Home > Standards & Accurracy

Section 6.0 Map Standards and Accuracy

In order for maps to be useful they must correctly represent real world entities both geometrically and geographically to some measurable degree. Local officials producing maps as public documents have a responsibility to adhere to good standards of map production. This responsibility applies to all types of spatial data both analog and digital. Standards provide rational for how spatial data may be used by defining to what degree the data represent the real world, or in this case how accurately they depict property boundaries. Conversion processes that do not adhere to the standards discussed in this section may present difficulties for use in combination with other spatial data for analysis and evaluation.

6.1. Definitions

Spatial data standards refer to the set of guidelines or specifications that define the structure and uses of spatially related information, including hardcopy maps and digital datasets. There are specific standards that refer to the production of both hardcopy tax maps and digital datasets. This chapter does not attempt to reproduce all of the specifications involved with tax mapping, but rather to discuss those standards that are most relevant to automated parcel maps.

Spatial data standards have been promulgated regarding the content, preparation, and accuracy of maps. Accuracy pertains to the quality of data and the number of errors in the dataset. It is the degree to which information in the dataset matches true or accepted values (Struck 1999, 1). Of particular importance to digital datasets are positional and attribute accuracy. For example, entities when measured on a map at the scale of 1":200" should not vary from their real world positions by more than six feet horizontally. Accuracy is not be confused with precision, which refers to the level of measurement used to compile the data (Floode and Huebner 1995, 2). Precision takes into account how well geographic or attribute data is recorded. For instance, engineers may measure a road with great precision to a fraction of an inch. High precision does not indicate high levels of accuracy nor does high accuracy imply high levels of precision (Floode and Huebner 1995, 2)

The Commonwealth of Massachusetts Department of Revenue, Division of Local Services’ (DLS) publishes a Guideline For Tax Mapping which addresses these issues. To obtain a copy, contact the DLS at 51 Sleeper Street, Boston, MA 02205-9490, (617) 626-2300, http://www.state.ma.us/dls.

6.2. The Importance of Accuracy

There are several facets to the issue of accuracy. First, users must consider the quality of data used to produce the spatial dataset. The conversion of existing records, without careful review of the accuracy of that information, may mean that the quality of the GIS parcel data is poor, out-of-date, or incomplete (Wright 1997). As the saying goes, "garbage in, garbage out." If the condition of hardcopy makes the transfer of accurate information into digital format highly questionable, users may wish to base the conversion on deeds or record plans held at the county or district Registry of Deeds. Of course this process does not mean that all adjoining parcels will fit together like a jigsaw puzzle. Cartographic expertise may be required to reconcile adjoining parcel boundaries that overlap or show gap.

Second, users must consider the utility of parcel spatial data. What tasks will be performed using the parcel dataset? Will it overlay with other coverages at the same scale for analysis or modeling? The GIS should not only make analysis easier, but it should also increase confidence in the outcomes of analyses. When combining datasets in an application, the outcome may be considered only as accurate as the least accurate dataset. For example, an accounting of properties with contaminated wells is accurate to the degree at which the coverages used were produced at the same scale, using the same standards for positional accuracy.

Also, it is highly probable that a parcel coverage will be used to create new coverages such as zoning or landuse. Users should note that spatial datasets created in this fashion will inherit the accuracy of the propagating coverage and any errors contained therein.

Finally, the public tends to accept information presented in a GIS as fact, without questioning the data or the rational behind its application. This perception is based on the fact that maps have an inherent "truthfulness" and the accuracy of information generated by computers often accepted without question. Therefore, doubts raised over the dimensions or location of parcel boundaries depicted in the coverage must be answered with review of the property deed or a preliminary survey, not by a computation of the GIS. Local officials should be aware that parcel maps in hardcopy or digital form, are a representation of parcel boundaries and not a legal determination of dimensions or ownership (Colleary 1999). Under Commonwealth law, only a legal conveyance such as a deed or record plan provides a complete and accurate description of parcel boundaries in metes and bounds (Davis 1956, 194-195; Burgess 1999 np).

6.3. Spatial Data Standards

Positional accuracy is a measurement of how close map features are to their true position on the Earth (Struck 1999, 1). The Commonwealth of Massachusetts requires that all tax maps meet National Map Accuracy Standards (Massachusetts 1987, 14). The standards were first promulgated by United States Bureau of the Budget in 1947, and require that ninety percent (90%) of all measurable points fall within 1/30^th of an inch from their true coordinate position for maps at scales larger than one inch equals twenty thousand inches (1":20,000") (US 1947). According to this standard, the horizontal positional accuracy of parcels maps at a scale of one inch equals twelve hundred inches (1":1,200") or one inch equals one hundred feet (1":100'), is +/- 3 feet (United States 1947).

