WEBVTT 00:00:06.319 --> 00:00:09.259 You're listening to a stage talk titled Monitoring 00:00:09.259 --> 00:00:11.919 the Environment from Space. This week we were 00:00:11.919 --> 00:00:14.300 joined by the technical director from Space for 00:00:14.300 --> 00:00:16.800 Good as they walked us through the possible practical 00:00:16.800 --> 00:00:19.460 applications for satellite imagery when it comes 00:00:19.460 --> 00:00:22.539 to environmental monitoring, explaining how and 00:00:22.539 --> 00:00:25.699 why satellite images may be the best way to oversee 00:00:25.699 --> 00:00:28.640 large -scale rapid impacts on our climate and 00:00:28.640 --> 00:00:31.420 wildlife. This talk was hosted by my colleague 00:00:31.420 --> 00:00:34.700 Giancarlo Fiorella on Thursday the 10th of April, 00:00:34.859 --> 00:00:50.320 2025 in the Bellingcat Discord server. Welcome 00:00:50.320 --> 00:00:52.579 everybody. My name is Giancarlo Fiorella and 00:00:52.579 --> 00:00:54.899 I'm the Director of Research and Training here 00:00:54.899 --> 00:00:57.600 at Bellingcat. I am filling in for Charlie this 00:00:57.600 --> 00:01:01.500 week. who is away at a conference. I'm really 00:01:01.500 --> 00:01:05.540 excited to introduce you to Lisa Bruckhausen. 00:01:05.879 --> 00:01:11.099 Lisa is the tech lead at Space for Good and is 00:01:11.099 --> 00:01:14.219 a remote sensing specialist. Lisa and her team 00:01:14.219 --> 00:01:17.680 use satellite imagery and remote sensing techniques 00:01:17.680 --> 00:01:21.480 to monitor changes in the environment from space. 00:01:22.459 --> 00:01:25.459 Now, previously we'd spoken about a tool from 00:01:25.719 --> 00:01:29.140 a Bellingcat contributor named Chris Giles. And 00:01:29.140 --> 00:01:31.599 the name of the tool, you might remember, was 00:01:31.599 --> 00:01:34.840 the OSINT Forest Area Tracker. This is a tool 00:01:34.840 --> 00:01:37.799 that looks at burned areas of forest in Ukraine. 00:01:38.579 --> 00:01:40.480 And the purpose of the tool is to give users 00:01:40.480 --> 00:01:45.219 the ability to look further afield about forest 00:01:45.219 --> 00:01:47.620 fires in this part of the world. Now, that tool 00:01:47.620 --> 00:01:51.269 uses something called Sentinel -2 data. You can 00:01:51.269 --> 00:01:53.549 find it on our GitHub. If you go to GitHub and 00:01:53.549 --> 00:01:55.390 you look for Bellingcat, this is one of the many 00:01:55.390 --> 00:01:57.810 wonderful tools you'll find there. Space for 00:01:57.810 --> 00:02:00.510 Good have just come out with their own application 00:02:00.510 --> 00:02:03.909 to help people on the ground monitor deforestation 00:02:03.909 --> 00:02:07.409 from space. And Lisa is here to talk about that 00:02:07.409 --> 00:02:10.310 application and more of her work at Space for 00:02:10.310 --> 00:02:13.090 Good. Now, if you'd like to ask a question, please 00:02:13.090 --> 00:02:16.349 do so in the chat as we talk. I'll be writing 00:02:16.349 --> 00:02:19.810 down your questions and then asking them to Lisa 00:02:19.810 --> 00:02:22.310 towards the end of the talk. If you don't want 00:02:22.310 --> 00:02:24.849 me to read your username when I ask your question 00:02:24.849 --> 00:02:28.150 to Lisa, please make a note of that in the chat. 00:02:29.250 --> 00:02:33.030 Lisa, thank you so much for being here this afternoon 00:02:33.030 --> 00:02:36.090 in the Netherlands. I'm going to hand over the 00:02:36.090 --> 00:02:40.280 microphone to you now. Thanks very much for having 00:02:40.280 --> 00:02:44.639 me. Thanks everyone for being here, for listening 00:02:44.639 --> 00:02:48.199 in and a special thanks for Giancarlo and Charlie 00:02:48.199 --> 00:02:50.800 for setting this up and of course everyone here 00:02:50.800 --> 00:02:54.680 for all the good work. So I'm here today to share 00:02:54.680 --> 00:02:58.759 some of the work that we are doing and share 00:02:58.759 --> 00:03:01.840 some ideas and some inspiration on hopefully 00:03:01.840 --> 00:03:05.439 open some eyes on how a lot more can be done 00:03:05.439 --> 00:03:07.599 with the eyes we have in the sky, the satellites. 00:03:08.340 --> 00:03:10.719 I'm Lisa. I'm technical director of Space for 00:03:10.719 --> 00:03:13.740 Good and today I will focus mostly on the practical 00:03:13.740 --> 00:03:16.219 use cases with satellites. So I'll talk a bit 00:03:16.219 --> 00:03:19.460 about that, like I do every day and hopefully 00:03:19.460 --> 00:03:23.900 after this so will all of you. At Space for Good 00:03:23.900 --> 00:03:26.780 we work with satellite data to monitor the earth. 00:03:27.080 --> 00:03:29.000 We do this for projects we believe have a positive 00:03:29.000 --> 00:03:31.860 impact on the planet. I'll talk through some 00:03:31.860 --> 00:03:33.560 of the work we've done, some big developments 00:03:33.560 --> 00:03:35.659 and some smaller ones which nevertheless I think 00:03:35.659 --> 00:03:39.310 have a good impact. and hopefully share some 00:03:39.310 --> 00:03:41.650 ideas and inspirations on what's possible out 00:03:41.650 --> 00:03:45.090 there. During this session I will not get technical, 00:03:45.370 --> 00:03:47.270 but rather provide with a satellite overview. 00:03:48.629 --> 00:03:50.550 However, you're more than welcome to ask technical 00:03:50.550 --> 00:03:54.469 questions at the end, hoping for those. And when 00:03:54.469 --> 00:03:57.370 I say monitor the planet with satellites, I don't 00:03:57.370 --> 00:04:00.009 just mean working with pictures from space. I 00:04:00.009 --> 00:04:03.590 mean uncovering the stories. Environmental stories, 00:04:03.909 --> 00:04:06.729 land use stories, sometimes very personal stories. 00:04:07.129 --> 00:04:10.629 where the start of the story starts with the 00:04:10.629 --> 00:04:15.370 data. So first of all I didn't want to talk about 00:04:15.370 --> 00:04:18.850 Forester, our product to help detect illegal 00:04:18.850 --> 00:04:22.970 logging near real time in the field. It's a tool 00:04:22.970 --> 00:04:26.310 to give rapid alerts and really help Foresters 00:04:26.310 --> 00:04:29.790 on the ground take action rapidly rather than 00:04:29.790 --> 00:04:34.689 having long -term visuals over trends or regional 00:04:34.689 --> 00:04:38.560 actions. We rolled this out first of all in Indonesia. 00:04:39.000 --> 00:04:42.620 We're working with local partners to take direct 00:04:42.620 --> 00:04:45.959 action. Sometimes we use Sentinel -2 and need 00:04:45.959 --> 00:04:48.120 the optical data. For this, we use Sentinel -1, 00:04:48.339 --> 00:04:50.819 which is a radar data set, which means we can 00:04:50.819 --> 00:04:53.879 penetrate the clouds. Now, if we look at tropical 00:04:53.879 --> 00:04:57.800 rainforest, then rain kind of in there. There's 00:04:57.800 --> 00:05:00.920 a lot of precipitation, a lot of cloud cover. 00:05:01.399 --> 00:05:03.959 That means that with optical data, we often have 00:05:04.110 --> 00:05:06.430 just see the clouds. With radar data, we can 00:05:06.430 --> 00:05:09.209 look through the clouds and see changes that 00:05:09.209 --> 00:05:10.949 are happening. It's a bit less intuitive to read 00:05:10.949 --> 00:05:12.790 the data, but that's what we have algorithms 00:05:12.790 --> 00:05:19.050 for. So we look for changes, rapid changes, within 00:05:19.050 --> 00:05:24.250 areas that we before classified as forest. Within 00:05:24.250 --> 00:05:27.550 those areas, we look whether it's likely illegal 00:05:27.550 --> 00:05:31.209 logging or a legal or natural event such as a 00:05:31.209 --> 00:05:34.000 landslide. If we think it's illegal logging, 00:05:34.360 --> 00:05:36.040 we send the WhatsApp alert out to people in the 00:05:36.040 --> 00:05:39.839 fields directly. They go and they take action. 