|Geographic Information Systems
The web is a major resource for data and information Geographic Information Systems (GIS). Spatial data are the backbone of a GIS, and the web is a limitless source of such data. This fact is both a blessing and a curse, because one needs to remember that a GIS is only as good as the data in it. Thus, going out on the web, locating a vast amount of data, and downloading it without a plan and a scheme to verify these data is NOT the way to construct a GIS. However, with a plan and an understanding of the data required for a specific project, the web is certainly an incredible source of information. In addition, there are many educational sites, tutorials, and overviews of data and systems that have been built by government agencies, companies, and research programs. However, the volume and variety of this information is often overwhelming. Below we have selected some key sites targeting those wanting to get started understanding and building a GIS. Our intent is to provide beginners with a manageable list of particularly useful sites. These are our picks and we would be happy to receive comments and suggestions regarding other useful sites.
There is a lot more to becoming proficient in the use of GIS than learning to manipulate computer programs such as ArcView and ArcInfo, but proficiency with GIS software is essential. However, learning how to move data around a computer system and perform simple tasks such as reformatting is also essential. For example, there is a considerable amount of data available on the web, but little of it can be input directly into GIS software without manipulation.
The data base is the most important element of a GIS, and it takes significant discipline expertise to choose wisely from existing data sources and to generate quality data. In addition, its takes discipline and expertise to conduct ground truth studies.
|Most data are inherently either raster or vector in nature. DEMs (digital elevation models) and images are the most common raster data and are mostly used as the first layer in a GIS (i.e., the layer on which vector data are overlain). Vector data are defined by either points, lines, or polygons. Links between vector data and tabular information are the key to differentiating between a GIS and computer cartography.
The generation of high quality graphic products requires a working knowledge of map projections and the basics of geodesy. Multiple conversions between projections, coordinate systems, and reference ellipsoids are an inevitable part of the production of a graphic product from a GIS.
Sometimes simply displaying data in an effective way geographically provides the information desired. However, further analysis of the data is often needed. This analysis may involve straightforward statistics or simple relations such as less than, greater than, equals, change over time, etc. Gridding point data to create a contour map or a surface using a TIN (triangulated irregular network) is an example of a more sophisticated mathematical operation. The input of data into a groundwater flow modeling program and displaying the results is another.
Vector or Raster ?
A basic issue that must be understood from a variety of perspectives is the difference between vector and raster data. Vector data, can be defined by points, lines, or polygons (areas) and are efficient in terms of computer resources required for storage and ease of plotting. Objects that one draws with a program such as Illustrator are stored as vectors. Raster data are stored as regular grids and are usually displayed as images. A scanned image is stored as raster information and Photoshop is an example of a program designed to manipulate such data. In the simplest case, the value at pixel location is either black (0) or white (1) and requires storage of only 1 bit of information. Shades of gray are often represented by values ranging from 0 (black) to 255 (white). The 256 possible values ranges over 2 to the 8th power or 8 bits. Thus, a gray-scale image requires about 8 times as much storage as a black and white one! Color requires 3 bands (or scans) of data so that red, green, and blue values can be created and plotted. With each band representing 8 bits of variation, the color image would require 24 bits (over 16 million) of storage for each pixel. GIS texts tend to discuss this issue without sufficient emphasis on its impact on hardware and software resources.