As stated previously, the accuracy required of a parcel dataset is a function of its intended application. If parcel map automation is based on the conversion of hardcopy tax maps to digital format, the positional accuracy is derived from the original map. There should be no loss in accuracy from a conversion process. Consequently, if parcel data is rectified to digital orthophotos, the positional accuracy will reflect the accuracy of the orthophotos.

It should be noted that the positional accuracy of a spatial dataset does not change as the scale of a map feature is increased or decreased visually within a GIS application like ArcView or Mapinfo.

For more information on National Map Standards see the United States Geologic Survey (USGS) web site http://mapping.usgs.gov/standards/

The Federal Geographic Data Committee (FGDC) defines cadastral data as, the geographic extent of the past, current, and future rights and interests in real property including the spatial information necessary to describe that geographic extent" (1996, 3). Cadastral maps or tax maps describe and record land ownership, or rights and interests. The FGDC's Cadastral Content Standard forms the basis for automating the parcel boundaries found in public records. It standardizes domains or the entries for feature attribute tables (FAT) and provides the definitions of all possible cadastres. The FGDC's specifications for automating parcel records depend in part on the information contained in the conveyance. Other rules are based on data integrity. One type of integrity is that all information must be referenced to a source document. Another relates to the relationship between geographic entities and attributes (FGDC 1996, 4). For instance, there should be a one-to-one relationship between the entity and the corresponding record or attributes within the FAT. This one to one relationship must extend to records of databases linked or joined to the FAT.

The Cadastral Data Content Standard is intended to support the automation of parcel records and is intended to be usable by all levels of government and the private sector (FGDC 1996, 4). The document is highly detailed and should be reviewed by a GIS enterprise undertaking the automation project. The Cadastral Content Standard for the National Spatial Data Infrastructure is available from the Federal Geographic Data Committee website http://www.fgdc.gov/standards/documents/standards/cadastral/.

To assure that other users can evaluate the accuracy and use the parcel coverage with other spatial data, it must be projected in Massachusetts State Plane Coordinates. The state coordinate system is based on the Lambert Conical projection, datum NAD83 (horizontal position), and is measured in meters.

Since no amount of accuracy should be lost during the conversion process, an attempt should be made to maintain the scale at which hardcopy maps were produced. Several sources specify the following divisions of scale based on the relative size and density of parcels in rural as opposed to urban areas.

Source: Massachusetts 1982, 15; IAAO 1988, 7

If positional accuracy proves reliable, dimensional accuracy may is assumed to be similarly reliable. Since dimensional accuracy also encompasses parcel acreage, it is advisable to select a number of samples and compare acreage to hardcopy maps or the metes and bounds description found in the deed or record plan (Struck 1998, np).

When developing the FAT, users should take into account guidelines for calculating and recording parcel dimensions promulgated by the Commonwealth Department of Revenue, Division of Local Services (DLS). According to the DLS, dimensions of property lines and acreage for parcels of one (1) acre or more shall be shown on official 24" x 36' tax maps. Where no dimensions exist, a scaled dimension may be show followed by a letter "s" to indicate scaling. Where deed dimensions do not agree with the amount of distance in the spatial dataset or base manuscripts, the discrepancy should be noted with a letter "d" following the dimension. This means that the FAT should contain fields to identify scaled and discrepancy parcels. Even when "best fit" practices are used, this information will help to qualify the dataset.

Completeness is a key element in assuring the creation of a quality dataset. At the most basic level, the issue of completeness refers to data omissions. For example, is there a "significant percentage" of polygons omitted from a "threshold", or prescribed, area? On a broader scale, it is a measure related to selection criteria, generalization, definitions used, and relevant mapping rules that are used to define the dataset (Completeness 99). This defines two schools of thought: 1) quantitative completeness, or the inclusion of an "appropriate" amount of data, and 2) qualitative completeness, the adherence of data to the database design.

From a quantitative perspective, completeness is an assessment of the dataset’s existing features against what should currently be located within the dataset. "Completeness may relate to a number of digital map features: annotation symbols, textual annotation, linework. Completeness will also relate to the attribute data, and whether all necessary attributes are accounted for." (GIS 1999). When outsourcing data conversion, a typical minimum data omission standard is 1%. For instance, if there are 250 roads in a geographical area and 2 are missing, then the dataset falls within the 1% minimum of required features that can be missing. If, however, only 225 roads (or 90%) are included, then the map is only 90% complete.