00:05:40.980 --> 00:05:43.120 So what we've seen before is that they come up 00:05:43.120 --> 00:05:45.120 with security guards to show the people that 00:05:45.120 --> 00:05:46.620 they have been spotted and to get the people 00:05:46.620 --> 00:05:49.720 out. Sometimes they use the reports we generate 00:05:49.720 --> 00:05:51.540 to go to local authorities and get the police 00:05:51.540 --> 00:05:54.139 involved. And with that, they can remove the 00:05:54.139 --> 00:05:57.379 people before a big conversion to agriculture 00:05:57.379 --> 00:05:59.959 happens, which means that the forest can recover 00:05:59.959 --> 00:06:03.990 faster and better. We now have a system that's 00:06:03.990 --> 00:06:07.009 over 98 % accurate, so that's really nice. In 00:06:07.009 --> 00:06:08.850 this area at least. In other areas we always 00:06:08.850 --> 00:06:12.089 need to retrain and revalidate because it's not 00:06:12.089 --> 00:06:13.870 a globally applicable model, but it's a model 00:06:13.870 --> 00:06:16.230 that really needs to be trained locally because 00:06:16.230 --> 00:06:18.970 the patterns and the nature is different in different 00:06:18.970 --> 00:06:25.730 locations. A bit about the stories. So one of 00:06:25.730 --> 00:06:28.490 the things we hear from our clients is that they 00:06:28.490 --> 00:06:30.589 can finally detect deforestation that happens 00:06:30.589 --> 00:06:34.209 away from roads. Imagine here it's a tropical 00:06:34.209 --> 00:06:36.949 rainforest condition. It's about 200 ,000 hectares 00:06:36.949 --> 00:06:40.509 of area to monitor. Roads tend to be washed away 00:06:40.509 --> 00:06:43.550 with rainfall or any other events and suddenly 00:06:43.550 --> 00:06:46.649 the road is gone. And even if the roads are perfect, 00:06:46.889 --> 00:06:49.129 yeah, it's a rainforest. You can't look into 00:06:49.129 --> 00:06:51.410 the forest and see very far. So you can only 00:06:51.410 --> 00:06:54.490 see a few meters. So before it was almost impossible 00:06:54.490 --> 00:06:58.850 to see these events because they happen outside 00:06:58.850 --> 00:07:01.110 of the direct road network, outside of direct 00:07:01.110 --> 00:07:04.629 fuel. And one more thing that we've seen, which 00:07:04.629 --> 00:07:09.230 I really like talking about. I don't like the 00:07:09.230 --> 00:07:13.389 story behind it. But we detect a full satellite 00:07:13.389 --> 00:07:15.930 view of the whole area, which means that we detect 00:07:15.930 --> 00:07:20.529 all deforestation events. What they noticed is 00:07:20.529 --> 00:07:24.329 with that data that around one security guard 00:07:24.329 --> 00:07:27.769 was positioned at different security posts. Whenever 00:07:27.769 --> 00:07:31.870 he was positioned, there would be a peak in illegal 00:07:31.870 --> 00:07:34.769 logging. So likely he let some family members 00:07:34.769 --> 00:07:37.750 through illegally but that's very difficult to 00:07:37.750 --> 00:07:39.790 assess if you don't have the full data. So by 00:07:39.790 --> 00:07:42.629 having that full map it could pinpoint hey here 00:07:42.629 --> 00:07:45.129 we need to do some more research to identify 00:07:45.129 --> 00:07:47.850 what's exactly going on. Quite a nice case to 00:07:47.850 --> 00:07:49.889 see like how this information we never thought 00:07:49.889 --> 00:07:51.730 something to be pinned down to an individual 00:07:51.730 --> 00:07:53.829 but that actually was possible with the help 00:07:53.829 --> 00:07:55.810 of the data and of course merging it with all 00:07:55.810 --> 00:07:59.829 sorts of other data sets. Now what we do here 00:07:59.829 --> 00:08:04.439 differs quite a bit from EUDR compliance because 00:08:04.439 --> 00:08:07.180 we are focused on near real -time direct action 00:08:07.180 --> 00:08:09.300 and I am assuming directly that almost everyone 00:08:09.300 --> 00:08:11.420 knows what EUDR is but some of you might not 00:08:11.420 --> 00:08:13.060 hear about it as much as I do so let me quickly 00:08:13.060 --> 00:08:16.759 expand EUDR is the European regulation for deforestation 00:08:16.759 --> 00:08:20.620 free supply chain starting next year probably 00:08:20.620 --> 00:08:24.699 you never know with EU it's intended to ensure 00:08:24.699 --> 00:08:27.579 that agricultural products entering the EU are 00:08:27.579 --> 00:08:30.870 not causing deforestation so that you can eat 00:08:30.870 --> 00:08:33.509 your bar of chocolate guilt free. Deforestation 00:08:33.509 --> 00:08:36.190 wise I'm not talking about fair pricing or human 00:08:36.190 --> 00:08:38.470 rights or calories today, but that it doesn't 00:08:38.470 --> 00:08:42.970 cause deforestation. EODR has a single definition 00:08:42.970 --> 00:08:47.789 of forest deforestation land cover where we rather 00:08:47.789 --> 00:08:50.309 look at trees disappearing. So we don't look 00:08:50.309 --> 00:08:53.870 at is a plot of forest converted to agriculture, 00:08:53.950 --> 00:08:56.610 but we look at there were trees here, there are 00:08:56.610 --> 00:08:59.639 none now, take action. rather than saying there 00:08:59.639 --> 00:09:02.740 were trees here, there's agriculture, two years 00:09:02.740 --> 00:09:06.000 later, so you can't import. And then we really 00:09:06.000 --> 00:09:07.679 focus on enabling people in the field to take 00:09:07.679 --> 00:09:12.440 direct action. That's a bit about forest application, 00:09:12.799 --> 00:09:15.580 but deforestation, but it's not all bad. There's 00:09:15.580 --> 00:09:17.879 also afforestation, the growth of new forest, 00:09:17.980 --> 00:09:21.620 where new forests are being planted. So, nice 00:09:21.620 --> 00:09:25.440 story here. without naming specific names I'd 00:09:25.440 --> 00:09:27.240 like to come into a little case study we did 00:09:27.240 --> 00:09:29.720 where we got a request from a small tree planting 00:09:29.720 --> 00:09:32.419 NGO to compare how some of their different sites 00:09:32.419 --> 00:09:35.299 have been doing they used to do a lot of field 00:09:35.299 --> 00:09:39.740 work but with environmental more and more information 00:09:39.740 --> 00:09:42.779 about the flying but also safety concerns for 00:09:42.779 --> 00:09:45.720 quite a long period they couldn't go into their 00:09:45.720 --> 00:09:49.879 afforestation areas anymore so we did some analytics 00:09:49.879 --> 00:09:52.350 for them to compare how different sites were 00:09:52.350 --> 00:09:54.490 doing, which one was healthy, which one was managed 00:09:54.490 --> 00:09:58.210 well etc etc and in some sites we found that 00:09:58.210 --> 00:10:01.789 there were no trees left so the local partners 00:10:01.789 --> 00:10:03.590 had been receiving money to take care of the 00:10:03.590 --> 00:10:07.889 trees but had failed to report the death of several 00:10:07.889 --> 00:10:10.169 full plots so we're gonna go by mistakes happen 00:10:10.169 --> 00:10:12.970 but thanks to this data the project could be 00:10:12.970 --> 00:10:16.850 changed around gain additional management and 00:10:16.850 --> 00:10:18.970 it was made clear to the people locally that 00:10:18.970 --> 00:10:21.950 there were checkups being done which is, first 00:10:21.950 --> 00:10:25.570 is good, but validation is better. And of course 00:10:25.570 --> 00:10:28.529 it's bad that those trees didn't make it, but 00:10:28.529 --> 00:10:31.129 with satellite data, they could make early interventions 00:10:31.129 --> 00:10:34.990 and changes into the process. Now in this project, 00:10:35.009 --> 00:10:38.029 we also made some nice analytics, some quite 00:10:38.029 --> 00:10:40.909 nice algorithms we developed to compare the different 00:10:40.909 --> 00:10:45.049 sites, the missing trees. I would like to say 00:10:45.049 --> 00:10:47.190 that there was this amazing algorithm in -depth 00:10:47.190 --> 00:10:49.700 assessment. But basically it was looking at a 00:10:49.700 --> 00:10:52.519 picture and going to a colleague and, do you 00:10:52.519 --> 00:10:55.320 see trees? I don't see trees. And having figured 00:10:55.320 --> 00:10:58.700 out that there was a major issue here. Satellite 00:10:58.700 --> 00:11:01.519 data analytics can be rocket science, but often 00:11:01.519 --> 00:11:06.000 it's not. And then having that layer of extra 00:11:06.000 --> 00:11:08.620 accountability, especially in an industry where 00:11:08.620 --> 00:11:12.500 so much greenwashing is happening, can give that 00:11:12.500 --> 00:11:16.159 additional trust. And not having that layer... 