Logo image
Back
Modeling Land Use/Cover Changes in the Pensacola Metropolitan Area Using CA-Markov and Machine Learning Techniques
Thesis   Open access

Modeling Land Use/Cover Changes in the Pensacola Metropolitan Area Using CA-Markov and Machine Learning Techniques

Asmita Karki
University of West Florida Libraries
Master of Science (MS), University of West Florida
2026

Metrics

1 Record Views

Abstract

CA-Markov Deep learning Land change modeler LULCC Urban growth Terrset Sustainability Urban Planning
Rapid urban expansion is reshaping environmental systems and landscape patterns in developing coastal areas. Land-use/land cover change assessment plays a critical role in determining both historical land transformation and the drivers of urban development. This study evaluated temporal and spatial land-use/land-cover change dynamics and projected future urban growth in the Pensacola Metropolitan Statistical Area, Florida, using multi-temporal Landsat imagery from 2006, 2016, and 2021. Random Forest and a deep learning-based U-shaped network architecture for semantic segmentation were used to produce land-use/land-cover maps depicting urban or built-up areas, forests, agriculture, wetlands, water bodies, barren land, and open space. The 2006 Random Forest classification achieved an overall accuracy of 50.49% and a Kappa coefficient of 0.35, whereas the U-shaped network architecture-based semantic segmentation classifications for 2016 and 2021 achieved overall accuracies of 84.86% and 78.57%, with Kappa coefficients of 0.82 and 0.75, respectively. Urban growth for 2030 and 2040 was simulated using the land change modeler within the Terr Set Geospatial Monitoring and Modeling System, and this integrated multilayered perceptron neural networks and cellular automata-Markov modeling. The validation of the predicted 2021 map yielded an overall accuracy of 70.82%, a kappa coefficient of 0.610, and an area under the curve of 0.985. Historical analysis showed significant urban expansion, with urban land at 4.97% in 2006, 15.36% in 2021, and projected to reach 19.42% in 2030 and 23.93% in 2040. There was a significant decline in agricultural land and open green space. Normalized difference vegetation index and Land surface temperature results showed decreased vegetation and increased surface temperature linked to urban expansion. These findings highlight environmental and sustainability concerns and provide critical policy insights on sustainable urban planning and resource management in fast-growing urbanized coastal regions.
pdf
Modeling Land Use/Cover Changes in the Pensacola Metropolitan Area Using CA-Markov and Machine Learning Techniques6.52 MBDownloadView
Preprint link to preprint Open Access

Details

Logo image