Urban growth analysis and prediction over six Indian cities and its association with pre-monsoon LST and rainfall
摘要
Seventy per cent of the global population is expected to reside in cities by 2050, and Indian cities will be home to 14% of the global urban population. While the population of India has doubled in the last fifty years, its urban population has increased by five times. Since urbanisation is an irreversible process, thus sustainable and accurate future planning is necessary. In this study, spatiotemporal change analysis of land use land cover (LULC), with special emphasis on the urban area, was carried out for the period 2006–2030. The future LULC scenario for 2024 and 2030 is predicted for six different Indian cities (Mumbai, Ahmedabad, Bengaluru, Chennai, Delhi-NCR (national capital region), and Hyderabad). The modules for land-use change simulation (MOULSCE) plugin is used in QGIS (Quantum Geographic Information System Version 2.18.12) for LULC change modelling purpose. The study used LULC data, six urban growth factors, and the CA-Markov approach for computing the transition potential and simulated LULC. The proposed approach achieved strong performance with overall accuracy > 89%, kappa coefficient > 0.78, Jaccard score > 0.80, and Hamming loss < 0.11 across all the considered cities. The LULC change analysis revealed consistent urban expansion across all the selected cities, primarily at the expense of non-urban land or water bodies. Between 2006 and 2018, the urban area expanded by ~1.6% (in Ahmedabad) to ~12% (in Delhi-NCR). However, the study found that the growth rate at which these cities are expanding is expected to decrease in the future, indicating an increase in the urban compactness. Moreover, the pronounced spatial variability in rainfall and land surface temperature (LST) suggests a possible association with the urban growth. In most cities, both variables exhibited noticeable increases, particularly within core urban areas.
HighlightsUrban areas expanded across six Indian metropolitan cities during 2006–18, with the highest growth observed in Delhi-NCR. An ANN-based CA-Markov model accurately (overall accuracy > 89% and Kappa coefficient > 0.78) simulated future LULC scenarios. Future projections indicate continued urban growth with declining or stagnant surface water areas in most cities. Spatial urban growth displays a strong association with increased pre-monsoon LST and spatial rainfall variability. The proposed GIS-based framework offers a cost-effective tool for urban growth monitoring and planning.