|Short-term training programme on GIS (Geographic Information System) for fisheries (1993)|
Chief Scientific Officer
Bangladesh Space Research and Remote Sensing Organization (SPARRSO), Agargaon, Sher-e-Banglanagar, Dhaka 1207
Before going to detailed discussions on the Geographic Information System (GIS), let us give a thought on what geography means. In the classical sense, the word geography may be defined in terms of its constituent parts: geo and graphy. Geo refers to the Earth and Graphy indicates a process of presentation or writing. Thus in this literal interpretation geography means writing about earth. Another definition of geography focuses on mans relationship with the land. In more complete sense the geography deals with earths features, objects, phenomenon or any other information which are spatial in character. A key tool in studying and presenting this spatial information is map which contains an abstract representation/portrait of the real-world features of the earth on a two dimensional flat surface, more-commonly a sheet of paper which can be easily handled. Naturally a real-world geographic feature, when transferred on a map, is scaled down by a factor called the scale of the map, so that the real-world surface of interest can be accommodated on the choosen sheet of paper. Thus the scale of the map is a very important map feature.
A map contains the following:
- Spatial data of one or more themes.
- dots, lines, polygons and symbols describing the features.
- Captions: Title, Legends or Keys.
- Textual information or attributes of the Geographic features.
- Map projection, Map scale and Tics showing the geographic co-ordinates along the border of the map.
A real-world consists of many information that are represented by maps. A collection of such maps used for preparation of a development plan may be called an analogue system of geographic information. But a GIS is certainly more than that.
2. Geographical Information System
A GIS is a hardware and software for solving complicated problems dealing with spatially referenced data. The modern GIS may be defined as:
An organized collection of computer hardware, software and personnel designed to efficiently capture, store, update, manipulate, analyze and display all forms of geographically referenced information. From experience, it can be emphasised that such an accurate and comprehensive definition of GIS might not help the new comers to GIS a great deal. Its meaning becomes clear when some one learns and works closely with a GIS for solving definite problems.
Many widely used computer programmes, such as spreadsheets (e.g. LOTUS 1-2-3), statistical packages (e.g. SAS) or drafting packages (e.g. Auto-CAD) can handle simple geographic or spatial data. Why, then, are they not usually thought of as a GIS? The generally accepted answer is that a GIS is a GIS if it permits spatial operation on data. The geographic information in a GIS consist of the spatial data and the corresponding tabular data connected by common IDs. A complete data base of a particular feature is called coverage.
The Geographic Information System have the potential for improving our understanding of the world around us and perform the planning and decision making activities on the land use, environmental and resource monitoring and management, etc. More efficiently and intelligently. The measurement, mapping, monitoring and modeling of environmental features and process can be enhanced through the use of a GIS. The geographic information processing begins and ends with the real world. Data is collected from the real world (project area), preprocessed and processed in a GIS for a definite decision making purpose and the results are applied on the project area for the subsequent activities relating to development.
3. COMPONENTS OF GIS
Several components constitute a GIS. They are
- Basic procedures
- A well defined project
- People (Project Manager, GIS system Manager, GIS Experts, User Scientists/Engineers and Maintenance team)
The collection of hardware required for establishing a GIS is shown below:
- Interactive graphics Workstation
- Tape drive
A full GIS implementation may take all the previously gained knowledge of the software, methods and problems and organize them into efficient, easy-to-use techniques and interfaces. It is worth noting that no single component by itself produces a fully-functioning GIS.
4. DATA STRUCTURE IN GIS
Geographic features on a map are represented as: points, lines and polygons. Some times it becomes useful to make a representation of a truly three dimensional surface, such as elevation data set. However, the representation of space in two dimensions, such as administrative boundaries of Bangladesh, are more common.
The contents of the spatial data base is a model of the Earth. Points are used to represent objects whose dimension is too small like tubewells, electric poles, market place, railway stations, etc., lines are used to represent the roads and streams and polygons are any region enclosed by lines (closed surface) that represents uniform features like water areas, agricultural lands, forests.
Besides the geometric data the attribute data characterizing the spatial data are also important.
There are two broadways of organizing the spatial data in a GIS: raster and vector data structures. While any spatial data can be expressed in either raster or vector format, but for the data representing a truely three dimensions surface, the raster format is treated to be suitable one.
