A. BASIC GIS CONCEPTS
B. USE OF GEOGRAPHIC INFORMATION SYSTEMS IN NATURAL HAZARD ASSESSMENTS AND INTEGRATED DEVELOPMENT PLANNING
C. GUIDELINES FOR PREPARING A GIS
This chapter introduces the planner to the concept and applications of geographic information systems (GIS) for natural hazard management in the context of integrated development planning. The chapter also discusses the acquisition of a GIS, the elements involved in making that decision, and the basics of how to select, establish, and use such a system
Natural events such as earthquakes and hurricanes can be hazardous to man. The disasters that natural hazards can cause are largely the result of actions by man that increase vulnerability, or lack of action to anticipate and mitigate the potential damage of these events. Previous chapters make clear that this book does more than describe hazards; it deals with how this information can be incorporated into development planning to reduce the impact of natural hazards. Planners are familiar with the bewildering array of disparate pieces of information that have to be analyzed and evaluated in the planning process. The process is complicated, however, by entirely new data sets on assessment of various natural hazards, separately and in combination, and by the need to analyze these hazards with respect to existing and planned development, choose among means of mitigating the damage the hazards can cause, carry out an economic analysis of the alternatives of mitigation versus no mitigation, and determine the impact of these alternatives on the economic and financial feasibility of the project.
Along with these added complications come techniques for managing the information so that it does not overwhelm the planner. Among these are geographic information systems (GIS), a systematic means of geographically referencing a number of "layers" of information to facilitate the overlaying, quantification, and synthesis of data in order to orient decisions.
This chapter demonstrates the effectiveness of geographic information systems, specifically personal-computer-based systems, as a tool for natural hazard management in the context of integrated development planning. The chapter is directed towards two different audiences. To planners it shows the utility of the tool by giving a number of practical examples of applications extracted directly from planners' experiences. To the decision-makers of planning agencies it sends the message that if their agency does not now have access to a GIS, they should certainly be thinking about it. Technical subordinates should find here the wherewithal to present the appropriate arguments for GIS use to uninformed decision-makers.
There are a number of reasons why planning agencies in Latin American and Caribbean countries would benefit from a GIS:
- It can be surprisingly cheap; very expensive equipment and highly specialized technicians can be avoided by proper selection of a system and its application. The main constraint may not be lack of funds but lack of appropriate personnel and equipment;
- It can multiply the productivity of a technician; and
- It can give higher quality results than can be obtained manually, regardless of the costs involved. It can facilitate decision-making and improve coordination among agencies when efficiency is at a premium.
On the assumption that some readers are unfamiliar with GIS, the chapter first reviews some basic concepts covering the operations, functions, and elements of a system. Next come a number of examples of applications for natural hazard management at the national, subnational, and local levels, to help the reader evaluate the benefits and limitations of a GIS. A three-step process is presented for reaching the decision to acquire or upgrade GIS capability: (1) a needs assessment, determining the agency's GIS applications and objectives and those of possible joint users; (2) analysis of the costs and benefits of the acquisition; (3) brief guidelines for the selection of appropriate hardware and software combinations. The chapter ends with a short discussion of how to set up a system.
The chapter does not attempt to replace the many technical manuals on how to select and operate a GIS. Once the agency has decided to consider the acquisition of a system, it will require more specific guidance in the form of supplementary literature and/or technical assistance.
1. What is a GIS?
2. GIS Operations and Functions
3. Elements of a GIS
The concept of geographic information systems (GIS) is not new. It was first applied conceptually when maps on the same topic made on different dates were viewed together to identify changes. Similarly, when maps showing different kinds of information for the same area were overlaid to determine relationships, the concept of GIS was actually in use. What is new and progressing rapidly is advancing computer technology, which allows the low-cost examination of large areas frequently, and with an increasing amount of data. Digitization, manipulation of information, interpretation, and map reproduction are all steps in generating a GIS that now can be achieved rapidly, almost in real time.
The concept of a GIS is basically analogous to a very large panel made up of similarly shaped open boxes, with each box representing a specified area on the earth's surface. As each element of information about a particular attribute (soil, rainfall, population) that applies to the area is identified, it can be placed into the corresponding box. Since there is theoretically no limit to the amount of information that can be entered into each box, very large volumes of data can be compiled in an orderly manner. After assigning relatively few attributes to the box system, it becomes obvious that a collection of mapped information has been generated and can be overlaid to reveal spatial relationships between the different attributes, i.e., hazardous events, natural resources, and socio-economic phenomena (see Figure 5-1).
There are many kinds of GIS, some more suitable for integrated development planning studies and natural hazard management than others. At the most elementary level, there are simple manual overlay techniques, such as the one proposed by McHarg in Design with Nature, which have proven to be very valuable tools. However, the information needed for hazard management and development planning can become so overwhelming that it is almost impossible to cope with it manually. At the other extreme are highly sophisticated computerized systems that can analyze baseline scientific data such as satellite imagery and can produce, by using plotters, large-scale maps of excellent cartographic quality. Such systems are very expensive, difficult to operate, and may exceed the needs of many planning offices.
Figure 5-1. OVERLAY CHARACTERISTICS OF A GIS
Among computerized GIS, PC-based GIS are most affordable and relatively simple to operate, capable of generating maps of varying scales and tabular information suitable for repeated analysis, project design, and decision-making. Even though PC-based GIS may not produce maps of cartographic quality or sufficient detail for engineering design, they are most viable for planning teams analyzing natural hazard issues in integrated development projects.
Data manipulated by a computer-based GIS are arranged in one of two ways: by raster or by vector. The raster model uses grid cells to reference and store information. An area for study is divided into a grid or matrix of square (sometimes rectangular) cells identical in size, and information-attributes represented as sets of numbers-is stored in each cell for each layer or attribute of the database. A cell can display either the dominant feature found in that cell or the percentage distribution of all attributes found in the same cell. Raster-based systems define spatial relationships between variables more clearly than their vector-based counterparts, but the coarser resolution caused by using a cell structure reduces spatial accuracy.
Vector data are a closer translation of the original map. These systems reference all information as points, lines or polygons, and assign a unique set of X, Y coordinates to each attribute. Usually, vector system software programs have the capability to enlarge a small portion of a map to show greater detail or to reduce an area and show it in the regional context. Vector data can offer a larger number of possible overlay inputs or layers of data with greater ease. The vector model does represent the mapped areas more accurately than a raster system, but because each layer is defined uniquely, analyzing information from different layers is considerably more difficult.
