Assessing the uncertainty of historical map information is a prerequisite for using this source type for land cover reconstructions. We assessed the accuracy of forest cover information from 1930 / 1940 Swiss topographic maps ( Siegfried maps ) by comparing them with forest cover gained from orthophotos and terrestrial photos from the same period. Orthophotos were considered to contain the true forest cover information and consequently inconsistencies between historical maps and orthophotos were considered as true errors ( overor underestimation of the forest cover on the map ). Terrestrial photos reflect the surveyors perspective in the field and we therefore used this source type to develop hypotheses on potential reasons for inaccuracy of map information. Generalized linear models were used to test the hypotheses. The results confirmed our expectation that disagreement between maps is more likely to occur near forest edges, at higher altitudes and in less accessible and visible areas. Contrary to our expectations terrain slope did not play a significant role to explain surprisingly high rates of inaccuracy occurred in forest core areas.
