Towards Large-scale Backcasting of Urban Built-up Surface Estimates to 1950 by Integrating Earth Observation Data and Multimodal Geospatial Data

Our knowledge on long-term urbanization trends prior to the 1970s is sparse, mainly due to a lack of harmonized, historical, digital geospatial data. However, such knowledge is crucial to better understand settlement dynamics, and to provide more reliable, quantitative data sources for long-term population modeling. Herein, we outline a data integration effort that leverages historical geospatial data to provide reliable estimates of built-up surface and its spatial distribution back to the 1950s, making use of recent Earth observation data, but also historical maps, aerial imagery, census and cadastral data and other data sources. In this context, we test the suitability of PanTex, a GLCM-derived textural metric to derive, in an unsupervised way, spatial proxy variables for settlement distributions in the 1950s from historical maps and aerial imagery, achieving AUC values of >0.93 in study areas in both, the data-abundant global North and in the global South. These results represent important and promising first steps towards an integrated, multimodal, long-term settlement model.