Using georeferenced text from social media to map the cultural ecosystem services of freshwater ecosystems

Cultural ecosystem services (CES) are vital for enhancing human well-being, including those provided by freshwater ecosystems such as recreation, aesthetic values, and education. However, assessing these services is challenging due to their intangible nature and personal perception. Text-based social media data offers a valuable source of information for assessing CES. In this study, we developed a novel methodological framework using georeferenced text from social media to map CES of specific ecosystems. This framework is implemented through TweetMyRiver, a tool designed to extract, analyze, and classify posts from Twitter/X related to freshwater CES. By combining expert knowledge with artificial intelligence (AI) models, we ensured robustness and scalability. We developed the tool in the Ter River basin and tested it in three other river basins: the Fluvia` basin in Catalonia, the Forth basin in Scotland, and the Scarce basin between France and Belgium. The results of the tool are analyzed descriptively and statistically to verify its accuracy, reliability, and applicability in different contexts. Our tool enables the analysis of CES across large areas and over time, providing insights into their distribution, drivers, and dynamics. It has the potential to inform decision-making, support conservation efforts, and contribute to sustainable ecosystem management. Future research should focus on customizing the tool for the analysis of CES in other ecosystem types, leveraging more accessible georeferenced text data, and incorporating different machine learning approaches.