In building the large models that are used today, collecting elevation data is often a time-consuming process. A good modeler wants to devote the appropriate level of effort to data collection that will yield the desired accuracy at a minimum cost. Some of the data collection options are:

• USGS Topographic Maps
• Surveying from known benchmarks
• Digital Elevation Models (DEMs)
• SDTS Digital Elevation Models
• Digital Ortho-Rectified Photogrammetry
• Contour Maps (contour shapefiles)
• As-built Plans
• Global Positioning Systems (GPS).

The data type used by the Elevation Extractor is Digital Elevation Models (DEMs). Digital Elevation Models, available from the USGS, are computer files that contain elevation data and routines for interpolating that data to arrive at elevations at nearby points. DEM data are recorded in a raster format, which means that they are represented by a uniform grid of cells of a specified resolution (typically 100 ft.). The accuracy of points interpolated from the grid depends on the distance from known benchmarks and is highly site-specific. However, it is usually on the order of 5 to 10 ft. when the ground slopes continuously. If there are abrupt breaks in elevation corresponding to road cuts, levees, and cliffs, the elevations taken from the DEMs can be inaccurate.

DEMs are raster files containing evenly spaced elevation data referenced to a horizontal coordinate system. In the United States, the most commonly used DEMs are prepared by the U.S. Geological Survey (USGS). Horizontal position is determined based on the Universal Transverse Mercator coordinate system referenced to the North American Datum of 1927 (NAD 27) or 1983 (NAD 83), with distances given in meters. In the continental U.S., elevation values are given in meters (or in some cases feet) relative to the National Geodetic Vertical Datum (NGVD) of 1929.

DEMs are available at several scales. For water distribution, it is best to use the 30-meter DEMs with the same spatial extents as the 7.5-minute USGS topographic map series. These files are referred to as large-scale DEMs. The raster grids for the 7.5-minute quads are 30 by 30 meters. There is a single elevation value for each 900 square meters. (Some maps are now available with grid spacing as small as 10 by 10 meters, and more are being developed.) Ideally, some interpolation is performed to determine the elevation value at a given point. The DEMs produce the best accuracy in terms of point elevations in areas that are relatively flat with smooth slopes but have poorer accuracy in areas with large, abrupt changes in elevation, such as cliffs and road cuts.

The Spatial Data Transfer Standard, or SDTS, is a standard for the transfer of earth-referenced spatial data between dissimilar computer systems. The SDTS provides a solution to the problem of spatial data transfer from the conceptual level to the details of physical file encoding. Transfer of spatial data involves modeling spatial data concepts, data structures, and logical and physical file structures. In order to be useful, the data to be transferred must also be meaningful in terms of data content and data quality. SDTS addresses all of these aspects for both vector and raster data structures.

The SDTS spatial data model can be made up of more than one spatial object (referred to as aggregated spatial objects), which can be thought of as data layers in the Point or Topological Vector profiles. A Raster Profile can contain multiple raster object record numbers, which are part of the RSDF module of a Raster Profile data set. Multiple raster object record numbers must be converted into separate grids by converting each raster object record number one at a time into an Output grid.

LIDAR is relatively new technology which determines elevation using a light signal from an airplane. LIDAR elevation data is collected using an aerial transmitter and sensor and is significantly more accurate and expensive than traditional DEM data. LIDAR data can be produced in a DEM format and is becoming more widely available.