  database
 Made a thematic map (coring & soil type)
 Interpolation of coring data
o Historical maps - georeferencing
o Aerial photography
o First where there are cables and pipes
o LIDAR
o Field survey  thematic maps
o Geophysical research (vectorised polylines)

WEEK 4: PREDICTIVE MODELLING

Scientific method in which we aim to simplify the world in order to understand, define, quantify,
visualise, or simulate reality.

(Conolly and Lake: Ch. 3, Ch. 8)

Predictive modelling Attempts to predict the value (or probability of occurrence) of a dependent
variable in an unsampled location using one or more independent variables.
i.e. the likelihood of settlements occurring in unsampled landscapes based
on the locational characteristics of a surveyed area.

Reasons to apply predictive modelling in archaeology:

 Academically – to gain insight into former human behaviour

 Cultural Resource management (CRM)
o Predict archaeological site location to guide future developments.
o Malta convention! (Treaty of Valetta)
 It provides a translation to policy makers and project developers. What are the appropriate
procedures to a project?

Predictive models in the Netherlands:

 IKAW – (Indicatieve Kaart van Archeologische Waarden) national coverage, land & sea.
o Semi-inductive model (expert judgement)
 Municipality Verwachtingskaart – developed by commercial companies, accurate to a local
level.
 Archeologische Waarden kaart

Predictive modelling and GIS:

 CRM – database/vector models

 GIS can use relatively simple techniques in a meaningful way to produce spatial models (as
apposed to statistical methods  raster/vector models
 GIS can produce a cartographic image  legal, convincing, aesthetic
 GIS LECTURE NOTES

How are predictive models made:

Variables:

 Environmental
 Social/manmade

Approaches:

 Inductive  model is constructed based on the correlation between known archaeological

sites and attributes from (mostly) the current physical landscape

 Deductive  model constructed on basis of general statement, or hypothesis about human

behaviour in the past

Assessing the quality of your model is dependent on the quality of independent testing.

Creating a predictive map:

1. Reclassify – suitability rank

2. Combining maps – overlay (easier with raster models (suitability mapping))

A good model includes:

 Must be able to explain and predict

 Must be valid (and verified)
 A model is always wrong!!! It is a simplification (but some are useful)

 As many known sites as possible

 Within a small area
 Optimising the size of the area with high expectation
 GIS LECTURE NOTES

Independent model quality:

 Use some of the known sites (split sampling) to determine whether the prediction is right for
these sites. (i.e. Chi-squared and Monte Carlo-simulations
 Expert judgement
 New discoveries/new research projects.

Disadvantages in predictive modelling:

 Use of incomplete data sets

 Biased selection of environmental parameters, often governed by availability of cheap data
(e.g. elevation)  neglect for the influence of cultural factors
 Neglect of the changing nature of the landscape
 Low spatial and temporal resolution
 Inappropriate statistical tools
 Little consideration for model validation
 Failure to consider how proxy variables really contribute to locational decisions

WEEK 5: SURFACE MODELS AND VISIBILITY

 Visualisation
o With special software you can render photorealistic landscapes
 Examples:
o Elevation models
o Hill Shading (shaded relief)
 Visibility analysis (viewshed)/intervisibility
 Mobility (i.e. cost of travel related to slope elevation  Cost-distance analysis)
 Predictive modelling (physical and social factors)
 Simulation of natural processes

Sources of Surface Models: Elevation data (DEM)

 Contour lines
1. Digitise map contours into a vector file
2. Rasterise the vector contours
3. Interpolate elevations between contour lines (often create terraces…)
 Spot heights/field measurements
o Lidar data (satellite data)  Raster DEM file
 Photogrammetry – structure from motion. Creates 3d representation of model visually

Interpolation techniques:

Plays a major role in the creation of elevation models. It is a way of calculating the value of unknown
elevation between known points.

 Not random
 Positive autocorrelation
 Cannot be assumed for anthropogenic phenomena
 Use algorithms
 Evaluating: any mismatch is prolongated into the derivatives
 GIS LECTURE NOTES

o Solution: Low pass filters (smoothing/moving average)

 Visually attractive
 Enhancing terrain models (removes interpolation artefacts
o High pass filter (sharpening/emphasising change)
 Edge detection (edge extraction)
 Weighted average

Methods:

1. Global methods – uses all points collectively to assert a trend

2. Local methods – provides information about local points as accurately as possible.

Inverse Distance Weighting (IDW):

 The higher the power the more local the pattern

 Easy, fast and simple
 Smoothing (averages out/does not pass through
measured points)
 Tendency for circular shapes

Splines: Estimates the unknown values by bending a surface through known values

 Uses a mathematical function that minimises overall surface curvature  smooth surface
that passes exactly through input points.
 Settings: Amount of acceptable tensions  higher = rougher surface
 Not applied very often

