Intergovernmental Information Systems Advisory Council

XI. PLAN FOR GIS IMPLEMENTATION - Part 2 of 3

I. GIS Design

  1. Standards
    You should set certain critical standards for your GIS data to maintain consistency both within your own data sets and with those maintained by others that you will use. Consistent use of good standards will help control your GIS maintenance and operations costs while greatly aiding you in exchanging data with others. As noted above, cooperative efforts in GIS development can yield important benefits, but cooperative efforts will fail without adequate standards followed by all parties.

    GIS standards setting efforts are currently being conducted by Minnesota Department of Transportation, Minnesota Department of Natural Resources, and the Metropolitan Council. Each of these efforts is focused on the internal needs of the organization. The Governor's Council on Geographic Information is active in setting more general GIS standards for Minnesota in cooperation with these groups and others. The Governor's Council can be contacted through LMIC. The Federal Geographic Data Committee is establishing GIS standards for the federal government. LMIC also has information about this effort.

    1. Coordinate Systems and Projections
      Coordinate systems are imaginary grids superimposed on the earth's surface that can be used to reference the exact or absolute location of a feature on the earth. Latitude/longitude (Lat/Long) is the fundamental geographic coordinate system, consisting of parallel lines of latitude circling the globe in an east-west direction and north-south lines of longitude (meridians) that converge at the poles. These are the lines you see on a globe dividing up the earth into sections. The distance covered by one degree of longitude varies by latitude because of the convergence of the meridians as they approach the poles. Therefore, latitude/longitude is a non-orthogonal system (not one in which the map features are represented as if latitude and longitude lines are at right angles to each other). Lat/Long is an excellent "lowest common standard" for the exchange of spatial data, but is not suitable for the mapping of this data due to the major distortions that would result.

      Maps are flat representations of the shapes, sizes, and locations of features on the earth. The curved surface of the earth cannot be represented on a flat map without distortions of shape, area, scale/distance, or direction. Map projections have been devised to choose, understand, and control this distortion.

      The choice of projection determines how features on the map look and what kind of distortion will be present. Different projections are appropriate for representing large, medium, and small areas on the earth's surface. When mapping a relatively small area, like a county or city, less distortion is apparent on the flat map.

      At the county and city scale, the most commonly used projections in Minnesota are State Plane, County Coordinate Systems (CCS) based on the State Plane System, and Universal Transverse Mercator (UTM).

      State Plane, based on a conic projection, avoids the distortion inherent in displaying the curved earth on flat surfaces by dividing the state into three zones that do not cover much distance north to south. This minimizes the effect of convergence of meridians across the zone, thus allowing the assumption that all the north-south and east-west lines cross at right angles and are scaled the same/are the same length. The unit of measure in State Plane has historically been the foot, however, NAD83 (North American Datum of 1983) State Plane data is published in meters.

      UTM, in which both Mn/DOT and LMIC distribute data, is also zonal. It is based on a Mercator projection that is useful for displaying map data that has more of a north-south orientation and covers a larger area than State Plane. Most of Minnesota is contained in UTM Zone 15, with the western-most parts of the state being in Zone 14. In many cases this difference is ignored, and the entire state is treated as if it were in an "extended" zone 15. Maps using the UTM projection, if showing a large enough area East to West, will have noticeable distortion.

    2. Datums
      As stated above, the earth is more or less round. The exact nature of the "more or less" can have significant effects on how well the absolute position of features can be represented on a map. The study of this "more or less" is called geodesy, and its products are spheroids. Spheroids are mathematical representations of the true shape of the earth, from which calculations of location can be made. These calculations are called "datums."

      Digital maps for GIS all exist in some datum, referenced to some spheroid. The most commonly encountered horizontal datums are North American Datum of 1927 (NAD27), based on the Clarke Spheroid of 1866, and the North American Datum of 1983 (NAD83), based on the GRS spheroid of 1980. For example, most United State Geological Survey (USGS) quadrangle maps are based on NAD27. Mn/DOT currently requires all its mapping to be in NAD83.

