Thanks so much for that kind introduction, Jack [Dangermon, Esri President], and for the invitation to speak here at the Senior Executive Summit this morning, on behalf of the National Geospatial-Intelligence Agency.
For those who aren’t that familiar with us, NGA is a U.S. Government agency of nearly 15,000 people. We serve as both a Combat Support Agency in the U.S. Department of Defense, and as one of the six principal agencies in the U.S. Intelligence Community. About half of us in are in the Washington DC area, a quarter of us are in St. Louis, Missouri, and a quarter are spread out all over the planet, co-located with military and international partners. Rather than those organizations reaching back for our expertise, we project our expertise forward.
Our mission is to provide geospatial intelligence – GEOINT for short – to national policy makers, military leaders, and first responders. And our motto is “Know the World, Show the Way…from Seabed to Space.”
Now, from a national security perspective, I could certainly spend my entire time telling you how our agency accomplishes those mission sets by supporting NATO efforts to counter Russian aggression in Ukraine. Or helping to unite the world by making sure there’s freedom of navigation in the South China Sea, the Taiwan Strait, and all international waters. Or creating greater awareness of conflicts around the planet – “from Seabed to Space,” as our motto states. And then sharing that awareness with our allies and partners.
But I’ll tell you that at NGA, we’ve also started to concentrate more on in the areas less traditionally associated with bread and butter intelligence work, and that’s what I’d like to talk to you about today. First: The necessity of incorporating AI into GEOINT software and workflow processes, because of the deluge of data we face. Second: The criticality of our GEOINT tradecraft, to include expertise in the science of Geodesy. And third: Leveraging Climate Security, to improve our Humanitarian Assistance and Disaster Relief capacity for both warning and operations.
At a higher level, if we could bring our traditional ethos that’s been primarily emphasized Warning, Targeting, and Safety, and apply it more rigorously than ever before to other areas, such as the prevention of some climate-related issues, and the ability to solve problems caused by natural disasters, considerably earlier, how could we help change the world? How could we get ahead of those situations?
Let’s start with Geospatial Intelligence Artificial Intelligence – GEOINT AI—and the upcoming “deluge of data” that the coming years will bring.
With all of the government and commercial satellite programs that will be capable in the next five to ten years, GEOINT data will at least triple. But there won’t be a corresponding growth of people or money – those will be relatively flat.
So GEOINT AI will be necessary in helping our analysts extract deeper insights and make more informed decisions, from the massive amounts of geospatial data they collect. The AI/GIS combination is key, because GIS excels at visualizing and analyzing spatial relationships, and AI can identify complex patterns and correlations within that spatial context. It thankfully automates tedious, repetitive tasks. And it also builds predictive models for Disaster Relief situations – like the paths and impact of hurricanes or wildfires, or the spread of disease outbreaks.
AI holds the potential to significantly alter our world – which is both exhilarating and frightening.
Before we can fully understand AI’s power to change the future, it’s worth looking at a few events that could have turned out differently with AI.
Think back to the 2022 Russian invasion of Ukraine. Remember the Russian Army’s massive traffic jam north of Kyiv, that stretched up to 35 miles long? If they had real-time traffic monitoring with AI to assist in routing adjustments, and AI-generated topographic data to provide alternative movement options, they would have been more likely to have reached Kyiv. Fortunately in this case, they failed – and they had to retreat.
Back in 2016, Hurricane Matthew hit the Carolinas very hard. FEMA needed detailed assessments of structures, damage levels, transportation and infrastructure conditions, and areas and extents of flooding, but there was a lot of cloud cover – not good for getting full imagery. Had we had GEOINT AI at work, the many hours of work to determine more than 1,300 segments of impassable road, and more than 3,200 areas of flooding, would have been done in minutes – and FEMA would have been able to respond much faster.
Looking at Humanitarian Assistance and Disaster Relief type missions, what does the near future hold thanks to GEOINT AI? Well, FEMA now does this quite well with its Modeling & Simulations and with NOAA predictions, but we’re looking at ways we could assist them considerably more, whenever an advancing cyclone or hurricane is forecasted, and we’re formally requested by civil authorities. Together, we can use higher fidelity geospatial, weather, and previous disaster data to predict how many people will be displaced from their homes, and where they’ll probably move. AI insights can help emergency personnel identify how much food, water, and medical care will probably be needed, and where to send it. And AI models can instantaneously assess flooding, building, and road damage – based on satellite images and weather forecasts. FEMA uses more than just satellite images – airborne, UAVs, and crowd-sourcing, etc. – and the accuracy is quite high, even 90 percent – but it currently takes humans viewing images to get to 100 percent. But AI allows for quicker and more effective distribution of emergency aid, and could assist Urban Search and Rescue missions for those still in danger – people who were isolated from escape routes.
I don’t need to tell those of you from here in California, with multiple fires burning throughout the state, that the same goes for wildfires. The National Interagency Fire Center uses wind and weather data to gage fire locations and their potential spread, which eventually helps the firefighters who put them out. More predictive analysis, using AI, could maximize effectiveness.
All right, let’s pivot to another imperative, and talk about all the ongoing changes within and on our planet Earth, which are the proximate causes for many disasters.
