Revealing youth-perceived cultural ecosystem services for high-density urban green space management: a deep learning spatial analysis of social media photographs from central Beijing
Haiyun Xu, Jiaxuan Duan, Mujie Ren, Guohan Zhao, Zhifeng Liu
Abstract
Abstract Context In high-density urban areas, managing urban green spaces (UGSs) is challenging due to limited space. Young people in these environments often experience mental health issues, exacerbated by insufficient exposure to green spaces. UGSs provide cultural ecosystem services (CESs), crucial for enhancing mental well-being and landscape sustainability in general. Objectives Focusing on central Beijing, our study addresses three key questions: (1) What is the spatial pattern of youth-perceived CESs? (2) Which CESs are most valued by youth? and (3) How do the various UGSs provided by different landscape features contribute to youth-perceived CESs ? Methods We used geo-tagged images from Little Red Book, the most active social platform for young people, combined with deep learning techniques to map the spatial distribution of nine types of CES perceived by youth. Random forest analysis was used to identify the contribution of different landscape features to these CESs. Results Our findings show that social recreation is the most valued CES, followed by nature appreciation. Social and artistic CESs are more clustered, while nature appreciation is more dispersed. UGSs in built environments play a more significant role in supporting social recreation, while urban parks contribute primarily to landscape and nature appreciation. Waterbodies and parks contribute less to social recreation. Conclusions This study highlights how crowdsourced imagery and deep learning methods can reveal the spatial distribution of CESs and their potential to improve youth mental health. The findings offer valuable insights for urban planners to enhance CESs in UGSs, addressing mental health concerns in high-density urban settings.