Latest / Stage Talks with Bellingcat / "Monitoring the environment from Space" with Lisa Broekhuizen
Transcript
- 0:06You're listening to a stage talk titled Monitoring
- 0:09the Environment from Space. This week we were
- 0:11joined by the technical director from Space for
- 0:14Good as they walked us through the possible practical
- 0:16applications for satellite imagery when it comes
- 0:19to environmental monitoring, explaining how and
- 0:22why satellite images may be the best way to oversee
- 0:25large -scale rapid impacts on our climate and
- 0:28wildlife. This talk was hosted by my colleague
- 0:31Giancarlo Fiorella on Thursday the 10th of April,
- 0:342025 in the Bellingcat Discord server. Welcome
- 0:50everybody. My name is Giancarlo Fiorella and
- 0:52I'm the Director of Research and Training here
- 0:54at Bellingcat. I am filling in for Charlie this
- 0:57week. who is away at a conference. I'm really
- 1:01excited to introduce you to Lisa Bruckhausen.
- 1:05Lisa is the tech lead at Space for Good and is
- 1:11a remote sensing specialist. Lisa and her team
- 1:14use satellite imagery and remote sensing techniques
- 1:17to monitor changes in the environment from space.
- 1:22Now, previously we'd spoken about a tool from
- 1:25a Bellingcat contributor named Chris Giles. And
- 1:29the name of the tool, you might remember, was
- 1:31the OSINT Forest Area Tracker. This is a tool
- 1:34that looks at burned areas of forest in Ukraine.
- 1:38And the purpose of the tool is to give users
- 1:40the ability to look further afield about forest
- 1:45fires in this part of the world. Now, that tool
- 1:47uses something called Sentinel -2 data. You can
- 1:51find it on our GitHub. If you go to GitHub and
- 1:53you look for Bellingcat, this is one of the many
- 1:55wonderful tools you'll find there. Space for
- 1:57Good have just come out with their own application
- 2:00to help people on the ground monitor deforestation
- 2:03from space. And Lisa is here to talk about that
- 2:07application and more of her work at Space for
- 2:10Good. Now, if you'd like to ask a question, please
- 2:13do so in the chat as we talk. I'll be writing
- 2:16down your questions and then asking them to Lisa
- 2:19towards the end of the talk. If you don't want
- 2:22me to read your username when I ask your question
- 2:24to Lisa, please make a note of that in the chat.
- 2:29Lisa, thank you so much for being here this afternoon
- 2:33in the Netherlands. I'm going to hand over the
- 2:36microphone to you now. Thanks very much for having
- 2:40me. Thanks everyone for being here, for listening
- 2:44in and a special thanks for Giancarlo and Charlie
- 2:48for setting this up and of course everyone here
- 2:50for all the good work. So I'm here today to share
- 2:54some of the work that we are doing and share
- 2:58some ideas and some inspiration on hopefully
- 3:01open some eyes on how a lot more can be done
- 3:05with the eyes we have in the sky, the satellites.
- 3:08I'm Lisa. I'm technical director of Space for
- 3:10Good and today I will focus mostly on the practical
- 3:13use cases with satellites. So I'll talk a bit
- 3:16about that, like I do every day and hopefully
- 3:19after this so will all of you. At Space for Good
- 3:23we work with satellite data to monitor the earth.
- 3:27We do this for projects we believe have a positive
- 3:29impact on the planet. I'll talk through some
- 3:31of the work we've done, some big developments
- 3:33and some smaller ones which nevertheless I think
- 3:35have a good impact. and hopefully share some
- 3:39ideas and inspirations on what's possible out
- 3:41there. During this session I will not get technical,
- 3:45but rather provide with a satellite overview.
- 3:48However, you're more than welcome to ask technical
- 3:50questions at the end, hoping for those. And when
- 3:54I say monitor the planet with satellites, I don't
- 3:57just mean working with pictures from space. I
- 4:00mean uncovering the stories. Environmental stories,
- 4:03land use stories, sometimes very personal stories.
- 4:07where the start of the story starts with the
- 4:10data. So first of all I didn't want to talk about
- 4:15Forester, our product to help detect illegal
- 4:18logging near real time in the field. It's a tool
- 4:22to give rapid alerts and really help Foresters
- 4:26on the ground take action rapidly rather than
- 4:29having long -term visuals over trends or regional
- 4:34actions. We rolled this out first of all in Indonesia.
- 4:39We're working with local partners to take direct
- 4:42action. Sometimes we use Sentinel -2 and need
- 4:45the optical data. For this, we use Sentinel -1,
- 4:48which is a radar data set, which means we can
- 4:50penetrate the clouds. Now, if we look at tropical
- 4:53rainforest, then rain kind of in there. There's
- 4:57a lot of precipitation, a lot of cloud cover.
- 5:01That means that with optical data, we often have
- 5:04just see the clouds. With radar data, we can
- 5:06look through the clouds and see changes that
- 5:09are happening. It's a bit less intuitive to read
- 5:10the data, but that's what we have algorithms
- 5:12for. So we look for changes, rapid changes, within
- 5:19areas that we before classified as forest. Within
- 5:24those areas, we look whether it's likely illegal
- 5:27logging or a legal or natural event such as a
- 5:31landslide. If we think it's illegal logging,
- 5:34we send the WhatsApp alert out to people in the
- 5:36fields directly. They go and they take action.
