A concept study found that an Earth observation digital twin could improve NOAA's weather monitoring/modeling capabilities
///

NESDIS Publishes Orion Space Solutions’ Earth Observation Digital Twin Study

2 mins read

An Orion Space Solutions concept study, published under a partnership with the National Environmental Satellite, Data and Information Service, found that the deployment of a machine learning-enabled Earth observation digital twin data ground processing and dissemination system could help improve the National Oceanic and Atmospheric Administration’s weather monitoring and modeling capabilities.

Under a Joint Venture Partnerships award, Orion used a sophisticated visualization engine; the ability to run, manage and visualize complex processes using high-performance computing in the cloud; and physics-informed machine learning technologies to prototype an automated Earth observational data processing, analysis and visualization system and explore how digital twins could be used for a variety of purposes, including space domain awareness, sea ice retrieval, weather forecast dissemination and advanced machine learning algorithm training, NESDIS said Monday.

As part of the study, which aims to extend and automate data processing and analytics as a next-generation ground processing enterprise, the company showed how the Earth observation digital twin prototype can be integrated with NOAA’s Unified Earth System Modeling Framework for scientists, engineers, forecasters and the public to better visualize, understand, and predict the past, present and future of the Earth environment.

Recommendations for Successful Digital Twin Implementation

Orion provided NOAA with recommendations on how to proceed should it decide to use the digital twin, including using open-source software tools, processes and engines that meet standards set by the geospatial community; automating the data processing pipeline to deliver information as it becomes available; and leaving source scientific data in original, container format, with a second visualization data format stored as modeled data points embedded in a hierarchical grid format for efficient storage and streaming while maintaining scientific fidelity.