APPROACHING QUANTITATIVE ACCURACY IN EARLY DUTCH CITY MAPS
J. Benavides, J. Nerbonne
RuG University, Faculty of Arts, Groningen, Netherlands
Accuracy is a matter concerning everyone involved in using cartography as basis for research. Spatial accuracy defines in principle how close a measured value on a map is to the true value (in the real world). Our interest focuses on accuracy of early maps. There are several facts that inevitably introduce error at some stage of mapmaking or even during accuracy assessment, which we include in our analysis as follows:
-Early maps were made using techniques spatially much less sophisticated but also much more inaccurate than those we use nowadays. However, modern accurate maps represent the closest sources to the real situation we can use as reference for measuring accuracy,
-maps are two-dimensional representations of a three-dimensional surface, and
-the methods we use for accuracy assessment involve comparisons between sources that differ in content, nature, function and maybe most important: in time.
Considering the big gap in time and content between old and modern maps we propose the use of an intermediate source that facilitates the interpretation of old maps and further comparisons and measurements from map to map. We suggest the use of cadastral maps of earlier 19th century as intermediate sources considering they are much more similar in content and representation to old maps than modern maps. As an attempt to measure accuracy in early Dutch city maps we used a set of six maps of the city of Zwolle of the 17th and 18th century. We tested map accuracy standards and studied distributions of positional errors (magnitude and direction) based on a sample data of Cartesian coordinates (x, y) of several points. Assuming as true position the one measured in the cadastral map of 1832, we compare true positions of sample points to correspondent positions in old maps.
By using point-based analysis and statistical tools we explore local variations of positional errors and localize areas of high and low errors. The goal is to identify whether the spatial distributions of the error within and between maps can reveal relationships between function of the map and processes of mapmaking, land surveying or copying. Results indicate that positional errors in magnitude show significant differences according to: 1) geographical location of the feature (inside, outside city) and 2) feature class of the point (e.g. building, fortification, water body). This probably related to the function of the map and priority of feature depiction.