- 5:40So what we've seen before is that they come up
- 5:43with security guards to show the people that
- 5:45they have been spotted and to get the people
- 5:46out. Sometimes they use the reports we generate
- 5:49to go to local authorities and get the police
- 5:51involved. And with that, they can remove the
- 5:54people before a big conversion to agriculture
- 5:57happens, which means that the forest can recover
- 5:59faster and better. We now have a system that's
- 6:03over 98 % accurate, so that's really nice. In
- 6:07this area at least. In other areas we always
- 6:08need to retrain and revalidate because it's not
- 6:12a globally applicable model, but it's a model
- 6:13that really needs to be trained locally because
- 6:16the patterns and the nature is different in different
- 6:18locations. A bit about the stories. So one of
- 6:25the things we hear from our clients is that they
- 6:28can finally detect deforestation that happens
- 6:30away from roads. Imagine here it's a tropical
- 6:34rainforest condition. It's about 200 ,000 hectares
- 6:36of area to monitor. Roads tend to be washed away
- 6:40with rainfall or any other events and suddenly
- 6:43the road is gone. And even if the roads are perfect,
- 6:46yeah, it's a rainforest. You can't look into
- 6:49the forest and see very far. So you can only
- 6:51see a few meters. So before it was almost impossible
- 6:54to see these events because they happen outside
- 6:58of the direct road network, outside of direct
- 7:01fuel. And one more thing that we've seen, which
- 7:04I really like talking about. I don't like the
- 7:09story behind it. But we detect a full satellite
- 7:13view of the whole area, which means that we detect
- 7:15all deforestation events. What they noticed is
- 7:20with that data that around one security guard
- 7:24was positioned at different security posts. Whenever
- 7:27he was positioned, there would be a peak in illegal
- 7:31logging. So likely he let some family members
- 7:34through illegally but that's very difficult to
- 7:37assess if you don't have the full data. So by
- 7:39having that full map it could pinpoint hey here
- 7:42we need to do some more research to identify
- 7:45what's exactly going on. Quite a nice case to
- 7:47see like how this information we never thought
- 7:49something to be pinned down to an individual
- 7:51but that actually was possible with the help
- 7:53of the data and of course merging it with all
- 7:55sorts of other data sets. Now what we do here
- 7:59differs quite a bit from EUDR compliance because
- 8:04we are focused on near real -time direct action
- 8:07and I am assuming directly that almost everyone
- 8:09knows what EUDR is but some of you might not
- 8:11hear about it as much as I do so let me quickly
- 8:13expand EUDR is the European regulation for deforestation
- 8:16free supply chain starting next year probably
- 8:20you never know with EU it's intended to ensure
- 8:24that agricultural products entering the EU are
- 8:27not causing deforestation so that you can eat
- 8:30your bar of chocolate guilt free. Deforestation
- 8:33wise I'm not talking about fair pricing or human
- 8:36rights or calories today, but that it doesn't
- 8:38cause deforestation. EODR has a single definition
- 8:42of forest deforestation land cover where we rather
- 8:47look at trees disappearing. So we don't look
- 8:50at is a plot of forest converted to agriculture,
- 8:53but we look at there were trees here, there are
- 8:56none now, take action. rather than saying there
- 8:59were trees here, there's agriculture, two years
- 9:02later, so you can't import. And then we really
- 9:06focus on enabling people in the field to take
- 9:07direct action. That's a bit about forest application,
- 9:12but deforestation, but it's not all bad. There's
- 9:15also afforestation, the growth of new forest,
- 9:17where new forests are being planted. So, nice
- 9:21story here. without naming specific names I'd
- 9:25like to come into a little case study we did
- 9:27where we got a request from a small tree planting
- 9:29NGO to compare how some of their different sites
- 9:32have been doing they used to do a lot of field
- 9:35work but with environmental more and more information
- 9:39about the flying but also safety concerns for
- 9:42quite a long period they couldn't go into their
- 9:45afforestation areas anymore so we did some analytics
- 9:49for them to compare how different sites were
- 9:52doing, which one was healthy, which one was managed
- 9:54well etc etc and in some sites we found that
- 9:58there were no trees left so the local partners
- 10:01had been receiving money to take care of the
- 10:03trees but had failed to report the death of several
- 10:07full plots so we're gonna go by mistakes happen
- 10:10but thanks to this data the project could be
- 10:12changed around gain additional management and
- 10:16it was made clear to the people locally that
- 10:18there were checkups being done which is, first
- 10:21is good, but validation is better. And of course
- 10:25it's bad that those trees didn't make it, but
- 10:28with satellite data, they could make early interventions
- 10:31and changes into the process. Now in this project,
- 10:35we also made some nice analytics, some quite
- 10:38nice algorithms we developed to compare the different
- 10:40sites, the missing trees. I would like to say
- 10:45that there was this amazing algorithm in -depth
- 10:47assessment. But basically it was looking at a
- 10:49picture and going to a colleague and, do you
- 10:52see trees? I don't see trees. And having figured
- 10:55out that there was a major issue here. Satellite
- 10:58data analytics can be rocket science, but often
- 11:01it's not. And then having that layer of extra
- 11:06accountability, especially in an industry where
- 11:08so much greenwashing is happening, can give that
- 11:12additional trust. And not having that layer...
- 11:16and lead to more scrutiny which at least leads
- 11:19to us all having something to do so that's also
- 11:22a win right? Looking at the current Belenka challenges
- 11:25there's also I also want to quickly point out
- 11:27some other things everything that has a location
- 11:30always grab a satellite image we did a project
- 11:33around illegal zoos small project there was no
- 11:37budget available so we had no commercial data
- 11:39but even looking at 10 meter resolution openly
- 11:43available satellite data It's the Sentinel -2
- 11:46program that's looking at only true color images.
- 11:49We could see the expansion of some of these wildlife
- 11:51breeding facilities as an extra source of information
- 11:56to help with the investigation of that party
- 11:58there. And animals or meat directly leads to
- 12:02biodiversity. We see a spike in biodiversity
- 12:06credits. where people are paying to support biodiversity
- 12:10within different areas, within different conservation
- 12:13sites or new nature developments. What is missing
- 12:18in a lot of cases is objective data. There is
- 12:22quite a lot we can see with satellites that relates
- 12:24to biodiversity. Very rarely you look at individual
- 12:28species, but often you get landscape features,
- 12:31diversity, landscape history, all those things
- 12:34can be mapped. That also brings us to an issue.
- 12:36If we can map multiple things, we can get different
- 12:40outcomes. And which outcome do you share when
- 12:44all of the different outcomes are correct, but
- 12:47also saying something different? I don't always
- 12:49have an answer to that, but I do know that the
- 12:51answer can't always be the most optimistic case
- 12:53or the most profitable case. Even if we don't
- 12:57look at the complexities of biodiversities, but
- 13:00look into carbon credits, which is something
- 13:03measurable theoretically, you can measure a tree.
- 13:05We see billions flowing into the carbon credit
- 13:08market, but are the protected areas even under
- 13:12threat? And that's why satellites are very important
- 13:15to generate a baseline to see what has been happening,
- 13:18because we can look back until the 70s of the
- 13:20last century to see evidence -based what has
- 13:23been going on. We can provide support whether
- 13:28a project is resulting in avoided deforestation
- 13:30or an aforestation, and identify good reference
- 13:34areas, or make strong claim why a reference area
- 13:38chosen, so an area where we compare the prediction
- 13:41site with chosen, is selected wrongly. And the
- 13:46net satellites can give a timeline rather than
- 13:49snapshots to tell the whole story instead of
- 13:51part of the story. We see that a lot with land
- 13:54cover conversion, where land is changing from
- 13:57forest to crop land, from pasture to urban, and
- 14:02we can detect whether that change is cyclic.
- 14:05So if you have shifting cultivation, every year
- 14:08part of a forest is cleared to convert to agriculture.