      Conversion routines do exist for changing datums. The best ones make use of "NADCON" calculations. The best GIS software will enable the user to change projections, coordinate systems, and datums. While these routines do introduce a small amount of error, these errors are not a problem for most GIS applications.

      Mn/DOT's standard is UTM meters, NAD83. LMIC provides most data in UTM meters, NAD27, often with a y-shift of -4,700,000. Fortunately, most GIS software can convert data between these common standards.

      The State Plane system with measurement units in feet or meters and County Coordinate Systems based on State Plane are the most commonly used in Minnesota local government. They seem to work best for local units of government because, as an orthogonal system, they are easily understood and widely used, and because most land ownership documents are denominated in feet or a compatible English system of measure. We recommend using State Plane, or a County Coordinate System based on it, in the North American Datum of 1983 (NAD83).

    3. Geodetic Control
      Features on digital maps used in a GIS should represent, as faithfully as possible, the true geographic locations of places on the earth and the true spatial relationships between these places on the earth. One important way to aid this endeavor is to register key points in the digital map to the real world coordinates of those points on the ground. These real world coordinates should be in one or another of the recognized projections, units, and datums. The map is then said to be "in control." Doing this will greatly aid the internal consistency and accuracy of the digital map and support the use of this digital map with any others that are similarly in control.

      This is not the same as having your base map tied to the Public Land Survey system (PLSS) of townships and sections. The PLSS is important in land ownership descriptions, but its section corners may or may not have geodetic control coordinates (coordinates representing a position on the earth in an established coordinate system) established for them.

      Geodetic control is critical if you ever use GIS maps from other sources or make yours available for use by others. Having your base map in geodetic control means it fits an established locational standard and other data that fit this standard will overlay correctly with it. Of course this fit is still subject to accuracy limitations inherent in the source of the data and the methods used to create the digital dataset.

      If your map layers are not in control, it may be possible to adjust them to more or less line up with control points. However, it takes extra work and always introduces additional error. The magnitude of this additional error is often considerable.

      Establishing adequate geodetic control will be an important part of your implementation costs. Many counties will improve and densify their geodetic control before moving forward on base mapping or parcel-mapping efforts. You should consult your civil engineering and survey staff about the density and quality of established control points for a GIS to meet your objectives. For parcel base map control, we recommend establishing real world coordinates at least for each Public Land Survey section corner in rural areas and for each quarter section point in developed areas.

      You should establish written standards pertaining to geodetic control in your GIS design. These standards will include projection, coordinate system, datum, required accuracies, and control grid. The quality and density of the grid that is formed by your control point network is another standard you must set. This will depend upon the level of positional accuracy you need to have in your base map. You should consult your county surveyor to help set this standard. Keep in mind that remonumentation and control densification can be expensive.

    4. Accuracy
      There are three kinds of accuracy that pertain to maps. These are topological accuracy, the correct pattern of relationships among map features; relative accuracy, the degree to which map features are shown with the correct distances and directions from each other; and absolute or positional accuracy, the degree to which map features are shown in the correct position on the map in relation to their real locations on the earth. These are discussed here using examples from board digitizing, that is, the hand guided electronic tracing of paper maps on a digitizing board.
      1. Topological Accuracy - Topological accuracy is best illustrated with an example.

        A schematic map of a location drawn on the back of a napkin will not have correct and consistent scaling or exact compass bearings. The relative distances between features shown will not be correct and the absolute position on the earth for the features shown will not be known. Yet this map can lead you to your destination if its features are shown in the correct relationships. That is, road intersections are encountered in the order shown, a lake is in fact more or less north of a certain road which runs more or less east-west, the bank does adjoin the post office on the side opposite a certain road intersection, etc. These patterns, largely disregarding scale and orientation, are what is meant by topology. The first requirement of a map, sometimes the only requirement, is that it be topologically correct, adequately representing the pattern of things on the earth.

