Thresholding-based cellular automata for transportation network derived future urban growth patterns in a peri-urban area

Abstract

The present study uses the standalone Cellular Automata (CA) model to predict the transportation network-derived future urban growth patterns of a rapidly expanding peri-urban area in the Mumbai Metropolitan Region. The Landsat satellite images of 1999, 2009, and 2019 are used to deduce the study region's land use/land cover (LULC) patterns. The road networks, suburban railway stations and proposed metro stations are considered the primary drivers of urban growth. The classification results show an increase in built-up area by 48 km2 (104%) from 1999 to 2019, decreasing the open land and vegetation by 21%. The fine-tuned CA gives an area under the relative operating characteristic curve of 0.907, locational accuracy >72% and quantitative accuracy of >89%, indicating a good model fit. The calibrated model is used to simulate the urban growth of 2029 for two scenarios: first a no-metro case and second considering the proposed metro's impact. While the built-up area increases by 28%–120 km2 for the no-metro case, including the proposed metro increases the built-up prediction to 133 km2. The enhanced built-up highlights the role of metro development in triggering significant growth in the region. Although the present study uses Kalyan and contiguous area, the general structure of model and its ability to function with a limited data makes it adaptable across diverse regions at different stages of development worldwide. Such an urban growth model can serve as a practical tool for planners in prompt decision-making.