- 14:13That's not always the same as a fool on deforestation
- 14:16because maybe a year later a different area is
- 14:19cleared and it's part of the system that's being
- 14:21used within the region. But often it's not. Often
- 14:25it's just a cascade effect. Like we see in Indonesia,
- 14:29where we see in certain areas a pattern from
- 14:33primary forest. to the graded forest, to small
- 14:36water fields, to commercial palm oil. And that
- 14:40whole pattern can happen within a few years.
- 14:43Also, I want to quickly talk about invasive species
- 14:45monitoring. We do quite a lot of work in Indonesia.
- 14:48We now started just out the project to map invasive
- 14:51grasslands to support suitable afforestation
- 14:53areas. For us, that's pretty much the same as
- 14:56looking at making a land cover map, saying what's
- 14:59there. It's a bit more complex if we look at
- 15:01very specific species or species group. Almost
- 15:04everything you can see with the naked eye, especially
- 15:07if you can look from above and see it. You can
- 15:10also model and map on large scale with satellites,
- 15:12but also develop triggers, codes that give a
- 15:15trigger if something happens. That'd be quite
- 15:18nice to give warnings. And then I think my current
- 15:21favorite project to talk about is environmental
- 15:24risks. Imagine you spend a lot of time and resources
- 15:28to map a minefield in Ukraine. Imagine a dam
- 15:32breach or storm causing a flood. Mines get shifted
- 15:37with the water. They're relocated. You don't
- 15:39know where. You need to remap possibly a much
- 15:42larger area to figure out where the mines are
- 15:45in order to be cleaning them. With that we're
- 15:47working with the Halo Trust to map environmental
- 15:50risk to support in mine clearance priority setting.
- 15:53So where is a fire likely to happen? or a flood,
- 15:57like a fire you can't exactly send a fire truck
- 15:59into a minefield if there's a forest fire. Or
- 16:02where is a drought likely to happen and what
- 16:04communities will that affect? That's all information
- 16:07that can be helped in this case for better decision
- 16:09making but that's all information that can also
- 16:12be generated globally for different purposes.
- 16:15So really looking into the different risks based
- 16:18on the patterns that we've seen and based on
- 16:20the patterns that we expect in the future with
- 16:23for example climatic differences. And if a fire
- 16:27happens, where, how big, the faster the information
- 16:32can reach the right people, the better the action
- 16:34can be taken. So I can really recommend everyone
- 16:37to, if you haven't already, to look at the NASA
- 16:39Global Database firms. It provides information
- 16:43on patterns and trends and historic fires. So
- 16:47that way we can map and also make predictions.
- 16:49We can give alerts, we can map recovery, always
- 16:52globally and everywhere available. and objective.
- 16:57I say objective, but in this time of deep fakes
- 17:02and fake data, how can we be sure that satellite
- 17:05data can be trusted? Or how can we know when
- 17:09it can't? So far, large datasets have had very
- 17:15few large -scale issues, and I should be careful
- 17:18what I say with this audience, because the satellite
- 17:22data can feel inherently objective. It's from
- 17:25space. It looks like the truth. But as soon as
- 17:29it hits the ground or website, it can be framed,
- 17:33processed, and cherry -picked to give a particular
- 17:36story or a particular angle. And we can't pretend
- 17:40like the data is always going to be correct.
- 17:47So how to make sure the data is not manipulated
- 17:51or misrepresented and how to make sure that you
- 17:54don't accidentally misrepresent the data. Quite
- 17:58simply put, there's a lot of different satellites
- 18:01out there, so don't rely on a single image. Look
- 18:06at the time series, look at what change happened.
- 18:10Is it a sudden or gradual change? Is it a seasonal
- 18:13change? For example, if you look at the deforestation
- 18:16events, you can make very impressive images about
- 18:20European deforestation. by comparing summer and
- 18:23winter images, because the trees lose their leaves
- 18:25and everything looks dead. It's not. Guys, the
- 18:29forest survives. So really look at how does this
- 18:35thread, how is what we're seeing normally looking
- 18:38over time? And how is it looking with multiple
- 18:41sources? So there's a lot of different satellite
- 18:44constellations out there, both public and commercial.
- 18:48Satellites are a witness. and you don't rely
- 18:51on one witness when you've got a dozen out there.
- 18:53We have, for example, very nicely Sentinel and
- 18:57Landsat, where Landsat is hosted by the NASA.
- 19:01Sentinel is from the European Space Agency. So
- 19:04different sources that can look at the same thing,
- 19:07different sensors, different resolutions, et
- 19:09cetera, et cetera, but they can see the same
- 19:11trends. So always have a look at, hey, can I
- 19:14validate this with a different data set? And
- 19:18then if you're truly concerned about the validity
- 19:20of the data, heading back to the original source.
- 19:24So a lot of the data you can find in different
- 19:26places has been processed, which is freaking
- 19:29amazing because I can tell you atmospheric correction
- 19:32is really annoying to do. And usually you don't
- 19:34really need to do it yourself. You can quite
- 19:37easily get that done automatically. But if you
- 19:42have reason of doubt, it's good to check the
- 19:44original data source and go back into that. and
- 19:49to check the full data source. So we're all used
- 19:51to look at true color images, red, green, blue,
- 19:54make a very nice picture. People working a lot
- 19:57with satellite data tend to work a lot with the
- 19:59near infrared band as well, make false color
- 20:01composites. Looks really nice. You can see a
- 20:03lot more changes. But it's a lot harder to fake
- 20:0723 bands of sentinel imagery, including atmospheric
- 20:09effects, than it is to fake a single picture.
- 20:12You'll look back to that single image. that original
- 20:18data set and see, can you validate the whole
- 20:20processing chain or not? And with false position,
- 20:25I think that's more of a misrepresentation usually
- 20:28than a true hacking into the data. If people
- 20:36make strong claims about data, then the models,
- 20:40then the evidence should back that up. If they
- 20:43make a... If they don't share any uncertainty
- 20:46estimates, that's not that relevant. I now share
- 20:49the accuracy of our model in Eurasia and I can
- 20:52tell you that's tested on one side and it's amazing
- 20:55there. But I can also tell you that I can make
- 20:57a global deforestation detection algorithm that's
- 21:0099 % accurate. I would do that. I would say that
- 21:05in the sea, there will never be deforestation.
- 21:09So that's already 60 % done. I will say that
- 21:12outside the forest, there will never be deforestation.
- 21:14So then we only have a small area to check. And
- 21:17actually saying there's less than 1%, there's
- 21:19only a couple of percent of deforestation. That
- 21:21means that if I say there's zero deforestation
- 21:23globally, I have a global model that's going
- 21:26to be 99 % accurate. It's also going to be absolutely
- 21:30useless, but it's going to be accurate. So know
- 21:35when people stroll around accuracy metrics, what
- 21:38to ask for. I can share the accuracy metrics
- 21:41about our model as well if you want, by the way.
