As extreme weather events and natural disasters place growing pressure on communities and emergency services, having access to the right information at the right time is becoming increasingly important.
Yet vast amounts of geospatial data are already being collected from satellites, drones, sensors and predictive models, raising a key question: how can this information be turned into intelligence that is genuinely useful during a crisis?
In this interview, Guy Schumann, CEO and Founder of RSS-Hydro, explores the role of geospatial intelligence in improving disaster response and supporting faster, more informed decision-making. He discusses the challenges surrounding data accessibility and awareness, as well as the importance of bringing together information from multiple sources to create a clearer picture of developing events.
Schumann also explains RSS-Hydro’s PIN approach, which is designed to transform complex geospatial information into concise, location-specific intelligence that can be delivered quickly to those who need it. He discusses the importance of trust, collaboration and accessibility, and considers how the geospatial intelligence market could evolve as demand for better disaster risk information continues to grow.
Disaster response is often reactive, and information could potentially be out of date by the time it reaches decision-makers. How can geospatial data help us move towards a more proactive approach?
That’s a very good question because, as I mentioned during our last interview, I’ve worked across many different sectors since leaving university, all focused on the same topic: how geospatial data can support disaster management.
I’ve sat through many workshops with disaster teams in the United States and the United Nations (UN), and they all have the same problem. Everybody knows the data is there on a global level. We have enough technology and enough data. But when a disaster is imminent or actually taking place, the information is not reaching the people who need it quickly enough.
That was always highlighted, whether you were talking to the Federal Emergency Management Agency (FEMA), NGOs or the UN World Food Programme. It was always the same message. It’s not necessarily a question of budget.
You might think that if you have the budget, the data will be available more quickly. I think that may be true for defence because they also have very serious contracts. But when there’s a disaster, that geospatial information isn’t there when you need it.
And your other question was how we can use geospatial technology to help make that disappear.
That’s where I think we need to completely rethink the market as such. When we talk about geospatial data or Earth observation, for example, we really need to change the way we deliver things. We also need to be able to provide information faster.
The data can be available quickly, but data itself isn’t necessarily helping anybody. You need to extract useful information from that data stream.
That’s where the “data cocktail” idea comes in. We need to start bringing together geospatial data from different sources, including Earth observation satellites, drones, and model data.
It’s only when we do that that we can potentially close the gap created by missing information when it’s needed. We have the ability to prevent that from happening by mixing or aggregating these different data sources and extracting the information in a consistent structure.
The structure of delivery should not change, but what feeds into that structure needs to change.
If we don’t have observational data, perhaps we can bring in a model simulation. Maybe we can use a forecast or another form of data, such as information from drones if satellites don’t pass over.
We shouldn’t be saying, “Sorry, we couldn’t deliver that because the satellite didn’t pass over Nepal,” for example. If a drone has acquired the data, we should somehow get access to that drone data and feed it into the same system as the other sources.
We then need to make sure the means of delivery is understood by the people who need the information, so that it can easily be transferred and interpreted.
I don’t think it’s true that data isn’t available. Nowadays, someone always has data somewhere. What’s missing is accessibility and interpretation, and those things aren’t happening quickly enough.
We’ve got vast amounts of data from satellites, sensors and models. So right now, what bottlenecks are preventing faster or better decisions?
One is access restrictions to some of that data. Clearly, some data isn’t intended for everybody to have, so we need to be careful. But accessibility, or access rights, is very often a bottleneck.
Another bottleneck is awareness of the data that exists.
RSS-Hydro is very specialised in Earth observation. We understand commercial and public missions; we understand drones, and we work with many technology companies. But in general, awareness of what data is available isn’t where it should be.
A lot of people aren’t aware that, for example, if you don’t have a low Earth orbit satellite available, you could potentially look at geostationary satellites, where one satellite can observe an entire continent at once. That provides very frequent data.
Those data are probably more difficult to use for a disaster requiring high resolution, but you can then look at downscaling algorithms and other technologies.
AI is also very helpful nowadays. Machine learning and AI can be used for very fast downscaling and very fast interpretation to support decision-making.
So I think the first bottleneck is data accessibility, and the other is data awareness. What is actually out there?
You can’t have one person who knows everything, but I think that’s where AI and large language models (LLMs) are very helpful. You can now query an LLM and ask, “There’s a flood somewhere. What data can I access and how can I access it?”