The science dealing with the size and shape of the rock we live on is called Geodesy. It includes the study of the gravitational field, coordinates, and reference frames. And at NGA, it also includes other foundational geophysical elements, such as the magnetic field and subsurface modeling. We all learned in school that the mass of the Moon and the Sun causes the tides on the Earth, and the daily movement of water. The mountains, valleys, crust thickness, and density changes of our planet cause even larger changes in gravitational fluxes. For navigation systems, this change of gravity can cause hundreds of feet – or more – in errors. Using 2.7 grams per cubic centimeter for rock density was good enough for your college physics or geology class – but rock density varies, based on the rocks, as well as changes over time. So we need precise and accurate numbers – five more decimals for gravity values for example – to ensure that our gravitational and magnetic systems will operate accurately in support. These accurate numbers come from performing precise geodetic surveys and measurements everywhere on this rock.
If you take it a step further, Geodesy also includes measuring the orbits of satellites. And it’s essential that the scientists in this field – geodesists – constantly provide updated geodetic products for satellite operations. Otherwise, the satellite applications that we now take for granted would degrade or fail. Satellites need new geodetic products uploaded to them constantly, because the Earth and the satellites are always moving. They’re pushed by solar radiation, and pulled by Earth’s gravitational force, which varies over time and location. And when you’re nearly 11,000 nautical miles away, moving at 7,000 miles an hour, tiny errors in geolocation become huge errors for users of the satellite data.
So there are ways that AI can help here, too – with high-precision digital elevation models – aligning them with topographic data and maps. That eliminates the need for costly and time-intensive stereo extraction. For example, the height of a building, or the slope of a road, can be calculated directly and programmatically from an overlapping elevation model.
Now, here’s the challenge: We have some terrific geodesists at NGA who work hard on all this. But China has embraced the Geodesy field to the extent that it now has more Ph.D. geodesists than the rest of the world combined. Geodesy skills are critical to expertise in our positioning, navigation, and timing missions. They’re essential to such diverse areas as precision missiles, autonomous platforms, next-generation synchronized atomic clocks, and celestial navigation. They’re also embedded into to our way of life – our phones, the GPS systems in our cars, our banking systems. There’s a great quote by the former Chief Scientist of Xerox 's Palo Alto Research Center from back in the early 1990s: “The most profound technologies are those that disappear. They weave themselves into the fabric of everything in life, until they’re indistinguishable from it.” Being the source for precision provides a national, technological, and economic advantage. So we all must look for ways to invest more in Geodesy – with respect to the U.S., our Allies, and our partners.
One final imperative – let’s talk about Climate Security.
We’re trying more than ever to channel our traditional ethos – which has been mostly associated with warfighting – into the humanitarian and disaster arena. We’re exploring future weather and climate conditions on U.S. and partner bases, the impacts from sea level rise on coastal bases (plus tides and storms), and the impacts of extreme heat on flight operations globally. Some of this gets done at the theater and combatant command level, but there’s no reason NGA couldn’t provide this as a global service – given our expertise and data holdings.
At the strategic level, using a 24/7 Watch approach could help us to better integrate our tradecraft across the Human Geography and GEOINT analysis disciplines. Taking that approach could provide more advanced warning for U.S. border security partners, and a closer look at the impacts of weather-related issues – like how food and economic security affect migration. We do work on this already, but there are valid arguments that we could do it more robustly, and take on a leading role in providing GEOINT data and services, to all our partners across the Title 50 and Non-Title 50 Agencies.
One area of concern is sea ice. It’s melting, and fast. And because it’s melting, it’s creating greater access opportunities to areas in the Arctic – like the Northern Sea Route and the Northwest Passage – which is a cause of concern for several nations’ operational security, not to mention maritime safety hazards for navigation.
So, NGA applied AI to this issue, to deliver ways to better track and map sea ice. We developed the Snowfox algorithm, through a partnership with the U.S National Ice Center – to automate monitoring Arctic sea ice characterization, to include extent, concentration, and stage of development – using Synthetic Aperture Radar imagery. Snowfox now helps our analysts process the huge amounts of data we get from the Sentinel 1 RADAR satellite constellation by: Capturing hundreds of pan-Arctic Sentinel 1 SAR images daily, applying a Machine-Learning algorithm, and producing a binary map of white ice and black open water – either “ice” or “no ice.” The output is quick and fast – and about 85 percent accurate – which cuts our time by more than half, and supports analysts at multiple centers and our mission partners. Our customers include the National Ice Center, the US Navy, the US Coast Guard, and the National Oceanic and Atmospheric Administration, or NOAA. And of course, we’ll keep refining Snowfox, with more and greater satellite imagery sources.
To sum it all up, GEOINT AI, Geodesy, and Climate Security are three vital areas where NGA – and by default GIS – can support global unity by improving our predictions, warnings and responses for Humanitarian Assistance and Disaster Response.
Overall, I’m quite hopeful for the future, because I so highly respect the GIS and geospatial community. I truly don’t think there’s a group of professionals on this planet doing more for this planet or more to unite our world than this community. And that applies whether we’re trying to save the world from bad weather, or protect it from bad actors, or both.
So together, we’re going to continue to build great relationships, and brainstorm about our next great collaborative projects. And I strongly believe that we’ll do that expertly, as we chart the way forward, and keep making impressive progress together. Thank you very much.