- 21:43Those are reliable. And ask those right questions.
- 21:48Or make sure you or someone you can depend on
- 21:52can do the actual analysis. And know that the
- 21:56data can be hacked. Processing pipelines can
- 22:00be tampered with. Cloud platforms can be compromised.
- 22:03Scripts can be manipulated. very easily putting
- 22:08in an altered timestamp, we've all seen that
- 22:11happening, is already giving a completely different
- 22:14view of what the data represents. So that's why
- 22:18open codes, audit trails, their reviewed methods
- 22:21are important. So satellite data is a powerful
- 22:26witness, but it does need cross -examination.
- 22:30It can be mayplated, it can be wrong, but it
- 22:33is objective. And with the right checks, it can
- 22:37also be a very strong ally in the truth. So to
- 22:40end a bit cheesy, the eyes in the sky are watching,
- 22:44but they're on our side if we know where to look.
- 22:47Where this forest lost ghost plantation or land
- 22:50grabbing satellite data can be an ally. So start
- 22:54exploring, ask your questions now or contact
- 22:57me later for more information. If you can't find
- 23:01me based on what I've just said then you shouldn't
- 23:03be in this group. So thank you all for listening.
- 23:08Thanks very much, Lisa, for your presentation,
- 23:11for sharing your knowledge and your experience
- 23:13with us. Folks, if you've got questions for Lisa
- 23:16based on what she just said, please type them
- 23:19in the chat. I'm watching the chat. I will copy
- 23:22your question and ask it to Lisa later on. Lisa,
- 23:26I've got a couple of questions here for you just
- 23:28right off the bat as people get warmed up here
- 23:32with their own questions. Can you tell us a bit
- 23:36about what got you personally interested in satellite
- 23:40data? What was your journey like into this field
- 23:43from maybe a young person in the university or
- 23:47high school to where you are today? How did you
- 23:49get interested in this and how did you end up
- 23:51working at Space for Good? Yeah, thanks. It's
- 23:54been a while since I thought about that. In high
- 23:58school I wanted to really go into nature conservation,
- 24:01going into biology, but then I saw that biology
- 24:04was going to be very much looking into cells
- 24:07in most universities and looking into the cell
- 24:09processes and that was not what drew me. I really
- 24:12wanted to look into ecology and the bigger processes
- 24:15and see how those things work. So then I started
- 24:18to study forest and nature conservation in Wageningen.
- 24:21I did my bachelor's there. I started my master
- 24:23and then I got a bit into GIS remote sensing
- 24:26and spatial modeling. And that was for me really
- 24:29what I was looking for. It's working with the
- 24:33computer, making some models, figuring out how
- 24:35things actually work and can we replicate that.
- 24:39Basically making major conservation into a game
- 24:41and seeing like, how can I alter that? And that's
- 24:44really drew me and then... slowly but surely
- 24:46I started to learn more about these satellites
- 24:49and everything I could see. And basically with
- 24:51everything I did, I was like, okay, but where
- 24:53can I see, what can I see that's going on? And
- 24:56that really got me excited to every time. But
- 24:58the data says, yeah, I'm kind of nerdy. Great.
- 25:03Thanks for that answer. And as a kind of a follow
- 25:06-up, if you're a person, maybe you're listening
- 25:10to this live here in our Discord server, you're
- 25:13listening to this later. and you are interested
- 25:15in getting into this kind of work or you're hearing
- 25:18for the first time, oh, wait a minute, there's
- 25:20these satellite imaging platforms and one of
- 25:22them is called Landsat and the other one is called
- 25:24Sentinel. If you're just learning about this
- 25:28now or you're becoming interested in this now,
- 25:31what are some recommendations that you have for
- 25:32somebody like that to learn more about this field
- 25:35and to maybe get their hands on some satellite
- 25:37imagery and start looking at them? Yeah, so the
- 25:41very, very first step is Google Earth, where
- 25:44you have the time view as well. So you can look
- 25:46back in the past and you can start playing around
- 25:48a bit and get hopefully excited. Then I would
- 25:53recommend to work mostly with Sentinel, Sentinel
- 25:56-2, which is optical data. It's a bit easy to
- 25:58work with and radar data, for example. And it's
- 26:0110 meter resolution, so you can see quite a lot
- 26:03of different things. You can download it either
- 26:05via one of the DS servers or Google Earth engine.
- 26:09Google Earth Engine is still freely available
- 26:14for certain use cases. So that's a nice win.
- 26:19Unfortunately for a lot, not anymore, which is
- 26:21too bad. But then Google Earth, you can, there's
- 26:23a lot of tutorials and a lot of different learnings
- 26:28out there to be looking at things. And yeah,
- 26:31start with seeing if you can map your own house
- 26:34and different bands in different views. Look
- 26:38at any article you're reading like, okay, what
- 26:42is the latest image I can get? Get really, really
- 26:45frustrated by cloud cover and figure out how
- 26:48to deal with it. And then at one point think,
- 26:50okay, now I'm going to need radar data to complement
- 26:52this or now I'm going to need Landsat data to
- 26:54complement this. Have a look and play around
- 26:57with it. It is coding based. You can also look
- 26:59at QGIS, Quantum GIS, which is a software you
- 27:02can install your computer and then just download
- 27:05the data and open it in there. But I recommend
- 27:07working coding based because you can all the
- 27:10things so much more easily. Thanks for all those
- 27:14tips there. By the way, I'm dropping the links
- 27:16to everything that Lisa's mentioning here in
- 27:19our Discord server chat. So if you take a look
- 27:23at the chat there, I've dropped a link to Google
- 27:25Earth Pro, which, yes, is maybe the most beloved
- 27:28and used tool of the open source researcher,
- 27:31whether you're working primarily with satellite
- 27:33imagery or... You're doing geolocations for human
- 27:37rights investigations. The Sentinel Google Earth
- 27:40Engine, I've also dropped a link in the chat
- 27:42there, and QGIS as well. If you're listening
- 27:45to this later and you don't have access to the
- 27:47chat, you just Google for those words that I
- 27:50just mentioned, and that Lisa just mentioned,
- 27:52and then you'll see links to these tools that
- 27:54you can download. We have a question from Fraser,
- 27:58community member Fraser. Hello Fraser, good to
- 28:00see you. Thanks for being here. Fraser has this
- 28:03question. Lisa, Fraser asks, have you had a look
- 28:06at the Leiden guidelines on digitally derived
- 28:10evidence and satellite imagery? Yes, I had a
- 28:16look at it, but not recently enough to be answering
- 28:19this in this. I do know there's a couple of different
- 28:22works out there at the moment, also with CivicSpace
- 28:25looking to launch their own satellite constellation
- 28:27that is not from any... any government or any
- 28:33agency or any commercial entity but it's really
- 28:35looking at evidence -based blockchain validated
- 28:39data. I've had a look at it and I definitely
- 28:42will have a look at it again and I recommend
- 28:44everyone to have a look at it and I will not
- 28:45go further into that. Yeah, oh great, that's
- 28:48an answer. Hey, that's a great answer. Thank
- 28:49you for that answer. Thanks for your question,
- 28:52Fraser. And by the way, if you have any questions,
- 28:55please type them in the chat here. I will ask
- 28:57them to Lisa. As we continue, we got about a
- 29:00half an hour left, so do not be shy. Lisa's an
- 29:03expert in this field. You don't often have a
- 29:05chance to talk to somebody like Lisa, so do not
- 29:08be shy. Ask your questions in the chat, please.
