Recognition and assessment of error in geographic information systems

Abstract

Computer-generated maps, manipulated through geographic information systems (GIS), are powerful tools for analyzing complex spatial interactions. Depending upon the level of error inherent in the source data and the error operationally produced through data capture and manipulation, GIS products may possess significant amounts of error. Errors inherent in (1) land-cover maps derived through classification of Landsat digital data; (2) slope-angle and slope aspect information derived from digital terrain tapes; and (3) soil-type data acquired from USDA Soil Conservation Service (SCS) soil survey reports were assessed. These data were compared to field information collected for 35 sample sites distributed throughout a four-square-mile test area. A 2.5-acre (IOO sqm) and IO.O-acre (200 sqm) resolution cell were utilized to capture data using both the cell midpoint and cell-dominant methods of encoding. Inherent error ranged from 43 percent for the 2.5-acre (100 sqm) land-cover map to 83 percent for the IO.O-acre (200 sqm) slope-angle map. Operational error was also examined. When two or more data layers were combined, errors were found to range from 13 to 29 percent. Combined inherent and operational error ranged from 71 to 83 percent. Smaller cell sizes produced lower inherent and operational error levels, as did the use of the cell dominant area method of data capture. Error significantly increased when more than two data layers were analyzed in concert.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 57 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.