00:11:16.429 --> 00:11:19.470 and lead to more scrutiny which at least leads 00:11:19.470 --> 00:11:22.009 to us all having something to do so that's also 00:11:22.009 --> 00:11:25.149 a win right? Looking at the current Belenka challenges 00:11:25.149 --> 00:11:27.769 there's also I also want to quickly point out 00:11:27.769 --> 00:11:30.210 some other things everything that has a location 00:11:30.210 --> 00:11:33.870 always grab a satellite image we did a project 00:11:33.870 --> 00:11:37.190 around illegal zoos small project there was no 00:11:37.190 --> 00:11:39.509 budget available so we had no commercial data 00:11:39.509 --> 00:11:43.149 but even looking at 10 meter resolution openly 00:11:43.149 --> 00:11:46.100 available satellite data It's the Sentinel -2 00:11:46.100 --> 00:11:48.600 program that's looking at only true color images. 00:11:49.240 --> 00:11:51.860 We could see the expansion of some of these wildlife 00:11:51.860 --> 00:11:56.000 breeding facilities as an extra source of information 00:11:56.000 --> 00:11:58.679 to help with the investigation of that party 00:11:58.679 --> 00:12:02.179 there. And animals or meat directly leads to 00:12:02.179 --> 00:12:06.480 biodiversity. We see a spike in biodiversity 00:12:06.480 --> 00:12:10.950 credits. where people are paying to support biodiversity 00:12:10.950 --> 00:12:13.870 within different areas, within different conservation 00:12:13.870 --> 00:12:18.850 sites or new nature developments. What is missing 00:12:18.850 --> 00:12:22.210 in a lot of cases is objective data. There is 00:12:22.210 --> 00:12:24.490 quite a lot we can see with satellites that relates 00:12:24.490 --> 00:12:28.529 to biodiversity. Very rarely you look at individual 00:12:28.529 --> 00:12:31.370 species, but often you get landscape features, 00:12:31.710 --> 00:12:34.389 diversity, landscape history, all those things 00:12:34.389 --> 00:12:36.370 can be mapped. That also brings us to an issue. 00:12:36.649 --> 00:12:40.330 If we can map multiple things, we can get different 00:12:40.330 --> 00:12:44.350 outcomes. And which outcome do you share when 00:12:44.350 --> 00:12:47.690 all of the different outcomes are correct, but 00:12:47.690 --> 00:12:49.690 also saying something different? I don't always 00:12:49.690 --> 00:12:51.590 have an answer to that, but I do know that the 00:12:51.590 --> 00:12:53.710 answer can't always be the most optimistic case 00:12:53.710 --> 00:12:57.570 or the most profitable case. Even if we don't 00:12:57.570 --> 00:13:00.610 look at the complexities of biodiversities, but 00:13:00.610 --> 00:13:03.110 look into carbon credits, which is something 00:13:03.110 --> 00:13:05.309 measurable theoretically, you can measure a tree. 00:13:05.549 --> 00:13:08.149 We see billions flowing into the carbon credit 00:13:08.149 --> 00:13:12.610 market, but are the protected areas even under 00:13:12.610 --> 00:13:15.690 threat? And that's why satellites are very important 00:13:15.690 --> 00:13:18.470 to generate a baseline to see what has been happening, 00:13:18.529 --> 00:13:20.750 because we can look back until the 70s of the 00:13:20.750 --> 00:13:23.750 last century to see evidence -based what has 00:13:23.750 --> 00:13:28.250 been going on. We can provide support whether 00:13:28.250 --> 00:13:30.830 a project is resulting in avoided deforestation 00:13:30.830 --> 00:13:34.490 or an aforestation, and identify good reference 00:13:34.490 --> 00:13:38.870 areas, or make strong claim why a reference area 00:13:38.870 --> 00:13:41.710 chosen, so an area where we compare the prediction 00:13:41.710 --> 00:13:46.149 site with chosen, is selected wrongly. And the 00:13:46.149 --> 00:13:49.149 net satellites can give a timeline rather than 00:13:49.149 --> 00:13:51.610 snapshots to tell the whole story instead of 00:13:51.610 --> 00:13:54.789 part of the story. We see that a lot with land 00:13:54.789 --> 00:13:57.490 cover conversion, where land is changing from 00:13:57.490 --> 00:14:02.289 forest to crop land, from pasture to urban, and 00:14:02.289 --> 00:14:05.090 we can detect whether that change is cyclic. 00:14:05.360 --> 00:14:08.840 So if you have shifting cultivation, every year 00:14:08.840 --> 00:14:12.919 part of a forest is cleared to convert to agriculture. 00:14:13.480 --> 00:14:16.860 That's not always the same as a fool on deforestation 00:14:16.860 --> 00:14:19.419 because maybe a year later a different area is 00:14:19.419 --> 00:14:21.899 cleared and it's part of the system that's being 00:14:21.899 --> 00:14:25.019 used within the region. But often it's not. Often 00:14:25.019 --> 00:14:28.500 it's just a cascade effect. Like we see in Indonesia, 00:14:29.759 --> 00:14:33.159 where we see in certain areas a pattern from 00:14:33.159 --> 00:14:36.659 primary forest. to the graded forest, to small 00:14:36.659 --> 00:14:40.299 water fields, to commercial palm oil. And that 00:14:40.299 --> 00:14:42.379 whole pattern can happen within a few years. 00:14:43.200 --> 00:14:45.500 Also, I want to quickly talk about invasive species 00:14:45.500 --> 00:14:47.899 monitoring. We do quite a lot of work in Indonesia. 00:14:48.039 --> 00:14:51.139 We now started just out the project to map invasive 00:14:51.139 --> 00:14:53.679 grasslands to support suitable afforestation 00:14:53.679 --> 00:14:56.659 areas. For us, that's pretty much the same as 00:14:56.659 --> 00:14:59.179 looking at making a land cover map, saying what's 00:14:59.179 --> 00:15:01.080 there. It's a bit more complex if we look at 00:15:01.080 --> 00:15:04.480 very specific species or species group. Almost 00:15:04.480 --> 00:15:07.440 everything you can see with the naked eye, especially 00:15:07.440 --> 00:15:10.100 if you can look from above and see it. You can 00:15:10.100 --> 00:15:12.259 also model and map on large scale with satellites, 00:15:12.340 --> 00:15:15.220 but also develop triggers, codes that give a 00:15:15.220 --> 00:15:18.159 trigger if something happens. That'd be quite 00:15:18.159 --> 00:15:21.259 nice to give warnings. And then I think my current 00:15:21.259 --> 00:15:24.220 favorite project to talk about is environmental 00:15:24.220 --> 00:15:28.320 risks. Imagine you spend a lot of time and resources 00:15:28.320 --> 00:15:32.649 to map a minefield in Ukraine. Imagine a dam 00:15:32.649 --> 00:15:37.490 breach or storm causing a flood. Mines get shifted 00:15:37.490 --> 00:15:39.690 with the water. They're relocated. You don't 00:15:39.690 --> 00:15:42.490 know where. You need to remap possibly a much 00:15:42.490 --> 00:15:45.529 larger area to figure out where the mines are 00:15:45.529 --> 00:15:47.769 in order to be cleaning them. With that we're 00:15:47.769 --> 00:15:50.149 working with the Halo Trust to map environmental 00:15:50.149 --> 00:15:52.870 risk to support in mine clearance priority setting. 00:15:53.769 --> 00:15:57.000 So where is a fire likely to happen? or a flood, 00:15:57.240 --> 00:15:59.379 like a fire you can't exactly send a fire truck 00:15:59.379 --> 00:16:02.799 into a minefield if there's a forest fire. Or 00:16:02.799 --> 00:16:04.500 where is a drought likely to happen and what 00:16:04.500 --> 00:16:07.759 communities will that affect? That's all information 00:16:07.759 --> 00:16:09.519 that can be helped in this case for better decision 00:16:09.519 --> 00:16:12.039 making but that's all information that can also 00:16:12.039 --> 00:16:14.559 be generated globally for different purposes. 00:16:15.340 --> 00:16:18.559 So really looking into the different risks based 00:16:18.559 --> 00:16:20.399 on the patterns that we've seen and based on 00:16:20.399 --> 00:16:23.139 the patterns that we expect in the future with 00:16:23.139 --> 00:16:27.129 for example climatic differences. And if a fire 00:16:27.129 --> 00:16:32.070 happens, where, how big, the faster the information 00:16:32.070 --> 00:16:34.830 can reach the right people, the better the action 00:16:34.830 --> 00:16:37.850 can be taken. So I can really recommend everyone 00:16:37.850 --> 00:16:39.870 to, if you haven't already, to look at the NASA 00:16:39.870 --> 00:16:43.549 Global Database firms. It provides information 00:16:43.549 --> 00:16:47.529 on patterns and trends and historic fires. So 00:16:47.529 --> 00:16:49.649 that way we can map and also make predictions. 00:16:49.889 --> 00:16:52.850 We can give alerts, we can map recovery, always 00:16:52.850 --> 00:16:56.340 globally and everywhere available. and objective. 