Data is a core component of Geographic Information Systems, learn more about the different aspects of geographic data including Vector, Raster, layers, 3-D data and the complexities of topology.
The very simplest advantage that GIS gives in comparison with paper maps is that you can change the appearance of the information to any style you like. In conventional mapping a large amount of time and effort is spent on deciding appropriate colours and styles for the depiction of features, to ensure that the image is as clear and informative as possible. The flexibility of GIS adds an extra dimension to this process: you can change the appearance depending on exactly what message you want to convey.
Changing the appearance of vector data
The greatest flexibility comes when using vector data; remember that all the computer stores is a set of coordinates which make up the shape of the object. Any GIS will let you choose the colour and style of how the features are represented on screen. These styles will then be reflected in the printed output from the system and, ironically, many organisations use GIS simply to create customised printed maps.
With point features you can change the symbol type and the colour, and for line features you can change the style and colour. For area features you can change the colour of the shape itself as well as its perimeter. The colouring of the body of the shape (the fill) can be made solid, patterned or even transparent.
Changing the appearance of raster data
The nature of raster data inherently defines how the mapping should appear. The raster map is an image of coloured pixels, and the fact that a road is depicted comes from a visual interpretation of adjoining pixels of the same colour, not from any information saying this is a road in the data structure itself. The only data entities are the pixels themselves, and the only intelligence stored about those pixels is their colour.
Having said that, it is still possible to alter the appearance of raster data in a GIS. In simple terms you could make all red pixels appear blue, for instance. Usually this is not a good idea, because the image was designed with the most appropriate colours in the first place.
However, in some circumstances this facility is useful because you can tone down the colour scheme of the raster image to allow other information to be placed on top and made more readable. The graphic below demonstrates some different ways in which the appearance of a raster image can be adjusted.
GIS really gets going as a powerful tool when you start to work with different layers of information about the same geographical area at the same time.
When compiling a conventional map, the cartographer has to draw a balance between displaying as much information as possible to make the map useful without adding so much detail that it becomes cluttered and confusing. With GIS, this problem is removed – many different layers of information can be added, and shown in different combinations and in a different order, depending on the particular message to be conveyed.
Using the power and flexibility of the computer, different data layers can be switched on and off at the click of a mouse, so that many different views can be created for the same location.
The image on the left represents the way in which a GIS can display many different layers of information at the same time. Using different combinations, the display can serve a wide range of purposes that could only otherwise be achieved by producing a whole set of different paper maps.
Referring to the graphic: step 1 creates a communication network map by switching on just the towns, roads and railways; step 2 generates a view of the relief of the area by switching on the contours and rivers; and step 3 gives a view of all these layers together which can help to analyse the spread of urbanisation in relation to networks and relief.
Identifying change over time
Another benefit of mixing and matching different layers is that by combining mapping for the same area surveyed at different times, you can identify any changes:
These are just simple examples of how a different message can be portrayed within a GIS by showing a mixture of feature layers. The most sophisticated GIS users are likely to be working with hundreds of layers, enabling them to create any kind of map display for a particular geographical location.
Now things start to get more exciting – GIS not only revolutionises the usefulness of map information, by allowing it to be shown in many different combinations, it also takes us beyond the realm of the flat, planar view of the landscape.
From 2D to 3D
To understand a two-dimensional (2D) representation of the real landscape you need a level of interpretation and imagination. The physical world exists in three dimensions and, unless you ignore those extruded plastic maps of the world with snow-capped lumps showing the main mountain ranges, the realm of conventional maps is uncompromisingly flat. The capability of GIS to produce dynamic and attractive three-dimensional (3D) maps is one of its most exciting benefits.
Map makers use a range of visual symbols to show height information and create the illusion of an undulating surface:
- Spot height symbols
- Hill shading
- Cliff and slope symbols
- Viewpoint symbols
- visualisation of the 3D landscape;
- calculation of gradients for roads and railways;
- environmental impact analysis for engineering projects;
- screening of objects such as power stations and wind turbines through line of sight analysis;
- radio wave propagation analysis – important to mobile communication networks;
- flood risk analysis;
- town planning; and
leisure products – many computer games use realistic landscapes based on GIS height data.