The choice of raster or vector-based GIS depends on the user's needs. Vector systems, however, demand highly skilled operators and may also require more time and more expensive equipment, particularly for output procedures. Vector-based GIS software is also much more complex than that for the raster system and should be checked for performance in all cases. It is up to the planner or decision-maker to choose what system is most appropriate.
a. Data Input
b. Data Storage
c. Data Manipulation and Processing
d. Data Output
Data input covers the range of operations by which spatial data from maps, remote sensors, and other sources are transformed into a digital format. Among the different devices commonly used for this operation are keyboards, digitizers, scanners, CCTS, and interactive terminals or visual display units (VDU). Given its relatively low cost, efficiency, and ease of operation, digitizing constitutes the best data input option for development planning purposes.
Two different types of data must be entered into the GIS: geographic references and attributes. Geographic reference data are the coordinates (either in terms of latitude and longitude or columns and rows) which give the location of the information being entered. Attribute data associate a numerical code to each cell or set of coordinates and for each variable, either to represent actual values (e.g., 200 mm of precipitation, 1,250 meters elevation) or to connote categorical data types (land uses, vegetation type, etc.). Data input routines, whether through manual keyboard entry, digitizing, or scanning, require a considerable amount of time.
Data storage refers to the way in which spatial data are structured and organized within the GIS according to their location, interrelationship, and attribute design. Computers permit large amounts of data to be stored, either on the computer's hard disk or in portable diskettes.
Data manipulation and processing are performed to obtain useful information from data previously entered into the system. Data manipulation embraces two types of operations: (1) operations needed to remove errors and update current data sets (editing); and (2) operations using analytical techniques to answer specific questions formulated by the user. The manipulation process can range from the simple overlay of two or more maps to a complex extraction of disparate pieces of information from a wide variety of sources.
Data output refers to the display or presentation of data employing commonly used output formats that include maps, graphs, reports, tables, and charts, either as a hard-copy, as an image on the screen, or as a text file that can be carried into other software programs for further analysis.
a. Hardware and Software Components
b. Users and Users' Needs
c. Information and Information Sources
Hardware components of a basic GIS work station consist of: (1) a central processing unit (CPU) where all operations are performed; (2) a digitizer, which consists of a tablet or table where analog data are converted to digital format; (3) a keyboard by which instructions and commands as well as data can be entered; (4) a printer or plotter to produce hard copies of the desired output; (5) a disk drive or tape drive used to store data and programs, for reading in data and for communicating with other systems; and (6) a visual display unit (VDU) or monitor where information is interactively displayed. Several GIS software packages are available representing a very broad range of cost and capability. The selection of the appropriate combination of hardware and GIS software components to match the user's needs is discussed in Section C.
Planners need to carefully evaluate their GIS needs and proposed applications before taking the decision to acquire an install a GIS. Once a positive conclusion has been reached, its hardware-software configuration should be designed based on those needs and applications, and within the constraints posed by the financial and human resources available to operate the system.
It is possible that the costs of establishing a GIS exceed the benefits to a single agency. Under these circumstances, it is worthwhile determining if several agencies might share the GIS. Appendix A gives a list of users of natural hazard data. The potential users must agree on the data to be compiled, the data formats, standards of accuracy, etc. As a result, the data requirements of a variety of users are made compatible, and the value of the data increases commensurately.
Sharing information has its costs as well as benefits. Negotiating with other users can be a painful task, and compromises inevitably ensure that no one user will get the equipment most precisely suited to his uses. In this regard, it is important to establish a comfortable working relationship among sharers.
General reference maps and information on natural hazards and natural resources should form a "library of knowledge" for any GIS. Most areas of Latin America and the Caribbean have general background sources of such data. Virtually all countries have topographic maps, road maps, generalized soils maps, some form of climate information, and at least the locational component of natural hazards information (e.g. location of active volcanoes, fault lines, potential flood areas, areas of common occurrence of landslides, areas of past tsunami occurrence, etc.). Natural hazards locational data can be made compatible in a GIS with previously collected information about natural resources, population, and infrastructure, to provide planners with the wherewithal for a preliminary evaluation of the possible impacts of natural events.
Even though some of this information is available in almost every country and can be supplemented with satellite data, the question remains, are there enough data to justify a GIS? The principal value of the GIS is in processing and analyzing masses of data that have become overwhelming for manual handling. In determining the applicability of a GIS, an agency must decide if it is data handling or merely the lack of data that is the main obstacle to hazard management.
1. GIS Applications at the National Level
2. GIS Applications at the Subnational Level
3. GIS Applications at the Local Level
4. Use of a Geo-referenced Database
GIS applications in natural hazard management and development planning are limited only by the amount of information available and by the imagination of the analyst. Readily available information on natural events (e.g., previous disaster records), scientific research (papers, articles, newsletters, etc.), and hazard mapping (seismic fault and volcano location, floodplains, erosion patterns, etc.) are usually enough to conduct a GIS preliminary evaluation of the natural hazard situation and guide development planning activities. (See Chapters 4 through 12 and Appendix A for sources of information.)
At the national level, GIS can be used to provide general familiarization with the study area, giving the planner a reference to the overall hazard situation and helping to identify areas that need further studies to assess the effect of natural hazards on natural resource management and development potential. Similarly, GIS can be used in hazard assessments at the subnational level for resource analysis and project identification. At the local level, planners can use a GIS to formulate investment projects and establish specific mitigation strategies for disaster prevention activities. The following examples of OAS applications are intended to demonstrate the versatility of the tool and suggest to planners applications that may fit their agencies' needs.
GIS APPLICATIONS AT A NATIONAL LEVEL
Sector Vulnerability Assessment
The managers of public and private sectoral agencies share a concern about the vulnerability of their sectors to hazardous events; Where are the weak links? Where might damage occur? What is the Impact of losing x service in y city for z days? What mitigation investment would resolve that problem? What is the cost-benefit of that investment? As an example, in 1989 the Costa Rica Sectoral Energy Directorate (DSE) requested OAS to assist in the analysis of the vulnerability of the energy sector to natural hazards. The study was conducted using two approaches: (1) field examinations and/or interviews with energy sector personnel; and (2) use of a GIS to overlay energy subsector Infrastructure information with selected hazard data.
The GIS exercise, confirmed by results obtained by field observations, dearly showed the potential disruption of Important segments of main transmission lines by landslides, and pointed out critical areas where mitigation or hazard reduction activities Should be undertaken (see Figure 1-5). Although GIS analyses were not done for all hazards and sub-sectors, it became obvious that the outcome would have been virtually the same as the field examination's results for coarse-grained hazards such as earthquakes, hurricanes, and droughts, but less accurate for fine-grained hazards such as flooding in narrow river valleys. It was believed that if hazard data had been available at a 1:50,000 scale, the GIS would have given the same results for all hazards (although additional time would have been required to input the data).
