Analyzing and predicting land use and land cover dynamics using multispectral high-resolution imagery and hybrid CA-Markov modeling
Xulong Duan, Muhammad Haseeb, Zainab Tahir, Syed Amer Mahmood, Aqil Tariq
Abstract
Rapid land use and land cover (LULC) change, driven primarily by urbanization, presents significant challenges to ecological conservation and sustainable development. Understanding and predicting these transformations is crucial for effective land management and policy formulation. This study investigates the dynamic LULC changes in Okara District, Pakistan, from 1994 to 2024 and projects future patterns for 2034 and 2044 using the Cellular Automata Markov (CA-Markov) model. Okara District is experiencing rapid urbanization, impacting its natural resources and environment. This research employs a hybrid CA-Markov model integrated with GIS techniques to analyze historical LULC changes and predict future scenarios. Landsat-5, 8, and 9 were used for the decision tree classifier (achieving high accuracy above 95 %). Vegetation decreased from 92.681 % (3998 km 2 ) to 88.160 % (3803 km 2 ), while built-up area increased from 1.697 % (73 km 2 ) to 8.437 % (364). Barren land also reduced from 4.999 % (215) to 2.719 % (117), with water bodies remaining relatively constant. The CA-Markov model, which has been validated with a kappa coefficient of 0.91, predicts the continuation of these trends. By 2033, vegetation is projected to decline to 85.852 % (3704 km 2 ), with the built-up area expanding to 11.119 % (480 km 2 ). These trends are predicted to continue until 2044, with vegetation decreasing to 81.799 % (3529 km 2 ) and built-up area reaching 14.886 % (642 km 2 ). Barren land is projected to decline to 2.185 % (94 km 2 ) by 2033 and 1.735 % (75 km 2 ) by 2044, while water bodies may slightly increase. These findings highlight the district's urgent need for sustainable land management practices. This research contributes to a better understanding of LULC dynamics in rapidly changing regions, supporting informed decision-making for sustainable development. • Investigated the dynamic LULC changes in Okara District, Pakistan, from 1994 to 2024. • Landsat images were used for the decision tree classifier (achieving high accuracy above 95 %). • Vegetation declined from 92.7 % to 88.2 %, projected to reach 81.8 % by 2044. • Built-up land expected to expand to 14.9 % (642 km²) by 2044. • Study aids planners in balancing urban growth with ecological preservation.