[MapHist] historical maps in GIS, accuracy, RMS error

[Martin Lukas]

Dear Map Historians,

Has somebody experiences with using historical maps in GIS and questions 
of accuracy? I am using historical dutch maps of Java, Indonesia, from 
the 18th and 19th century to reconstruct the changing size of a lagoon. 
In order to overlay them with more recent satellite images, I 
georeferenced them using ArcGIS (standard cartographic projection UTM 
WGS84, 3rd order polynomial transformation). For the maps from the 19th 
century, I received RMS errors around 200m; for those from the early 
18th century RMS errors of about 1000m. Compared with modern maps these 
are, of course, extremely high values, and it is visible that the 
georeferenced maps do overlay in some areas but not in others. Where 
they don't overlay well, I use visual interpretation based on landmarks 
and in this way estimate the course of the shoreline. Given this degree 
of inaccuracy, is it methodologically tenable to use the maps for that 
purpose? Has anybody used historical maps in GIS and could share 
experiences? Or do you know about other studies?

I would be extremely grateful for any hint, and of course any helpful 
advice will be awarded with a note in the acknowledgments of the article 
if it will be published ;o).

Thanks very much for your time.

I am looking forward to hearing your suggestions.

Sincerely,
Martin




-- 
Martin Lukas
Zentrum für Marine Tropenökologie (ZMT)
(Center for Tropical Marine Ecology)
Fahrenheitstr. 6
28359 Bremen, Germany

Tel.:     +49 421 23800 65
Fax:      +49 421 23800 50
E-mail:   martin.lukas at zmt-bremen.de
Homepage: www.zmt-bremen.de

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.geo.uu.nl/pipermail/maphist/attachments/20090304/03806ed0/attachment.html