While not intended to supplant field observations, the GIS approach had nevertheless some striking advantages in technician time, especially for this case in which only existing information was used. In addition, the GIS also provided full color maps showing the potential impact of landslide events on the electricity sub-sector, which were very useful In explaining the results and mobilizing follow-up action
Use of a GIS to combine information on natural hazards, natural resources, population, and infrastructure can help planners identify less hazard-prone areas most apt for development activities, areas where further hazard evaluations are required, and areas where mitigation strategies should be prioritized. A seismic hazard map for example, even at this level, can give planners the location and extent of areas where heavy capital investments should be avoided and/or areas where activities less susceptible to earthquakes, tsunamis, or volcanoes should be considered.
Similarly, in hazard-prone areas, use of a GIS to overlay hazard information with socio-economic or infrastructure data can reveal the number of people or type of infrastructure at risk. This sort of exercise was done in 1989 by OAS/DRDE, for several OAS member states. It was shown, for example, that in Peru more than 15 million people were living in earthquake-prone areas with a seismic intensity potential of VI or greater, that close to 930,000 people were potentially at risk of a tsunami wave height of 5 meters or more, and that 650,000 people were living within a 30 km radius of active volcanoes. Overlaid with infrastructure information, this same kind of analysis identified lifelines or vital resources in high-risk areas, and with adequate sectoral information, it can be further expanded to calculate potential losses in capital investment, employment, income stream, and foreign exchange earnings.
Little time was necessary to produce the maps: two days were required to code, digitize, and edit the maps, and only minutes were necessary to do the analysis. Moreover, with the information in the system, additional requests or changes in parameters (e.g., a 40 instead of a 30 km radius around a volcano) can be processed in a few minutes, while an entirely new set of drawings and calculations would be required if manual techniques were used. Figure 5-2 gives some examples of applications of GIS at the national and subnational levels.
At a subnational level of planning, GIS technology can be used for natural hazard assessments to show where hazardous natural phenomena are likely to occur. This, combined with information on natural resources, population, and infrastructure, can enable planners to assess the risk posed by natural hazards and to identify critical elements in high-risk areas. This information can then be used to formulate less-vulnerable development activities and/or mitigation strategies to lessen vulnerability to acceptable levels.
Figure 5-2 - EXAMPLES OF GIS APPLICATIONS FOR NATURAL HAZARD MANAGEMENT AT THE NATIONAL AND SUBNATIONAL LEVEL OF PLANNING
||Information source, data display||Considering land form, slope, land use, vegetation cover, and wind direction, what area is likely to be affected if this volcano erupts? How many people could be affected?|
|Index of information||List all available hospitals located not within 30 km radius of the volcano|
|Status reporting||Periodic assessment of volcanic activity|
|Monitoring change||How has the savanna desert boundary changed in the last 5 years? What changes in climate and land use could account for the on-going desertification process?|
||Research support||What factors determine landslide activity in this area? According to these factors, what zones are susceptible to landslides?|
|Forecasting||What population centers are likely to be affected by this hurricane? What is the most likely lava flow path in case a volcanic eruption occurs?|
|Policy development||What areas in this growing urban region should be restricted to low-density development?|
|Aid allocation||Where should mitigation strategies be prioritized?|
|Project evaluation||If erosion trends continue, what will be the economic impact on the project? What are the costs and benefits of instituting or not instituting erosion control measures?|
Source: Adapted from United Nations Environmental Program (UNEP). GRID (1985).
GIS APPLICATIONS AT A SUBNATIONAL LEVEL
Urban Expansion in an Area Prone to Landslides
Tegucigalpa, capital of Honduras, is a hill/city on geologically unstable ground, constantly suffering damaging landslides. In 1987, an OAS/DRDE study identified more than 300 landslides occupying an area of approximately 1,350 ha within the metropolitan area and determined that 20 percent of that area presented high to extreme landslide hazard susceptibility. The situation has since been aggravated by increasing rural migration, frequently occupying steep areas of questionable stability. City officials had two urgent tasks: identify landslide-hazard free urban expansion areas for new settlement and resettlement programs, and delineate priority areas where hazard mitigation should be considered.
By entering data on land use, landslide hazard susceptibility, topography, slope, and protected areas, a GIS database was created to identify areas potentially suitable for expansion. City officials could then set minimum criteria for areas of new development (i.e., no more than 5 percent of the area can be subject to landslide hazard, no access road can be located within 1000 ft, of a 20 percent slope, etc.). Using the GIS, areas meeting the criteria could be identified. The number of people living in extreme and high landslide hazard areas could also be determined, providing the basis for selecting priority areas for implementing prevention measures (relocation, construction, retrofitting, etc.).
For this exercise, the advantages of using GIS as compared to manual mapping techniques are obvious. Not only does GIS afford great time savings (for the overlay, display, assessment, and analysis of hazardous areas), but GIS also offers flexibility in selecting the minimum standards. Tentatively selected standards can be tested for feasibility and adjusted. Using a GIS, this process would take minutes, while with manual methods, it would take a week of redrafting and recalculations
In a landslide study for example, data on slope steepness, rock composition, hydrology, and other factors can be combined with data on past landslides to determine the conditions under which landslides are likely to occur (see Chapter 10). To analyze all possible combinations with manual techniques is a virtually impossible task; thus, typically only two factors are analyzed, and the composite units are combined with the landslide inventory map. With GIS, however, it is possible to analyze an almost unlimited number of factors associated with historical events and present conditions, including present land use, presence of infrastructure, etc. OAS/DRDE has used this technology to overlay maps of geology, slope steepness, slope orientation, hydrology, and vegetation, and then overlaid the results with a landslide inventory map to identify the factors associated with past and present landslides. The resultant landslide hazard zonation map provides planners with a designation of the degree of landslide propensity for any given area.
For floods, GIS and remotely-sensed data can be used to identify flood-prone areas, map floods in progress, delineate past floods, and predict future ones (see Chapters 4 and 8). GIS can combine information on slope, precipitation regimes, and river carrying capacity to model flood levels. Synthesis information obtained from such an integrated study can help planners and decision-makers determine where to construct a dam or reservoir in order to control flooding.
Likewise, a map depicting volcano locations may be entered into the GIS; volcano attributes such as periodicity, explosivity index (VEI), past effects, and other attributes may be ascribed to each volcano record in a relational database. Combining these data with information on human settlements or population density, land use, slope, presence of natural barriers, and other natural resource or socio economic data, the GIS can generate maps and/or tabular reports depicting hazard-free areas (e.g., areas outside a certain radius or impact area of an active volcano, areas with less than 25% slope and high vegetation cover, etc.). Finally, information on other hazards can be combined to create new sub-sets of data, each one complying with different pre-established minimum standards for development.
At this level, GIS can be used in prefeasibility and feasibility sectoral project studies and natural resource management activities to help planners identify specific mitigation measures for high-risk investment projects and locate vulnerable critical facilities for the implementation of emergency preparedness and response activities. In population centers, for example, large scale GIS databases (resolutions of 100 m2 per cell or less) can display the location of high-rise buildings, hospitals, police stations, shelters, fire stations, and other lifeline elements. By combining these data with the hazards assessment map-previously collected or generated through GIS-planners can identify critical resources in high-risk areas and adequately formulate mitigation strategies. (See Figure 5-3).
