|
||||
|
5.6 GEOGRAPHICAL INFORMATION SYSTEMS When stormwater utilities in Florida were initiated with the City of Tallahassee in the mid-1980s, database management systems (DBMS) and geographic information systems (GIS) were relegated to mainframe and other larger computers with complicated languages and special computer operators. Spreadsheets were available with Lotus 1-2-3, Quattro Pro and others; however, only limited GIS applications were available to stormwater managers. As a result, most of the early stormwater utilities in Florida were developed as a combination of a spreadsheet or database management system (e.g., dBASE) and hand-drawn or computer-drawn reference areas shown on paper maps. Information handling was improved if reference areas were digitized using computer aided drafting (CAD) software; data access was only slightly intelligent. That is, special routines had to be coded to access area measurements and to connect graphical information to the utility database. In recent years, GIS and DBMS systems have significantly improved for general users and spatial information is more readily available to stormwater staff. Concurrent with the technical innovations, GIS costs have decreased so that high-powered computers are now on every desk and are usually connected through local area networks (LANs), wide area networks (WANs), and the Internet (world wide web). Large database, visual-oriented processing, and graphical user interfaces (GUI) are available on personal computers so management of data is at a much higher level. New technology offers stormwater managers much greater access to GIS fee/assessment calculation database information. This means that the development of GIS information for the stormwater utility database should be discussed early in the implementation process. Ultimately, use of GIS processing will speed the measurements and reduce implementation costs. Most computer mapping and analysis systems available are either CADD-based systems or GIS systems. Although CADD and GIS share some common characteristics, they are different technologies, developed for different purposes. Understanding these differences is important in order to take advantage of what each has to offer, realize why GIS is so popular, and appreciate the rapid growth and change in computer mapping and analysis technology. 5.6.1 CADD CADD systems are the combination of hardware and software that enable the computerization of drawings. Although CADD is a drafter/designer tool, it allows for computerized drawings of anything, from a section view of a storm sewer system, to community street maps showing parcel boundaries. Automating drawings using a CADD system offers several distinct advantages:
CADD is a drafting tool, designed with that specific purpose in mind. For example, most construction plans today are created by CADD software. CADD can also be used for general purpose map-making. Many counties maintain tax or parcel maps on a CADD system. However, one of CADD's limitations is its inability to retrieve information from, or conduct analyses based upon, the mapped information. While it is possible to annotate a CADD-drawn map with descriptive information (e.g., frontage or parcel number), this annotation is not in any way connected to the graphic entities it describes. Neither is it possible to see spatial relationships between features, or their descriptive attributes (e.g., parcels that are within 100 feet of a wetland). Geographic information systems, by contrast, provide the ability to conduct such attribute and spatially related analyses. 5.6.2 GIS A GIS is a combination of computer hardware and software that allows for the management, analysis and mapping of spatial information, and associated attributes (descriptive data) with cartographic accuracy. Another commonly used name for GIS is "land information system." A GIS has the ability to store and analyze information describing the location of entities and areas on the earth's surface, such as land parcels, stormwater management facilities, or water bodies. This locational information is often referred to as spatial information or spatial data. In addition to spatial data, a GIS can store descriptive data about those same entities. Examples of such information include the installation date of a storm sewer pipe, or the owner of a parcel of land. The geographic data and the descriptive data that describe them are integrally linked and can be analyzed relative to one another in much the same way one would analyze standard tabular data from a relational database. A GIS is unique because it is the only computer mapping and analysis tool that provides a database interface (for the creation and maintenance of descriptive data), a graphic interface (for the creation and maintenance of spatial data), and the ability to relate these data to one another through either. GIS data are stored in layers, similar to CADD, but the layers are "seamless." This means that the layers (e.g., parcels) are not related to specific maps, but cover the entire geographic extent of the area of interest. The layers also carry information about themselves (descriptive data), meaning that even though all parcels (for example) are in a single layer, a user can locate, map, or otherwise analyze only a specific type of parcel (e.g., only parcels with frontage greater than 300 feet, within 100 feet of a wetland, and with no sewer service) simply by asking the system to locate it. The seamless nature of the data, and its inherent intelligence (its attached descriptive data), make it possible to create maps of different geographic coverages, which display different information, depending on the intended purposes or the maps. Thus, a user working with a digital database, the source of which was a