Mapping Urban Greenspaces: Classification and Accessibility Analysis Across Bavarian Cities

Kurzfassung

With climate change intensifying urban heat and extreme weather events, German policymakers and administrative local authorities have expressed a commitment to the monitoring and development of their urban greenspaces. The provision status and the accessibility of urban greenspaces in cities across Germany’s federal states such as Bavaria have not yet been evaluated. Therefore, a statewide multi-temporal land cover classification composite with a focus on intra-urban vegetation was created with a random forest model. This approach explored the use of novel 3 m-resolution daily available Planet Fusion Monitoring data. With the classified vegetation data, urban greenspaces were detected and sorted by type, and for 76 Bavarian cities, their accessibility metrics were calculated with a network analysis. The random forest model had an overall accuracy of 83.9%, and classified vegetation with an accuracy of 92.5%, demonstrating its value for intra-urban and small-scale vegetation detection. Results regarding the greenspaces showed that accessibility to urban greenspaces varies drastically across Bavarian cities and urban greenspace types, except for the population’s consistently close proximity to passive urban greenspaces like private gardens or greenspaces on public land. Notably, consistently low numbers of under-served population were found in a cluster of cities in the north-east of the metropolitan region of Munich. The highest proportions of the population within 300 m of walking distances to the nearest recreational greenspaces were found to live in the cities of Waldkraiburg, Bad Kissingen, and Starnberg, while lowest proportions were found in Weilheim i.OB. The findings of this thesis provide a novel contribution to the overall status of urban greenspace distribution and accessibility in Bavaria, especially regarding intra-urban small-scale vegetation detection. The detailed but comprehensive state-wide classified land cover information can act as a baseline and provides policymakers and implementing authorities with a basis for targeted monitoring and development of their urban greenspaces.