Integrated Digital Exploratory Analysis System (IDEAS) – An Open-Source Software Framework for Digital Twins

An Earth System Digital Twin (ESDT) is a dynamic, interactive, digital replica of the state and temporal evolution of Earth systems. It integrates multiple models along with observations, and connects them with analysis, AI, and visualization tools. Together, these enable users to explore the current state of the Earth system, predict future conditions, and run hypothetical scenarios to understand how the system would evolve under various assumptions. To establish a sustainable, extensible, and expandable ESDT solution, a formal software architecture is needed. Through the collaboration across multiple NASA centers, the teams partnered with the Apache Software Foundation to establish the professional open-source Integrated Digital Exploratory Analysis System (IDEAS) framework for digital twins. IDEAS is designed with three basic concepts. 

• Design to Exploit– ESDT architect needs to be aware of existing freely available provisioned data, model, and visualization services that the ESDT could benefit from. In computer science, we instructed our students the simple Don’t Repeat Yourself (DRY) principle. The is especially true for ESDT, because ESDT is aiming for operational big data solutions, not just a quick demonstration. If an organization is already actively making their curated data, analysis, or model outputs available, the ESDT should find ways to exploit them. Software reuse is another form of exploitation. Why reinvent? Exploiting existing services and software reuse would significantly reduce future operation costs. An ESDT should also be prepared to be exploited by other ESDTs. It will be discussed in the Design to Expand principle. 
• Design to Extend– An ESDT needs to be extensible to support new measurements, models (both numeric and AI-based), and interfaces. As we are connecting our digital assets, we will encounter gaps and limitations, both in data and technology. As we identify new climate phenomena or scientific needs, the ESDT needs to be prepared for these changes in technology and requirements. 
• Design to Expand– It would be unrealistic to expect a single ESDT is capable to replicate the complex Earth System that is equipped with all the past and present observations, to drive all possible global and regional numerical models, and process all AI capabilities concerning instruments, data, and predictions. We could try to acquire all publicly available data, but an ESDT should also be able to integrate private, commercial data. The Babel-like ESDT would require significant computing and store, and well as staff (science and technical) to operate. This is the motivation behind the open-source IDEAS framework, to encourage collaboration to set up common public-facing architecture. Imagine being able to orchestrate a federation of ESDTs using exploration tools like a Jupyter notebook without having to setting up a local ESDT instance. We think federated ESTDs is the answer to making actionable science available to Disadvantaged Communities. 

The presentation will discuss IDEAS architecture, its latest progress, and its growing portfolio of digital twins for the Earth including air quality, wildfire, hydrology, and coastal zone.