- 29:10I got another question for you here, Lisa. The
- 29:14question is this. Towards the end there, you
- 29:19talked about deep fakes and whatever we want
- 29:22to call AI -generated images and fake data being
- 29:25out there. You mentioned... that there were,
- 29:28I think you said, very few large -scale issues
- 29:30out there with imagery. I know in our own work
- 29:34at Bellingcat, we're often asked, can you think
- 29:37of examples of AI images in the wild that you've
- 29:40detected, that you've worked with? And certainly
- 29:43in our field, we've seen quite a few. So my question
- 29:45to you is, can you think of examples of AI -generated
- 29:49satellite imagery? Are they out there? Have you
- 29:51come across them? Are there people out there
- 29:53who are trying to trick others into looking at
- 29:56a satellite image and saying it shows something
- 29:58but really it's generated by AI? Is that something
- 30:02that you've come across? And if not, is that
- 30:04something that you fear might happen soon? Yeah,
- 30:07I have not come across any big ones I come came
- 30:10across when doing some research for this to also
- 30:12find some an area where someone made a nice fake
- 30:16post about a fire in In a big park which was
- 30:24like obviously fake as in it looked nice, but
- 30:27it was also more made with Photoshop So then
- 30:29there's no data behind it. I do think it's likely
- 30:33to be happening in the future if not now already,
- 30:36especially with high resolution data where there
- 30:39is active tasking. So there's often only one
- 30:41image or very few images to be altered basically.
- 30:47And then if you look at war zones, for example,
- 30:50there's very high stakes. So I can imagine it
- 30:52would be naive to say that it's never going to
- 30:54happen. It's likely to happen. But I think with,
- 30:58especially in the case where you have open data,
- 30:59we have time series of the same data sets, it's
- 31:02easy enough to disprove. oh wait that's not I
- 31:06think that are easy to disprove still get sometimes
- 31:08picked up by the crowd but it is at least easy
- 31:10to disprove so it is something to be conscious
- 31:13of but that's not yeah for the deforestation
- 31:18I don't see it as the direct biggest risk factor
- 31:21yeah thanks for that and you know it's interesting
- 31:24when you were going through the workflow or the
- 31:27suggestions on how to work with images satellite
- 31:31images in this time of fake images, of AI images
- 31:35being out there. The steps that you outlined
- 31:37are the same, many of them are the same as we
- 31:40recommend to people who ask us about, hey, how
- 31:42do I know that an image from a protest is really
- 31:45a real image and not an image from, or how do
- 31:48I know that this picture from a war zone really
- 31:50is an actual picture from a war zone and not
- 31:53a fake picture? You know, you talked about looking
- 31:55for other images, right? So you called it the
- 31:57time series, like look at other pictures that
- 31:59apparently show the same thing. Because I think
- 32:03I suspect that as with pictures from war zones
- 32:07or from protests or from anywhere that are generated
- 32:10by AI, the platforms can't, at least I'm not
- 32:13aware of this, can't replicate the same scene
- 32:16from different angles. But as you said, it's
- 32:20kind of naive to think it's not going to happen.
- 32:22I suppose for satellite imagery, something that'd
- 32:25be interesting to think about is satellite images
- 32:30are... They show intricate patterns right like
- 32:34roads and houses and and and I wonder if the
- 32:39AI image generation technology is so good that
- 32:41it can actually accurately mimic those and I'm
- 32:44thinking specifically about the cover of the
- 32:47Berkeley protocol book from the Human Rights
- 32:51Center at Berkeley. The cover of that book is
- 32:53actually an AI -generated image of a satellite
- 32:57view. But if you take a close look at it, and
- 32:59an expert eye like yours could definitely tell,
- 33:02hey, this is weird. I've looked at a million
- 33:03satellite images. None of them look like this.
- 33:06But as you say, it's one of those things you
- 33:08have to think about as being potentially possible
- 33:10in the next little while. There, indeed, my risk
- 33:14is there is especially with open satellite data,
- 33:18there is a lot of data available to train on.
- 33:20And a couple of years ago, the generation of
- 33:25deepfakes was very, very early on. And if you
- 33:28see how fast that goes, how good video material
- 33:31we can get. I'm not sure why I'm here. I could
- 33:33have just sent an AI. How good that all gets.
- 33:38Then I think, yeah, the satellite data, there
- 33:40is a lot of data out there. is going to be feasible
- 33:42as well. The patterns at the moment, I don't
- 33:44think yet, but I'm talking to some people who
- 33:46are working on how to connect the roads, which
- 33:49is also very useful if you want to map a road
- 33:52network in an area. It's nice to be able to map
- 33:54those roads. But yeah, it can be used to generate
- 33:57fakes as well. And if we talk about the wrong
- 33:59data, I see at the moment a bigger issue in misrepresented
- 34:05data. having a data set and saying it's and that's
- 34:09something that we do see what people say claim
- 34:11it's in a different area and again you can say
- 34:14or a different time and you can say no wait i've
- 34:16got the evidence that it's not but if the information
- 34:19is out there coming with the evidence isn't always
- 34:21enough yeah thanks for that lisa i have another
- 34:25question here since we're talking about you know
- 34:27we're talking about what may be coming in the
- 34:29future in terms of ai image generation you've
- 34:32just mentioned this ability to make realistic
- 34:35roads, for example. Are there any advances in
- 34:39remote sensing technology that you think are
- 34:42coming soon or in the next couple of years that
- 34:44you're really excited about? A lot, a lot. So
- 34:49what's happening right now is Google has just
- 34:52announced a better version of a system to be
- 34:56talking to maps, basically. So rather than doing
- 35:00all the investigation yourself, being able to
- 35:02just ask, hey, for this area of interest, how
- 35:05much deforestation has there been? I think there's
- 35:07a big risk there, but I also think there's a
- 35:08lot of opportunities there to take that first
- 35:11step into how can we actually see what's going
- 35:14on on our planet? I think the roads, that's for
- 35:17me to make the system of automatically connecting
- 35:22different features. based on a lot of data, but
- 35:26also based on quite complex models. Having a
- 35:29detailed road network can mean so, so much for,
- 35:33for example, detecting illegal logging, but also
- 35:36predicting it. We see that illegal logging almost
- 35:38always happens near roads and almost always first
- 35:41a new road is made and then logging happens and
- 35:44it expands. If you can see that earlier by having
- 35:46just a few points. then you can really map out
- 35:49a lot so that's really interesting to me and
- 35:52yeah biodiversity making some sort of integrated
- 35:56solution how can we map biodiversity to well
- 36:00it's not really clear what biodiversity there's
- 36:03not one definition of biodiversity but to be
- 36:05able to say hey this is going well and this is
- 36:09not going well Thanks, Lisa. We have a question
- 36:13from Chris. And if you have a question, folks,
- 36:15please type it in the chat. We've got about 20
- 36:17minutes left here with Lisa from Space for Good.
