Highlights • We reviewed 103 scenario analysis-based urban ecosystem services (SA-UES) studies from around the world. • SA-UES studies focused on the urban landscape as a whole (Type I) or green infrastructure only (Type II). • Studies from Asia were mostly of Type I while those from Europe and the Americas were evenly split between Type I and Type II. • SA-UES studies pursued mainly city-scale, exploratory, and spatially-explicit scenarios, mostly with regulating services. • Future SA-UES research needs to focus more on climate change, policy intervention, and cultural services.

Abstract Scenario analysis (SA) provides a useful tool to envision the future conditions of urban ecosystem services (UES), but our understanding of how SA has been used in UES research is rather limited. Thus, this study was intended to (1) review scenario analysis-based urban ecosystem services (SA-UES) studies; and (2) synthesize the main findings that can help improving urban planning and management. By adopting a systematic review procedure, we identified and analyzed 103 relevant articles from Web of Science. We found that SA-UES research comprises primarily studies geared towards urban landscape planning/management (Type I) and green infrastructure (GI) planning/management (Type II), with the former exploring the impacts of multiple land use/cover types on UES and the latter evaluating the potential of GI to provide UES. About 93% of the examined cities were located in Asia, Americas, and Europe, with Type I research accounting for 80%, 46%, and 41% of the total number of publications in each continent, respectively. SA-UES research pursued mainly city-scale, spatially-explicit, and exploratory scenarios, with regulating services mostly considered. The most often considered scenario drivers were urban land demand and natural land protection in Type I research, and GI quantity and spatial location in Type II research. UES scenarios were mainly represented quantitively using five types of approaches: adapting existing scenarios, GIS-based mapping, land use/cover change model, tree growth model, and optimization method. UES were most often evaluated by adopting biophysical approaches (64%), followed by monetary approaches (22%). SA-UES studies have shown that the same drivers may have contrasting impacts on UES in different contexts. Thus, sustaining UES requires context-specific solutions, which can be facilitated by the use of SA. To move forward, SA-UES research should further promote the holistic landscape approach that integrates both green and non-green infrastructures, with more emphasis on climate change, policy intervention and stakeholders-participatory scenarios, and cultural services.