Spatiotemporal Assessment of Urban Growth Dynamics in the Hill City of Shimla, India, from 1991 to 2041 Using Machine Learning Algorithms

Rapid urbanization in hill cities brings forth unique environmental and infrastructural challenges, necessitating data-driven strategies for sustainable development. This chapter presents an integrated geospatial framework to assess urbanization trends in the hill city of Shimla, India. The random forest (RF) algorithm is applied to LANDSAT data to produce high-accuracy land use and land cover (LULC) maps for 1991, 2001, 2011, and 2021, with Kappa coefficients ranging from 0.888 to 0.902. Results indicate a significant 283% increase in built-up areas, expanding from 7.90 km 2 in 1991 to 22.39 km 2 by 2021. Conversely, bare land decreased by 32%, and vegetation cover increased slightly to 69.21 km 2 . Urban growth patterns initially focused on the east–southeast and west–northwest, later shifting to the east–northeast and south–southeast. Further, the multi-layer perceptron–Markov chain analysis (MLP–MCA) model, trained using 2001 and 2011 LULC maps, predicted a 28.42% increase in built-up areas by 2041, with corresponding reductions in bare land (8.80%) and vegetation (4.78%). The projected growth will be concentrated in the east–northeast and west–northwest, raising concerns about increased congestion, pollution, and groundwater depletion. Effective urban planning, including regulated tourism, enhanced public transport, and sustainable infrastructure, is critical to managing these challenges and fostering sustainable development in Shimla.