- 36:19Chris is asking this question here, Lisa. Now,
- 36:23there's some jargon in here that I don't know.
- 36:25So I'm going to read it out as it is written.
- 36:28And maybe this is the jargon that you know. And
- 36:31the question is this. I'm curious to hear more
- 36:33about how you are ground -truthing satellite
- 36:36data to help detect distortions. So we ourselves
- 36:42are not earlier, but let me first like phrases
- 36:45a bit for the rest. So ground rooting that's
- 36:47looking at what's on the ground and perceiving
- 36:49that as the truth and you can use that to retrain
- 36:52models. I think here the distortion, the distortions,
- 36:56but Chris, chat if it's wrong, is really looking
- 37:00at the distortions of the images, which is why
- 37:04you're looking at, for example, something that
- 37:07we've seen in the past quite a lot, the earlier
- 37:10generations of planet data had a lot of misalignment
- 37:14of different bands now i'm getting a bit technical
- 37:17here so that's when you have a picture of the
- 37:20red green and blue and they're all not exactly
- 37:23aligned so you could have a situation where the
- 37:27red value of a certain pixel was 10 meters below
- 37:30blue value so you would have a shift there with
- 37:34ground rooting you can basically align pictures
- 37:37so they're all exactly at the same spot. And
- 37:40then you can look for example, changes and then
- 37:42you know where something is exactly. Usually
- 37:45that is done by the satellite data providers.
- 37:47And I think right now, most of those have good
- 37:49enough on -train models that we can focus on
- 37:52our own strengths, which is mostly the data analytics
- 37:57and going into depth on the AI models. Thank
- 38:03you for that question. And thanks for that answer,
- 38:05Lisa. I have another question. What is your kind
- 38:11of biggest frustration about the limitations
- 38:13of the current technology? What's the thing that
- 38:15you run across the most that makes you go, oh
- 38:17my gosh, I wish it wasn't like this. I wish we
- 38:20could fix this issue with the technology. What's
- 38:23something that's missing in the technology or
- 38:25a capability that's missing at the moment? Yeah,
- 38:29it really depends on use case to use case. What
- 38:32is, for me, I think my biggest frustration is
- 38:36not as much in the technology, but in the technology
- 38:38communication. So I'm spending a lot of time
- 38:42with different users and different potential
- 38:43users talking about, hey, what can be done and
- 38:46what cannot be done. And then sometimes you have
- 38:48people who say, who think we're working with
- 38:51satellites and AI. So I want to see this individual
- 38:55tree and how well it's doing. And sometimes you
- 38:58have people who still think like satellites and
- 39:00AI, no, that's not working in forest, forest
- 39:02is special. It's somewhere in the middle and
- 39:05that's always really annoying. If we go back
- 39:07to really technical, I think a lot is possible.
- 39:12For me, the challenges with working on a large
- 39:17scale with high resolution data, especially with
- 39:20time series, are still really annoying because
- 39:23I know that there's a lot of data out there,
- 39:25but most of it is commercial and then it tends
- 39:28to be nuts. available for a lot of the applications
- 39:30that I feel need it most. You can do a lot if
- 39:33you want to go into oil and gas, but if you want
- 39:36to make sure there's a world left, then it's
- 39:39a constant uphill battle that way. Again, maybe
- 39:43not the most technical issue, so I think, yeah.
- 39:47I kind of like the technical challenges. Yeah,
- 39:49clouds are annoying, but it's part of the fun,
- 39:52right? Gives us a puzzle to solve. Yeah, whenever
- 39:56we work with, we have a really wonderful network
- 39:59of volunteers and sometimes we get satellite
- 40:04imagery from a satellite image provider that
- 40:07they request. They'll be working on a case and
- 40:08they'll go, hey, I need a picture from here.
- 40:10And if the volunteer, if the individual isn't
- 40:13familiar with satellite imagery, we have lots
- 40:16of folks who've never worked with it in the past
- 40:18and are now working with it for the first time
- 40:21at BellyCat. you know, 95 % of the time we get
- 40:23the image, we send it to them and we go, it's
- 40:25cloudy. Like we could, you know, here's the result,
- 40:27or you can't see anything. And they go, oh, of
- 40:28course, like there's clouds. Yeah, there's lots
- 40:30of clouds. And depending on the place, there's
- 40:32clouds like every day. So, yeah, definitely sympathize
- 40:36with that. We did a project in Ivory Coast and
- 40:40I think in a year's time we could get one image.
- 40:43Yeah, that sounds about right. Yeah. We have
- 40:46another question here from another community
- 40:49member named BQN. And BQN asks, Lisa, what is
- 40:54the most interesting or innovative or unusual
- 40:57or creative application of satellite image data
- 41:01that you have come across? So I kind of want
- 41:09to talk about this project we've done. It wasn't
- 41:11technically with satellite image because here
- 41:13we could get through a partner, really like the
- 41:16project as well, some aerial data. But basically
- 41:20for us, the processing is pretty much the same.
- 41:22It's a bit less atmospheric correction needed,
- 41:24but data is data, right? So I was working with
- 41:28aerial data and that was to map out latrines.
- 41:32Where MREF held Africa, this project, this wash
- 41:34intervention to support, to give out micro credits
- 41:39for people to build latrines. And they wanted
- 41:42to know how effective the project was. Now they
- 41:46needed to do a comparison to a different site.
