There are several sources of error that can produce inaccuracy in the spatial data. As stated early, inaccuracies within and the physical condition of the source data are obvious contributors. Errors may also occur throughout the automation process in the conversion of scales, projection into a new coordinate system, orthophoto rectification, and by human error. Whether or not the automation occurs in-house or is outsourced, check plots are necessary to confirm the accuracy and completeness of the process.
One check plot per map sheet should be prepared for review by a team well acquainted with the municipality's parcel system. Check plots may be produced on high quality paper or Mylar at a scale of 1:100. If there is strong confidence in the accuracy of the source maps, the team may also wish to produce plots at a scale that will overlay the source maps. Grid ticks indicating state plane coordinates should be displayed in each corner of individual maps. The review team should check to see if
· boundaries match the original source maps;
· all parcels are accounted for;
· right-of-ways are consistent and parallel; and
· parcel identification numbers are present along with other annotation that was specified for the project.
The team should record any gaps, slivers, or parcels with boundary dimensions that do not match adjoining parcels for correction. A second check plot may be required if extensive corrections are necessary. Any parcel that cannot be located or its boundaries determined through deed research should be recorded on an errata list and included in the metadata (see Section 8.0. Metadata).
Parcel mapping that is highly accurate in all other respects is almost worthless if queries of the attribute data continually return the wrong values (Struck 1998, np). Therefore a review of the link between the geographic data and the tax assessors' database (attribute data) should be performed as part of the check plot procedure. Specially, the team should identify:
· parcels without a match to the database;
· database records without a matching graphic parcel; and
· records referencing more than one parcel graphic.
There should be a one-to-one relationship between the parcel identification number for each digital parcel and the corresponding record within the assessors' database. Well constructed database queries and a random sampling of parcels by their unique identifier are acceptable methods to test for attribute accuracy.
Next Section: Metadata Home