number of paper plan maps, is not limited to working with one plan or another. Instead, the entire system can be analyzed and/or plots created, or of any subsection within it, regardless of the boundaries of the paper maps from which the data were derived. This spatial analysis capability can be applied to a multitude of data layers, simultaneously. For example, even though parcels, municipal boundaries, and storm sewers are on different layers, it is possible, using a GIS, to identify all parcels less than 5,000 square feet, in City X, which are located further than 500 feet from a lateral storm sewer. It is also possible to maintain descriptive data about each parcel, pipe, municipality, etc., or to merge existing descriptive data (from, say a spreadsheet or database file) into the GIS. Additional GIS capabilities include network modeling, described in the section on AM/FM, and topographic analysis (i.e., the ability to view and compute volumes, land surface areas, and lines of site). At its inception, GIS' main advantage was its ability to combine both graphic and descriptive data into one system, perform spatial analysis, and produce cartographically accurate maps. At that time, the graphic editing tools provided by many CADD systems were superior to those provided by most GIS'. Additionally, some relational database systems provided much more power and flexibility in manipulating and processing tabular (descriptive) data than did the existing GIS'. A class of systems began to emerge that combined CADD and relational database systems together into a rudimentary GIS. AM/FM systems, discussed later, evolved this way. These systems have some GIS capability (network analysis and graphic/tabular entity relationships), but do not possess true spatial analysis capability. As it matures, however, GIS will continue to incorporate more of the graphic editing power of "stand-alone" CADD systems, and more of the query and data management power of stand-alone relational database systems. These increases in capability occur in one of two ways: enhancing functionality within a GIS, or providing strong links to existing CADD or database systems and their interfaces. 5.6.3 AM/FM Automated mapping/facilities management (AM/FM) systems are usually prepackaged systems which use the linear, network relationships between facilities (e.g., how storm sewers, manholes, or pipes are connected, or the geography of street networks) to perform inventory and work management, and, sometimes, basic modeling functions. Many of the AM/FM systems on the market today were designed before the advent of user-friendly GIS', and are therefore based on a combination of CADD and relational database management systems (RDBMS). Most are designed for public works and utility purposes, although many newer packages are being offered for fire, police, and school department uses. The spatial data is created and maintained in the CADD system, and transferred to the database system as coordinate data (e.g., latitude/longitude and elevation) via text files. The RDBMS then provides the interface to create and maintain tabular data about the facilities and to analyze them. Data storage is an important feature of AM/FM systems. These systems typically allow the tracking of existing facilities (e.g., manholes, sewer pipes, tanks) and any information that may be associated with those facilities (e.g., construction material, construction date, or size). In addition, AM/FM systems often facilitate tracking of dynamic information, such as the status of a sewer pipe (active, disconnected, etc.), work completed over time, or resources required to perform maintenance. AM/FM systems are also capable of analysis. Once an organization has computerized its facility inventory and has started maintaining the inventory and logging ongoing work on facilities, it is ready to analyze that information. AM/FM systems provide analysis results in both tabular and graphic formats. Whereas a tabular analysis may, for example, include a monthly or yearly summary of the type and location of flooding complaints, a thematic map could be created from the same data showing all serious structure flooding in red and other flooding in green. Such graphic illustration facilitates the pinpointing of problem areas, and often leads to diagnosing the causes of problems. 5.6.4 Basic Principles of GIS An understanding of the basic principles associated with computer mapping and analysis technology is essential to its appropriate and accurate use. The following discussions summarize these principles. Coordinate Systems/Projections A map of an area creates a two-dimensional representation of three-dimensional entities. As a result, some type of mathematical algorithm must be used to convert the latitude and longitude of any point on a spherical surface to a set of coordinates representing that same point on a flat map. The type of conversion used is called a projection. When performing such a conversion, some distortion is unavoidable. Since spatial data is stored using the converted coordinates, and since different projections result in different distortions, it is essential that the projection used for a particular project be decided upon at the inception of the project, before any data is gathered. GIS has the ability to mathematically transform map features from one projection to another, allowing map layers derived from different sources to be successfully used together for visual or analytical purposes. CADD systems have no projection capabilities. Scale The term "map scale" can be used to refer to either the scale of the source map from which the computerized data were developed, or the scale of a digital map when it is plotted. In either case, map scale is usually expressed as either a ratio (e.g., 1:5,000 