00:16:57.919 --> 00:17:02.539 I say objective, but in this time of deep fakes 00:17:02.539 --> 00:17:05.960 and fake data, how can we be sure that satellite 00:17:05.960 --> 00:17:09.180 data can be trusted? Or how can we know when 00:17:09.180 --> 00:17:15.380 it can't? So far, large datasets have had very 00:17:15.380 --> 00:17:18.259 few large -scale issues, and I should be careful 00:17:18.259 --> 00:17:22.660 what I say with this audience, because the satellite 00:17:22.660 --> 00:17:25.630 data can feel inherently objective. It's from 00:17:25.630 --> 00:17:29.829 space. It looks like the truth. But as soon as 00:17:29.829 --> 00:17:33.750 it hits the ground or website, it can be framed, 00:17:33.869 --> 00:17:36.589 processed, and cherry -picked to give a particular 00:17:36.589 --> 00:17:40.430 story or a particular angle. And we can't pretend 00:17:40.430 --> 00:17:43.430 like the data is always going to be correct. 00:17:47.009 --> 00:17:51.490 So how to make sure the data is not manipulated 00:17:51.490 --> 00:17:54.640 or misrepresented and how to make sure that you 00:17:54.640 --> 00:17:58.920 don't accidentally misrepresent the data. Quite 00:17:58.920 --> 00:18:01.259 simply put, there's a lot of different satellites 00:18:01.259 --> 00:18:06.599 out there, so don't rely on a single image. Look 00:18:06.599 --> 00:18:10.640 at the time series, look at what change happened. 00:18:10.680 --> 00:18:13.119 Is it a sudden or gradual change? Is it a seasonal 00:18:13.119 --> 00:18:16.019 change? For example, if you look at the deforestation 00:18:16.019 --> 00:18:20.480 events, you can make very impressive images about 00:18:20.480 --> 00:18:23.740 European deforestation. by comparing summer and 00:18:23.740 --> 00:18:25.779 winter images, because the trees lose their leaves 00:18:25.779 --> 00:18:29.900 and everything looks dead. It's not. Guys, the 00:18:29.900 --> 00:18:35.039 forest survives. So really look at how does this 00:18:35.039 --> 00:18:38.079 thread, how is what we're seeing normally looking 00:18:38.079 --> 00:18:41.640 over time? And how is it looking with multiple 00:18:41.640 --> 00:18:44.900 sources? So there's a lot of different satellite 00:18:44.900 --> 00:18:47.460 constellations out there, both public and commercial. 00:18:48.859 --> 00:18:51.230 Satellites are a witness. and you don't rely 00:18:51.230 --> 00:18:53.210 on one witness when you've got a dozen out there. 00:18:53.829 --> 00:18:57.230 We have, for example, very nicely Sentinel and 00:18:57.230 --> 00:19:00.670 Landsat, where Landsat is hosted by the NASA. 00:19:01.170 --> 00:19:04.630 Sentinel is from the European Space Agency. So 00:19:04.630 --> 00:19:06.549 different sources that can look at the same thing, 00:19:07.650 --> 00:19:09.829 different sensors, different resolutions, et 00:19:09.829 --> 00:19:11.750 cetera, et cetera, but they can see the same 00:19:11.750 --> 00:19:14.430 trends. So always have a look at, hey, can I 00:19:14.430 --> 00:19:18.109 validate this with a different data set? And 00:19:18.109 --> 00:19:20.319 then if you're truly concerned about the validity 00:19:20.319 --> 00:19:23.200 of the data, heading back to the original source. 00:19:24.440 --> 00:19:26.380 So a lot of the data you can find in different 00:19:26.380 --> 00:19:29.740 places has been processed, which is freaking 00:19:29.740 --> 00:19:32.500 amazing because I can tell you atmospheric correction 00:19:32.500 --> 00:19:34.700 is really annoying to do. And usually you don't 00:19:34.700 --> 00:19:37.539 really need to do it yourself. You can quite 00:19:37.539 --> 00:19:42.559 easily get that done automatically. But if you 00:19:42.559 --> 00:19:44.680 have reason of doubt, it's good to check the 00:19:44.680 --> 00:19:49.349 original data source and go back into that. and 00:19:49.349 --> 00:19:51.930 to check the full data source. So we're all used 00:19:51.930 --> 00:19:54.690 to look at true color images, red, green, blue, 00:19:54.869 --> 00:19:57.529 make a very nice picture. People working a lot 00:19:57.529 --> 00:19:59.589 with satellite data tend to work a lot with the 00:19:59.589 --> 00:20:01.430 near infrared band as well, make false color 00:20:01.430 --> 00:20:03.789 composites. Looks really nice. You can see a 00:20:03.789 --> 00:20:07.369 lot more changes. But it's a lot harder to fake 00:20:07.369 --> 00:20:09.930 23 bands of sentinel imagery, including atmospheric 00:20:09.930 --> 00:20:12.029 effects, than it is to fake a single picture. 00:20:12.869 --> 00:20:18.269 You'll look back to that single image. that original 00:20:18.269 --> 00:20:20.329 data set and see, can you validate the whole 00:20:20.329 --> 00:20:25.609 processing chain or not? And with false position, 00:20:25.829 --> 00:20:28.750 I think that's more of a misrepresentation usually 00:20:28.750 --> 00:20:36.450 than a true hacking into the data. If people 00:20:36.450 --> 00:20:39.650 make strong claims about data, then the models, 00:20:40.930 --> 00:20:43.789 then the evidence should back that up. If they 00:20:43.789 --> 00:20:46.400 make a... If they don't share any uncertainty 00:20:46.400 --> 00:20:49.339 estimates, that's not that relevant. I now share 00:20:49.339 --> 00:20:52.059 the accuracy of our model in Eurasia and I can 00:20:52.059 --> 00:20:55.039 tell you that's tested on one side and it's amazing 00:20:55.039 --> 00:20:57.539 there. But I can also tell you that I can make 00:20:57.539 --> 00:21:00.559 a global deforestation detection algorithm that's 00:21:00.559 --> 00:21:05.819 99 % accurate. I would do that. I would say that 00:21:05.819 --> 00:21:08.700 in the sea, there will never be deforestation. 00:21:09.019 --> 00:21:12.890 So that's already 60 % done. I will say that 00:21:12.890 --> 00:21:14.630 outside the forest, there will never be deforestation. 00:21:14.750 --> 00:21:17.490 So then we only have a small area to check. And 00:21:17.490 --> 00:21:19.650 actually saying there's less than 1%, there's 00:21:19.650 --> 00:21:21.890 only a couple of percent of deforestation. That 00:21:21.890 --> 00:21:23.650 means that if I say there's zero deforestation 00:21:23.650 --> 00:21:26.670 globally, I have a global model that's going 00:21:26.670 --> 00:21:30.670 to be 99 % accurate. It's also going to be absolutely 00:21:30.670 --> 00:21:35.250 useless, but it's going to be accurate. So know 00:21:35.250 --> 00:21:38.869 when people stroll around accuracy metrics, what 00:21:38.869 --> 00:21:41.289 to ask for. I can share the accuracy metrics 00:21:41.289 --> 00:21:43.029 about our model as well if you want, by the way. 00:21:43.069 --> 00:21:48.529 Those are reliable. And ask those right questions. 00:21:48.690 --> 00:21:52.130 Or make sure you or someone you can depend on 00:21:52.130 --> 00:21:56.910 can do the actual analysis. And know that the 00:21:56.910 --> 00:22:00.250 data can be hacked. Processing pipelines can 00:22:00.250 --> 00:22:03.109 be tampered with. Cloud platforms can be compromised. 00:22:03.369 --> 00:22:08.950 Scripts can be manipulated. very easily putting 00:22:08.950 --> 00:22:11.529 in an altered timestamp, we've all seen that 00:22:11.529 --> 00:22:14.369 happening, is already giving a completely different 00:22:14.369 --> 00:22:18.250 view of what the data represents. So that's why 00:22:18.250 --> 00:22:21.970 open codes, audit trails, their reviewed methods 00:22:21.970 --> 00:22:26.869 are important. So satellite data is a powerful 00:22:26.869 --> 00:22:29.349 witness, but it does need cross -examination. 00:22:30.670 --> 00:22:33.990 It can be mayplated, it can be wrong, but it 00:22:33.990 --> 00:22:37.200 is objective. And with the right checks, it can 00:22:37.200 --> 00:22:40.839 also be a very strong ally in the truth. So to 00:22:40.839 --> 00:22:44.279 end a bit cheesy, the eyes in the sky are watching, 00:22:44.920 --> 00:22:46.980 but they're on our side if we know where to look. 00:22:47.700 --> 00:22:50.259 Where this forest lost ghost plantation or land 00:22:50.259 --> 00:22:54.200 grabbing satellite data can be an ally. So start 00:22:54.200 --> 00:22:57.920 exploring, ask your questions now or contact 00:22:57.920 --> 00:23:01.180 me later for more information. If you can't find 00:23:01.180 --> 00:23:03.200 me based on what I've just said then you shouldn't 00:23:03.200 --> 00:23:06.950 be in this group. So thank you all for listening. 