- 41:49to see how many latrines people would normally
- 41:51build, at least would build us outhouses. And
- 41:55it's kind of invasive to be knocking on people's
- 41:57door and ask, hey, can I see your toilets? So
- 42:00they were looking for a different solution and
- 42:02then we developed a model to see, hey, where
- 42:05can we see outhouses and how can we distinguish
- 42:08latrines from chicken coops, which is very challenging.
- 42:12And I really like talking about that, like looking
- 42:14at the different angle, like, yeah, sure, sure,
- 42:16we can map latrines, why not? Yeah, I don't think
- 42:20I've ever heard of anything like that. That's
- 42:22fascinating. Thanks for sharing that. We have
- 42:25a question here from Aikan. Hello, Aikan. Thanks
- 42:31for your question. And the question is, do you
- 42:34have any tips for using platforms like Google
- 42:36Earth or Sentinel for identifying and confirming
- 42:41smaller scale environmental sites like artisanal
- 42:46mines, for example? So artisanal minds, we've
- 42:51also done some work in. I believe there might
- 42:54even be some tutorials using it in Google Earth
- 42:57Engine. Otherwise, what I would really look at
- 43:01is look for bare soil index. So really look at
- 43:05bare soil and mapping changes into that. So looking
- 43:09at the time series analysis, you can look into,
- 43:11for example, BFAST, B -F -A -S -T. that's normally
- 43:18done on the NDVI but you can also do it on for
- 43:21example bare soil index and using that to map
- 43:25where you see for example bare soils popping
- 43:27up in certain area sizes so you want to look
- 43:31at then quite small area sizes where there's
- 43:33quite a rapid change ideally you make some some
- 43:38smart selections about your area of interest
- 43:40where you expect it to happen so for example
- 43:42close to rivers and then Second point of advice
- 43:46is don't be afraid to get dirty and just click
- 43:50through images and look at things. So use these
- 43:52algorithms. If you don't have a lot of training
- 43:54data, don't spend too much time on building complex
- 43:57models. If you don't have the data to put in
- 43:59there and the model is not going to be great
- 44:01anyway and click through things, look at the
- 44:05different images and use your own eyes as well
- 44:08because our own eyes are very good at change
- 44:10detection and pattern recognition. Use the computer
- 44:15to take the first step, but don't try to get
- 44:17everything done perfectly Rather just look and
- 44:20click and do the dirty work yourself as well
- 44:23We have a question now from wit Fox and you've
- 44:29talked You know, we just talked a little bit
- 44:32about this. Let's let's see if we can get more
- 44:35On this Lisa. You've talked a bit about deforestation
- 44:39biodiversity fires mines. We talked about latrines
- 44:42a moment ago What are some things that you think
- 44:45satellites are being underutilized for? Or what
- 44:49are some other perhaps non -obvious use cases
- 44:53for satellite imagery data, either optical or
- 44:58multispectral? Thanks. Take your moment. Because
- 45:06I know satellite data is being used for loads
- 45:10of different things. even for mapping penguin
- 45:13colonies by looking at their shit you can do
- 45:15with satellite so what they're not being used
- 45:18for is kind of difficult to look at yeah I think
- 45:26a lot of places a lot more can be done I think
- 45:30for mapping these environmental claims for big
- 45:35organizations I think they can be used a lot
- 45:38more because I feel like for a lot of these cases
- 45:42companies will make a claim, for example, in
- 45:46the offsetting. They will share some sort of
- 45:49location with it or they will share some sort
- 45:50of report with it. But the data tends to be based
- 45:53on field work, which is nice, but which must
- 45:56be validated. We've seen it a lot where even
- 45:58companies we work with who do field work professionally
- 46:01and who do it for the project we asked them,
- 46:06give us data where there's just errors in there.
- 46:10That makes me wonder how much validation is being
- 46:14done to support these claims for, for example,
- 46:18the regulated carbon market. I know it's being
- 46:21done there, and I know it's kind of an open and
- 46:23short case, but I think more can be done there.
- 46:28Thanks, Lisa. We have about 10 minutes here left
- 46:30with Lisa Brughausen from Space for Good. If
- 46:33you have any questions, please type them in the
- 46:35chat. You've got 10 minutes to ask her. A question,
- 46:39we've got one here from SirSkippy. And SirSkippy
- 46:42asks, do you make the data that you collect available
- 46:46on OpenStreetMap for ECUs by others? No, we do
- 46:53not. And the data we collect, we don't have our
- 46:56own satellites. So the data we collect is the
- 46:58satellite data itself, the raw satellite data.
- 47:01And the data, the results, the information we
- 47:04generate We don't make available and I think
- 47:08so we also work for, for example, commercial
- 47:10entities to take their direct action. I think
- 47:14in most of those cases, it would not be beneficial
- 47:17to them to publish it. And then we also need
- 47:19to be cautious with that and rather focus on
- 47:21the direct action and on the mapping. Plus in
- 47:25OpenStreetMap, if we do, for example, deforestation
- 47:27detection. So like I said, what we do is we look
- 47:30at deforestation as there was a forest it's gone
- 47:33now there's no more forest if you look at most
- 47:37land cover maps they look at what is the land
- 47:41cover we the official land cover that's here
- 47:45the land use that's here so if you cut a forest
- 47:48and a new forest grows it is still the land cover
- 47:51class of forest and then there's no rapid change
- 47:55for change that we detect and then in the other
- 47:58land cover mapping we tend to have a bit more
- 48:00classes than are relevant for This road networks
- 48:03could make sense. But even then I think that
- 48:08working with the GPS tracker so people have locally
- 48:10are probably going to be more, uh, more directly
- 48:14applicable also because it's in private areas.
- 48:17So they don't want the date. They don't want
- 48:18people on the roads. So no, we do not. The short
- 48:22answer. Thanks, Lisa. We got another one here
- 48:24from, um, uh, this one's from Charlie. Um, are
- 48:30you Charlie from, from Bellingham? And Charlie
- 48:32asks, I'm going to paraphrase here. Imagine you're
- 48:36listening to this and you go, wow, Space for
- 48:38Good sounds incredible. I think I have some skills
- 48:40that might make me a candidate. I'm going to
- 48:43look for postings on the Space for Good website.
- 48:45Maybe if they're hiring somebody, I'm going to
- 48:47apply. And the question to you, Lisa, is what
- 48:49kinds of skill sets do you look for in team members?
- 48:52Just in case somebody in the audience here wants
- 48:55to maybe give a shot at working with Space for
- 48:59Good. Yeah, thanks. So what we're looking for
- 49:03is really in -depth remote sensing and modeling
- 49:08skills. So if we look for a data analyst or data
- 49:10scientist and we look for people who really have
- 49:12experience in the data, but also have experience
- 49:14in one or two fields of interest. So we have,
- 49:19I myself have a mixed background with forestry.
