Global mapping of urban climate adaptation derived from text-mining of local plans

Cities are increasingly faced with intensifying climate impacts and natural hazards such as floods, droughts, and wildfires. Despite ongoing adaptation efforts to improve social resilience, knowledge about adaptation progress is scattered. Municipal climate plans contain a wealth of information about local adaptation planning and policies, but are seldom studied at a large scale due to their unstructured nature. Here, we use a series of natural language processing (NLP) techniques to extract information on planned adaptation measures for 548 cities with over 1 million inhabitants worldwide. Results reveal a bias toward flood hazards, with cities in the Global South underrepresented, covering only 50% of the target cities. Using the BERTopic seeded topic model, we found that measures related to water management and nature-based solutions were predominant, with some variation across regions. This global mapping provides a starting point for understanding adaptation progress and its gaps, offering a scalable methodology for analyzing municipal adaptation efforts across diverse, multilingual contexts.