GE3502GE5502 Geographic and Land Information Systems

1
GE3502/GE5502Geographic and LandInformation
Systems
Lecture 2 A Review of Some of the Basics
SCHOOL OF TROPICAL ENVIRONMENT STUDIES AND
GEOGRAPHY
2
Lecture Plan

• Dimensions of geographic information
• Spatial data models - Raster
• Spatial data models - Vector

3
1. Dimensions of Geographic Information

• Geographic information has 3 properties
• i) Position and spatial properties where ?
• Absolute precise spatial coordinate
• Relative in relation to something else,
• i.e. beside, inside, outside.
• ii) Non-spatial attributes or descriptive
  properties what?
• iii) Temporal properties when ?

4
i) Position and Spatial properties Where?

• Absolute
• Objects usually described within a cartesian
  (x,y) coordinate system such as
• Transverse Mercator (UTM) or
• Australian Map Grid (AMG) or
• Global referencing LATITUDE/ LONGITUDE system
• Maps, images, censuses, field data all have to
  have a common coordinate system before they can
  be entered into a GIS and before any spatial
  analysis can be performed.
• This is a critical area of GIS development
  (registration of datasets to a common coordinate
  system).

5
i) Position and Spatial properties Where?
Relative

• Things can be described in terms of their
  position in relation to each other.
• They can be adjacent, they can overlap and they
  can be contained within.
• These are topological descriptions of an object
  ie. spatial relationships between objects.
• Both absolute and relative coordinate systems
  form the mapping element of GIS. There are
  numerous digital mapping packages, but only those
  that recognize spatial topology (relative spatial
  location) can function as GIS (well follow this
  up in lecture 4).

6
ii) Non spatial attributes What?

• The identity or properties of an object or theme
  such as colour, cost, name, area, pH, vegetation
  type, etc. .
• raster GIS represented as a numerical key
• vector GIS stored in a separate database

7
iii) Temporal properties When?

• Temporal relationships
• before
• after
• at the same time
• change at a particular rate, etc..
• These are all important properties, and have
  consequences for the analysis of spatial
  patterns.

8
iii) Temporal properties When?
How old is the information data? How up-to-date
is the data? Can we rely on its temporal
accuracy? Depends on the type of data e.g.
Analysis of wildlife habitat. If aerial photos
are 5 years old, can we be sure that the
vegetation in the landscape hasnt changed.
However, we could be pretty sure that a 5 year
old DTM would probably be OK.
9
iii) Temporal properties When?

• GIS widely used in temporal modelling, i.e.
  measuring and predicting change.
• Before it can function as a modelling tool, a GIS
  must have a historical component
• Change may be episodic or continuous.

10
iii) Temporal properties When?

• Episodic Change
• e.g. Australian census. Provides us with a
  range of demographic information. Every 5 years
  in Australia.
• CENSUS (t1) CENSUS (t2) CENSUS (t3). Three
  different versions of the same theme.
• allow analysis of historical trends in the same
  theme, and this data is represented in an
  episodic form (although actual change is
  continuous).

11
iii) Temporal properties When?

• Continuous Change
• Continuous change occurs in environmental
  phenomena, such as
• the weather, crop growth or condition
• Surface water distribution timing, extent,
  depth of water bodies in floods or in drainage
  systems
• Biomass measurements of relative greenness and
  bareness (primary productivity) in relation to
  local grazing impacts, seasonal change or longer
  term climatic trends, e.g. El Niño.

12
iii) Temporal properties When?

• The Key to measuring continuous change is
  Remotely Sensed data, which can be used to
  monitor at rates of between every 12 hours to 16
  days, and arrives in raster data format.
• I will devote an entire lecture to the
  measurement of change using GIS later in the
  semester.

13
Spatial Data Models

• Objects and surfaces are represented in
  completely different ways in a GIS in vector and
  raster spatial data structures.

