Abstract
<title>Abstract</title> <p>Pedotransfer functions (PTFs) have become indispensable tools in soil science for estimating difficult-to-measure soil properties from readily available data. Traditional PTFs have largely relied on empirical and statistical relationships, which often exhibit limited transferability beyond the conditions under which they were developed. In response, mechanistic pedotransfer functions (MPTFs) have emerged as a promising paradigm that integrates process-based understanding of soil physical, chemical, and biological processes with predictive modeling. This review synthesizes current knowledge on the conceptual foundations, theoretical frameworks, data requirements, modeling approaches, applications, and future prospects of mechanistic pedotransfer functions. Particular emphasis is placed on the representation of soil structure, pore-network dynamics, water flow, solute transport, carbon cycling, and root–soil interactions as fundamental drivers of soil property prediction. The review further examines process-based models, hybrid mechanistic–machine learning approaches, physics-informed artificial intelligence, digital soil twins, and emerging sensing technologies that are reshaping predictive soil science. Applications of mechanistic pedotransfer functions in hydraulic property estimation, irrigation management, land degradation assessment, soil carbon modeling, climate change studies, precision agriculture, and digital soil mapping are discussed. Additionally, major challenges related to data availability, scaling, uncertainty quantification, computational demands, and model transferability are critically evaluated. The synthesis demonstrates that mechanistic pedotransfer functions provide greater interpretability, physical realism, and extrapolation capability than conventional approaches, while recent advances in artificial intelligence and digital technologies offer new opportunities for enhanced prediction. Future developments are expected to focus on integrated mechanistic–AI frameworks, multi-source data fusion, and real-time soil monitoring systems that support sustainable land management and global food security.</p>