Introduction
In late 2021, Rumo Logística and EMerald Geomodelling used airborne geoscanning to perform large-scale ground investigations along 170 km of the Lucas do Rio Verde railway, a new, 730-km-long railway being planned in Mato Grosso, Brazil. Airborne geoscanning combines airborne geophysical data with limited intrusive ground data to generate 3D ground models, often with the help of machine learning algorithms. These rapidly acquired data allowed us to quickly characterize geological heterogeneity in the area, predict the depth to the soil-rock interface, and identify key groupings of soil types.
While results were limited by poor geophysical contrast between sand and sandstone in some portions of the survey area, they allowed the project owner to focus the next phase of ground investigation on critical areas. Early access to such information resulted in direct cost-savings in follow-up investigations (which are estimated to be about 33% for the next phase of investigation), but also indirect savings through reduction of risk. For example, we show that we reduce the uncertainty in estimating rock excavation costs by 30% compared to a conventional, borehole-only workflow. This case study is the first successful example of airborne geoscanning in an infrastructure project in a tropical climate.
Conclusion
This case study demonstrates the value of using airborne geoscanning in large-scale infrastructure projects and is the first example of a project in a tropical climate, to our knowledge. This sandstone basin-based infrastructure project adds to the growing body of literature and demonstrates the versatility of the method. Airborne geoscanning was valuable for characterizing the geologic heterogeneity, modelling the interface between soil (Category 1) and rock (Category 2/3), and predicting soil types in an early phase of this railway construction project.
Using these ground models and uncertainty estimates, we identified priority zones for follow-up ground investigations. In certain areas, frequency-domain airborne geoscanning struggled with delineation between sand and sandstone; however, when looking at the project overall, the insights it provided is expected to reduce the number of drillings required in the next phase of investigations by 33%. Moreover, longer-term reduction of risk in the project is expected. Although some of these risk reductions are difficult to quantify, we illustrated that the uncertainty in the rock excavation cost for the project was reduced by 30% compared to conventional, borehole-only methods. These results are representative of the potential gains that airborne geoscanning can provide for similar infrastructure projects.