GIS APPLICATIONS AT THE LOCAL LEVEL
Farmer Settlement and Resettlement Planning in an Erosion Vulnerable Area
Land settlement projects commonly involve multiple and complex objectives. When equitable distribution of land is defined in terms of earning capacity instead of parcel size, land capability and management practices have to be factored into the equation. of natural hazards are also considered, as they must be if the project is to be sustainable and equitable in the long term, the number of factors becomes too unwieldy for manual analysis. In 1985, an OAS study prepared GIS data flies for the Mabouya Valley Project, located in east-central Saint Lucia. The project, which involved settling a large number of farmers on former plantation land subject to erosion, sought to identify current land uses in conflict with land capability and/or erosion risks in order to improve the management of some parcels, resettle remaining farmers on redesigned parcels, and improve the equity of land distribution.
Eight maps were coded into the system: Ecology, Human Settlements, Land Capability, Life Zones, Water Resources, Erosion Risk, Present Land Use and Vegetation, and a proposed development strategy. Three synthesis maps were produced by overlaying present land use with land capability, present land use with erosion risk, and the development strategy with erosion risk.
The GIS exercise showed that large commercial parcels occupied 76 percent of ail the lands suitable for unrestricted or moderately restricted cultivation, while 99 percent of the land occupied by small farms was classified as severely restricted or worse. When compared with severe erosion hazard, the synthesis maps showed that 2% of the area devoted to commercial agriculture vs. 30 percent of the area of the small mixed farms was affected.
This modest GIS exercise, using readily available information, constituted a tiny fraction of the overall study, yet it demonstrated clearly that land redistribution would be required to achieve the project's objective of equitable distribution, and provided the data required to guide the redistribution and to introduce improved soil management practices.
The decision on the type of information to be used for depicting the variables included in the database-whether real-scaled or symbolic dimensions-becomes a critical decision at this level. Real-scaled data should prevail over symbolic information, especially at this level of planning, when precise information is required to assess the risk posed to specific investment projects. Floodplain elevations, for example, represented in scales smaller than 1:50,000, will show only approximate location. Any GIS calculations or operations that include cell measurements (area, perimeter, distance, etc.) need to be accurate enough to provide planners with a clear and precise illustration of the overall and project-to specific hazard situation of the study area. Floodplain hazard assessments combine thematic maps (e.g., soils, geology, topography, population, infrastructure, etc.) and need a precise cell representation of floodplain elevation in order to indicate where the probable flood areas are and what are the probable population, natural resource and infrastructure components that might be affected by a flood event. Figure 5-4 gives examples of GIS applications undertaken by the OAS/DRDE.
A geo-referenced database (GRDB) is a microcomputer-based program that combines data management with map display, allowing planners and emergency managers to graphically display hazard impact areas, and relate them to people and property at risk. Although a GRDB also uses points, lines, and polygonal symbols to represent data, it differs from a GIS in that it does not have overlaying capabilities. However, GRDB's ability to manage and combine large databases with map display, text relating displayed elements (hazard impact areas, location of shelters, health centers, fire stations, police stations, etc.) to their respective descriptive information, makes it suitable for emergency planning and post-disaster rehabilitation and reconstruction work.
Figure 5-3. EXAMPLES OF GIS APPLICATIONS FOR NATURAL HAZARDS MANAGEMENT AT THE LOCAL LEVEL OF PLANNING
||- Aid in the analysis of spatial distribution of socio-economic infrastructure and natural hazard phenomena||- What lifeline elements lie in high-risk areas?|
|- Use of thematic maps to enhance reports and/or presentations||- What population could be affected?|
|- Link with other databases for more specific information||- Where are the closest hospitals or relief centers in case of an event?|
|Land Information Storage and Retrieval
||- Filing, maintaining, and updating land-related
data (land ownership, previous records of natural events, permissible uses, etc.)
||- Display all parcels that have had flood problems in the past|
|- Display all non-conforming uses in this residential area|
|Zone and District Management
||- Maintain and update district maps, such as zoning maps or floodplain maps||- List the names of all parcel owners of areas within 30 m of a river or fault line|
|- Determine and enforce adequate land-use regulation and building codes||- What parcels lie in high and extreme landslide hazard areas?|
|Site Selection||- Identification of potential sites for particular uses||- Where are the hazard-free vacant parcels of at least x ha lying at least y in from a major road, which have at least z bed-hospitals within 10 km radius?|
|Hazard Impact Assessment||- Identification of geographically determined hazard impacts||- What units of this residential area will be affected by a 20-year flood?|
|Development/Land Suitability Modelling||- Analysis of the suitability of particular parcels for development||- Considering slope, soil type, altitude, drainage, and proximity to development, what areas are more likely to be prioritized for development? What potential problems could arise?|
Source: Adapted from Levine J., and Landis, J. "Geographic Information Systems for Local Planning" in Journal of the American Planning Association (Spring, 1989), pp. 209-220.
Figure 5-4 - ORGANIZATION OF AMERICAN STATES/DEPARTMENT OF REGIONAL DEVELOPMENT AND ENVIRONMENT EXAMPLES OF APPLICATIONS OF GIS IN HAZARD ASSESSMENT AND DEVELOPMENT PLANNING
|Colombia Puerto Bogota, Department of Cundinamarca||1,3,000 (16.81 m2 per cell)||Identification of hazard-free urban areas suitable for the relocation of 34 families presently under high landslide risk.||- Base map
- Urban perimeter map
- Urban census map
- Geologic map
- Natural hazards map
- Risk zones map
- Land use information
- Population density
|Identification of possible relocation sites for 34 families. Sites in question had to comply with the following requirements: not in hazardous zone, 100m. away from the river, within urban limits, and in unoccupied areas or with low population density.|
|Ecuador: Agricultural Sector Vulnerability Study||1:2,000,000 (1 km per cell)||Determination of the vulnerability of the agricultural sector in terms of income, employment, foreign exchange earnings, and food possible mitigation strategies.||- Political map
- Road network and storage facilities map
- Flood, erosion, drought, landslide, seismic, and volcanic hazard map
- Crop producing areas (26 cultivation systems)
- Socio economic data
|49 possible critical events selected for further study and/or profile level mitigation strategy formulation. Follow-up institutional support delineated.|
|Honduras: Jesús de Otoro Valley, Department of Intibuca||1:50,000 (2.08 ha per (2.08 ha per||Identification of flood- and erosion-prone areas for the selection of agricultural production projects.||- Present land use
- Projected land use
- Human settlements
|66 percent of the land presently occupied or planned for irrigated agriculture investment was found to be in flood-susceptible areas.|
|Paraguay: Southwestern section of the Paraguayan Chaco||1:500,000 (208 ha per cell)||Identification of hazardous areas for landuse capability definition and agricultural project selection.||- Soils map
- Forest tipology
- Alternative forest uses
- Agricultural zones
- Landuse capability
|Identification and quantification of areas under varying degrees of limitations or restrictions in areas previously recognized as best suited for their respective production|
|Saint Lucia Mabouya Valley Project||1:10,000 (2.1 ha per cell)||Identification of current and proposed land uses in conflict with land capabilities and/or erosion risks; selection and distribution of farming resettlement sites.||- Human settlements
- Land capability
- Present land use
- Erosion risk
- Water resources
- Life zones
- Development strategy
|99 percent of the land occupied by small farms was classified as severely restricted or unsuited for cultivation. 2 percent of the land for commercial agriculture vs. 30 percent of the land for small farms was affected by severe or critical erosion hazard.|
USING A GEO-REFERENCED DATABASE IN THE AFTERMATH OF A DISASTER
After a disaster, quick response in analyzing the situation and formulating a workable rehabilitation program is essential. In 1988, after Hurricane Gilbert struck Jamaica, the Government was faced with the overwhelming task of allocating a great variety of relief resources to agencies and population, and of coordinating the rehabilitation effort among all the institutions and agencies involved. At the Government's request, the OAS helped to install a geo-referenced database system to organize the compilation and analysis of damage assessment records which would then be used to help manage rehabilitation and reconstruction efforts.