4.1 Raster data structure
In raster format, the spatial data is represented by regular cells of uniform shape and size where the attribute value is located at the centre of the cell. The raster data consists of the arrays of such cells, called the picture element or pixel each of which has a position and a value representing the feature. The elements of the rows are called the samples and are numbered from the left to the right. The rows (some times called lines) are numbered from top to bottom. Thus the upper left corner of the raster file is considered as the origin and represents the first sample of the first line (1, 1).
For a raster data base, the accuracy of the measurement is limited by the size of the pixel. The smaller the pixel size, the higher is the spatial accuracy. In this data structure, the points can not be located exactly, it belongs to either one cell or another and there is nothing in between. More-over, a point belonging to any location of the pixel is attached only with the centre of the pixel. More commonly used tessellations of the raster structure are the square cells, where the neighboring cells are not equidistant because the diagonal cells have higher distance. Other tessellations of the raster structures are regular squares, triangles and hexagons. In hexagonal tessellations, the cells are equidistant. However, the use of a hexagonal or even a triangular data structure create two problems. First, the cells can not recursively be subdivided in to smaller cells of the same shape as the original cells, as in the case of square system. Conversely, a hexagon made up of smaller hexagons will not be the same shape as those smaller ones.
The individual cells of the data base does not commonly refer to the real-world coordinate, but may be georeferenced using a ground control point file where the real-world positions corresponding to the same locations at the raster data base are registered. Occasionally a raster data base has a corresponding header file which contains the coordinate of the upper left corner, cell size in X and Y directions, map projection and number of overlays.
The examples of raster data base are: Digital Elevation Model (DEM) and computer classified satellite data.
4.2. Vector data structure
In a GIS having a vector data structures, The X value and Y value of each point is encoded; together these are the (X, Y) coordinate pairs in a two-dimensional rectangular coordinate system. Thus the geographic features: points, lines and polygons are represented as follows:
- Points such as wells, telephone poles, archaeological sites etc. are represented by pairs of X and Y coordinates (X, Y).
- Line features such as streams, streets, etc. are represented by a number of line segments called arcs constituted by a streams of such coordinate pairs. Starting and ending of a line segment or arc is designated as nodes. Each coordinate pair of the arc is called a vertex. The line features are displayed by joining the vertices together.
- Polygon features such as soils, landuse, water-bodies, administrative boundaries, etc. are represented by a stream of vertices enclosing a particular surface. In another word a polygon may be constructed by one or more arcs.
Each of the features in a GIS is identified by a unique ID. Lines and polygons are displayed by joining the vertices.
4.3. Concept of Topology
The spatial data in some of the GIS are topologically linked which makes geographic data intelligent. Topology determines the relationship between the spatial objects. The topological relationships have the following basic characteristics:
- Arcs join at nodes.
- Arcs join to make polygon.
- Arcs have directions (From-node to To-node) and left and right sides.
Topology helps to store data more efficiently, process larger data set, process data faster, combine adjacent polygons with similar characteristics and overlay geographic features. The topology of point, line and polygon coverages are different and each topology are built separately in a GIS like ARC/INFO.
Some of the topological features are given below with particular reference to ARC/INFO - GIS of ESRI.
- Arc-node topology tells which arcs are connected to each other
- Polygon-arc topology tells which arcs make up a polygon Left-right topology tells which polygons are adjacent to the arcs
A GIS coverage consists of geographic features topologically linked and their associated attributes stored in an automated map. A coverage have the following features:
- Tic that links coverage to the real-world coordinates. In another word Tics may be referred as ground control points.
- Arc - Node - Feature IDS - Label Points at which the Label - Ids are positioned - Polygons - Annotations
4.4. Attribute data structure
Feature attribute tables are the dBase data files which have specific items relative to the spatial data base. The spatial data base together with this attribute data base is called a coverage. Every coverage has a feature attribute table which is automatically created when topology is created. The default attribute items are the same within each feature types (points, arcs or polygons) but additional user defined items can be added.
4.5. Layer concept in GIS
For most applications data base of the project area is created with many features and each feature type is stored in a separate coverage. The coverages are geometrically registered and have the same tics and boundaries and are called data layers. Such layers of spatial data describe many geographies of the real-world and allows complicated spatial, logical and arithmetic operations among the data layers to obtain totally new knowledge or develop environment models which were not possible before the development of computer based GIS technology.