Kriging: Estimates the power (function) and search radius based on the actual data

 Moving average (similar principles to IDW)

 Sensitive to error margins
 Pros:
o More accurate interpolations
o Measure of confidence: Error of prediction
 Cons:
o Complex calculations
o User choices, model fit
o Slow

Triangular Irregular Network (TIN)

 Nodes can be placed irregularly over a surface  have a higher resolution in areas where a
surface is highly variable or where more detail is desired. Vice versa
 Goes through all known values (no averaging)
 CON:
o slow for large regions
o Triangular shapes (crude)

Derivatives:

First-order
 GIS LECTURE NOTES

 Slope = steepness (maximum rate of change of the elevation at a given location. Calculated
as % and decimal degrees)
 Aspect = orientation (rate of change in downhill direction)

Second-order

 Curvature = change in slope shape (convex vs. concave)

Visibility & viewsheds (Raster GIS analysis):

 Monumentality
 Intervisibility
 Overview

Settings: Observation point and viewer height

 Single viewshed
 Multiple viewshed (binary)
 Cumulative viewshed (combined)
 Total viewshed  cumulative viewshed for all cells
o Sum of all possible viewsheds
o Count number of cells in viewshed and record this in the viewpoint
o Takes a long time (days)

Problems with viewshed analysis:

 Quality of elevation model

 Algorithm used
 Curvature of earth
 Vegetation
 Visibility is not binary
o Contrast
o Maximum distance (haze)
o Brightness
 Edge effects
o Parts of viewshed outside study area
o Relevant viewsheds from viewpoints outside study area
 Intervisibility is not always the same!

CASE STUDY – Jordan. Udhruh Archaeological Project

CASE STUDY – Geoglyph visibility

Archaeological context of visibility analysis:

 Orientation and navigation

 Territories and control
 Symbolism and perception

WEEK 6: LEAST COST PATHS

Movement through the landscape: Cost Surfaces – Least Cost Paths (Raster Analysis)
 GIS LECTURE NOTES

Minimum Total Cost – Energy or Time taken to reach a location. Impacted by:

 Physical factors – rivers, slope, terrain

 Cultural factors – territories, danger, existing paths, visibility.
 Mode of transport
 Shortest route is not necessarily the most efficient

Purpose:

 Replication of routes. Understanding location of routes (sea mobility)

 Prediction of routes
 Site catchments (delimiting, analysing)

Effective slope is different from slope:

1. Slope maps show maximum rate of change of elevation across a cell  you do not always
traverse a cell in this direction
2. Uphill vs. Downhill

Friction map – directionality:

Isotopic algorithms Cost is the same in all directions (e.g. terrain roughness, vegetation)

Partially anisotropic algorithm Cost is dependent on direction of travel, but maximum cost is the
same (e.g: wind)

Anisotropic algorithms Cost of traversing a location is dependent on the direction it is being

used (upslope, downslope) as the magnitude of friction.

Cost Surfaces: Basis for calculating the least cost path. Cumulative cost surface.

Calculation is important:

 Weighing several factors  i.e. land-use and slope

 Combined weighted outcomes
 Disadvantage: only relative values
 Barriers and corridors:
o River – both barrier and corridor (centre low friction surrounded by buffer with high
friction)
o Absolute and relative barriers

PROBLEMS:

 Innacuracy of DEM
 Calibration by field experiment is difficult
 Algorithm artefacts: cardinal directions get priority
 Anisotropic cost surfaces are made without a sense of direction
 Ratio scale (values need to be meaningful and have an absolute zero)

WEEK 7: METADATA AND CARTOGRAPHY

Metadata: Information about data

 Data management
 Data retrieval
 GIS LECTURE NOTES

 Data preservation
 Rights & ownership management
 Data discovery
 Data sharing
 Data context

FAIR principles:

 Findable
o Described well
o Registered data
 Accessible
o Open and free data
 Interoperable
o Should use consistent and identifiable vocabularies
o Broadly applicable language
 Reusable
o Released with clear usage license

 Dublin Core Metadata record – not designed for geographical data

 ISO standard – extensive and detailed (400 fields)

Basics of Cartography:

Essential map items:

 Map (Thematic, Topographic)

 Legend
 Title
 Orientation
 Scale
 Citation (creator, date, metadata)

Types:

 Choropleth
 Proportional Symbols
 Dot density (random points)
 Isometric (isolines)

 Figure-ground concept – The main message of your map (figure) vs. the background (ground)
which serve to emphasise data and indicate orientation respectively.
 Visual hierarchy!
 Limited in B&W illustrations. It becomes difficult to present varied elements
 Carefully consider size and position of graphics and labels
 Infographics