ANNOUNCEMENTS
This study analyses the spatio-temporal dynamics of urban expansion in Vijayawada and predicts future growth using geospatial methodologies. Landsat 8 satellite imagery from 2014, 2019, and 2024 was classified by supervised classification in Erdas Imagine 2014, with further accuracy verification performed by visual interpretation on Google Earth. Additional geographical variables, including a 30-meter resolution Digital Elevation Model (DEM) from USGS and Euclidean distances to roads, urban areas, and streams (sourced from OpenStreetMap), were processed and standardised. The layers have been included into the QGIS MOLUSCE plugin for modelling purposes. The MOLUSCE program was created with Artificial Neural Networks (ANN), using the classified maps from 2014 and 2019 as training data and the 2024 Land Use/Land Cover (LULC) as a validation layer to simulate land use change. A validation run was performed by modelling the 2024 Land Use/Land Cover (LULC) and comparing it to the actual 2024 map, resulting in an overall accuracy of 60.81% and a Kappa coefficient of 0.48659. Despite important geographical consistency, the model showed enhanced precision in predicting class quantity distributions (Kappa histogram: 0.93186) but indicated only moderate accuracy in spatial allocation (Kappa location: 0.52217). The 2029 simulation suggested significant urban expansion, particularly along transportation routes and low-gradient areas, highlighting the importance for sustainable land use planning. The study highlights the effectiveness of open-source methodologies and remotely sensed data in predicting urban growth in rapidly expanding Indian cities.
Keywords: MOLUSCE, CA- ANN, Erdas Imagine, Land Use/ Land Cover Simulation, Spatio-Temporal Analysis.