Initial configuration of the system consisted of eight computer maps, ranging from 1:1 million (entire country) to 1:44,000 (enlarged Kingston area) scale maps, with main road network and individual records for each town and settlement. It took a team of three people four days to build the database and train its users. The system was put into service immediately, providing the basis for coordination among all the agencies participating in the emergency relief program.
Afterward, the system was expanded to include critical facilities location (health centers, shelters, police, fire) and lifeline networks (water and electricity) for the Kingston area. With assistance from the United Nations Development Programme (UNDP), eleven more systems were installed in key government departments directly involved in relief distribution and reconstruction. Direct phone and radio links were also installed between all systems to allow easy consultation and information interchange. Since then the base map has expanded to more than 130 maps covering the entire country at a scale of 't:50,000» with larger scales for population centers and key economic zones.
Although it will take some time to quantify the benefits of this system, It is clear that Jamaica now possesses a powerful information system which can be used not only as a decision-support system for emergency management offices, but also as a planning tool that can assist government agencies to better plan and coordinate development planning and emergency preparedness and response activities.
Through a GRDB, information can be accessed for data update and utilization by all involved agencies. In this way, emergency management offices can have almost immediate access to an updated inventory of settlements, lifelines, hazard impact areas, and special emergency needs, facilitating inventory and deployment of emergency resources; sectoral ministries and utility companies can prepare more effective plans and projects by having access to updated population and infrastructure data; and central planning agencies can use the system as a tool for reconstruction planning coordination.
This kind of system was used in Jamaica after Hurricane Gilbert as a mechanism for coordinating disaster relief (see box above), and in Costa Rica, the Ministry of Natural Resources and Mines requested the OAS to provide a GRDB to monitor the vulnerability of the country's energy infrastructure to natural events. Although there are clear benefits in using a GRDB in emergency management, its transformation as a tool in development planning will need time, cooperation, and support from all agencies involved.
1. Conduct a Needs Assessment, Define Proposed Applications and Objectives
2. Execute an Economic Analysis for GIS Acquisition
3. Select among Alternative Systems and Equipment
4. Establish a Database
Benefits of a GIS may be so compelling that the decision to acquire a system can be made with little hesitation. In most cases, however, the decision can only be reached after a thorough analysis. The following section introduces a systematic process for reaching a decision about acquiring a GIS. Potential users must remember that a GIS is not always the right tool for a given situation, and it may not necessarily pay for itself.
Before deciding to acquire or use a system, planners need to make a meticulous evaluation of their GIS needs. This must include a definition of how their planning activities and decisions will be assisted by using a GIS. Specific objectives and applications of the GIS should be defined. Answers to the questions outlined in the box below can help.
QUESTIONS PLANNERS NEED TO FORMULATE TO ASSESS THE NEED OF A GIS
- What planning decisions need to be made?
- Which decisions involve the use of mapped information and information susceptible to map display?
- What information cannot be managed efficiently with manual techniques?
- What information management activities will be supported by the proposed GIS?
- What are the number and types of decisions that will be supported with a GIS?
Is the GIS principally for analysis? Is cartographic quality output needed?
- To what extent will a GIS help achieve the desired objectives?
- Who will be the users of the information generated with a GIS? How many user groups will there be? In terms of information, time, and training needs, what is required to obtain the desired results? Is there budget and staff support? What agencies are participating in similar projects? To what extent would a GIS help to attract the interest of other agencies and facilitate cooperation?
QUESTIONS THAT HELP IN EVALUATING THE SUITABILITY OF AN AVAILABLE GIS
- What kind of system is it?
- What hardware and software are used?
- Are its capabilities compatible with the needs of the new users?
- Is the in-house technical expertise capable of serving the new users?
- What are the institutional arrangements that would allow the use of this GIS?
- Who are the current users? To what extent is the current user network compatible with the network envisioned?
- What data does it contain? To what extent does the data presently in the system cover identified needs?
If this preliminary investigation indicates that obtaining and using a GIS is a good option for an agency, it should seek the most cost-effective method of doing so. A frequently neglected option is to determine if an existing system is available. If the existing GIS is underutilized, the current owner might find a time-share offer attractive, particularly if the new agency brings data and analyses to the partnership. If no suitable GIS exists, another alternative is for a group of agencies to establish a GIS that meets their common needs. Obviously, the trade-off in both these options is lower cost vs. independence of action, but if the partnership also brings improved working relationships and compatible data to a group of agencies that work on common problems, these benefits may exceed the independence cost. The questions in the box above offer planners some guidance as to whether an existing system is suitable to their needs.
Another opportunity for reducing investment cost is the use of existing equipment. If a computer is available, is it compatible with the GIS envisioned? What are the economic and institutional costs of timesharing and inconvenience?
KEY ELEMENTS NEEDED WHEN PLANNING A GIS ACQUISITION COST CALCULATIONS:
- What is the software purchase cost?
- What hardware configuration is needed to fit the software requirements?
- Is a new computer needed? What options have to be included? What is the cost of acquiring a new computer versus upgrading an existing one?
- What are the anticipated hardware repair and maintenance, and software support costs?
- What are the personnel requirements for the installation and operation of a GIS?
- Will existing personnel be used or will new personnel have to be hired? Is a computer programmer needed? What training costs are anticipated?
- What is the cost of allocating personnel to hardware and software maintenance?
- What is the expected cost for the data input process? How many staff need to be hired or assigned to digitize the information? What is the cost involved in maintaining the data generated for and by the system?
