Enhancing multiattribute seismic facies analysis: The effects of dimension reduction on unsupervised clustering
Interpretation · 01.08.2026 · 14 (3) · T267-T302
DOI: 10.1190/int-2025-0015 ↗Kurz gefasst
Fachartikel in der Zeitschrift Interpretation, online erschienen am 10.07.2026. Verfasst von April Moreno-Ward und 1 weiteren Autor:innen. Die Arbeit nennt 41 Literaturangaben.
Abstract (Originalsprache)
The integration of machine learning (ML) into 3D seismic workflows has shown improved geologic interpretability of deepwater channels in the Taranaki Basin, New Zealand. The integration of two novel dimension reduction (DR) techniques, kernel principal component analysis (K-PCA) and uniform manifold approximation and projection (UMAP), showed improved interpretability for these channel systems. K-PCA yielded higher-resolution seismic images for identifying internal channel architectural elements (AEs), whereas UMAP demonstrated improved resolution for identifying lateral AEs when compared to an established DR technique, principal component analysis (PCA). Building on these findings, the focus shifted to determine how the implementation of DR on seismic attributes before conducting unsupervised clustering affects cluster performance. Our results demonstrated that pre-DR unsupervised k-means clustering produced nonunique clusters that exhibit low statistical performance, whereas post-DR clustering with K-PCA and UMAP showed noticeable improvement in statistical metrics and visual separation of clusters. The interpretation of cluster performance required careful consideration of dimensional perspective; data that appeared highly overlapped in 2D visualizations often reveals distinct cluster separation and better cohesion when examined in 3D space. A balanced evaluation framework was proposed that integrated both quantitative metrics (silhouette score and Davies–Bouldin Index) with a qualitative geologic assessment to ensure meaningful interpretation. This study presented a workflow for adopting and using advanced statistical methods from a geoscience perspective. As ML adoption grew within the field, workflows and interpretational considerations presented by geoscientists for geoscientists became increasingly valuable.
Zitieren
APA
Moreno-Ward, A., & Bedle, H. (2026). Enhancing multiattribute seismic facies analysis: The effects of dimension reduction on unsupervised clustering. Interpretation, 14(3), T267-T302. https://doi.org/10.1190/int-2025-0015
Vancouver
Moreno-Ward A, Bedle H. Enhancing multiattribute seismic facies analysis: The effects of dimension reduction on unsupervised clustering. Interpretation. 2026;14(3):T267-T302. doi:10.1190/int-2025-0015
BibTeX
@article{MorenoWard2026,
author = {Moreno-Ward, April and Bedle, Heather},
title = {Enhancing multiattribute seismic facies analysis: The effects of dimension reduction on unsupervised clustering},
journal = {Interpretation},
year = {2026},
volume = {14},
number = {3},
pages = {T267--T302},
doi = {10.1190/int-2025-0015}
}
EUVIMED