Sometimes it gives you a wrong response, but very often it’s actually not that bad and isn’t far off from telling you what you should do and what data is available.
So, indeed, it’s awareness and accessibility.
How do you think we could increase that awareness?
In my opinion, there are already a lot of organisations dealing with both of these issues.
The Open Geospatial Consortium, or OGC, is one of them. They build standards for geospatial data, data streams and workflows. They follow the FAIR principles, as does the Committee on Earth Observation Satellites, or CEOS, for example.
They look at different datasets and work to make sure they are findable, accessible, interoperable and reusable. I think that’s the right mechanism.
A lot of private-sector companies are now also complying with those principles. They’re making sure data is analysis-ready, reproducible and usable by AI. These are very important things.
However, data accessibility is still not solved.
Awareness has improved significantly, and there has been a lot of progress. We’re also getting very good at standardising geospatial datasets from different sources through the organisations I mentioned.
Space agencies around the world are also working according to the same principles, using the same formats and focusing on interoperable standards.
The accessibility issue is a big problem because a lot of the instrumentation, whether it’s a satellite or an IoT sensor, is often owned by a national entity or sits within a particular jurisdiction.
Making that data open to the world can be more difficult than people think. The asset measuring the data is owned by a state or by some body using taxpayer money.
If you’re the European Commission, the US, Australia, the UK, or another government that is happy to make those data available to the world, that’s fine. But I also perfectly understand that some countries may not want to do that.
You then get into datasets such as river water levels and river flows that may be extremely important but aren’t freely or openly available. Sometimes you can’t even access them if you’re willing to pay for them.
A lot of these data are regarded in some countries as almost state-owned property, or even as information that should remain secret.
Take a reservoir, for example, where water flow is controlled by one entity because it needs the water. That entity may measure the water levels but not want to share those levels or flows with another country. That’s understandable.
Whether it should be that way isn’t a question for me. That’s a question for geopolitical and legal discussions. But data accessibility isn’t easy to solve.
You mentioned it’s a very complex issue, but as the effects of climate change and increasing natural disasters become more global, do you think it could become increasingly important for countries to collaborate on these systems in the coming years?
Yes, I think so, and we’re already seeing that.
We’re seeing stronger cross-agency collaboration between different countries. We’re also seeing more partnerships between private-sector entities, as well as partnerships between NGOs across countries and at the government level.
The shame is that, over the last five years or so, this progress has been overshadowed somewhat by the geopolitical changes that are happening. But cross-collaboration is still taking place, and it’s still very strong.
I would say it’s maybe even stronger than ever. What you’re seeing is countries perhaps going more bilaterally than before. There may be stronger ties between individual countries, with data exchange happening between them.
There’s also a lot of pressure from the general public to get more information about the climate risks we’re now seeing, including wildfires.
Security and safety are also becoming more prominent. This is perhaps being framed more in terms of national security and public safety, whereas in the past you would have heard more about climate and climate change.
It’s the same collaborative effort, but it’s increasingly being framed around national security and safety, which I think is absolutely needed.
If you have wildfires or floods impacting critical infrastructure, and these impacts are increasing, that’s a safety issue and a security issue.
I therefore think there is more cross-collaboration. I don’t see it slowing down. I see an increase in collaboration, including around data sharing.
I think it’s actually happening more than ever. It’s perhaps just less visible because of the geopolitical situation and the difficult circumstances people are facing.
Moving on to how we get this geospatial data into the hands of the people who need it most, RSS-Hydro brings together Earth observation, predictive modelling and real-time data. How does this combination change what emergency managers can do during a developing extreme weather event?
I think it’s probably not the only way. It’s just the way I’m thinking about it, and RSS-Hydro is starting to implement it that way. There are definitely other solutions.
The way we approach it is in two ways.
If we want to get something into the hands of people very quickly and in a very simple way, we still need to actually reach the people who need it.
Let’s say, for example, these are search and rescue teams in another country. We can’t easily contact them and say, “Here’s our data; please use it. It costs you X amount.” Even if they have the budget, we can’t necessarily reach them directly.
Instead, we find partners already serving those people or entities. We then partner with them to eventually get our data into the right hands.
Commercially, we build partnerships and go to market together with others.
The other way we do it is through humanitarian work. Last year, we created our own NGO as a separate entity.