00:23:08.390 --> 00:23:10.910 Thanks very much, Lisa, for your presentation, 00:23:11.869 --> 00:23:13.569 for sharing your knowledge and your experience 00:23:13.569 --> 00:23:16.710 with us. Folks, if you've got questions for Lisa 00:23:16.710 --> 00:23:19.990 based on what she just said, please type them 00:23:19.990 --> 00:23:22.190 in the chat. I'm watching the chat. I will copy 00:23:22.190 --> 00:23:26.690 your question and ask it to Lisa later on. Lisa, 00:23:26.710 --> 00:23:28.269 I've got a couple of questions here for you just 00:23:28.269 --> 00:23:32.650 right off the bat as people get warmed up here 00:23:32.650 --> 00:23:36.450 with their own questions. Can you tell us a bit 00:23:36.450 --> 00:23:40.329 about what got you personally interested in satellite 00:23:40.329 --> 00:23:43.650 data? What was your journey like into this field 00:23:43.650 --> 00:23:47.309 from maybe a young person in the university or 00:23:47.309 --> 00:23:49.549 high school to where you are today? How did you 00:23:49.549 --> 00:23:51.150 get interested in this and how did you end up 00:23:51.150 --> 00:23:54.809 working at Space for Good? Yeah, thanks. It's 00:23:54.809 --> 00:23:58.420 been a while since I thought about that. In high 00:23:58.420 --> 00:24:01.519 school I wanted to really go into nature conservation, 00:24:01.980 --> 00:24:04.039 going into biology, but then I saw that biology 00:24:04.039 --> 00:24:07.259 was going to be very much looking into cells 00:24:07.259 --> 00:24:09.460 in most universities and looking into the cell 00:24:09.460 --> 00:24:12.000 processes and that was not what drew me. I really 00:24:12.000 --> 00:24:15.259 wanted to look into ecology and the bigger processes 00:24:15.259 --> 00:24:18.420 and see how those things work. So then I started 00:24:18.420 --> 00:24:20.759 to study forest and nature conservation in Wageningen. 00:24:21.299 --> 00:24:23.900 I did my bachelor's there. I started my master 00:24:23.900 --> 00:24:26.940 and then I got a bit into GIS remote sensing 00:24:26.940 --> 00:24:29.859 and spatial modeling. And that was for me really 00:24:29.859 --> 00:24:33.039 what I was looking for. It's working with the 00:24:33.039 --> 00:24:35.660 computer, making some models, figuring out how 00:24:35.660 --> 00:24:37.700 things actually work and can we replicate that. 00:24:39.259 --> 00:24:41.799 Basically making major conservation into a game 00:24:41.799 --> 00:24:44.180 and seeing like, how can I alter that? And that's 00:24:44.180 --> 00:24:46.710 really drew me and then... slowly but surely 00:24:46.710 --> 00:24:49.049 I started to learn more about these satellites 00:24:49.049 --> 00:24:51.190 and everything I could see. And basically with 00:24:51.190 --> 00:24:53.289 everything I did, I was like, okay, but where 00:24:53.289 --> 00:24:56.150 can I see, what can I see that's going on? And 00:24:56.150 --> 00:24:58.730 that really got me excited to every time. But 00:24:58.730 --> 00:25:03.710 the data says, yeah, I'm kind of nerdy. Great. 00:25:03.710 --> 00:25:06.890 Thanks for that answer. And as a kind of a follow 00:25:06.890 --> 00:25:10.549 -up, if you're a person, maybe you're listening 00:25:10.549 --> 00:25:13.410 to this live here in our Discord server, you're 00:25:13.410 --> 00:25:15.660 listening to this later. and you are interested 00:25:15.660 --> 00:25:18.859 in getting into this kind of work or you're hearing 00:25:18.859 --> 00:25:20.640 for the first time, oh, wait a minute, there's 00:25:20.640 --> 00:25:22.859 these satellite imaging platforms and one of 00:25:22.859 --> 00:25:24.640 them is called Landsat and the other one is called 00:25:24.640 --> 00:25:28.480 Sentinel. If you're just learning about this 00:25:28.480 --> 00:25:30.279 now or you're becoming interested in this now, 00:25:31.160 --> 00:25:32.759 what are some recommendations that you have for 00:25:32.759 --> 00:25:35.579 somebody like that to learn more about this field 00:25:35.579 --> 00:25:37.900 and to maybe get their hands on some satellite 00:25:37.900 --> 00:25:41.890 imagery and start looking at them? Yeah, so the 00:25:41.890 --> 00:25:44.410 very, very first step is Google Earth, where 00:25:44.410 --> 00:25:46.329 you have the time view as well. So you can look 00:25:46.329 --> 00:25:48.329 back in the past and you can start playing around 00:25:48.329 --> 00:25:53.349 a bit and get hopefully excited. Then I would 00:25:53.349 --> 00:25:56.009 recommend to work mostly with Sentinel, Sentinel 00:25:56.009 --> 00:25:58.630 -2, which is optical data. It's a bit easy to 00:25:58.630 --> 00:26:01.289 work with and radar data, for example. And it's 00:26:01.289 --> 00:26:03.089 10 meter resolution, so you can see quite a lot 00:26:03.089 --> 00:26:05.849 of different things. You can download it either 00:26:05.849 --> 00:26:09.329 via one of the DS servers or Google Earth engine. 00:26:09.799 --> 00:26:14.380 Google Earth Engine is still freely available 00:26:14.380 --> 00:26:18.980 for certain use cases. So that's a nice win. 00:26:19.859 --> 00:26:21.819 Unfortunately for a lot, not anymore, which is 00:26:21.819 --> 00:26:23.259 too bad. But then Google Earth, you can, there's 00:26:23.259 --> 00:26:28.880 a lot of tutorials and a lot of different learnings 00:26:28.880 --> 00:26:31.660 out there to be looking at things. And yeah, 00:26:31.799 --> 00:26:34.980 start with seeing if you can map your own house 00:26:34.980 --> 00:26:38.700 and different bands in different views. Look 00:26:38.700 --> 00:26:42.079 at any article you're reading like, okay, what 00:26:42.079 --> 00:26:45.519 is the latest image I can get? Get really, really 00:26:45.519 --> 00:26:48.640 frustrated by cloud cover and figure out how 00:26:48.640 --> 00:26:50.140 to deal with it. And then at one point think, 00:26:50.140 --> 00:26:52.460 okay, now I'm going to need radar data to complement 00:26:52.460 --> 00:26:54.240 this or now I'm going to need Landsat data to 00:26:54.240 --> 00:26:57.380 complement this. Have a look and play around 00:26:57.380 --> 00:26:59.920 with it. It is coding based. You can also look 00:26:59.920 --> 00:27:02.539 at QGIS, Quantum GIS, which is a software you 00:27:02.539 --> 00:27:05.279 can install your computer and then just download 00:27:05.279 --> 00:27:07.779 the data and open it in there. But I recommend 00:27:07.779 --> 00:27:10.099 working coding based because you can all the 00:27:10.099 --> 00:27:14.039 things so much more easily. Thanks for all those 00:27:14.039 --> 00:27:16.339 tips there. By the way, I'm dropping the links 00:27:16.339 --> 00:27:19.859 to everything that Lisa's mentioning here in 00:27:19.859 --> 00:27:23.180 our Discord server chat. So if you take a look 00:27:23.180 --> 00:27:25.059 at the chat there, I've dropped a link to Google 00:27:25.059 --> 00:27:28.720 Earth Pro, which, yes, is maybe the most beloved 00:27:28.720 --> 00:27:31.559 and used tool of the open source researcher, 00:27:31.740 --> 00:27:33.240 whether you're working primarily with satellite 00:27:33.240 --> 00:27:37.250 imagery or... You're doing geolocations for human 00:27:37.250 --> 00:27:40.069 rights investigations. The Sentinel Google Earth 00:27:40.069 --> 00:27:42.750 Engine, I've also dropped a link in the chat 00:27:42.750 --> 00:27:45.950 there, and QGIS as well. If you're listening 00:27:45.950 --> 00:27:47.869 to this later and you don't have access to the 00:27:47.869 --> 00:27:50.910 chat, you just Google for those words that I 00:27:50.910 --> 00:27:52.430 just mentioned, and that Lisa just mentioned, 00:27:52.450 --> 00:27:54.730 and then you'll see links to these tools that 00:27:54.730 --> 00:27:57.730 you can download. We have a question from Fraser, 00:27:58.049 --> 00:28:00.309 community member Fraser. Hello Fraser, good to 00:28:00.309 --> 00:28:03.049 see you. Thanks for being here. Fraser has this 00:28:03.049 --> 00:28:06.740 question. Lisa, Fraser asks, have you had a look 00:28:06.740 --> 00:28:10.920 at the Leiden guidelines on digitally derived 00:28:10.920 --> 00:28:16.519 evidence and satellite imagery? Yes, I had a 00:28:16.519 --> 00:28:19.220 look at it, but not recently enough to be answering 00:28:19.220 --> 00:28:22.640 this in this. I do know there's a couple of different 00:28:22.640 --> 00:28:25.519 works out there at the moment, also with CivicSpace 00:28:25.519 --> 00:28:27.880 