- Is a secure facility available suitably equipped for protection of computers and data files?
- What are the production or revenue losses mostly associated with lack of information? How does this compare with the information that would be available if a GIS were present?
- What are the costs savings from substituting labor-intensive d rafting processes with a GIS?
- What are the benefits of integrating more timely information in the decision-making process, and of being able to perform sensitivity analysis on proposed development plan options?
Once an agency has reached tentative decisions to acquire GIS capability, alone or in partnership, it should undertake an economic analysis of the proposition.
Acquiring a GIS system is a capital investment that may represent several thousand U.S. dollars. As contended by Sullivan (1985), standard investment appraisal methods can be applicable to information technologies such as GIS. The questions in the box above will help planners to roughly estimate and compare the major cost and benefits associated with a GIS acquisition.
The cost of maintenance and repair of all components of a GIS must also be considered in the investment analysis. The more sophisticated the system, and the more remote the home base of operation, the higher its maintenance cost. Software demands maintenance too, and arrangements should be made to subscribe to effective support from the provider of the software. The hiring of expertise to modify the software according to the project should be expected. A GIS is a dynamic tool; there will always be new data and new capabilities to be added, requiring additional efforts and expenses.
When a new system must be established, planners must carefully select the appropriate hardware and software. The system should be simple and must, of course, fit the budget and the technical constraints of the agency. Large digitizers and plotters, which are capable of producing maps of cartographic quality, are expensive and difficult to maintain. Small equipment, which can be as effective as the larger models for map analysis, is becoming increasingly available at affordable prices. Figure 5-5 presents some of the criteria that should be considered in a GIS acquisition.
There are many GIS packages available, some more expensive and more powerful than others. Some cheaper software's have good analytical capabilities, but lack computer graphics. Based on objectives, budget, and personnel constraints, planners should investigate the alternatives for GIS software with a simple interface, strong analytical and graphical capabilities, and an affordable price. Regardless of the selection, GIS software must be tested, and its claims must be verified against the needs of the user. As the software for GIS projects can cost more than the hardware it is designed to run on, the testing should be done on the hardware configuration to be used.
Figure 5-6 reviews most of the GIS software currently available. The systems are ranked by cost, and information is provided on type of operating system, type of output device supported (directly related to the kind of output maps produced, raster or vector), and other capabilities such as area measurement, statistical analysis, and geo-referenced overlaying.
a. Determination of Proposed Applications of the System
b. Determination of Data Needs and Sources for the Applications Selected
c. Design of the Data Files
Once the GIS has been acquired, an information system must be designed. Typically, first-time GIS users tend to put lots of seemingly appropriate data into the system, trying to develop some application immediately. Usually, systems designed on a data-supply rather than on a information-demand basis result in a disarray of data files and a chaotic and inefficient database.
A systematic approach to building an efficient and practical database includes i) a careful determination of users' needs, defining intended applications of the needs, and, if possible, iii) a design evaluation and/or testing in a pilot study (see the GIS design procedure outlined in Figure 5-7).
Small planning agencies or specific hazard mitigation projects may need a simple analysis of what has worked elsewhere to define what the GIS will be used for and what products it is expected to produce. Large organizations or more comprehensive projects, however, need to develop a standard and systematic approach, usually requiring interviews with management, users, and existing system support staff. Answers to the questions below can orient planners in identifying potential applications.
Data on natural hazards, demographic data, and location of population, are the prime concern of natural hazards management and should be defined very early in the process. Infrastructure and settlement sites provide the logical links that make a GIS useful in identifying population locations. When this information is combined with recent data detailing changes in land use, a clear understanding of where the people are located and the kind of activities they are undertaking and how they may be affected by natural hazards can be obtained. With this information, disaster prevention and preparedness actions can be initiated.
QUESTIONS THAT HELP PLANNERS IDENTIFY POTENTIAL GIS APPLICATIONS FOR HAZARDS MANAGEMENT
- What hazard management decisions will be made that could be improved by the use of a GIS?
- How will GIS help to identify the hazards that pose a significant threat and to evaluate the risk involved?
- How could GIS help determine mitigation measures for investment projects and lifeline network elements for disaster prevention activities?
Figure 5-5 - CRITERIA TO BE CONSIDERED WHEN PLANNING FOR A GIS ACQUISITION
a. CPU/System Unit
- Compatibility with standards
- Memory capacity (RAM)
- Disk drives
- Backup system
- Expansion capacity
- I/O channels
- Communication ports
- Warranty terms
b. Features and Peripherals
- Monitors (terminals)
- Power supply
- Networking capacity
a. System Software
- Compatibility with standards
- Special features
b. Utilities Software
- Ease of use
- Integration with total system
- Peripheral control
c. Applications Software
- Appropriateness to needs
- Performance (capacity, speed, flexibility)
- Interface capability
- Upgrade potential
- Training and other user services
- Initial hardware price (CPU, monitor, printer, etc.)
- Additional components (peripherals, digitizers, adapters, etc.)
- Availability with standards duty-free components
- Maintenance agreement and other service
- Transportation/ delivery
- Software price
- Updates/ upgradings
- Maintenance staff (size, experience)
- Existing customer base
- Service facilities
- Inventory of components
- Guaranteed response time
- Capacity to deal with entire system
- Range of courses offered
- Staff experience
- Documentation/ aids
Source: Adapted from USAID, Information Resources Management. Guidelines for Managing Automation Assistance in AID Development Projects, Version 1 (1986).
Figure 5-6 GIS SOFTWARE REVIEW a/
OPERATING SYSTEM COMPATIBILITY
IBM (PC DOS)
|LESS THAN $500
||IBIS(V)||IBIS||IBIS||A C G S|
|SAGIS||SAGIS||SAGIS||A C S|
|OSU MAP||OSU MAP||A C S|
|IDRISI||IDRISI||A G S|
|Atlas Graphics||Atlas Graphics||A S|
|EPPL7||EPPL7||A C G S|
|GEOVISION||GEOVISION||GEOVISION||A C G|
|GRASS||GRASS||GRASS||A C G S|
|PMAP||PMAP||PMAP||A C G S|
||ETAK Geocoder||ETAK Geocoder||ETAK Geocoder||C|
|FMS/AC||FMS/AC||FMS/AC||A C G S|
|Landtrak||Landtrak(V)||Landtrak||Landtrak||A C G|
|Geopro||Geopro||Geopro||A C G S|
|ILWIS||ILWIS||ILWIS||A C G S|
|MOSS||MOSS||MOSS(AO, P)||MOSS||MOSS||A C G S|
|MAPLE||MAPLE||MAPLE||MAPLE||A C G S|
|Mapgrafix (M)*||Mapgrafix||A G|
|MORE THAN $10,000
|SPANS||SPANS||SPANS||SPANS||A C G S|
|TerraPak||TerraPak||TerraPak (P)||TerraPak||C G S|
|System 9||System 9 (SU)||System 9||A C G S|
|Gee-Graphics||Goo-Graphics||Geo-Graphics (V)||Geo-Graphics||A G S|
|VIPERS||VIPERS||A C S|
|Infocam (V)||Infocam||A C G S|
|UltiMap (AE)||UltiMap||UltiMap||A C G S|
|System 600||System 600 (V)||System 600||System 600||A C S|
|GeoVision GIS||GeoVision GIS (V)||GeoVision GIS||A C G|
|DeltaMap||DeltaMap||DeltaMap||A C G S|
|ERDAS||ERDAS||ERDAS (AV, P, V)||ERDAS||A G S|
|ARC/INFO||ARC/INFO||ARC/INFO (AO, P, V, VC)||ARC/INFO||ARC/INFO||A C G S|
"OTHER SYSTEM" KEY
AE - AEGIS
AO - AOS
AV - AOS VS
M - Mac
P - PRIMOS
SU - Sun OS
V - VMS
VC - VM/CMS
X - XENIX
"OTHER CAPABILITIES" KEY
A - Area measurement
C - Command language user interface
G - Geo-referenced overlying
S - Statistical analyses
a/In each section software is listed in order of increasing cost.
