• 
• Japanese
• 
• IETC's Focal Area
• Waste Management
• Publications & Tools
• Meetings & Workshops
• Global Partnership on Waste Management
• Background
• Objectives
  & Expected Outcomes
• Framework
• Sub-focal Areas
• Information Platform
• Meetings
• Participation / Membership
• Water and Sanitation
• Publications & Tools
• Meetings & Workshops
• Resources
• Publications & Tools
• About IETC
• Background and History
• Director
• Activity Reports
• Supporting Foundations
• Internship
• Contact Us

C. Features of spatial data

The natural landscape, with its heterogeneity and complexity, needs to be made less complex for management purposes. Rarely does the manager have the facility or luxury to accurately conform a management program to a diverse and complex ecosystem, within the fiscal and time constraints imposed upon management agencies. Therefore, for a scientific purposes, we simplify these ecosystem processes by modelling, which effectively uses physical or mathematical representations to mimic the natural environment. In this process, we delineate certain entities (landscape units) which can be described by:

• location in space,
• type,
• specific attributes,
• relationships between landscape units,
• time.

Location in space means that every landscape unit can be described by a system of co-ordinates (cartographic or on a computer screen). Typing helps to classify landscape units and merge them in to larger sets. This entails grouping landscape units by specific attributes. These attributes describe specific properties of the landscape units, expressed in a qualitative sense or on a quantitative scale. Relationships between landscape units can be of the following types: one-to-one, one-to-many, or many-to-many. Specific types of relationships are formed by a common border between units. Time is inherently part of a landscape unit, and can be seen as a fourth dimension which describes the evolution of objects within the landscape.

The geometric representation of an object in a given co-ordinate system is described by points, lines, and surfaces. Points do not have a topological dimension, but, in GIS programs, points can be displayed as symbols (for example, a sampling point can describe the source of a pollutant). Lines are used to represent one-dimensional objects. A line is a sequence of points (segment), in which we may distinguish beginning and ending points, and nodes at the intersection of lines. Lines can be additionally described by other coded information, as, for instance, elevation on a contour line. Surfaces are used to represent two-dimensional objects (e.g., lakes, forests, etc.). Three-dimensional objects are represented as a surface which has an attribute. For example, contours on a map can provide information on elevation that can be used in digital terrain models.

Models of the natural environment, stored in a GIS, are comprised of objects (landscape units) and information on the relationships between them. In computer graphics, there are two models that represent the geometry of objects - vectors and rasters. Vector models intuitively are closer to our way of describing the environment, as a set of landscape units with sharp boundaries (for example, geological or soil maps). In contrast, in a natural landscape, most often we observe fuzzy boundaries between landscape units that are best represented by the raster model. Sharp boundaries are rarely found in nature, and are more typical of constructed environments (for instance, in the urban areas). An advantage of the raster data model is a better representation of a fundamental natural landscape property; namely, continuity. The data structure is simple because every thematic layer can be seen as a rectangular matrix.

In vector models, every line has an assigned code which helps to save an object within a layer (in CAD programs), or in a thematic layer (in GIS programs). Information about connections between, and intersections of, vectors is written in the form of a topology. Thanks to topology, it is clear which objects (landscape units) share a boundary, and which points form nodes. Common methods of coding vector-based topological models are the DIME (dual independent encoding), and POLYVERT (arc-node structure) coding methods. Tables, storing information on vector topology, are used to calculate distances and areas of polygons, determine intersection points on lines, and analyse networks of lines.

Raster models represent objects by the use of evenly distributed elementary surfaces, which, in computer graphics, are called pixels (picture elements). The object representation is approximated by the spatial resolution of the applied raster. The location of raster elements is expressed by matrix co-ordinates (defined by row and column) which correspond to computer screen co-ordinates.

Coding of objects in raster models is done by assigning a numerical attribute to the pixels. Every pixel is assigned one value. Pixel values can have the following degrees of accuracy:

• 1-bit picture (Boolean, black and white) - values 0 or 1;
• 4-bit picture (16 colours or grey tones) - values 0-15;
• 8-bit picture (256 colours or grey tones) - values 0-255;
• 16-bit picture (65,536 colours or grey tones) - values 0-65,536; and,
• 24-bit picture (16.7 million colours or grey tones) - values 0-16,777,216.

In many GIS programs, the range of pixel values at the 8-bit scale is sufficient. Remote sensing data are often stored in this format, as pixel values are stored as a digital numbers (DN). Nevertheless, because of some calculations that are performed in GIS programs require other data formats, other domains are used:

• integer values 0 to 10⁹ (4 byte);
• values in the range -1 to 1 with a step of 0.01 (1 byte);
• percent values 0 to 100 with a step of 0.1 (2 byte); and,
• real values - 10¹² to + 10¹² (8 byte).

Table of Contents

• Brochure
• 
• 



• International Year of Forests
• 
• 



• World Environment Day
• 
• 



• UNEP Campaign
• 
•