Creating Urban Landscape Typology That Facilitates the Integration of Urban Ecological Knowledge into Practice

Urban landscapes are highly heterogeneous but possess recurrent, distinguishable features known as urban landscape types, which exhibit distinct social and environmental performance, and collectively form an urban landscape typology (ULT). ULT is useful as they can aid decision-making in urban planning, predicting the impact of various development proposals. Past attempts to develop ULT overlooked important land features including land use. We developed a conceptual framework based upon the recent literature, and operationalized it using Singapore, an archetypical compact city. Fourteen geospatial variables representing five main components of urban landscapes (land cover composition and configuration, land use, building density, and elevation) were used in a heuristic process to construct a series of models using varying clustering algorithms, coupled with adjustments utilizing contextual knowledge. Models were then tested for their performance regarding practical and environmental relevance. Our results identify ten distinguishable landscape types in Singapore. The six variables that were used in characterizing the UTL were: percent impervious surface, land cover diversity, land use diversity, building density, elevation, and percent water. Initial inspection against Google Earth, followed by a systematic validation exercise by urban planning and design students and professionals confirmed that the ten landscape types adequately correspond to distinguishable urban forms. The ULT were also examined for their ability to estimate land surface temperature (LST) as an environmental performance example; our results demonstrated that all landscape types had significantly different LSTs. We believe that our ULT and the methodological approach we adopted can assist in integrating urban ecological knowledge into practice.