Forecasting Spatio-Temporal Dynamics of Land Use Land Cover Land Surface Temperature and Precipitation in Northern Pakistan Using a Hybrid CA-Markov Modeling Approach

Abstract This study employs the advanced CA-Markov model to project future LULC changes, LST, and PPT patterns for 2040 and 2060. Utilizing historical LULC data from 2000, 2010, and 2020, derived from GEE with high classification accuracy, the model integrates spatial dynamics with temporal transition probabilities. Accuracy assessments for LULC maps consistently exceeded 90% overall accuracy, ensuring robust input data. The analysis of historical trends (2000–2020) revealed significant urbanization, with built-up areas increasing by 4.7%, while vegetation declined by 4.75%. Concurrently, LST consistently rose across all LULC classes except water bodies, with maximum LST exceeding 50.95 ⁰C 50.95 ⁰C by 2020. PPT patterns exhibited variability, with an increase in intensity observed between 2000 (47.46 mm) and 2010 (67.65 mm), followed by a notable decrease in 2020 (21.06 mm). These changes reflect the combined pressures of population growth, urbanization, and climate variability. Future projections indicate a continued increase in built-up areas by 2.39% in 2040 and 0.007% in 2060, alongside a sustained decrease in vegetation cover by 8.05% in 2040 and 0.247% in 2060. LST projected to intensify further, reaching over 47⁰𝐶47 ⁰C in 2040 and over 54⁰𝐶54 ⁰C in 2060. PPT patterns show regional variability, with southern and central areas experiencing higher levels. The study linear correlation between LST and PPT (𝑅² values of 0.019 for 2040 and 0.015 for 2060), suggesting other factors, such as LULC changes, are primary drivers of LST variations. These findings provide critical insights for sustainable urban planning, environmental management, and climate change adaptation strategies in Northern Pakistan.