Can we compare 19th century topographical and modern geodata when assessing land dynamics? – Reflections on a Danish case study

Historical maps are valuable sources of past land use and land cover (LULC). Numerous studies compare historical with contemporary maps to assess LULC-dynamics. Yet, maps are not objective representations and categorisation of space. Rather, spatial categories in maps reflect the need for spatial information of the organisation or individual requesting production of the map. This also implies that definitions of categories can change over time according to changing purposes of mapmaking. Therefore, analyses of LULC dynamics must account for categorisation of space in historical maps by investigating the mapping practices of the mapmaker. We apply results from a Danish case study (Levin et al, 2020) to reflect on how changing purposes of mapmaking affect assessed LULC-dynamics. For two study areas covering around 300 km² in northern and central Jutland, Denmark, we applied automated production techniques to extract land categories from Danish 1:20,000 topographical maps from the late 19th century. A comparison with a contemporary map shows overall dynamics, characterised by a decrease in open, non-forested habitat categories, such as heath, dune sand and wetlands at the gain of afforestation and an increasing area of agricultural land. However, a more detailed examination reveals two specific types of dynamics, which cannot be explained by this process of intensification but are, we submit, biased by changing purposes for mapmaking. First: Less than 8 % of the area, which in the historical maps was mapped as dune sand was in the contemporary map mapped as sand, while around 56 % was mapped as grass or herb vegetation. Part of this change can be the consequence of efforts to prevent sand drift or of increased airborne nutrient loads leading to increased vegetation cover. However, this change is also the consequence of a shift from mapping for mainly military interests, where sand was mapped as a category hindering mobility, to modern mapping, focusing on nature conservation and hence vegetation types. Second: Around 17 % of the area of non-forested semi-natural habitat types contained in the contemporary map was not mapped in the historical maps. The majority of these areas are characterised by dry grassland habitats, which today are critical to nature conservation, but historically did not influence military mobility or visibility and consequently were not included as a category in the historical maps. We conclude that in order to prevent misinterpretations of LULC dynamics, biases caused by changing mapping purposed between historical and contemporary maps must be addressed, not least when applying 19th century topographical maps, which predominantly where developed for military purposes.

Levin, G., Groom, G.B., Svenningsen, S.R. & Perner, M.L. (2020). Automated production of spatial datasets for land categories from historical maps. Method development and results for a pilot study of Danish late-1800s topographical maps. Aarhus University, DCE.