refers to a map where one unit on the map represents 5,000 of those same units on the ground), or an equivalent (e.g., 1" = 200' refers to a map where one inch on the map represents 200 feet on the ground). Larger scale maps, in which the ratio of map units to actual units is larger (e.g., 1:2,400), show features with greater detail and accuracy, but represent less overall area. Smaller scale maps, in which the ratio is smaller (e.g., 1:100,000) represent larger overall areas (on the same size sheet), but show less detail and are less accurate. Scale must first be considered as it pertains to source maps, since this has a greater impact on the quality of data, and thus, the success or failure of a project. When deciding upon the scale at which to automate maps, it is essential that one take into account the size of the area to be automated, the level of detail and accuracy required by the analysis, and the funding available for the project (as a general rule, the cost of data automation is directly proportional to map scale). Smaller scale source materials will usually suffice for projects covering large areas that require less accuracy, while projects at a town, village, city or site level that cover less area and require more accuracy will usually need larger scale maps. While engineering analysis often requires very large-scale maps or plans (1:480 or 1:1,200), planning applications, at the community level, are usually mapped at a 1:5,000 scale. Once spatial data has been computerized, it is possible to create maps plotted at any scale. However, caution must be taken to clearly indicate the scale of the source material on any digital map plotted at a different scale. For example, if a statewide project has resulted in a 1:100,000-scale base map, it would be inappropriate to use those data to develop a base map for a particular town, village, or city at 1:1,000. At such a scale, the implied accuracy would be much greater that the actual accuracy. It would not be inappropriate, however, to simply plot a map of the town, village or city at 1:1,000, as long as it was clearly noted that the source data were at 1:100,000. Accuracy As the previous section implied, accuracy of spatial data is directly related to the scale at which the data were automated. Additional factors that can affect accuracy include the quality of the original source data and the skill of the technician who automated the data. The National Map Accuracy Standards (NMAS), which provide general guidelines for spatially located data, specify that 90% of all mapped data should be within 0.02 to 0.03 inches of their true location (depending on the scale), when plotted at the source scale. Maps that incorporate many data layers, or different types of information, are only as accurate as the least accurate data layer. It is much more efficient and cost-effective to create data using stringent accuracy standards, than to use loose standards and attempt to correct the data later. Therefore, it is essential that one determine the desired level of accuracy before creating data. The cost of creating GIS data increases proportionally with the accuracy level; therefore, it is important that one define the smallest map scale that will satisfy the needs of the majority of users. A final aspect of accuracy, often taken for granted, is classification accuracy. Classification accuracy refers to the accuracy with which data have been classified or identified. If, for example, a number of bolted and gasketed manhole covers have been classified as standard frames and covers, a field crew might leave on a maintenance run without the proper equipment, or a standard inventory of replacement gaskets would be in error. Classification accuracy is just as important as spatial accuracy in completed data. Data Sources A wide variety of sources can be used for creating digital spatial data, including paper maps, aerial photographs and satellite imagery, existing digital data, surveyed data, and tabular files. Each of these sources provides specific advantages and disadvantages in terms of availability, accuracy, and cost. Paper maps often constitute the primary source of digital data. These maps can include USGS topographic maps, assessors' maps, soils maps, stormwater facility maps, etc. Digitizing is the most common method of automating, or converting paper maps into digital form. Digitizing entails tracing over the paper map with a hand-held device connected to a computer. Different types of maps will be available at different scales. Thus, it is important to use source maps that provide an adequate level of accuracy for the project, while avoiding a data creation task that exceeds budget due to digitizing an unnecessary level of detail. Scanning is another form of data conversion. Scanning entails "taking a picture" of a paper map, the image of which is directly received by the computer. The image appears as a "raster image," one made up of hundreds of little graphic dots. To be usable for most GIS applications, these raster images must be converted into "vectors" or arcs (described later in this section). Coordinate Geometry (COGO) is another source of data. COGO is a means of "creating" mappable information from survey data that indicates metes and bounds of properties. COGO is the most accurate possible means of mapping property boundaries. Its use is often dependent upon the quality of COGO source information and the need for the precision that COGO can provide. It is usually more expensive than paper digitizing means. Aerial Photographs and satellite imagery provide another common source for GIS data. Each must be carefully post-processed to be usable in a GIS, since spatial distortion is inherent in the raw images. This post-processing, usually referred to as "rectification," removes the distortions and "rectifies" the images to a proper projection and coordinate system. The process as applied to aerial photographs results in an image known as an orthophotograph, or a photograph that is planimetrically correct. These orthophotos will include tic marks (places on the map marked with accurate coordinates) for one or more coordinate systems, allowing the proper registration and appropriate projection of the data. Orthophotographs, or their digital equivalents, are a common source of digital data for community GIS'. Satellite images, because of their small scale (generally 1:24,000 or smaller) are not typically used for community systems, but are often used for regional systems or districts. Global Positioning Systems (or GPS) are also sources of GIS data. GPS' are mobile (often hand-held) units that can locate where they are on the earth's surface by real-time communication with satellites. Despite the very complex nature of GPS units, they are very easy to use. They are often deployed to collect point data (e.g., locating a cross-country manhole or a groundwater well), but can also be used to delineate areas (e.g., wetland boundaries), or to record routes of travel (e.g., plotting the course traveled by a stream between two culverts). Many GIS systems can directly read GPS data, so a field crew can take GPS readings, bring the unit back to the office, and a GIS technician can enter the data directly into a compatible GIS. As GIS and mapping applications become more prevalent, a growing amount of digital data is readily available for purchase. Many state and regional organizations have developed a variety of data layers. Examples of these may include pollution sources, watershed areas, land use areas, or soil types. In addition, a number of digital databases are available at the federal level, including digital elevation models (DEMs), which include topographic data, and digital line graphs (DLG) which include roads, rivers and water bodies. A very popular data source is the topologically integrated geographically referenced (TIGER) demographic data available from the U.S. Census Bureau. These files contain U.S. population and housing statistics, at the block and tract level. Most federal data are at scales smaller than 1:24,000. Users of existing digital data should expect to perform some scale and projection changes and some re-coding in order to prepare the data for their own applications. Finally, an endless supply of computer files containing tabular data is available from a variety of sources. These can be incorporated into GIS applications provided there is information that allows the tabular data to be spatially located. For example, an existing database of customer complaints, if referenced by address, can be merged into a GIS containing parcels, building footprints or sheets, as long as these data layers contain addresses as descriptive (or attribute) data. Attribute Data As discussed earlier, a GIS is capable of storing both spatial data and attribute, or descriptive, data. While the spatial data, on which we have concentrated most of our discussion, provides the placement of the entities being analyzed, the attribute data provides the information about those graphic entities (e.g., pipe size, material, and date of construction). For example, when mapping land use areas, spatial data allows us to geographically locate each land use area; attribute data tells us what each land use is (residential, commercial, etc.). Similarly, we can create a map containing political boundaries, but without attribute data, we cannot determine if they are town, village, city or county boundaries or what town(s), village(s), city(ies) and county(ies) they describe. CADD systems allow one to add text to maps that might describe the graphic entity; however, in a GIS, the information is "attached" to the graphic entity. Thus, for example, one can query the GIS to determine the total length of the pipe in a given sub-area, or the specific pipe sizes, between any two locations in a storm sewer system. Topology Topology is a fundamental and powerful feature of GIS, one that distinguishes it from all other computer mapping and analysis tools. Topology refers to the ability of a GIS to recognize the relationships between entities in space, and to use those relationships to perform spatial analyses, including the ability to actually create new data. Entities in a GIS can be represented by three basic features: (1) points which have no dimension, (2) lines which have length, and (3) polygons which have area. Correct topology requires three conditions to be met:
An example of analysis and data creation made possible by correct topology would be the intersection of land use polygons with zoning polygons to create a new data layer indicating areas in which new housing could be developed. Standards As more GIS data are created, it is increasingly important to establish some standards for those data and to strictly adhere to them. Adequate standards will ensure the accuracy of the data created, and the compatibility of the data with data from surrounding towns, villages or cities, and states in terms of format and attribute coding. Data creation, in particular, is one aspect of GIS that requires standards. Items to be addressed include accuracy of source maps, minimum acceptable number of tic marks, log files describing the steps taken to create the data, and topological accuracy. Finally, it is important that all data layers be adequately and consistently documented. If this is not done, the person who created the data will probably be the only one who can understand it. The data documentation should include a history of the data layer and a definition of the file structure for each file in the GIS database.
|