14
2. Spatial Data Models Raster

• Raster is basically a grid structure, matrix of
  cells (pixels)
• The region is split into uniform, regular cells
  that cover the entire area

• Cells usually square but may be rectangular and
  occasionally triangular or hexagonal
• Common sizes are 10m2, 30m2, 100m2 or 10km2 or
  1000km2

(Bernhardsen 1999, p. 68)
15
Objects in Raster Format

• Each cell (pixel) on the grid has a value that
  corresponds to the geographical theme being
  represented. For instance, the LANDUSE overlay
  represented in the previous diagram has 3 values
  for differing land cover types.
• An object is represented by a cell or connected
  group of cells.
• If the majority of a cell contains a particular
  feature, then the entire cell will usually be
  given a value representing that feature.
• Therefore, the resolution of the analysis is
  limited by the size of the cell
• Raster GIS does not record precise boundaries it
  only records the presence or absence, or a range
  of values for a property in a cell.

16
Raster Data Storage

• Cells in a raster model are given in a specific
  sequence from the matrix (starting from the upper
  left corner).
• Data is usually stored in two databases.
• Cell number and the value of the cell.
• An attribute table that assigns real world
  attributes to each value.
• Therefore, attributes are assigned to every cell.

17
(Bernhardsen 1999, p. 70)
18
Multiple Layers in Raster Format

• Each cell can have only one value, so different
  types of objects must be assigned to different
  raster layers.
• Each raster layer deals with one theme.

• Where multiple attributes correspond to a certain
  suite of objects, multiple layers must still be
  used.
• Therefore, raster data models usually have more
  layers than vector models.

(Bernhardsen 1999, p. 70)
19
Applications of Raster GIS

• Different values can be allocated to each cell.
  This can represent a continuous gradient very
  well, e.g. temperature, rainfall, vegetation.
• Therefore, can represent continuous data, e.g.
  topography (DEMs).
• Particularly important to environmental data
• Forms of environmental data such as depth,
  quality or mineral content of water, soil,
  biomass, or greenness in crop productivity, and
  changes in the structural characteristics of
  vegetation.
• These are continuous, have fuzzy boundaries, and
  some may require temporal analysis.
• Later on, well look at why raster is better at
  dealing with this type of data.

20
Raster GIS

• The most important thing to remember about
  raster GIS at this stage is that one of the most
  important types of geographical information -
  remotely sensed data - is captured by
  environmental monitoring satellites in raster
  form.
• Some of the techniques used in raster analysis
  are similar to those used in digital image
  analysis. Therefore, raster GIS is sometimes
  called image-based analysis.

21
3. Spatial Data Models Vector

• Vector GIS format does not use grid cells.
• Objects are recorded as points, lines (arcs), or
  polygons.
• Lines consist of two or more points that are
  connected, while polygons are made up of 3 or
  more lines that join to enclose an area.
• Spatial relationships between objects are
  recorded through the use of topology.
• Topology describes the relationships between
  elements in space, and is necessary for advanced
  vector analysis.
• Stores info in a continuous co-ordinate system
  (i.e. not limited by the size of the pixels).
  (The graphics are a much higher quality than that
  of a raster system.)

22

• The spatial information for each object in the
  map or each map unit is stored in a separate
  database from the attribute information.
• Spatial relationships between objects are
  recorded through the use of topology (covered in
  detail later).

(DeMers 1997, p. 110)
23
Vector GIS Cont.

• Spatial information is stored in a continuous
  co-ordinate system where each object in a map is
  represented by its geographic co-ordinates,
  rather than as a series of cells.
• Therefore spatial resolution not limited by the
  size of pixels.
• It has virtually infinite scaling limited only
  by the precision of the computer.

24
Vector GIS Cont.

• An object is represented as closely as possible
  to reality in the continuous co-ordinate space,
  so the graphics are a much higher quality than a
  raster system.
• Vector GIS is a cartographic, and database GIS.
  You cannot use it as a platform for image
  analysis, terrain modelling, or environmental
  modelling. The spatial data structure does not
  permit it.

25
(No Transcript)
26
Readings

• Chapters 1, 3 5 of Longley et al. 2001