Topology – it's all about relationships
While the term topography describes the precise physical location and shape of geographical objects, the term topology is more concerned with the logical relationships between the position of those objects. For example, in a topographic map of Hyde Park, London you would show an accurate depiction of the shape of the park and a precise alignment of the shape of the objects within it – Serpentine lake, for instance.
In a topological map the precise shape of the objects is not important – there will be a shape called Hyde Park and a shape called Serpentine lake, but most importantly the Serpentine lake object will be entirely contained inside the Hyde Park object.
It is the knowledge of this spatial relationship which is key. This may seem a dry and obscure point, but topology is critical to understand how the computer is able to analyse the relationships between objects. If the topology of a set of data is wrong then the GIS cannot analyse how objects relate to each other: are they next to each other? Do they overlap? Do they form a connection? Does one lie completely within another?
Geospatial data will have topology inherited from the source material. Hence, when you digitise a map, the topology, which is implicit in the visual interpretation of the map, is built into the data. However, care is needed. Unless the data is topologically correct the computer will not necessarily pick up the relationships.
One of the commonest errors when digitising data occurs when there is a slight inaccuracy in the start or end point of a line. This can result in the linework not being correctly joined up. The line can form an undershoot or an overshoot (see diagram below).
Although these errors can be difficult to detect by the human eye, they prevent the GIS from understanding the fact that these two features are actually joined to each other.
A GIS does more than just show the positions of objects on a computer screen – it can also be used to model real-world events. One of the most important examples of this is the ability to model networks. There are many networks in geographical data, such as water courses and street maps. A GIS can analyse the potential flow around these networks, a useful ability in flood analysis or route finding. It can only do this if the data has correct network topology – the joining of the lines at exactly the same point in the data. Lines in a GIS network are usually called links, the points which define the shape of the link are called vertices, and the points at which they join are called nodes.
Area features are defined in a GIS by the linear shape of the perimeter and some kind of reference point indicating that the space enclosed by those lines relates to a geographical feature. This reference point is referred to as a centroid, a seed or a label point.
Geospatial data is often captured in the form of linework showing the extent of physical features on the ground, like fences, roads, and rivers. The area features then have to be identified by assigning seed points to each bit of space in the resulting map data. For example, a wall feature may, at one point of its length, define the perimeter of a school playground, but further along form the edge of someone's front garden. It is the seeds that store the information about which links make up the edge of an area feature and what it is. It is very important to avoid undershoots in the data, otherwise the system cannot tell whether an area is closed at a particular point.
3D GIS data
Height information can be captured in a GIS in exactly the same way as the shape and location of objects. The spectacular ability of today's computers to perform calculations means that 3-D models of the ground surface can be constructed from data recording the height at different points across an area. The typical way this information is stored is an extension of the conventional grid coordinate system: as well as recording the latitude (the x axis) and longitude (the y axis) for a given point, the elevation (the z axis – usually as height above sea level in metres) is also stored. Thus the height information for an area is often referred to in terms of z values. The fluctuations in ground height across an area are a continuous phenomenon – every point on the ground has a z value irrespective of whatever physical features are present.
Point height information can be collected by surveyors out in the field, or more commonly by using remote sensing, including photogrammetry. Points of the same height can be joined to form a line or contour.
Once created, most 3D GIS data is stored as a grid of points, with x, y and z values stored as attributes, often referred to as a digital terrain model (DTM) or digital elevation model (DEM). From this grid a computer can build a 3D model.
An alternative method of representing a surface is to create a triangulated irregular network (TIN). A TIN model forms a continuous surface by connecting irregularly spaced spot heights to form triangles, keeping a flat surface within each triangle.
These 3D models can be made to look very realistic by applying colour to the surfaces. It is even possible to drape raster images of maps or aerial photos over the surface with quite stunning effect. Furthermore, if the heights of physical objects like buildings, forests and electricity pylons are known, these can also be built into the 3D model. Hence it is possible to create computer models of entire towns and villages which relate directly to the real world.
The ability of GIS to operate in three dimensions has many useful applications, for example: