We are a home for Earth science data and computing professionals. Our sessions bring together the community for hands-on, interdisciplinary deep dives as we explore "Innovation to Impact" this year. Learn more about ESIP: esipfed.org
The Open Science Data Federation (OSDF, https://osg-htc.org/services/osdf) is an NSF-funded infrastructure which connects disparate datasets into a single, nation-wide data distribution network that can democratize and deliver data to a wide range of computational environments. This session is aimed at two different groups of participants - 1) Open Data providers who might be interested in learning how to leverage the capabilities of OSDF to provide distributed access to their geoscience datasets and 2) Researchers who are interested in using these datasets. Through lightning talks, demonstrations, and open discussion, we will showcase OSDF-enabled workflows, provide practical guidance on how to contribute to or leverage the OSDF ecosystem, and run a short demonstration of using NCAR’s data through the NSF’s OSPool (https://osg-htc.org/ospool) resource. Value to Session Participants: Open data providers will learn how OSDF can help increase the visibility and impact of their datasets while minimizing redundant data transfers and cloud ingress/egress costs. Scientists will gain insights into accessing curated datasets more efficiently, with delivery optimized for their preferred computational platforms. Both groups will benefit from shared workflows, best practices, and opportunities to shape future infrastructure development.
Recommended Ways to Prepare for this Session: None.
Session Description At the May 2025 Cloud Native Geospatial Conference Lowndes gave a keynote titled “Crossing the chasm: change & resilience within organizations. Diffusion of Innovation theory in practice with NOAA Fisheries Openscapes”. It was about the awesome staff at NOAA Fisheries who are responsible for the stewardship of the nation’s ocean resources and their habitat.
Claiming we’ve crossed a big scary chasm is a bold claim, especially for a big goal: data modernization and workforce development across the agency. And, while the work is still hard and ongoing, this crossing is a story to celebrate. To amplify. To point and say “this is possible”. To repeat/fork in new places. To join. To grow. For you, us, to do these things. This is work that has been building for decades, by many dedicated people inside and outside of NOAA Fisheries, with many different job titles and contributions. This chasm crossing is a big deal because new data workflows are more efficient and robust, but they take real time to adopt and take new skills. It is not easy in a large organization. But through steady work over the last three years, it’s happening. We are out of the early adopter phase and into the early majority. We shared how we planned, using the Diffusion of Innovation theory by EM Rogers (1962). And then how we operationalized, using the Openscapes Flywheel (Robinson & Lowndes 2022).
In this session we will give an abbreviated version of this talk to set the tone for others in and across institutions to share their stories of crossing the chasm.
Value to Session Participants We encourage participants to learn and turn around and amplify stories, fork them, and be inspired to tell stories of their own.
Recommended Ways to Prepare for this Session No recommendations provided
Dr. Julia Stewart Lowndes is Openscapes founding director and co-leads NASA Openscapes and NOAA Fisheries Openscapes projects. I am a marine ecologist working at the intersection of actionable science, data science, and open science. My main focus is mentoring teams to develop technical... Read More →
As climate change drives increasingly destructive wildfire behavior, actionable, real-time intelligence is no longer a luxury, it’s a necessity. This session will showcase WindTL, an AI-powered wildfire modeling platform developed by SkyTL, designed to transform Earth observation data into operational insights for emergency responders, insurers, utilities, and land managers.
The session will explore how WindTL blends physics-informed neural networks (PINNs), machine learning, and real-time sensor fusion to predict wildfire behavior and ember spread — responsible for up to 90% of structure losses during wildfires. We'll share how WindTL operationalizes research-grade fire modeling and delivers user-friendly risk maps and simulations in minutes, even in low-connectivity environments.
We’ll also demonstrate how we’ve scaled WindTL using Google Cloud technologies, including Vertex AI and Kubernetes, to deliver national-scale impact. Real-world examples will include how WindTL accurately predicted ember crossing during the Thompson Fire, enabling faster resource deployment and protection of high-value homes.
The session will conclude with a live demo of the WindTL platform, a discussion on model transparency and community trust, and a call for cross-sector collaboration to accelerate innovation in wildfire resilience.
Value to Session Participants: Participants will gain practical insight into how cutting-edge AI, machine learning, and geospatial data integration are being used in the field today to support wildfire prediction, planning, and response. Through real-world examples, live demonstrations, and collaborative discussion, attendees will see how operational tools like WindTL turn research into action, bridging the gap between innovation and impact.
For scientists and researchers, the session offers an opportunity to explore how their datasets, models, and tools can be integrated into operational platforms. For emergency managers and risk professionals, it provides a look at how actionable intelligence can be deployed in real time to guide decisions. For technologists, it opens up pathways to scale similar models across other climate and disaster domains.
Recommended Ways to Prepare for this Session: Some discussion topics: What barriers still exist in translating wildfire science into real-time operational use? How can AI models be responsibly integrated into emergency decision-making workflows? Where are there opportunities for collaboration in wildfire data collection, validation, or deployment? What tools or datasets should be prioritized for integration into platforms like WindTL?
