A Concise Introduction to Geographical Information Systems and Science M Morad and T Connolly Kingston Centre for GIS Kingston University, UK Contents Section 1 An Overview of Main Concepts Section 2 GIS Terminology Section 3 Major GIS Acronyms Section 4 Recommended Further Reading 2 Section 1 An Overview of Main Concepts 3 GIS and geospatial data A Geographical Information System is, technically speaking, a computerised system for capturing, processing, enhancing, querying, analysing and visualising spatially-referenced data. GIS draws on several related disciplines, and its cross- disciplinary nature is illustrated below. GIS draws upon several related disciplines What distinguishes GIS from other forms of information systems, such as databases and spreadsheets, is that GIS deals with spatial information. GIS has the capability to relate layers of data for the same points in space, combining, analysing and, finally, mapping out the results. Spatial information uses location, within a coordinate system, as its reference base. The most common representation of spatial information is a map on which the location of any point could be given using latitude and longitude, or local grid references such as the British National Grid (BNG) or the New Zealand Map Grid (NZMG). See the diagram below for an example of a special engineering survey grid. y Location of North Sea wells 5 Well id x y depth f g a 1 1 2000 4 e b 2 2 3000 c 4 2 2900 3 d d 3 3 5400 e 2 4 1750 2 b c f 1 5 2500 g 3 5 6000 1 a 0 0 12345x Point data and related attribute information Some applications of GIS are obvious, for example water supply companies use GIS as a spatial database of pipes and manholes; local governments can use GIS to manage and update property boundaries, emergency operations and environmental resources. GIS may also be used to map out the provision of services, such as health care and primary education, taking into account population distribution and access to facilities. Increasingly, GIS is being used to 4 assist businesses in identifying their potential markets and maintaining a spatial database of their customers. In broad terms, a Geographic Information System could be defined as a set of principles and techniques employed to achieve one (or both) of the following objectives: 1. Finding suitable locations that have the relevant attributes. For example, finding a suitable location where an airport, a commercial forest or a retail outlet can be established. This is usually achieved through the use of Boolean (logical) operations, as illustrated schematically in the following figure. OR = (union) AND = (intersection) XOR = NOT = Boolean operations 2. Querying the geographical attributes of a specified location. For example, examining the roads in a particular locality, to check road density or find the shortest path, and so on. This is often achieved by ‘clicking’ onto the location or object of interest, and examining the contents of the database for that location or object. The following figure shows examples of spatial query functions in GIS. 5 Some of the basic spatial query functions GIS data are usually stored in more than one layer in order to overcome the technical problems caused by handling very large amounts of information at once. It is easier to work with complex spatial problems one layer at a time, to enable the revision of the data without having to overhaul the entire information system. This is a fundamental aspect of GIS, and working with layers of geographical information is known as data integration (see the figure on the next page). For example, a comprehensive GIS of a UK county might consist of several layers of data: • Layer 1 property boundaries and land use types (area maps) • Layer 2 road and railway networks (line maps) • Layer 3 terrain characteristics (spot-height or contour maps) • Layer 4 hospital and school locations (point maps) Area data Line data Surface data Point n data property boundaries road and rail n contours n n schools Examples of layers of area, line, surface and point data 6 Spatial data may be represented in GIS in one (or both) of the two following formats: • Vector model, as geometric objects: points, lines, polygons • Raster model, as image files composed of grid-cells known as pixels The following figure outlines this idea. Vector representation lake Real world Map Raster representation Examples of vector and raster data models for the same object 7 The importance of geographical data The nature of any natural or economic activity with a spatial dimension cannot be properly understood without reference to its spatial qualities. Spatial data have two essential parts: location and attributes. A GIS requires locational references. Typical locational references are latitude and longitude and national grid references such as the British National Grid. However, other geospatial codes can also be used to identify location, such as postcodes. Any locality would have a number of characteristics or properties associated with it. These attributes are usually kept in tables, containing such information as vegetation types, population, annual income, and so on. GIS systems store and process data in two formats, vector and raster (or as the filed and object views of reality, as some experts would prefer to say). In the vector data model (see the next Figure), the world is represented as a mosaic of points and interconnecting lines representing the location and boundaries of geographical entities. In vector data models, the data are represented as: • lines (arcs) • polygons (traversed areas) • nodes (intersection points) • points (labelled nodes) The raster (or grid-cell) data model has arisen from aerial and satellite imaging technology, which represents geographical objects as grid-cell structures known as pixels. See the following figure. 10 7,10 5,9 9,8 4,7 polygon 1,6 8,6 6,6 5 2,5 line 5,4 point 2,2 7,2 0 0 510 Geographic entities represented in a vector format Each data model has particular strengths and weaknesses, and the type of model used is determined by the nature of the work being undertaken and the data available. 8 The main advantage of the vector data format is that it permits a neat representation of points, boundaries, and linear features. This makes it particularly useful for analysis tasks that require accurate positioning, for example in engineering or cadastral boundary databases. It is also possible in a vector-based GIS to define the spatial relationship (ie the connectivity and adjacency) between coverage features. This aspect of GIS is known as topology, and is important for such purposes as network analysis (for example to find an optimal path between two nodes in a complex transport network). By contrast, raster-based GIS defines the position of features in terms of (x, y) co- ordinates where topological associations are more difficult to identify. However, the main disadvantage of vector data is that the boundaries of the resulting map polygons are discrete (enclosed by well-defined boundary lines), whereas in reality, assuming maps scales are not an issue, the map polygons may represent continuous gradation or gradual change, as in soil maps. Cadastral Vegetation boundaries River Thematic network Utility Soil diagram Advantages of Vector and raster data Grid Data Models use a raster matrix (a grid of image cells) to represent information. Grid data are also known as raster data. The resolution (visual definition) of the raster depends on its pixel (cell) size. In other words, pixel resolution represents the size of the ground area covered by each pixel in the image. The smaller the cell size, the higher the resolution. The raster data model is, therefore, good for representing indistinct boundaries, such as thematic information on soil types, soil moisture, vegetation, ground temperatures, and so on. Furthermore, as reconnaissance satellites and aerial surveys use raster-based scanners, the information (ie scanned images) can be directly incorporated into GIS programmes capable of working with raster data. However, the higher the grid resolution, the larger the data file is going to be. This is the main limitation of raster based GIS. 9 50 m pixel 25 m pixel (grid 20 x 16 = 320 pixels) (grid 40 x 32 = 1280 pixels) Raster resolution The question of which data model to use in GIS depends on the nature and objective of the GIS project. Primarily the model type will depend on the nature of the data. Issues of concern are the volume of the data generated, ease of analysis and accuracy. Generally, vector data sets are economical in terms of file size (ie do not take up much memory space) and have a high level of positional precision, but are relatively difficult to use in mathematical computations. On the other hand, grid data sets tend to take up more file space and have a coarser resolution, but are easier to work with mathematically (ie easier to process computationally). 10 Geospatial data processing There are several techniques for capturing, processing and analysing geospatial data in GIS. Some of these are explained below. Data capture. A GIS cannot analyse the information in a map, if the data are not already in digital form which the computer can recognise. Maps can be digitised (hand-traced with computer mouse) to collect the coordinates of the map features. Electronic scanning devices can also be used to convert map lines and points to digital information. PC or Workstation Digitising tablet connected via serial port A typical digitiser Information retrieval. With a GIS we can point at a location, object, or area on the screen and retrieve recorded information about it from the Database Management System (DBMS) which holds the information abut the map’s features.