Source: Adapted from "The 1988 GIS Software Survey" in GIS World, vol. 1, no. 1 (Fort Collins, Colorado: July, 1988).
Figure 5-7. GIS Design Procedure
Source: Adapted from Chambers, Don. "Overview of GIS Database Design" in GIS Trends, ARC News Spring 1989. (Redlands, California: Environmental Systems Research Institute, 1989).
Once the information requirements are identified, sources that will provide this information should be distinguished. Usually, a number of firsthand sources of information already exist, including maps and other documents (discussed in Appendix A), field observations, and remote sensors (discussed in Chapter 4). Figure 5-8 lists usually available natural hazard information that can be incorporated, into a GIS data file.
In concept, GIS programs should be developed to accept all kinds of data that will eventually be needed. Data may be available in the form of satellite images, weather satellite data, aerial photographs, generalized global or regional topographic or soils maps, or population distribution maps. Data such as these are sufficient to build an initial GIS. Once the framework is developed, new items can be added at any time.
The next step is to design the cartographic layers to be entered into the system, and the spatial attributes to be assigned to them. In this regard, detail of the database, input scale, and resolution must be considered.
Cartographic layers are the different "maps" or "images" that will be read into the system and later overlaid and analyzed to generate synthesis information. For example, cartographic layers depicting past landslide events, geological characteristics, slope steepness, hydrology, and vegetation cover were entered and overlaid in a GIS to create a landslide hazard map, as described in Section B.
There are three basic types of layers, and many different possible combinations among them: polygons (floodplains, landslide hazard areas), lines (fault lines, rivers, electrical networks), and points (epicenters, well locations, hydroelectric facilities). Selection of the correct layer type for a database depends on anticipated uses and on the scale and resolution of the source data. A volcano, for example, may be represented as a point at 1:250,000 scale, but it could well be a polygon at 1:20,000. Similarly, flood-prone areas may be represented as lines bordering rivers at scales smaller than 1:50,000, but as polygons on 1:10,000 scale maps. Planners must keep in mind that point and line representations may well be used for depicting variable locations, but they are seldom used for GIS operations involving cell measurement.
Figure 5-8 - NATURAL HAZARD INFORMATION TO BE USED IN A GIS
INTERMEDIATE THEMATIC INFORMATION
||Epicenters||Maximum recorded intensity, magnitude||Seismic zoning (strong ground motion data, maximum expected intensity or magnitude, recurrence interval)|
|Fault lines||Frequency distribution and gap data|
||Previous event impact||Potential affected area (ash, lava, Pyroclastic flow, lahar)|
|History of eruptions|
||Landfall map||Previous event impact||Design event (surge tide elevation and flood elevation|
|Precipitation||Landfall frequency distribution|
||Bed rock geology||Previous event impact||Hazard susceptibility|
||Precipitation||Previous event impact||Design event (flood elevation and recurrence interval)|
|Stream flow||Maximum stream elevation|
|Biomass production||Population density|
|Vegetation cover||Animal density|
Spatial attributes are identifiable characteristics of the resource information assembled for the GIS. For example, attributes considered for infrastructure can include roads, bridges, dams, etc. For land use, the different land-use map units can identify the attributes. All GIS input data are filed as attributes and can be recovered as individual items or aggregated into groups.
A soils map provides a good illustration of attribute designation. One attribute in the soils "layer" of data would be sand. All occurrences of sand would be located on the map. Once the attribute has been recorded, relevant descriptive material from the accompanying text should be included in the database, not just the legend. This greatly expands the usefulness of the information available to planners.
This same procedure, when used to prepare data for more than one point in time, provides the user with the information needed to measure changes overtime. The most frequent failure of time sequence data is due to the lack of details on the description of the attribute for the different time periods. Thus, it is important to include that information in text form within the GIS system.
Many attributes in some of the well known and frequently used mapped information sources can provide ample information for hazard management in the typical GIS. Six particularly useful sources are:
- land-use and soil surveys
- climatic data
- location of volcanoes, landslide areas, and major geological faults
- natural features (rivers, floodplains)
- human features (infrastructure, population),
- and topographic information (which provides elevation data, terrain complexity, and watershed information)
Natural hazard management decisions based only on the above six sources of data can serve the GIS requirements in many situations. As an example, soils information can provide saturation and runoff characteristics; topography provides watershed area and topographic relief, and combined with soils data can help to identify floodplains; climatic records are particularly useful when combined with run-off characteristics from the soils survey to provide information on flooding and erosion; and lifezone maps are useful in assessing desertification hazards. The number of people located on a floodplain, what urban support centers exist, the location of roads, airports, rail systems, etc., can all be put into the system and analyzed in map form. This information is also useful in the preparation of emergency response plans.
The correct combination of attributes for particular decisions based on a GIS may call for a surprisingly small number of data input sources. Almost all natural hazard situations will be strongly influenced by one or two combined features. Mud slides, for example, usually occur in areas having steep terrain and soils high in clay content. New volcanic eruptions are most likely to occur in areas of historically high seismic activity. Planners or GIS users must understand that the purpose of a GIS is not to procure and incorporate all possible data. That is costly, time consuming, and provides users with an over abundance of mapped data that can be counterproductive. What is important is the acquisition of an appropriate amount of data that provides the necessary information for rapid, effective decision-making for natural hazard management.
Too much detail may unnecessarily add to the cost of the GIS. If a data source is detailed beyond the point of usefulness, then generalized data should be used. If, for example, topographic data are mapped at 5m contours, but some basic decisions will be reached using 50m contours, then input and retrieval of topographic complexity can be reduced by a factor of ten. Careful study of the classification systems of the input data, combined with analysis of critical points of differentiation in the physical data sources can reduce the volume of data input without affecting the utility of the analysis.