      2. Relative Accuracy - Relative accuracy refers to the correctness of the distances and directions between map features. Problems of relative accuracy can be caused by poor source data or incorporating data from different map sources, particularly if they have different scales or levels of positional accuracy. For instance, a stream, from a USGS map, might be on the correct side of a road obtained from a county road map (correct topology), but at the wrong distance from the road. The greater the similarity between the source maps in scale and positional accuracy, the greater the relative accuracy is likely to be. However, relative accuracy does not require that features on the map be referenced to their precise location on the earth's surface. A dataset can have excellent relative accuracy but no relationship whatever to a known real world coordinate system. This situation is sometimes encountered with engineering or architectural drawings.

        The greater the quality of geodetic control and the positional accuracy of the features for the maps used the greater the relative accuracy is likely to be.

      3. Positional Accuracy - Positional or absolute accuracy is the extent to which a map feature is shown in its true location on the earth. It is often represented by a statement such as "plus or minus 10 feet." This means the mapped features' GIS coordinates are within 10 feet of their absolute or true coordinates on the ground. The quality and density of geodetic control will determine absolute accuracy. If criteria for positional accuracy are satisfied, topological and relative accuracy will be high as well.

        We recommend you set absolute and relative accuracy standards that truly serve your needs, but do not exceed them. For parcel base maps this is often +/- 10 feet for regional planning purposes, +/- 5 feet for local planning purposes, and +/- .5 feet for engineering purposes. The cost of creating the parcel base will increase rapidly as the accuracy standards are tightened.

      4. Determinants of Accuracy
          Data Quality - Data quality will be affected by the methods used to determine the locations of the features. Was the data obtained by field survey, aerial photography, calculations from written records, schematic maps, or some other means? Quality of geodetic control used in obtaining the data is an important determinant of data quality. Data quality will also be determined by the degree of care used in recording and transmitting information. In the case of map sources, it would depend on how carefully the maps were drawn and reproduced. For COGO, how accurate were the original measurements? Were they recorded correctly? How carefully were landmarks and monuments used in the creation of the legal descriptions? Have any of these landmarks (such as stream channels, lake shores, fence lines, etc.) moved? How precisely was the surveying done? Are there errors in the descriptions?
        1. Original Scale of Data Sources - If data is captured from hard copy maps, the scale of the source map can be a determinant of positional accuracy. It is very hard to distinguish anything smaller than about 1/50th of an inch on a map (the theoretical "minimum mapping unit"). A thin pencil line, or the smallest identifiable movement of the digitizing puck, is about 1/50th of an inch. For this reason, digitizing a USGS 7.5 minute quad map can not yield a positional accuracy of better than plus or minus 40 feet. At this scale, one inch on the map equals 24,000 inches on the ground. Thus, 24,000 inches divided by 50 equals 480 inches or 40 feet on the ground. If you digitize a road center from this map, you can say with certainty that its positional accuracy is no better than 40 feet to the right or the left of where you show it in the digital map.
        2. Quality of Geodetic Control - The quality of geodetic control for digital mapping depends upon how precisely and accurately the real world coordinates for your control points are known and how many of these control points can be used. The closer map features are to well established control points in the digital map, the better the control will be. In general, a denser control network yields better accuracy.

        3. Source Shape - For hard copy sources, accuracy may also depend on the extent to which your source map faithfully represents the true shape of the features represented. A common example of this kind of problem is the commercial plat book map that shows all PLSS sections as exactly one square mile each. No PLSS sections are square, being exactly one mile on each side. Some deviate a great deal. You must be aware that many hard copy maps local units of government will be converting to digital form have no indication of what projection they are in. Likewise, many air photos are not fully rectified or orthogonalized. These conditions will introduce errors of shape when they are used as sources for digital maps. The magnitude of these errors can be considerable.