Our society is developing tools, but if some of those tools need to be deployed to reach a sector that can’t pay much for those services because it needs the money elsewhere, and our data and information are equally important, we’ve decided to give that information to the NGO.
The NGO can then determine what’s possible without paying for those data streams.
So those are the two ways we’re operating.
One is very NGO-based and focused on humanitarian work, so we aren’t expecting societal revenue from that channel.
The other is through commercial partnerships, unless we can reach those entities directly. That’s very rare because you can’t easily reach civil protection, search and rescue or disaster management organisations at state, government or county level by yourself.
We’re very strong in building partnerships in the commercial sector.
As you said, your company is doing this, but do you think efforts are underway, especially in Europe, to increase accessibility for underserved communities or those without the budget for these technologies?
Yes, and I think a lot of those initiatives are actually being driven by B2B partnerships.
I think it’s the private sector stepping up in a lot of cases. Whether that’s because there’s a marketing idea behind it and they’re looking to increase visibility, in the end, it doesn’t really matter as long as those who need the data receive it.
Of course, that may not be forever.
I’ll give a good example, and I’m not afraid to name these companies because they’re all doing it. Kudos to those that can easily do it.
I think all the big commercial space companies selling Earth observation data, such as Planet Labs, Vantor, and Capella Space, have had, or still have, open data programmes for when there’s a disaster.
You can actually get free data from these entities. They post it, and you can download images showing the disasters.
Another very good example is the OpenStreetMap community, which uses satellite imagery to map infrastructure. While the imagery itself isn’t available to download, users can access it visually and use it to trace roads, buildings and other features.
These are WorldView images, which are made freely available for this purpose. They can’t be downloaded or owned, but they provide the visual information needed to build and update OpenStreetMap.
I think there’s a lot happening in the private sector that isn’t recognised often enough by governments because the private sector is stepping up in a lot of cases.
Again, whether it’s a marketing opportunity or not, as I said, I don’t mind. If Vantor was providing those images to OpenStreetMap as a marketing opportunity, good for them, because they also helped create an OpenStreetMap community.
Or whether it’s the Airbus Foundation providing free images to the UN. Of course, it’s a foundation. Google is doing it as well, and Microsoft. They all have foundations.
Yes, you may get a tax break, but they should if they make that donation of images, right?
Just because they’re a very large company making a lot of money doesn’t mean they shouldn’t get a tax break when they support those foundation initiatives, because some of them are really good and help a lot.
So I think there are a lot of these things happening in Europe as well as elsewhere in the world.
We spoke about getting this information to the areas that need it most. Your PIN approach turns complex geospatial information into more localised, actionable intelligence. Could you explain the PIN approach and why this shift from data and maps to clear, location-specific information matters during a crisis?
At RSS-Hydro, we cover floods and also fires. Location is always critical, and especially with floods, it’s very obvious.
You can have a flow of water coming down a street. The water may go into one building but not another. It’s actually very local. It’s always very localised.
Any disaster is, in the end, a local problem. Even when you have a continental-scale disaster, it will still affect different localities differently.
So, first of all, localisation, or location, is always a big thing.
Time is also a big factor. Whether you can provide that information to someone quickly or not makes a huge difference.
Then the information content needs to be at a level where somebody sitting behind the wheel of an emergency vehicle can understand it just as quickly and easily as somebody sitting in front of a geographic information system (GIS) screen and interpreting the data coming in.
The PIN is therefore trying to do three things at once.
First, it’s trying to take all the complex information as quickly as possible and reformat it into a very easy-to-understand format.
The parameters we base it on can also be defined by the customer or by the entity that needs the data. They can say, “I want these categories in this X, Y and Z way.”
We then build the algorithms to transform the complex data sources into that format.
The PIN is always location-specific. It can be a summary of the surrounding area, or it can be directly related to a specific asset or location.
The other thing we’re trying to do with the PIN is make the data less heavy.
Instead of sending somebody 10 images in one hour, each one gigabyte in size, we compress that information into the PIN. It can be 1,000 times less heavy to transfer, whether that’s through an S3 bucket, an API or even on somebody’s alphanumeric pager.
We want to get that PIN everywhere without any problem, so we need to reduce the volume of information.
But reducing the volume of information shouldn’t mean that you know less about your disaster because we couldn’t feed all the information through.
The information content needs to retain the same level of excellence. It’s simply presented in a way that is less heavy, less cumbersome and less convoluted to understand.