Qualitatively, completeness involves the structure of data and how it connects to the lay out of the database. In order for a dataset to be complete in this sense, it must correspond to recognized standards for precision, table structure, topology, projection and other specific requirements of the data model. All well-designed standards for quality assurance include completeness as an important component (Categories 99).

The parcel identification system should be easy to understand and flexible. In terms of a GIS, identifiers must be suitable links to other attribute databases and must be capable of being sorted into a logical sequential order and queried. Three examples of parcel identification systems are discussed in this chapter.

In the Guideline for Tax Mapping, the DLS suggest two standards for numbering parcels that may be used within the feature attribute table as the primary key for identifying the digital (1982, 17). The first standard uses map, block, and lot numbers. The block and lot numbers are two digits each. When parcels are subdivided, each new parcel is assigned a sequential suffix or three-digit number to the right of the decimal point of the lot number. The second standard uses the geographic coordinate locator number as the unique identifier. Coordinate identifiers provide superior information about the parcel's geographic location. The identifier is the easting (x) and northing (y) coordinate from the Massachusetts State Plane Coordinate System recorded to the nearest ten feet of the parcel's center point. The center point or tax parcel centroid may be derived through an operation of the GIS.

A third option is to use street addresses as a primary or secondary identifier. Urban and Regional Information Systems Association and the International Association of Assessors have developed a standard for parcel addressing as street address data are used as a common geographic key at the local level (IAAO and URISA 1992, 57). The organizations recommend the following format:

Figure 6.1: Standardized Address Format Source: IAAO and URISA 1992, 58

To minimize redundancy and maximize the use of the spatial dataset, street names, whether they appear on the map or not, should be spelled only one way. Regardless of whether an address is used as the primary identifier, it is recommended that governments adopt this standard for use by all departments.

At a minimum annotation appearing on a parcel map produced by a GIS should include the title, date of parcel conversion, date of revision(s), parcel identification numbers, legend, north arrow, scale, street names, and date of printing. Lettering and line work should remain consistent and uniform for all parcel maps produced by the GIS. (IAAO and URISA 1992, 57)

6.4. Unresolved Parcels and Errata List

Instances of parcel boundaries that do not conform to the space allotted are considered unresolved parcels. It is common to establish an errata list of unresolved parcels detailing their location and along with documentation as to why a parcel made the list. An errata list should be included with the documentation regarding the conversion process and construction of the coverage (see Section 8.0. Metadata).

The method of resolution is dependent upon the level of accuracy desired. Municipalities may choose to "best fit" unresolved boundaries or conduct deed research perform field checks, or conduct new surveys to create a dataset that is accurate to their needs. Users may also choose to leave discrepancies until surveys or GPS produce new findings (Struck 1998, np). The methodology for handling unresolved parcels should remain consistent. It is also important to determine how frequently unresolved boundaries occurred during the automation process. The ratio unresolved to the total number of parcels is a helpful indicator of the accuracy of the dataset, especially those produced through COGO (Struck 1999, 2). The industry standard is that the number of unresolved parcels should not exceed three percent of the total number of parcels automated.

6.5. Parcel Inventory

In the Guide To Contracting For Tax Mapping Services, the DLS recommends an inventory of all parcels for the creation of tax maps (1987, 13). A thorough inventory will require a measurable amount of staff time and may significantly increase the cost of automation. The precision to which this task is done will be a measure of accuracy goals for the dataset. For parcel data automation, this is the most accurate method. An inventory may include the following:

• Compiling a permanent copy of the most recent conveyance (property deed) for every parcel appearing on tax maps;

• Determining the boundaries and dimensions of each parcel using the most recent conveyance;

• If a conveyance does not exist, using surveys, aerial photographs, or contacting the property owner to determine boundaries;

• Resolving discrepancies between adjoining parcel boundaries using surveys, aerial photographs, or by contacting the property owner; and

• An errata list.

There are several benefits to conducting an inventory. First it provides the GIS development team the option to produce the dataset using COGO (see Section 2.3. COGO). Because each conveyance is defined in terms of geometric distances and angles from control points, these data can be input into a COGO application to construct each parcel. COGO is considered the most accurate method of parcel automation. Second, if some other digital parcel automation method is used, the inventory can be used to resolve discrepancies between adjoining parcels boundaries with greater certainty. Third, an inventory that is maintained as the dataset evolves establishes a readily accessible legal record to address questions or concerns by land owners regarding the depiction of parcel boundaries within the GIS.