4.6. Link between geographic features and attributes
In a GIS the geographic features and attributes are linked. The spatial data and the attribute data are related through the unique ID of the feature. Thus ID is stored in two places - with coordinate data and with attribute data (in attribute table).
An additional attribute data base can be created and linked to the attribute table of the GIS-coverage via a common item.
5. GIS FUNCTIONAL ELEMENTS
There are five functional elements that a GIS must have (based on Knapp, 1978 and Jeffrey Stone and John Estes 1980). These are data acquisition, preprocessing, data management, manipulation and analysis and product generation.
5.1. Data Acquisition
Data acquisition is the process of identifying and gathering of data required for application. This may be done though a number of procedures. One procedure might be to gather new data by preparing maps of required features using satellite observations, aerial-photographs or field observations. The another procedure of data acquisition is to locate and acquire the existing data in the form of maps, aerial or ground photography, surveys reports and documents. A GIS is of no use unless the relevant data has been identified, located and acquired.
The preprocessing consists of the following:
- Check the quality of the source data.
- Data entry: spatial and attribute data
- Digitize a map
- Scan a hard copy document
- Use the key board to enter coordinates
- Buy data in a commercial format (e.g. diskette, tape) and load it in your computer.
- Obtain digital copies from another department or agency.
- Enter attribute data from the key board
- Use existing files already on the computer
- Read existing data stored on tape, cartridge or diskette.
- Error Detection and Editing
- Edge matching
5.3. Data management
Data management consists of storage, delection and retrieval of data. In most larger installations the data is managed by a systems administrator. If the works in the computer is done without managing the data, at one time a message will appear DISK IS FULL. At the end of the day it is always good to delete the unnecessary files without pileing these up and take a complete backup of the work that has already been done. A large volume of data storage is to be well indexed and managed.
The data base management systems (DBMS) with the GIS may be of additional help in managing the attribute data base. A data base management system is the software that permits one or more users to work efficiently with the data. The essential components of the system must provide the means to define the contents of a database, insert new data, delete old data and ask about the contents and modify the contents. The DBMS may be used to create additional data base relative to the spatial data.
5.4. Data manipulation and analysis
The data manipulation and analysis in a GIS mainly consists of the following operations:
- Spatial operation: proximity analysis (buffering around a feature), neighborhood analysis, routing, classification and chloropleth analysis, spatial operations using logical selections based on the attribute data items and modification of the spatial data by any other maens.
- Logical, arithmetic and statistical operations/analysis of the attribute data.
- Modelling using GIS data base.
- Create overlays of coverages with different features. The output coverage contains the spatialand attribute information of the input coverages.
5.5. Product Generation
Finally the results obtained from the GIS processing, manipulation and analysis are presented in the form of maps and tables. The production of the maps are interactively done in the monitor and a plot file is created after the finalization of the map composition. The plot file is sent to the plotter or printer for creating the hard copy output.
6. REMOTE SENSING AND GIS
Remote sensing is the technology that has close link to the geographic information systems. The aerial photographs contain detailed information about the earths resource, environment and landuse which are interpreted for preparation of various thematic maps. These maps are then used as input to the GIS. The imagery from the satellites also produce similar spatial information which can be linked to GIS. In both these cases remote sensing represents a powerful technology for providing the input data for measurement, mapping, monitoring and modelling within GIS contex. In most practical sense, it is believed that neither remote sensing of earths resources nor geographic information systems can reach their full potential unless the two technologies are fundamentally linked.
The remotely sensed data in the digital form are the spatial information in the raster data structure. The image processing systems handle multispectral raster data and process or manipulates the data using various techniques for producing useful thematic information. The image processing and interpretation of these data require additional information (such as elevation, landuse/landcover) to achieve high levels of thematic classification accuracy, which naturally brings us to work in the GIS environment.
The remote sensing is the major source of input data to the GIS. When working simultaneously with an image processing system and a raster GIS it is easy to move data between them. Once the remotely sensed data has been converted to a meaningful data type, transferring this data to raster GIS is relatively simple. Header and trailer records or files may need to be modified during the conversion process.
