A promising alternative technique for ground improvement is microbially induced carbonate precipitation (MICP), harnessing microbial processes to induce calcium carbonate precipitation. While extensive research has focused on modelling the multiphysical processes of MICP, field-scale benchmarking remains limited. This study employs a 2D axisymmetric multiphysical modelling framework to simulate an upscaling experiment with varied injection strategies using a gravity flow approach in an unsaturated soil domain. A comprehensive monitoring campaign, including piezometer readings, dynamic penetrometer testing results, and calcite content measurements, enabled direct comparison with model results. This confirmed the model's ability to capture key precipitation patterns for treatment variations at a field-representative scale. After benchmarking, the model is used to analyse how injection strategy components, such as injection rate, geometry, volume, timing, and number of cycles affect treatment outcomes. The modelling approach provides deeper insight into the complex multiphysical mechanisms governing this process. The findings highlight that higher injection rates increase treatment homogeneity, while maintaining consistent bacteria and cementation solution injection rates improves efficiency and reduces reactant waste, resulting in a 75 % increase in average calcite content for an investigated case. Novel injection geometries demonstrate improved precipitation distribution compared to the traditional grouting equipment, particularly under lower flow rate conditions (∼1.5 L/min). Such insights contribute to optimising MICP treatment strategies, providing a framework for effective and scalable field applications.