USD-Based 3D Archiving Framework for Time-Series Digital Documentation of Natural Heritage

Abstract. Natural heritage assets, including ancient trees, ecosystems, and geological formations, are dynamic and evolve biologically and environmentally over time. Traditional documentation methods, typically static and fragmented, fail to comprehensively capture these temporal changes. Unlike built heritage, which benefits from standardized models such as Historic Building Information Modeling (HBIM), natural heritage lacks an integrated digital framework capable of representing its inherent variability and continuous transformation.
This study proposes a conceptual time-series documentation framework that unifies spatial geometry and contextual metadata within a single hierarchical model using Universal Scene Description (USD). Leveraging USD’s variantSet and in-file metadata capabilities, the framework represents multiple temporal states of natural heritage objects, enabling effective tracking, comparison, and interpretation across time. A flexible metadata schema is designed with a two-layer structure: common core fields (assetInfo) and type-specific extended attributes (customData), tailored to diverse natural heritage categories such as plants, animals, geological features, and scenic sites. Additionally, a dynamic visualization model supports interactive exploration, multi-temporal comparison, and condition-based filtering, facilitating both qualitative and quantitative heritage analysis. By shifting documentation from static capture to dynamic record-keeping, this framework addresses critical gaps in natural heritage preservation and aligns with international heritage charters emphasizing transparency, interpretability, and sustainability. The proposed approach provides a robust, scalable, and extensible foundation for preserving the evolving condition and meaning of natural heritage assets, ultimately fostering informed conservation strategies and sustainable heritage management.