Digital Soil Mapping in Environmental Studies and Smart Agricultural Practices – A Systematic Review

Digital Soil Mapping (DSM) afford high-resolution spatial output, real-time data updating capabilities, and has proved to be a soil resource and information management system. This review provides a systematic assessment of peer-reviewed journal articles published from 2011 to 2025, focusing on applications of DSM. From over 200 selected articles, the results show that greater attention has been directed towards agricultural land use (73.5%) while forests (7.5%), wetlands (1.5%), and urban areas (1.0%), remain underrepresented. Additionally, 33.7% of studies sampled top soil layer (0–20 cm range) while 17% of the the literatures were inconsistent regarding their sampling depth, displaying possible concerns about how well the vertical dimensions were captured. The nutrient composition (16%), soil organic carbon (18%), pH (14%), and texture of the soil (16%) were the most frequently modelled soil properties, leaving out other significant soil health parameters such as bulk density (4%), cation exchange capacity (5%), and electrical conductivity (6%). The modelling landscape was dominated by machine learning methods, primarily random forest (151) and ensemble techniques (97), while hybrid and uncertainty frameworks remained largely untapped. Alongside these observable gaps, parent material types and compositions, including spatial variability, were sparsely reported. The methodological sophistication and applicability of DSM will remain untapped without addressing these gaps and others, including property heterogeneity and validation issues. Closing these gaps using cross-disciplinary synergy, depth-function modelling, and open data frameworks will be vital to unlocking the full power of DSM in advancing climate change adaptation, food insecurity, and land degradation neutrality.