looking to launch their own satellite constellation 00:28:27.880 --> 00:28:33.640 that is not from any... any government or any 00:28:33.640 --> 00:28:35.640 agency or any commercial entity but it's really 00:28:35.640 --> 00:28:39.059 looking at evidence -based blockchain validated 00:28:39.059 --> 00:28:42.940 data. I've had a look at it and I definitely 00:28:42.940 --> 00:28:44.279 will have a look at it again and I recommend 00:28:44.279 --> 00:28:45.859 everyone to have a look at it and I will not 00:28:45.859 --> 00:28:48.160 go further into that. Yeah, oh great, that's 00:28:48.160 --> 00:28:49.960 an answer. Hey, that's a great answer. Thank 00:28:49.960 --> 00:28:52.180 you for that answer. Thanks for your question, 00:28:52.220 --> 00:28:54.920 Fraser. And by the way, if you have any questions, 00:28:55.019 --> 00:28:57.299 please type them in the chat here. I will ask 00:28:57.299 --> 00:29:00.309 them to Lisa. As we continue, we got about a 00:29:00.309 --> 00:29:03.329 half an hour left, so do not be shy. Lisa's an 00:29:03.329 --> 00:29:05.529 expert in this field. You don't often have a 00:29:05.529 --> 00:29:08.089 chance to talk to somebody like Lisa, so do not 00:29:08.089 --> 00:29:09.589 be shy. Ask your questions in the chat, please. 00:29:10.250 --> 00:29:14.490 I got another question for you here, Lisa. The 00:29:14.490 --> 00:29:19.930 question is this. Towards the end there, you 00:29:19.930 --> 00:29:22.250 talked about deep fakes and whatever we want 00:29:22.250 --> 00:29:25.130 to call AI -generated images and fake data being 00:29:25.130 --> 00:29:28.039 out there. You mentioned... that there were, 00:29:28.039 --> 00:29:30.380 I think you said, very few large -scale issues 00:29:30.380 --> 00:29:34.859 out there with imagery. I know in our own work 00:29:34.859 --> 00:29:37.059 at Bellingcat, we're often asked, can you think 00:29:37.059 --> 00:29:40.819 of examples of AI images in the wild that you've 00:29:40.819 --> 00:29:43.049 detected, that you've worked with? And certainly 00:29:43.049 --> 00:29:45.690 in our field, we've seen quite a few. So my question 00:29:45.690 --> 00:29:49.089 to you is, can you think of examples of AI -generated 00:29:49.089 --> 00:29:51.150 satellite imagery? Are they out there? Have you 00:29:51.150 --> 00:29:53.369 come across them? Are there people out there 00:29:53.369 --> 00:29:56.390 who are trying to trick others into looking at 00:29:56.390 --> 00:29:58.089 a satellite image and saying it shows something 00:29:58.089 --> 00:30:02.210 but really it's generated by AI? Is that something 00:30:02.210 --> 00:30:04.069 that you've come across? And if not, is that 00:30:04.069 --> 00:30:07.599 something that you fear might happen soon? Yeah, 00:30:07.599 --> 00:30:10.500 I have not come across any big ones I come came 00:30:10.500 --> 00:30:12.680 across when doing some research for this to also 00:30:12.680 --> 00:30:16.319 find some an area where someone made a nice fake 00:30:16.319 --> 00:30:24.380 post about a fire in In a big park which was 00:30:24.380 --> 00:30:27.619 like obviously fake as in it looked nice, but 00:30:27.619 --> 00:30:29.559 it was also more made with Photoshop So then 00:30:29.559 --> 00:30:33.200 there's no data behind it. I do think it's likely 00:30:33.200 --> 00:30:36.700 to be happening in the future if not now already, 00:30:36.900 --> 00:30:39.380 especially with high resolution data where there 00:30:39.380 --> 00:30:41.700 is active tasking. So there's often only one 00:30:41.700 --> 00:30:46.819 image or very few images to be altered basically. 00:30:47.380 --> 00:30:50.380 And then if you look at war zones, for example, 00:30:50.400 --> 00:30:52.720 there's very high stakes. So I can imagine it 00:30:52.720 --> 00:30:54.720 would be naive to say that it's never going to 00:30:54.720 --> 00:30:57.579 happen. It's likely to happen. But I think with, 00:30:58.200 --> 00:30:59.880 especially in the case where you have open data, 00:30:59.900 --> 00:31:02.700 we have time series of the same data sets, it's 00:31:02.700 --> 00:31:06.759 easy enough to disprove. oh wait that's not I 00:31:06.759 --> 00:31:08.380 think that are easy to disprove still get sometimes 00:31:08.380 --> 00:31:10.539 picked up by the crowd but it is at least easy 00:31:10.539 --> 00:31:13.039 to disprove so it is something to be conscious 00:31:13.039 --> 00:31:18.079 of but that's not yeah for the deforestation 00:31:18.079 --> 00:31:21.940 I don't see it as the direct biggest risk factor 00:31:21.940 --> 00:31:24.380 yeah thanks for that and you know it's interesting 00:31:24.380 --> 00:31:27.799 when you were going through the workflow or the 00:31:27.799 --> 00:31:31.779 suggestions on how to work with images satellite 00:31:31.779 --> 00:31:35.480 images in this time of fake images, of AI images 00:31:35.480 --> 00:31:37.640 being out there. The steps that you outlined 00:31:37.640 --> 00:31:40.460 are the same, many of them are the same as we 00:31:40.460 --> 00:31:42.839 recommend to people who ask us about, hey, how 00:31:42.839 --> 00:31:45.819 do I know that an image from a protest is really 00:31:45.819 --> 00:31:48.380 a real image and not an image from, or how do 00:31:48.380 --> 00:31:50.799 I know that this picture from a war zone really 00:31:50.799 --> 00:31:53.339 is an actual picture from a war zone and not 00:31:53.339 --> 00:31:55.279 a fake picture? You know, you talked about looking 00:31:55.279 --> 00:31:57.640 for other images, right? So you called it the 00:31:57.640 --> 00:31:59.960 time series, like look at other pictures that 00:31:59.960 --> 00:32:03.079 apparently show the same thing. Because I think 00:32:03.079 --> 00:32:07.779 I suspect that as with pictures from war zones 00:32:07.779 --> 00:32:10.420 or from protests or from anywhere that are generated 00:32:10.420 --> 00:32:13.519 by AI, the platforms can't, at least I'm not 00:32:13.519 --> 00:32:16.059 aware of this, can't replicate the same scene 00:32:16.059 --> 00:32:20.779 from different angles. But as you said, it's 00:32:20.779 --> 00:32:22.359 kind of naive to think it's not going to happen. 00:32:22.819 --> 00:32:25.420 I suppose for satellite imagery, something that'd 00:32:25.420 --> 00:32:30.599 be interesting to think about is satellite images 00:32:30.599 --> 00:32:34.009 are... They show intricate patterns right like 00:32:34.009 --> 00:32:39.309 roads and houses and and and I wonder if the 00:32:39.309 --> 00:32:41.769 AI image generation technology is so good that 00:32:41.769 --> 00:32:44.390 it can actually accurately mimic those and I'm 00:32:44.390 --> 00:32:47.650 thinking specifically about the cover of the 00:32:47.650 --> 00:32:51.700 Berkeley protocol book from the Human Rights 00:32:51.700 --> 00:32:53.539 Center at Berkeley. The cover of that book is 00:32:53.539 --> 00:32:57.420 actually an AI -generated image of a satellite 00:32:57.420 --> 00:32:59.819 view. But if you take a close look at it, and 00:32:59.819 --> 00:33:02.019 an expert eye like yours could definitely tell, 00:33:02.079 --> 00:33:03.940 hey, this is weird. I've looked at a million 00:33:03.940 --> 00:33:05.720 satellite images. None of them look like this. 00:33:06.819 --> 00:33:08.299 But as you say, it's one of those things you 00:33:08.299 --> 00:33:10.940 have to think about as being potentially possible 00:33:10.940 --> 00:33:14.200 in the next little while. There, indeed, my risk 00:33:14.200 --> 00:33:18.109 is there is especially with open satellite data, 00:33:18.170 --> 00:33:19.869 there is a lot of data available to train on. 00:33:20.869 --> 00:33:25.730 And a couple of years ago, the generation of 00:33:25.730 --> 00:33:28.730 deepfakes was very, very early on. And if you 00:33:28.730 --> 00:33:31.589 see how fast that goes, how good video material 00:33:31.589 --> 00:33:33.210 we can get. I'm not sure why I'm here. I could 00:33:33.210 --> 00:33:37.630 have just sent an AI. How good that all gets. 00:33:38.509 --> 00:33:40.529 Then I think, yeah, the satellite data, there 00:33:40.529 --> 00:33:42.619 is a lot of data out there. is going to be feasible 00:33:42.619 --> 00:33:44.680 as well. The patterns at the moment, I don't 00:33:44.680 --> 00:33:46.420 think yet, but I'm talking to some people who 00:33:46.420 --> 00:33:49.099 