More work is involved when transferring the raster data derived from remote sensing systems to a vector based GIS. A few image processing softwares provide the facilities of raster to vector conversions and so does a few GIS as well. The US system 600 has the capability of raster to vector and vector to raster conversion reference to some particular GIS systems. The ERIM-GIS of SPARRSO in the raster based one having some capability of vector data handling. A key area in the joint applications of remote sensing technology and geographic information systems is to make historic analysis and identify changes. Remote sensing provide an excellent tool for detecting change, while a GIS is perhaps the best analytic tool for the quantifying the process of change and conduct indepth study of the process as a function of space and time.
7. PROJECT SEQUENCE
For successfully completing a GIS project, it is necessary to follow a sequence of activities which are discussed here. Prior to going into the GIS business it is necessary to formulate the objectives of the GIS project, identify the required data and locate the sources of data. It is also required to prepare a considerably elaborate working plan and proceed step by step following this work plan. All the activities are to be well documented to avoid the confusions. At the end of the day a complete backup of the work is to be taken so that your data is not lost due to any reason. The sequence of GIS activities are described below:
Build the data base
- Create work space for the project
- Design the data base structure
- Input spatial data (digitize or online ingest)
- Edit and create topology
- Input attribute data
Manage and manipulate the data
- Transfer the data from digitize coordinate to real world coordinate
- Join the coverages created from the series of mapsheets covering the study area.
- Create overlays of individual coverages
- Clip the areas of interests for analysis is required based on clip-coverage
Analyzing the data
The analysis of the data is done according to the requirement of the project. Depending on the objectives, a processing and analysis plan is to be worked out. This will make the job easier. Sometimes it is recommended to write a programme or a number of programs using command language to run it sequentially to complete the processing works. The analysis comprises of both spatial and statistical manipulations.
Presentation of the results of analysis
The final step of the GIS project is the presentation of the results of analysis. Generally, this consists of production of high quality maps and tables containing the output results of the GIS. The hard-copy of these products are taken for ultimate use and preparation of the report.
It may be concluded that GIS is an important modern tool that handles spatially referenced data for monitoring, management, modelling and decision making purpose. The function of GIS in an organization is manifold. Prior to setting up a GIS for a particular objective, it is necessary to gather the experience of other people and organizations who have already installed and used GIS. The major problem in setting up and operate a GIS in Bangladesh comes from the scantiness of the reliable data and lack of hardware maintenance facilities in the country.
There are many different GIS presently being used in Bangladesh. In SPARRSO we have mainframe and PC ARC/INFO, IDRISI, and ERIM-GIS. Besides, there are a number of image processing systems US operating microvax 3400 and VAX 11/750 computer and a PC-Erdas. SPARRSO has a data base of a good numbers of remote sensing imagery in the CCTs as well as negative hard-copy films which are used for different environmental and resource applications and also act as valuable input to the GIS.
For avoiding duplication of digitization there should be a good up-to-date inventory of the GIS data base created at different departments in the country so that the data may be exchanged among the departments. For this purpose, the a standardization of data from the various sources and data bases is to be setup and the convertibility of data in different formats are to be worked-out to facilitate the GIS applications in the country.
9. Bibliographic references
Calkins H. W., and R. P. Tomlinson 1977 Geographic Information Systems: Methods and Equivalent for Land Use Planning. International Geographic Union Commission on Geographical Data Sensing and Processing. Resource and Land Investigation (RALI) Program, U.S. Geological Survey, Reston, Virginia.
Chowdhury A. M. and D. A. Quadir 1993 Data-base Management and Geographic Information System (GIS): Monitoring Adjustment and Poverty in Bangladesh (CIRDAP Project), Report submitted to CIRDAP, Bangladesh.
ESRI 1991 Understanding of GIS: ARC/INFO Method. Environmental System Research Institute, Inc. Redlands, CA, USA.
Meaden G. J. and J. M. Kapetsky 1991 Geographical Information Systems and Remote Sensing in Inland Fisheries, Technical Paper-318, Food and Agricultural Organization of the United Nations.
Quadir D. A. (coordinator), M. A. Rahman, M. J. Islam, M. Nessa, I. U. Ahmed, N. Nahar and M. Hossain 1993, Monitoring of Forest Cover in Chittagong Forest Division, A Report of the Sectoral Study K under Service Oriented Application of Remote Sensing Technology in Agriculture, Water Resources, Fisheries and Forestry Sectors UNDP Project No. BGD/85/031
Star J. and J. Ester 1990 Geographic Information System: An Introduction, Prentice Hall, Englewood Cliffs, New Jersy.