| 12 May 2025
Application of Self-Organizing Maps to characterize subglacial bedrock properties in East Antarctica based on gravity, magnetic and radar data
Abstract. Subglacial bedrock properties are a key to understand and predict the dynamics and future evolution of the Antarctic Ice Sheet. However, the ice sheet bed is largely inaccessible for direct sampling. Therefore, it is crucial to efficiently combine various attributes derived from satellite and airborne geophysical surveys to characterize subglacial properties. To reduce subjective choices in the joint analysis of data and related biases, we evaluate a Self-Organizing Map (SOM), an unsupervised machine learning technique. The concept of SOMs, an unsupervised machine learning approach, is briefly discussed, but we focus on data selection and their associated attributes for the case at hand. For this, we analysis the correlation between attributes in order to provide a validation of an appropriate choice. The SOM is trained on attributes derived from gravity, magnetics and ice-penetrating radar data for the Wilkes Land area in East Antarctica, a region where basal conditions may be of high importance to ice sheet flow and corresponding sea level rise, and where also suitable data sets for the application of the SOM exist. In contrast to the earlier studies, our approach uses original line data as far as possible, which have much higher resolution/sampling than in smooth gridded products, which were used for previous analyses. Previous analysis indicated the presence of both crystalline basement and sedimentary basins in the area, and our SOM shows a remarkable agreement, but suggests some points of difference, as for example, some highlands appear similar on previous interpretations, but have quite dissimilar physical settings, which is also expressed in our results. These variations can potentially be exploited further in describing subglacial properties and the coupling between bed and overlying ice-sheets.
The paper demonstrates several notable strengths, particularly in its methodological approach. The use of self-organizing maps (SOM) stands out as a positive aspect. SOMs are well-suited to combine multivariate data into meaningful clusters. Compared to more traditional methods, SOMs often excel at revealing underlying patterns and relationships. The authors use a reasonable range of attributes to identify such patterns, thereby capturing the nature of the subglacial geology in Wilkes Land, East Antarctica. Furthermore, SOMs operate in an unsupervised manner, reducing potential bias from manual classification and handling noisy or incomplete data more robustly than some alternative approaches. Their transparent structure and intuitive results make them especially valuable for insights into the structure of heterogeneous data, where the interpretability and simplicity of SOMs can be more beneficial than the raw predictive power of deeper, more complex models. The subglacial Antarctica serves as an experimental sandbox for method development, and we certainly urgently need to get a better understanding of the tectonic structure to provide ice sheet models with boundary conditions that define the ice sheet’s stability over the coming centuries.
Another strength of the paper is the decision to (mainly) work directly with flight line data rather than relying on interpolated grids. This approach preserves the original measurement fidelity and avoids the artifacts or smoothing effects. By analysing the flight line topography data directly, the study ensures that its results more accurately reflect the true spatial variability, which is crucial when dealing with complex or heterogeneous terrain. The discussion regarding the attributes and their relationship to geology is insightful and engaging.
Despite these methodological strengths, the paper currently suffers from weaknesses in its writing and presentation. The manuscript often lacks clarity, making it challenging for readers to follow the narrative. The overall structure feels incomplete, and the paper appears to have undergone insufficient editing prior to submission—it reads more like a first draft than a finished manuscript. As a result, key points and the significance of the findings are sometimes difficult to discern. In particular, the Introduction, Abstract, and Discussion would benefit from substantial revision, while the more mathematically driven Methods and Results sections are in better shape. The figures and tables are generally okay and useful; however, I strongly recommend changing the colour maps, as they are not perceptually uniform.
To strengthen the paper, I recommend a thorough revision to improve clarity, ensuring that the background, objectives, and discussion are well explained and logically organized, and that any unnecessary digressions are avoided. Careful attention to language and grammar will also help bring the manuscript up to the journal’s standards. Additionally, the paper would benefit from a bit more context and justification for the choice of methods, particularly in clarifying how Self-Organizing SOMs) differ from other potential approaches. That said, I agree that an extensive theoretical background is not required here.
In summary, while the methodological choices are strong, the manuscript requires significant revision for clarity and completeness before it can be considered suitable for publication. I am providing an edited and commented Word file (with apologies for any formatting issues caused by converting from PDF). My suggestions are not exhaustive, but I hope they offer a useful starting point for revising the manuscript and addressing some of the recurring issues. Additionally, some edits may have altered the authors' intentions and should be disregarded, but they are helpful in highlighting where the text is unclear.
Once these improvements are made, I believe Solid Earth can be a suitable journal for sharing these findings.