An Integrative Quantitative Approach to Assess the Impact of Mega Infrastructure on Agricultural Land Conversion and Fragmentation in Sumedang Regency

Mega-infrastructure development is a major driver of agricultural landscape transformation in Indonesia. This study quantitatively assesses the impacts of strategic projects—the Cisumdawu Toll Road and three major reservoirs—on agricultural land dynamics in Sumedang Regency. The methodology integrates multi-temporal Landsat imagery (2013-2024) with MLPNN-Markov-based predictive modeling to project land use/land cover to 2038. The predictive model was validated with a high degree of accuracy (kappa = 0.82). Spatial fragmentation was evaluated using landscape metrics synthesized via Principal Component Analysis (PCA). Results show a significant decline in agricultural land, primarily converted to built-up areas and water bodies, alongside intensified landscape fragmentation around infrastructure corridors. Projections indicate these trends will continue, revealing two distinct impact mechanisms: a rapid "pulse" impact from reservoirs and a sustained "press" impact from the toll road. These findings underscore the urgent need for integrated spatial planning and agricultural land protection policies in future infrastructure strategies.