Detail of the database must be directly correlated with the planning team's needs and it should be dynamic in nature. A planning team assigned to assess vulnerability to natural hazards could begin by looking at hazards at the national level, then shifting to more detailed studies in local areas of high risk. On the other hand, if an area is selected for regional development planning, the study of hazards can begin at the regional or local level. For example, if the development study is concerned with the transportation sector of a city and the area suffers frequent losses to landslides, the database established should obviously reflect this issue.
Regarding scale, planners or GIS users can take advantage of the flexibility some GIS offer by entering data at various scales and later requesting the system to adjust the scale to fit the particular purpose or stage of planning: small to medium scales for resource inventory and project identification; medium scales for project profiles and pre feasibility studies; and large scales for feasibility studies, hazard zone mapping, and urban hazard mitigation studies.
Resolution or spatial accuracy of the database will be reflected in the number of cells (columns and rows or Xs and Ys) making up the database. The greater the number of cells used to cover a given area, the higher the resolution obtained. However, high resolution is not always necessary, and the tradeoff between what is gained in terms of analytical capacity and what is lost in terms of consumption of computer's memory and input time must be considered. The type of graphic adaptor, the size of computer's memory, and the user's preference as to whether a full or partitioned screen should be used, are determining factors in this respect.
Finally, the design of the database should be tested for performance. Following a pilot test, it is not uncommon to obtain a sizable set of database design rectification's. Guidelines are usually not only directed at the spatial accuracy of data and layer design, but also at the identification of possible obstacles for final system implementation, and the development of procedures or a methodology for performing tasks under normal operational conditions.
The wide array of GIS applications presented in this chapter illustrates the value of GIS as a tool for natural hazards management and development planning. As demonstrated, geographic information systems can improve the quality and power of analysis of natural hazard assessments, guide development activities, and assist planners in the selection of mitigation measures and in the implementation of emergency preparedness and response actions.
As enticing as GIS may look, it is not a suitable tool for all planning applications. Much of the benefit of such an automated system lie, in the ability to perform repeated spatial calculations. Therefore, before making the decision to acquire a GIS, planners need to determine what planning activities could be supported with the system and carefully assess if the amount of spatial calculations and analysis to be performed justifies automating the process. If only a few calculations are foreseen, it will probably be more cost-effective to rely on local draftsmen to draw and overlay maps and calculate the results.
PC-based GIS are the best option for a planning team. Even so, planners will have to select between scores of available hardware configurations and software capabilities, prices, and compatibility's. Given the typical financial and technical constraints that prevail in Latin America and the Caribbean, the hardware configuration must be simple and affordable. For IBM-compatible systems, for example, a standard central processing unit (CPU), a high-resolution monitor, a small digitizer, and an optional color printer are usually effective enough for a development planning agency's needs, and can be easily purchased at affordable prices in most countries of the region. Large and sophisticated equipment requires more technical skills, is difficult to maintain and repair locally, and the added capabilities may not be significant for the planning agency's needs.
Similarly, there are many GIS software packages to choose from and, accordingly, a wide variety of capabilities and prices are available. Usually the more expensive the software, the more powerful the analytical capability and sophisticated the output options. However, added capability, particularly in the area of cartographic quality output, is not always necessary, and may not pay for itself. Prices range from one hundred to more than fifty thousand U.S. dollars. Although inexpensive systems lack certain features present in more expensive ones, they have functional capabilities sufficient to meet the basic analysis needs of natural hazard management activities. It is wise to start with some of these modest systems and later expand them according to the agency's needs.
Other aspects that should be considered are data availability and institutional support. For a GIS to be effective as a planning tool, any problems and difficulties in obtaining data from institutions with different mandates and interests must be resolved. A good understanding for sharing information between the different agencies involved in collecting, generating, and using data must be established to insure the dynamic nature of a GIS.
One last issue planners will have to face is the difficulty they will encounter in implementing GIS results. When it comes to translating GIS results into planning guidelines or mandates, it is not uncommon to see them rejected for political, economical, or other reasons. This may become more complicated at the local level. When local data needs are generalized and included in a GIS for a larger area, conflicts due to people's detailed knowledge of the area may arise.
Natural hazard management requires cooperation at all levels to be successful. Convincing local staff and decision makers that the GIS can provide timely, cost effective, and correct information is a critical step that needs support and attention for every program addressing natural hazard management issues.
Alexander, R., et al "Applying Digital Cartographic and Geographic Information Systems Technology and Products to the National Earthquake Hazards Reduction Program." Final Report Atlas, Appendix B to Research Project RMMC 86-1 in Proceedings of Conference XXXVIII: A Workshop on "Earthquake Hazards Along the Wasatch Front, Utah," Salt Lake City, Utah, May 14-16,1886, Open File Report 87-154 (Reston, Virginia: U.S. Geological Survey, 1987).
Berry, J.K. "Learning Computer Assisted Map Analysis" in Geographic Information Systems Report, Part III (October 1986), pp. 39-43.
Burrough, P.A. Principles of Geographical Information Systems for Land Resources Assessment (Oxford: Clarendon Press, 1986).
Carstensen, L.W. "Developing Regional Land Information Systems: Relational Databases and/or Geographic Information Systems" in Surveying and Mapping, vol. 46, no. 1 (March 1986).
Chambers, D. "Overview of GIS Database Design" in GIS Trends, ARC News Spring 1989. (Redlands, California: Environmental Systems Research Institute 1989).
Devine, H.A., and Field, R.C. "The Gist of GIS" in Journal of Forestry (1986).
Fox, J., and Chow, J. Geographic Information Systems for Rural Development: Appropriate Technology or White Elephant? (Honolulu, Hawaii: Environment and Policy Institute, 1988).
Frank, A. "Integrating Mechanisms for Storage and Retrieval of Land Data" in Surveying and Mapping, vol. 46, no. 2 (June 1986), pp. 107-121.
GIS World. "The 1988 GIS Software Survey" in GIS World, vol. 1, no. 1 (Fort Collins, Colorado, 1988).
Rojas, E., et al "Land Conservation in Developing Countries: Computer Assisted Studies in Saint Lucia" in Ambio, vol. 17, no.4 (1988), pp. 282-288.
United States Environmental Protection Agency (EPA). Geographic Information Systems (GIS) Guidelines Document. (Washington, D.C: EPA, 1988).
White, M.S., Jr. "Technical Requirements and Standards for a Multipurpose Geographic Data System" in The American Cartographer, vol. 11, no. 1 (1984), pp. 15-26.
Zwart, P. "User Requirements in Land Information System Design-Some Research Issues" in Surveying and Mapping, vol. 46, no. 2 (1986), pp. 123-130.