J63.4 Improving Algorithm Communication and Data Cognizance through Standardizing Documentation

Thursday, 16 January 2020: 11:15 AM
157C (Boston Convention and Exhibition Center)
Aaron Kaulfus, Univ. of Alabama, Huntsville, AL; and K. Bugbee, A. Harris, S. Bailey, R. Ramachandran, S. Harkins, and A. Barciauskas

Communicating the scientific and physical theories that make Earth observation algorithms and dataset generation possible is key to establishing trust between users and data repositories. Within the NASA Earth Science community, this information is communicated via Algorithm Theoretical Basis Documents (ATBDs). While most of NASA’s ATBDs contain common pieces of information such as the physical theory underpinning an algorithm, any assumptions made when implementing the algorithm and validation results, these documents have no formal structure. The lack of a standardized structure makes it difficult for end users to easily compare algorithmic methods and the resulting data products but also makes it challenging for algorithm developers to know definitively if all necessary information has been provided. The lack of a repository for ATBDs and their common distribution as static PDF or Word documents, makes search and discovery of not only the documents, but content within the documents, difficult for end users. To address these challenges for users, algorithm developers and data repositories, NASA has prototyped the Algorithm Publication Tool, or APT. The APT is a centralized cloud-based publication tool that standardizes the ATBD content model and streamlines the ATBD authoring process. The APT generates ATBDs as online, searchable web pages that supports dynamic updates as well as downloadable PDFs. The APT features a centralized database that supports search and discovery of ATBDs. This presentation will describe the development of the content model and supporting metadata model for ATBDs, the cloud-based implementation of APT and our vision for APT within the broader NASA Earth science data system.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner