Discussions surrounding ground-penetrating radar (GPR) have dominated many of the conversations around the search for unmarked graves at former Indian Residential Schools (IRS). While GPR has always been, and remains, a valuable tool in the search for unmarked graves, it is only one of many techniques suitable for this work. Due to the limitations of GPR,  it has long been argued that multiple technologies should be used in combination to produce more holistic results. However, the combination of multiple datasets can lead to interpretive challenges and requires a deep understanding of how different techniques complement or contradict each other. The Institute of Prairie and Indigenous Archaeology (IPIA) has worked to address this challenge, and here we present a potential matrix for evaluating the combined use of multiple technologies in the search for unmarked graves. Our aim is to highlight the myriad of ways multiple technologies can be combined and evaluated for a particular goal. We also discuss the importance of consistency, reproducibility, and accountability in this important work.

Digital methods for site and feature identification are rapidly advancing archaeological practice, particularly through the integration of remote sensing technologies. However, many approaches remain difficult for archaeologists to interpret and implement, often requiring specialized expertise that is not widely accessible within the discipline. This is especially evident in investigations related to Indian Residential Schools in Canada, where transparent methodology and careful interpretation are critical, and where development and validation of findings must be grounded in quantifiable data.

Building on this foundation, this presentation outlines the development and application of machine learning models, pipelines, and workflows designed to detect and characterize burial features using multi-modal datasets. Integrating LiDAR-derived terrain models, RGB orthomosaics, and other remotely sensed data, supervised models are trained using a human-in-the-loop approach with annotated datasets. Current detection is driven primarily by image-based features, while patterns in grave morphology, geometry, and feature characterization are being documented for future integration with computer vision approaches.

Rather than relying on generalized “black box” workflows, this work emphasizes task-specific analytical strategies tailored to archaeological objectives. Results highlight both the potential and limitations of automated detection in complex environments, and underscore the importance of transparency, reproducibility, and careful interpretation in culturally sensitive investigations.

Locating unmarked graves while working with Indigenous communities requires sensitivity, as the goal is to preserve the sanctity of these sites. As the search for unmarked graves continues, new technologies are continuously being developed for minimally invasive evaluations of the presence of human remains. One of these new technologies is the Subterra Grey (S4), a mobile soil spectroscopy and penetrometer unit. The purpose of this presentation is to present the current understanding of the S4’s capabilities based on data collected in collaboration with the Research on Experimental and Social Thanatology (REST[ES]) at the Université du Québec à Trois-Rivières. Soil spectroscopy can identify by-products of fat decomposition known as fatty acid salts, but they are not human decomposition-specific, leading to the problem of false positives. In combination, penetrometry detects soil compaction by measuring the force needed to penetrate the soil, providing insight into whether a burial was dug, but this is not grave-specific. Combined, penetrometry and spectroscopy can detect evidence of potential human decomposition and disturbed soil, but the parameters and limitations remain unclear. Through this presentation, the importance of clearly stating what a new technology can and cannot achieve, especially when working with communities, will be emphasized.