are working on how to connect the roads, which 00:33:49.099 --> 00:33:52.059 is also very useful if you want to map a road 00:33:52.059 --> 00:33:54.519 network in an area. It's nice to be able to map 00:33:54.519 --> 00:33:57.299 those roads. But yeah, it can be used to generate 00:33:57.299 --> 00:33:59.460 fakes as well. And if we talk about the wrong 00:33:59.460 --> 00:34:05.920 data, I see at the moment a bigger issue in misrepresented 00:34:05.920 --> 00:34:09.260 data. having a data set and saying it's and that's 00:34:09.260 --> 00:34:11.159 something that we do see what people say claim 00:34:11.159 --> 00:34:14.360 it's in a different area and again you can say 00:34:14.360 --> 00:34:16.619 or a different time and you can say no wait i've 00:34:16.619 --> 00:34:19.179 got the evidence that it's not but if the information 00:34:19.179 --> 00:34:21.619 is out there coming with the evidence isn't always 00:34:21.619 --> 00:34:25.320 enough yeah thanks for that lisa i have another 00:34:25.320 --> 00:34:27.019 question here since we're talking about you know 00:34:27.019 --> 00:34:29.099 we're talking about what may be coming in the 00:34:29.099 --> 00:34:32.639 future in terms of ai image generation you've 00:34:32.639 --> 00:34:35.420 just mentioned this ability to make realistic 00:34:35.420 --> 00:34:39.739 roads, for example. Are there any advances in 00:34:39.739 --> 00:34:42.179 remote sensing technology that you think are 00:34:42.179 --> 00:34:44.039 coming soon or in the next couple of years that 00:34:44.039 --> 00:34:49.159 you're really excited about? A lot, a lot. So 00:34:49.159 --> 00:34:52.719 what's happening right now is Google has just 00:34:52.719 --> 00:34:56.019 announced a better version of a system to be 00:34:56.019 --> 00:35:00.510 talking to maps, basically. So rather than doing 00:35:00.510 --> 00:35:02.389 all the investigation yourself, being able to 00:35:02.389 --> 00:35:05.030 just ask, hey, for this area of interest, how 00:35:05.030 --> 00:35:07.510 much deforestation has there been? I think there's 00:35:07.510 --> 00:35:08.969 a big risk there, but I also think there's a 00:35:08.969 --> 00:35:11.110 lot of opportunities there to take that first 00:35:11.110 --> 00:35:14.230 step into how can we actually see what's going 00:35:14.230 --> 00:35:17.969 on on our planet? I think the roads, that's for 00:35:17.969 --> 00:35:22.750 me to make the system of automatically connecting 00:35:22.750 --> 00:35:26.530 different features. based on a lot of data, but 00:35:26.530 --> 00:35:29.809 also based on quite complex models. Having a 00:35:29.809 --> 00:35:33.369 detailed road network can mean so, so much for, 00:35:33.369 --> 00:35:36.090 for example, detecting illegal logging, but also 00:35:36.090 --> 00:35:38.309 predicting it. We see that illegal logging almost 00:35:38.309 --> 00:35:41.250 always happens near roads and almost always first 00:35:41.250 --> 00:35:44.630 a new road is made and then logging happens and 00:35:44.630 --> 00:35:46.829 it expands. If you can see that earlier by having 00:35:46.829 --> 00:35:49.789 just a few points. then you can really map out 00:35:49.789 --> 00:35:52.690 a lot so that's really interesting to me and 00:35:52.690 --> 00:35:56.110 yeah biodiversity making some sort of integrated 00:35:56.110 --> 00:36:00.710 solution how can we map biodiversity to well 00:36:00.710 --> 00:36:03.050 it's not really clear what biodiversity there's 00:36:03.050 --> 00:36:05.690 not one definition of biodiversity but to be 00:36:05.690 --> 00:36:09.050 able to say hey this is going well and this is 00:36:09.050 --> 00:36:13.429 not going well Thanks, Lisa. We have a question 00:36:13.429 --> 00:36:15.269 from Chris. And if you have a question, folks, 00:36:15.429 --> 00:36:17.010 please type it in the chat. We've got about 20 00:36:17.010 --> 00:36:19.610 minutes left here with Lisa from Space for Good. 00:36:19.929 --> 00:36:22.989 Chris is asking this question here, Lisa. Now, 00:36:23.289 --> 00:36:25.909 there's some jargon in here that I don't know. 00:36:25.949 --> 00:36:27.670 So I'm going to read it out as it is written. 00:36:28.210 --> 00:36:31.610 And maybe this is the jargon that you know. And 00:36:31.610 --> 00:36:33.969 the question is this. I'm curious to hear more 00:36:33.969 --> 00:36:36.989 about how you are ground -truthing satellite 00:36:36.989 --> 00:36:42.739 data to help detect distortions. So we ourselves 00:36:42.739 --> 00:36:45.000 are not earlier, but let me first like phrases 00:36:45.000 --> 00:36:47.179 a bit for the rest. So ground rooting that's 00:36:47.179 --> 00:36:49.420 looking at what's on the ground and perceiving 00:36:49.420 --> 00:36:52.380 that as the truth and you can use that to retrain 00:36:52.380 --> 00:36:56.139 models. I think here the distortion, the distortions, 00:36:56.340 --> 00:37:00.760 but Chris, chat if it's wrong, is really looking 00:37:00.760 --> 00:37:04.559 at the distortions of the images, which is why 00:37:04.559 --> 00:37:07.300 you're looking at, for example, something that 00:37:07.300 --> 00:37:10.530 we've seen in the past quite a lot, the earlier 00:37:10.530 --> 00:37:14.809 generations of planet data had a lot of misalignment 00:37:14.809 --> 00:37:17.050 of different bands now i'm getting a bit technical 00:37:17.050 --> 00:37:20.369 here so that's when you have a picture of the 00:37:20.369 --> 00:37:23.690 red green and blue and they're all not exactly 00:37:23.690 --> 00:37:27.230 aligned so you could have a situation where the 00:37:27.230 --> 00:37:30.570 red value of a certain pixel was 10 meters below 00:37:30.570 --> 00:37:34.690 blue value so you would have a shift there with 00:37:34.690 --> 00:37:37.570 ground rooting you can basically align pictures 00:37:37.570 --> 00:37:40.570 so they're all exactly at the same spot. And 00:37:40.570 --> 00:37:42.170 then you can look for example, changes and then 00:37:42.170 --> 00:37:45.090 you know where something is exactly. Usually 00:37:45.090 --> 00:37:47.530 that is done by the satellite data providers. 00:37:47.590 --> 00:37:49.650 And I think right now, most of those have good 00:37:49.650 --> 00:37:52.829 enough on -train models that we can focus on 00:37:52.829 --> 00:37:57.389 our own strengths, which is mostly the data analytics 00:37:57.389 --> 00:38:03.369 and going into depth on the AI models. Thank 00:38:03.369 --> 00:38:05.190 you for that question. And thanks for that answer, 00:38:05.809 --> 00:38:11.150 Lisa. I have another question. What is your kind 00:38:11.150 --> 00:38:13.230 of biggest frustration about the limitations 00:38:13.230 --> 00:38:15.489 of the current technology? What's the thing that 00:38:15.489 --> 00:38:17.369 you run across the most that makes you go, oh 00:38:17.369 --> 00:38:20.789 my gosh, I wish it wasn't like this. I wish we 00:38:20.789 --> 00:38:23.130 could fix this issue with the technology. What's 00:38:23.130 --> 00:38:25.570 something that's missing in the technology or 00:38:25.570 --> 00:38:29.610 a capability that's missing at the moment? Yeah, 00:38:29.670 --> 00:38:32.590 it really depends on use case to use case. What 00:38:32.590 --> 00:38:36.699 is, for me, I think my biggest frustration is 00:38:36.699 --> 00:38:38.860 not as much in the technology, but in the technology 00:38:38.860 --> 00:38:42.019 communication. So I'm spending a lot of time 00:38:42.019 --> 00:38:43.719 with different users and different potential 00:38:43.719 --> 00:38:46.320 users talking about, hey, what can be done and 00:38:46.320 --> 00:38:48.599 what cannot be done. And then sometimes you have 00:38:48.599 --> 00:38:51.880 people who say, who think we're working with 00:38:51.880 --> 00:38:55.840 satellites and AI. So I want to see this individual 00:38:55.840 --> 00:38:58.260 tree and how well it's doing. And sometimes you 00:38:58.260 --> 00:39:00.360 have people who still think like satellites and 00:39:00.360 --> 00:39:02.159 AI, no, that's not working in forest, forest 00:39:02.159 --> 00:39:05.280 is special. It's somewhere in the middle and 00:39:05.280 --> 00:39:07.860 that's always really annoying. If we go back 00:39:07.860 --> 00:39:11.860 to really technical, I think a lot is possible. 