That’s the big idea behind the PIN. It’s location-based information first.
I imagine that this sector is continuously innovating, especially with AI and machine learning. The technology is constantly evolving. How do we go about building trust with those who will be relying on and using this information? I imagine it’s quite a big shift for them as well, with many of these new technologies.
That is the absolute most difficult thing to do because it needs to be based on trust.
As I said, if we stop sending the satellite image or the actual map and instead send information that is simply text and numbers, we need to make absolutely sure that the entity receiving it trusts us enough to say, “Okay, they really processed the data that they claim they did.”
We also provide the source and other information, so anybody could still look at and question the data behind it if they want.
But you’re absolutely right. Especially in disasters, trust is critical.
We need to make absolutely sure that this trust is built from the ground up. That’s not easy, especially today.
We make the PINs easy to understand because we eventually want LLMs to help interpret them as well.
But building trust has always been difficult. It’s difficult for humans to trust one another, and there can even be cultural differences.
We need to be careful about what data we use as a source.
I’m always saying that if there’s a disaster in a particular country, perhaps we should first look at what that country already has as a national data asset. We should start using those data to create the PIN.
Then, if the PIN goes to that same entity, there’s a greater connection.
Very simply put, we shouldn’t go to the US and say, “Look at all these European Space Agency (ESA) satellites we used,” while completely ignoring the country’s own assets.
We should use everything. We should include the US assets as well.
The same applies in Europe. We can’t go to Europeans and say, “Look, we’ve created this PIN using only American asset data, but you should like it now,” while ignoring all the Copernicus assets.
No, we should use everything.
That’s why I was mentioning this cocktail of information before.
The assumption is that the more information you have, or the more data sources you have, the better you can create that information.
You also naturally increase trust in the data because, if four different assets confirm the same information, trust and confidence increase.
There’s also the trust level between our business and other organisations. That’s difficult to build, but it needs to happen as well. It simply takes a lot of time.
So the issue isn’t necessarily the data itself. As you’ve mentioned before, it’s about how we condense it without losing the quality of the data, because that’s obviously important, and then deliver it in a way that’s tailored to the needs of an individual company, region or agency.
Yes, absolutely.
Everything you’re saying is very logical, but this is really a case of “easier said than done.”
Of course, there’s a huge amount of data available, so why can’t we simply know where a flood is going to happen? First of all, it’s not that easy to do. And secondly, you would need a world where everything is open and available, and ideally freely available.
But somebody is always paying for that data. If it’s free to me, it simply means somebody else has already paid for it, whether that’s a taxpayer or somebody else. There is no such thing as free.
Some data should remain openly accessible because it has already been paid for by someone. But the private sector also provides a huge amount of information and data, and there is a limit to how cheaply those companies can provide it.
I think this is also partly what the AI community has created. You can now write a prompt and expect data or information, but the underlying data and infrastructure have required significant investment to create.
So while information may be freely accessible, that doesn’t mean the data itself was free to produce. That’s an important distinction.
Looking ahead, how do you see geospatial intelligence transforming disaster response? As we’ve said, climate impacts are increasing. Over the next 10 years, how big a role is geospatial intelligence going to play?
I think it’s huge because, if you look at how many satellites are being put up now by the private sector to serve one sector, which is clearly defence, it’s significant.
Then, if you think about what I said before, disasters should have an equal weight in terms of national security impact as defence or an attack.
It should be regarded at the same level. A fire or a flood is probably more destructive than anything, or at least it could have that potential.
If you say that, then you have to say the geospatial intelligence market, I think, will totally explode over the next five years.
I’m asked by a lot of people, whether investors or governments, to provide projections.
They ask, “What’s the total market? What’s the market I can get? What’s the market I can serve? What’s the market I can finally obtain in the next five to 10 years?”
Well, I don’t know if I can give those numbers correctly anymore because I think the world will change so much in two years that a five-year projection could become meaningless.
The only thing I can say is that the geospatial intelligence market, for civilian markets as well as defence, will absolutely explode.
Maybe we’ll reach a trillion-dollar market faster than we think.
SpaceX is a big contributor to those numbers because making launches more accessible is a major reason why geospatial intelligence is expanding.
As I say, I think the numbers you see out there are totally wrong. I think they’re completely underestimated, and this will happen much faster than you’d think.
