Agenda

Networking with breakfast and coffee provided.

The emergence of Generative AI is perhaps the greatest technology advancement since the last SatSummit. Large Earth Models are emerging that draw from similar approaches as Large Language Models (LLMs). These Large Earth Models can potentially revolutionize how we derive earth insights. This session will review the Foundation Models approach, which differs from prior AI approaches, and share several promising implementations.

NASA Lifelines’ Scientist "Speed Dating" program is a networking series bringing together Earth scientists, social scientists, data scientists, economists, and other experts to meet new collaborators and identify novel research ideas to address humanitarian challenges with assistance from satellite data and tools.

In the session, participating experts will receive guidance from humanitarian representatives as to what challenges are most pressing for their missions, along with crucial context and constraints that the scientific community would need to come up with effective, application-ready research ideas. We will provide scientists and researchers with in-kind commercial imagery that can be used alongside NASA assets to address humanitarian use cases related to a number of issues in a given context, e.g. urban heat impact, mapping vulnerable populations, energy access, and rapid conflict zone assessments. Experts are encouraged to adopt a systems thinking approach with their research brainstorming and align closely with humanitarian needs, before they present a joint hypothesis at the end of the session that uniquely addresses a humanitarian challenge.

• More Than a Map: USAID's First Ever Geospatial Strategy (Carrie Stokes)
• Deep Data Diving: Exploring the depth of Water Diplomacy (Madeeha Yasmin Merchant)
• Eyes in the Sky, Grounded in Community – Moving from Imagery to Insight (Kate Fickas, PhD)
• Data, Justice, and Heart: How Data Visualisation Can Empower Positive Change (David González)
• Assessing the Impacts of Satellite Data - Approaches, Challenges, and Opportunities (Seamus Lombardo, PhD)

Networking break with beverages and snacks.

This is a dynamic moment for GHG Monitoring. Multiple ambitious efforts have recently been launched to create the most accurate picture of our GHG reality. These include new satellite constellations, data platforms, and broad coalitions. This panel will include representatives of some of the leading efforts across GHG data collection, publishing, and use.

Drawing on recent research on Geospatial Value Pipelines in Africa, the panel will discuss the state of geospatial in Africa, the challenges of accessing and applying earth data, and how various actors can support the development of Geospatial across the Continent.

• Modern Geospatial Foundational Models Are Not Foundational (Ritwik Gupta)
• Human in the Loop Learning for Geospatial AI for Good (Caleb Robinson, PhD)
• Building Open and Transparent GeoAI Frameworks for Scalability (David Saah)
• To show and tell: geospatial LLM agents as active collaborators for Earth observation (Steve Brumby)
• Sand Storm Detection and Monitoring using Machine Learning (Muhammad Adeel Hassan)
• Identifying Seasonality in Earth Observations for Engineering Features (Ali Filipovic)

As agriculture continues to be a leading driver of global deforestation, we are seeing a shift from voluntary company commitments to mandatory regulatory compliance to remove deforestation from supply chains. The landmark European Union Deforestation Regulation (EUDR) is currently spurring activity across the sector. How can we use this moment as an opportunity to improve sustainable sourcing and forest conservation? EO and geospatial data have a huge role to play in solving this complex challenge. We will discuss the current state of work in this space and the remaining gaps to be filled, as the community works towards feeding the planet without destroying the planet.

Networking break with beverages and snacks.

• A Light in the Dark: Detecting Mini-Grid Electrification with Remote Sensing in Sub-Saharan Africa (Isaiah Lyons-Galante)
• Energy Access Explorer: A Digital Public Good to Enable Climate-friendly and Viable Energy Transitions for All (Santiago Sinclair Lecaros)
• United We Observe: Collaborative EO Satellite Efforts in Tackling Environmental Challenges (David "Chalky" Petterson)
• Leveraging remote sensing data for environmental project diligence and monitoring transparency (Katie Tyler)
• Leveraging Earth Observation to Track Conflict-linked Environmental Damages (Wim Zwijnenburg)
• Finding the 140,000 impacted by Hurricane Iota in Honduras (Ian Coady)
• The science and practice of war damage impact assessment using satellite radar (Corey Scher)

A look at the use of earth data by the communities at the front lines of conservation and climate adaptation.

This is a session on accelerating the cadence of imagery workflows for crisis response. The world is entering an uncertain period of conflicts and climate crises. How are US federal agencies adapting to the faster tempo around remote sensing workflows? What are decision-makers getting in their briefs, and what do they find useful? Where are the emerging gaps?

Many people who access public earth data access it via a commercial cloud provider. This is particularly likely true among new data users outside the traditional science research community. The accessibility of cloud services promises to accelerate the adoption and impact of Earth imagery, much like open licensing did decades ago. This movement comprises various efforts, some led by government agencies and others by cloud providers. Each has a slightly different view of its target users and has solved access problems differently. Most have adopted a similar set of emerging cloud-native standards. This panel will explore the current landscape of public data on commercial clouds, including different efforts' viewpoints, roadmaps, aspirations, and challenges.

This series of talks will explore innovative applications of earth observation in water-related fields. We will discuss how satellite imagery is transforming water quality monitoring, risk management, and watershed modeling.

Lunch is provided and we hope you are able to catch up with colleagues and meet new people.

Networking break with beverages and snacks.

Deriving insights from satellite imagery using AI/ML relies heavily on ground reference data (or labels). These data are scarce outside the world's developed regions primarily due to a lack of investment and infrastructure to collect them. Foundation Models, however, promise to change this paradigm and break the barrier for everyone to benefit from AI/ML advancements, even with limited label data. We will discuss how end users can apply new Large Earth Foundation Models to solve the data scarcity problem.

• A Vision for Cloud Optimized Data Delivery (Aimee Barciauskas)
• A Brief History of Digital Earth (Alex Leith)
• Serverless and Cloud-Native Geospatial (Isaac Brodsky)
• One STAC Catalog to Rule Them All (Chris Stoner)
• Why GeoParquet (Kyle Barron)

As heat emergencies become more frequent, cities will have an outsized role in mitigation and response. However, most cities lack a strategy or devoted resources to deal with urban heat. Climate Change is moving faster than city bureaucracy and the victims of this situation are likely to be the most vulnerable. In this working session, we will discuss how to encourage cities to adopt an urban heat risk strategy. What data and expertise will cities need? What new data and tools are coming online? How can we assist city managers to make the business case for a heat strategy? How can we build or engage processes that lead to meaningfully adoption of these strategies? This is a facilitated conversation among participants.

There has been a dramatic shift in the shape of the space & geospatial industry since the last SatSummit. The commercial market has been characterized by consolidation, layoffs, a decrease in Venture Capitol, and falling enterprise values. New space and new geo startups continue to have difficulty finding market traction outside of Defense and Intelligence. In this environment, we’ve noticed some companies shifting away from prior commitments to global development and environmental stewardship. This working session will be an open discussion the current state of affairs and speculation on what lies ahead.

The past year has seen an unprecedented uptake of EO data for crisis monitoring. Conflict in Ukraine and Gaza shows an acute need for EO-driven documentation of the consequences of armed conflicts on cities, communities, and landscapes. But these efforts face real challenges, including ethical concerns, uneven data access and data restrictions, and political sensitivities of public and commercial data providers. Further, this work's urgent and sensitive nature makes it prone to siloing and duplication of efforts. This working session will discuss the current state of EO in conflict monitoring and efforts to advance the state of the practice.

Networking break with beverages and snacks.

The shift to cloud storage of earth data has created opportunities for new modes of data access and distribution. This transition requires the design and implementation of new cloud-optimized file formats and services. Broadly, this means that data storage and services are optimized to minimize the amount of requests for data and size of data to be transferred. But what does this look like in practice? What is our shared vision for the architecture of a cloud-native future? This session will dive into what a cloud-native future could look like. We will determine what choices are clear (use object storage) and which are unclear (do you need to reformat your data). Ultimately we hope to develop a shared vision for optimal delivery of visualization and analysis via the data in the cloud.

Should the Humanitarian Sector have its own satellites? A lively debate between alternative worldviews.

Lately, Foundation Models (FMs) have grown increasingly prominent, particularly in scientific applications. Yet, while constructing FMs is a significant achievement, effectively employing them in operational contexts poses its own set of challenges. Our aim in this breakout session is to explore strategies for deploying FMs in operational settings that are both cost-effective and scalable, ensuring their sustainability in the long run. Let's delve into how we can collaborate to address these considerations together.