Use Case Overview
The NVIDIA Omniverse Blueprint for Earth-2 Weather Analytics is a reference architecture, meant for independent software vendors and their application developers to build AI augmented simulation and visualization pipelines for climate and weather domain. This reference can be used to accelerate the development of climate tech applications for weather and climate analysis, planning and risk mitigation and for developing digital twins of weather and climate. The blueprint showcases:
- How the NVIDIA Omniverse platform can be used to visualize geo-spatial data at scale
- How a federated data architecture provides an adapter framework to connect partner platforms and open data repositories
- How AI weather models can be deployed via NIMs to produce accelerated weather forecasts
Experience Walkthrough
The blueprint experience showcases the global canvas built and rendered in Omniverse. Various data streams are layered on it to provide a virtual digital twin environment for interactive analysis. By integrating data from traditional weather simulation pipelines from agencies like NOAA, ECMWF with AI generated weather forecasts and observation data from providers like ESRI, Tomorrow.io, one can build next generation climate information systems.
In this experience, users can pick any time frame of interest and turn on or off the different layers to visualize global weather data like in a weather forecasting and analytics application.
Architecture Diagram
Included NIM
The following NIM microservices are used in this blueprint:
What's Included in the Blueprint
This blueprint comes with an open-sourced repository with documentation on how to replicate and customize this blueprint. The blueprint highlights the end to end pipeline that includes the following key components:
- Earth-2 NIMs for generating weather forecasts using AI models
- Data federation layer that provides the data translation interface and adapter mechanism to connect different data layers
- Omniverse Kit application and extensions for rendering and streaming visualized geo-spatial data on a global canvas
Minimum System Requirements
Hardware Requirements
- GPU: L40S, A6000
- RAM: 32 GB
- CPU: x86-64 architecture, 8 Cores (Intel Core i7 (7th Generation) or AMD Ryzen 5)
- Storage: 64 GB
Software Requirements
- OS: Ubuntu 22.04
- Deployment: Kubernetes
There are multiple deployment configurations that may require more compute requirements. For example, full deployment of the Omniverse application and NIM on the same machine will require 2 GPUs. For complete details, see the blueprint documentation on Github.
Example Walkthrough
The user interacts with the Omniverse Kit application, Earth-2 Command Center, which offers a high-fidelity interactive globe. Using either the WebRTC APIs of the Omniverse Streaming application, or more commonly the GUI of the desktop application the user can select the following:
- Date-time to fetch data for (e.g. 2025-01-01 at 12:00:00 UTC)
- Data source or AI weather model to query (e.g. Global Forecast System, GFS from NOAA)
- Weather fields / variable (e.g. surface temperature)
This fetch request is then converted into a data federation mesh pipeline by an Omniverse Kit application extension that will retrieve the requested data for a set time frame starting at the provided date time. The data federation then executes this pipeline that processes the raw weather data into textures that the Omniverse application can render on the globe. This data is then presented to the user as a feature layer in Omniverse that can be visually customized based on use case.
Ethical Considerations
NVIDIA believes Trustworthy AI is a shared responsibility, and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their supporting model team to ensure the models meet requirements for the relevant industry and use case and address unforeseen product misuse. For more detailed information on ethical considerations for the models, please see the respective NIMs Model Card++ Explainability, Bias, Safety & Security, and Privacy Subcards. Please report security vulnerabilities or NVIDIA AI concerns here.
License
Use of the models in this Omniverse Blueprint for Earth-2 Weather Analytics is governed by the NVIDIA AI Foundation Models Community License. Use of the Omniverse SDKs and microservices is governed by the NVIDIA Omniverse License Agreement
Terms of Use
GOVERNING TERMS: The trial service is governed by the NVIDIA API Trial Terms of Service. Use of this model is governed by the NVIDIA Community Model License. The software and materials are governed by the NVIDIA Software License Agreement and the Product-Specific Terms for NVIDIA Omniverse.
