Modelling urban mixed land-use prediction using influence parameters

Abstract

Abstract Mixed land-use is a popular concept in urban planning due to its expected role in improving environmental sustainability as well as citizen’s quality of life. Land use planning and regulations are not stringent in many cities like those in India, and policies are liberal towards mixed land uses. In these cities, mixed land-uses are a natural phenomenon manifesting under various influencing parameters. However, for studies on mixed land-uses, these cities pose data insufficiency challenges, as vital comprehensive spatial information related to land-uses is not available. Moreover, there is no standardised methodology established to assess the spatial distribution of mixed land-uses at the city level. This research has developed a GIS-based model using Weighted Overlay Analysis to predict and visualise the probability of mixed land-use at the macro or city level for the case of Pune, India. The model uses the easily available spatial data of influencing parameters of mixed land-use as input for prediction instead of comprehensive real land-use data. The model is validated by comparing the predicted mixed land-use intensities with established indicators of mixed land-use for four neighbourhoods. It is found that parameters that influence mixed land-use such as connectivity, grain pattern, population density and access to amenities can be used to predict the probability of mixed land-use. Around 35 per cent of the city area of Pune has more than 0.67 probability of mixed land-use. The model can produce the probable mixed land-use distribution across the city and can be used to compute mixed land-use intensities for neighbourhoods. Highlights for public administration, management and planning: • Mixed land-use probability distribution for Pune City, India is generated using Weighted Overlay Analysis in GIS. • As vital spatial data of land-use was unavailable, the prediction model uses data of influencing parameters of mixed land-uses such as population density, connectivity, grain pattern and access to amenities. • The mixed land-use probabilities predicted can be used to compute mixed land-use intensities of neighbourhoods. It is validated by comparing with traditional mixed land-use indicators.