Decadal landscape dynamics and future land system simulation using cloud-based geospatial intelligence and cellular automata modelling

Understanding land use and land cover (LULC) transformations in coastal regions is critical for environmental management and development planning. The present study quantifies LULC changes across Kanyakumari District, Southern India over a ten-year period (2015–2025). Random Forest classification applied to Landsat 8/9 imagery within Google Earth Engine identified seven LULC categories. Exported classified maps were analysed in Google Colab for transition matrices, trend evaluation and Markov-based future projections. Results demonstrate significant landscape transformation: scrub land declined by 99.48 km² (38.8%), built-up areas increased by 56.72 km² (14.2%) and forest cover expanded marginally by 26.80 km² (3.4%). Built-up area persistence increased from 90.9% (2015–2020) to 96.5% (2020–2025), indicating rapid urban consolidation. Descriptive rates of change were 5.67 km²/yr for built-up expansion and − 9.95 km² /yr for scrub land. Markov projections indicate built-up areas reaching 500.58 km² by 2030 and 542.93 km² by 2035, because the two observed intervals differ in the amount of change they contain, the calibration-dependent range for 2030 is 457.38-500.58 km², so these figures are scenario envelopes rather than forecasts, with corresponding forest contraction and continued loss of scrub land. Cellular Automata-Markov (CA-Markov) simulation reproduces these totals (500.68 km² in 2030) while differing in the spatial allocation of change. Hindcast validation returns a Kappa of 0.76 and a figure of merit of 0.025. Transition matrices identify scrub-to-forest conversion as the dominant pathway (79.29 km² during 2015–2020), with forest to built-up conversion of 25.54 km². The study provides quantitative baseline information essential for sustainable land management, urban planning and conservation prioritisation in the Western Ghats-coastal transition zone. Results demonstrate the utility of cloud-based geospatial processing for landscape-scale LULC monitoring in data-rich yet computationally demanding applications.