    5. Precision in GIS Software
      Precision in GIS software is the degree to which the software can maintain the detail of a feature's location. The data may be maintained in either "single" or "double" precision.

      Single precision can store a coordinate of seven digits without rounding. Thus, a point located at X=1,234,565 feet and Y=4,701,114 feet would remain precise through many processing operations down to the nearest foot. However, if any of the coordinate pairs were more than 7 digits long, rounding would occur and 1 foot of precision would be lost. Thus 44,701,114 would be rounded to 44,701,110. In this example, location in the GIS would be recorded precisely to the nearest 10 feet.

      Double precision can handle 14 digits in a coordinate without rounding. This will prevent rounding error on very large coordinate pairs.

      Some GIS software can create and use databases in both single and double precision. Some software can handle only one or the other. Double precision may be more expensive to implement and maintain because of higher software costs and greater demands on data processing and storage capacity.

      While this is a controversial point, single precision may be good enough for many local governments. High levels of precision can be maintained in a single precision system by using a coordinate shift. For example, LMIC usually performs a shift of Minnesota data in UTM meters by subtracting 4,700,000 from all y-coordinates. Although this may require additional steps when exchanging data, it can make a big difference in software and hardware cost. You should investigate the issue, with your particular needs in mind, by consulting your civil engineer or surveyor, as well as GIS vendors and other users.

  2. Data Layers
      Base Map Data Layers
      What map features would form the basic map of the jurisdiction, to which all other maps in the GIS would be aligned and referenced? This base map should contain at least the following kinds of features:
      • Geodetic control points of known coordinates in a standard projection and datum.
      • Public Land Survey section corners with township, range and section numbers
      • Public right of way boundaries
      • Parcel, survey, and land ownership lines and boundaries with unique parcel identifiers
      • Major water features that may be associated with parcel or political boundaries.
        **COMMENT** Consideration should be given to include all water features.
      • Political boundaries

      1. Aerial Photography - Many jurisdictions obtain recent aerial photography products to aid in base map creation and to provide layers of planimetric features - such as building footprints, edge of pavement, power lines, field edges, water features, fence rows, manholes, and catch basins - to use with the parcel base. These products can also include elevation contours and digital images to use as a backdrop for vector features. These products can be useful in helping locate, orient, and/or proof parcel line drawing, provided they are made from fully rectified air photos.

        Simple aerial photography contains distortion of shape, scale, and area due to several factors: curvature of the earth, differences in elevation on the ground, tilt of the aircraft, and parallax of the camera lens. Photogrammetrists can rectify the photo images to correct some or all of these distortions.

        While very useful as data sources, aerial photography and photogrammetry (the correction of distortions inherent in aerial photography) can be expensive. Discuss your objectives with vendors in this field to streamline and rationalize your planned expenditures for photographic products.

    1. Other Map Data Layers
      Geodetic control, the parcel base, and planimetric features captured in GIS format can help you create GIS maps of other features, such as land use and zoning, wetlands, and watersheds. Your list will be determined by the nature of your identified objectives, your source information, and your available resources. Much of the data you want or need may be available from state agencies. Land Management Information Center (LMIC), Mn/DNR and Mn/DOT are good sources of Minnesota data from a number of different federal and state agencies.

      For instance, the Minnesota Department of Transportation has available on CD-ROM a statewide GIS dataset of highway centerlines, water features, and political boundaries which is available to other governmental organizations. LMIC has federal data including the National Wetlands Inventory in a GIS format. LMIC also has Digital Elevation Models and Digital Orthophotography available for much of the state. LMIC and the Minnesota Geological Survey have county well index, soils, and various geologic datasets for many parts of the state. These are only a few of the datasets that may be useful in the development of your GIS.

      Contact LMIC to learn more about a wide range of Minnesota digital data for GIS.

E-mail comments or questions to IISAC at iisac@state.mn.us.

| GIS Guide Index | Continue - GIS Plan Part 3 | HOME|