- 49:22We have someone who has a background in doing
- 49:24carbon projects in the fields, getting that access
- 49:27with someone who's done a lot of work around
- 49:29fires in the past, agriculture. So knowing, and
- 49:32that's where I think most value lays in, combining
- 49:35the data sites with direct action in the field
- 49:39with something, someone who knows the background.
- 49:41If you don't know how a tree grows, then monitoring
- 49:44from above is going to be a lot harder and monitoring
- 49:47it when it's gone is going to be very difficult
- 49:50to assess. So that's ready for the data. They're
- 49:54the analytics, they're the scientist position.
- 49:57Then we have, if we work with, we have, for example,
- 50:00at the moment, one full stack developer that
- 50:03was looking for someone with cloud experience,
- 50:06scalability, et cetera, et cetera. But for us,
- 50:09the soft side is also very important where you've
- 50:11got a very cool team, if I may say so, and really
- 50:15working together as a team to build something
- 50:17that's better. And then, of course, around commercial,
- 50:20well, we've got a very small team around that,
- 50:22but that's also always good to have people who
- 50:24can try to get it to the final users and clients.
- 50:31Thanks very much. We have about eight minutes
- 50:32left here with Lisa. Get your questions in, folks.
- 50:35We have a question from Dusty Wieners. By the
- 50:38way, Wieners, if you folks don't know. is the
- 50:42skin on your elbow is the weenus. I'm not sure
- 50:45if you are familiar with that term, but Dusty
- 50:48Weenus is asking the following question. They
- 50:54say, Dusty Weenus says, is there any way to verify
- 50:57direct carbon emissions from industrial sites
- 51:00from space? For example, A company claims a certain
- 51:06amount of scope one emissions, but the satellite
- 51:08data actually indicates a larger amount. Is that
- 51:11possible? For example, methane mapping. Thanks,
- 51:16Dusty or Venus, I'm not sure how to refer to
- 51:19you. For example, for methane, there's some satellites
- 51:22out there that can really map at different resolutions
- 51:24the direct methane emissions. It's always important
- 51:27to look at then the full scale of events. If
- 51:30they report CIRTOSCOPE -1 emissions, you need
- 51:32to have the correct overpass times, the correct
- 51:34models behind it. And there's always the issue
- 51:36that you look with the satellite at the full
- 51:38column. So at the entire atmosphere, basically.
- 51:42So there is a potential for a mix -up. However,
- 51:47these satellites have been used, for example,
- 51:49to identify when there have been methane leaks
- 51:51or large -scale methane emissions where they
- 51:54shouldn't have been due to any technical issues.
- 51:58So yes, it can be done for larger scale operations.
- 52:02The farmer next door with 20 cows doesn't exist
- 52:06anymore. But anyway, you're not going to be able
- 52:08to find that. But for the large factories, yes.
- 52:12And for the pipelines where there might be maybe
- 52:17slight issues with the maintenance as well. I'm
- 52:21being positive here. Thanks. Thanks Lisa. I think
- 52:24this is going to be the last question here. Somebody
- 52:26notices we're talking about, you know, work in
- 52:28the field and maybe working space for good. Somebody
- 52:31noticed that there's an astropreneur position
- 52:34on the website. Are you familiar with that one?
- 52:36I'm asking it because Chris is asking about this,
- 52:38but also I've never heard of that term. Do you
- 52:40know about the position, the vacancy? Can you
- 52:42talk about it? It might be a word we made up.
- 52:50More on the data side than that side. So let
- 52:55me quickly, where is that? It's in the, yeah,
- 53:02spaceforgo .com forward slash vacancies. Marketing
- 53:06and sales, market research and analysis, finance
- 53:10and investment. Oh yeah, talented and engaged
- 53:12entrepreneurs. So that's basically an open, always
- 53:16open vacancy where people who are interested
- 53:18can. let us know and we can see whether there
- 53:21is there is anything out there uh that always
- 53:23depends a bit on projects on current needs on
- 53:27skills whether we can fit something in or not
- 53:29um lately usually not i have to admit because
- 53:33we're with quite quite a strong team at the moment
- 53:36uh but quite frequently we can find figure out
- 53:39something so um yeah entrepreneur combination
- 53:44of uh astro and entrepreneur so also looking
- 53:47at people who want to We do work as well with
- 53:49a shell of freelancers who take specific projects
- 53:52or who we support in giving specific projects
- 53:56when there are specific skill sets we don't have
- 53:59in the team, but we need in certain projects
- 54:01or that just want to be somehow connected because
- 54:05we are a lot of fun. Great. Thank you so much
- 54:08for that. And we've got this last question here,
- 54:10folks, again, from SIRS -KP, has the change in
- 54:12US government funding priorities affected Space
- 54:16for good at all. We get that question about again
- 54:17all the time. We're passing it on to you Lisa
- 54:21space for good. Yeah, you don't have to ask the
- 54:23book Affected us. I mean emotionally financially
- 54:30directly or indirectly. I think it's affecting
- 54:32affecting the entire world At the moment, it
- 54:36has not affected any of the current of the projects
- 54:39that we do. And that's also where we try to have
- 54:41a mix between commercial projects, projects with
- 54:44NGOs and governmental projects. We do know that
- 54:47a lot of our partners have been affected quite
- 54:49strongly and we do see quite some risk here because
- 54:52the budgets are also affecting the data sets.
- 54:56And then we're looking at the climatological
- 54:58data sets, which we use if we make predictive
- 55:00models. It's important to have those available
- 55:03globally rather than only in certain regions
- 55:06because the climate is kind of a global thingy.
- 55:11So there we do expect it to affect us on the
- 55:15data set quite a bit and on the projects not
- 55:18on the active projects we have but on the yeah
- 55:22future ones unfortunately I think that's yeah
- 55:26that's great well On that depressing note, we're
- 55:31going to call it. Thank you so much, Lisa. Brooke
- 55:34Hellsend from Space for Good. Thanks for your
- 55:37wonderful presentation. I learned a lot. I know
- 55:39folks here learned a lot as well. Thank you very
- 55:40much, Lisa, for spending this hour with us. Thank
- 55:43you. And have a good evening and day, everyone.
- 55:45Thank you for listening to the Stage Talk. If
- 55:48you'd like to catch a Stage Talk live where you
- 55:51can ask the guest questions, join the Bell and
- 55:54Cat Discord server by visiting www. The music
- 56:02you've heard is titled Dawn by Newer Self and
- 56:06is courtesy of Artlist.