00:39:12.480 --> 00:39:17.440 For me, the challenges with working on a large 00:39:17.440 --> 00:39:20.000 scale with high resolution data, especially with 00:39:20.000 --> 00:39:23.280 time series, are still really annoying because 00:39:23.280 --> 00:39:25.440 I know that there's a lot of data out there, 00:39:25.739 --> 00:39:28.019 but most of it is commercial and then it tends 00:39:28.019 --> 00:39:30.960 to be nuts. available for a lot of the applications 00:39:30.960 --> 00:39:33.659 that I feel need it most. You can do a lot if 00:39:33.659 --> 00:39:36.179 you want to go into oil and gas, but if you want 00:39:36.179 --> 00:39:39.980 to make sure there's a world left, then it's 00:39:39.980 --> 00:39:43.219 a constant uphill battle that way. Again, maybe 00:39:43.219 --> 00:39:46.639 not the most technical issue, so I think, yeah. 00:39:47.519 --> 00:39:49.360 I kind of like the technical challenges. Yeah, 00:39:49.559 --> 00:39:52.340 clouds are annoying, but it's part of the fun, 00:39:52.559 --> 00:39:56.110 right? Gives us a puzzle to solve. Yeah, whenever 00:39:56.110 --> 00:39:59.489 we work with, we have a really wonderful network 00:39:59.489 --> 00:40:04.150 of volunteers and sometimes we get satellite 00:40:04.150 --> 00:40:07.170 imagery from a satellite image provider that 00:40:07.170 --> 00:40:08.929 they request. They'll be working on a case and 00:40:08.929 --> 00:40:10.329 they'll go, hey, I need a picture from here. 00:40:10.769 --> 00:40:13.670 And if the volunteer, if the individual isn't 00:40:13.670 --> 00:40:16.449 familiar with satellite imagery, we have lots 00:40:16.449 --> 00:40:18.829 of folks who've never worked with it in the past 00:40:18.829 --> 00:40:21.090 and are now working with it for the first time 00:40:21.090 --> 00:40:23.909 at BellyCat. you know, 95 % of the time we get 00:40:23.909 --> 00:40:25.429 the image, we send it to them and we go, it's 00:40:25.429 --> 00:40:27.110 cloudy. Like we could, you know, here's the result, 00:40:27.170 --> 00:40:28.690 or you can't see anything. And they go, oh, of 00:40:28.690 --> 00:40:30.710 course, like there's clouds. Yeah, there's lots 00:40:30.710 --> 00:40:32.550 of clouds. And depending on the place, there's 00:40:32.550 --> 00:40:36.710 clouds like every day. So, yeah, definitely sympathize 00:40:36.710 --> 00:40:40.309 with that. We did a project in Ivory Coast and 00:40:40.309 --> 00:40:42.750 I think in a year's time we could get one image. 00:40:43.469 --> 00:40:46.449 Yeah, that sounds about right. Yeah. We have 00:40:46.449 --> 00:40:49.889 another question here from another community 00:40:49.889 --> 00:40:54.260 member named BQN. And BQN asks, Lisa, what is 00:40:54.260 --> 00:40:57.360 the most interesting or innovative or unusual 00:40:57.360 --> 00:41:01.079 or creative application of satellite image data 00:41:01.079 --> 00:41:09.460 that you have come across? So I kind of want 00:41:09.460 --> 00:41:11.880 to talk about this project we've done. It wasn't 00:41:11.880 --> 00:41:13.480 technically with satellite image because here 00:41:13.480 --> 00:41:16.780 we could get through a partner, really like the 00:41:16.780 --> 00:41:20.219 project as well, some aerial data. But basically 00:41:20.219 --> 00:41:22.280 for us, the processing is pretty much the same. 00:41:22.340 --> 00:41:23.820 It's a bit less atmospheric correction needed, 00:41:24.000 --> 00:41:28.199 but data is data, right? So I was working with 00:41:28.199 --> 00:41:31.820 aerial data and that was to map out latrines. 00:41:32.500 --> 00:41:34.900 Where MREF held Africa, this project, this wash 00:41:34.900 --> 00:41:39.260 intervention to support, to give out micro credits 00:41:39.260 --> 00:41:42.900 for people to build latrines. And they wanted 00:41:42.900 --> 00:41:46.659 to know how effective the project was. Now they 00:41:46.659 --> 00:41:48.739 needed to do a comparison to a different site. 00:41:49.150 --> 00:41:51.269 to see how many latrines people would normally 00:41:51.269 --> 00:41:55.309 build, at least would build us outhouses. And 00:41:55.309 --> 00:41:57.489 it's kind of invasive to be knocking on people's 00:41:57.489 --> 00:42:00.670 door and ask, hey, can I see your toilets? So 00:42:00.670 --> 00:42:02.750 they were looking for a different solution and 00:42:02.750 --> 00:42:05.250 then we developed a model to see, hey, where 00:42:05.250 --> 00:42:08.250 can we see outhouses and how can we distinguish 00:42:08.250 --> 00:42:10.869 latrines from chicken coops, which is very challenging. 00:42:12.349 --> 00:42:14.349 And I really like talking about that, like looking 00:42:14.349 --> 00:42:16.889 at the different angle, like, yeah, sure, sure, 00:42:16.949 --> 00:42:20.800 we can map latrines, why not? Yeah, I don't think 00:42:20.800 --> 00:42:22.099 I've ever heard of anything like that. That's 00:42:22.099 --> 00:42:25.440 fascinating. Thanks for sharing that. We have 00:42:25.440 --> 00:42:31.960 a question here from Aikan. Hello, Aikan. Thanks 00:42:31.960 --> 00:42:34.260 for your question. And the question is, do you 00:42:34.260 --> 00:42:36.880 have any tips for using platforms like Google 00:42:36.880 --> 00:42:41.500 Earth or Sentinel for identifying and confirming 00:42:41.500 --> 00:42:46.159 smaller scale environmental sites like artisanal 00:42:46.159 --> 00:42:51.150 mines, for example? So artisanal minds, we've 00:42:51.150 --> 00:42:54.489 also done some work in. I believe there might 00:42:54.489 --> 00:42:57.510 even be some tutorials using it in Google Earth 00:42:57.510 --> 00:43:01.190 Engine. Otherwise, what I would really look at 00:43:01.190 --> 00:43:05.429 is look for bare soil index. So really look at 00:43:05.429 --> 00:43:09.050 bare soil and mapping changes into that. So looking 00:43:09.050 --> 00:43:11.369 at the time series analysis, you can look into, 00:43:11.510 --> 00:43:18.500 for example, BFAST, B -F -A -S -T. that's normally 00:43:18.500 --> 00:43:21.179 done on the NDVI but you can also do it on for 00:43:21.179 --> 00:43:25.119 example bare soil index and using that to map 00:43:25.119 --> 00:43:27.659 where you see for example bare soils popping 00:43:27.659 --> 00:43:31.500 up in certain area sizes so you want to look 00:43:31.500 --> 00:43:33.820 at then quite small area sizes where there's 00:43:33.820 --> 00:43:38.500 quite a rapid change ideally you make some some 00:43:38.500 --> 00:43:40.119 smart selections about your area of interest 00:43:40.119 --> 00:43:42.500 where you expect it to happen so for example 00:43:42.500 --> 00:43:46.420 close to rivers and then Second point of advice 00:43:46.420 --> 00:43:50.079 is don't be afraid to get dirty and just click 00:43:50.079 --> 00:43:52.880 through images and look at things. So use these 00:43:52.880 --> 00:43:54.460 algorithms. If you don't have a lot of training 00:43:54.460 --> 00:43:57.079 data, don't spend too much time on building complex 00:43:57.079 --> 00:43:59.380 models. If you don't have the data to put in 00:43:59.380 --> 00:44:01.440 there and the model is not going to be great 00:44:01.440 --> 00:44:05.659 anyway and click through things, look at the 00:44:05.659 --> 00:44:08.420 different images and use your own eyes as well 00:44:08.420 --> 00:44:10.400 because our own eyes are very good at change 00:44:10.400 --> 00:44:15.179 detection and pattern recognition. Use the computer 00:44:15.179 --> 00:44:17.760 to take the first step, but don't try to get 00:44:17.760 --> 00:44:20.920 everything done perfectly Rather just look and 00:44:20.920 --> 00:44:23.480 click and do the dirty work yourself as well 00:44:23.480 --> 00:44:29.380 We have a question now from wit Fox and you've 00:44:29.380 --> 00:44:32.219 talked You know, we just talked a little bit 00:44:32.219 --> 00:44:35.360 about this. Let's let's see if we can get more 00:44:35.360 --> 00:44:39.320 On this Lisa. You've talked a bit about deforestation 00:44:39.320 --> 00:44:42.320 biodiversity fires mines. We talked about latrines 00:44:42.320 --> 00:44:45.739 a moment ago What are some things that you think 00:44:45.739 --> 00:44:49.400 satellites are being underutilized for? Or what 00:44:49.400 --> 00:44:53.039 are some other perhaps non -obvious use cases 00:44:53.039 --> 00:44:58.059 for satellite imagery data, either optical or 00:44:58.059 --> 00:45:06.960 multispectral? Thanks. Take your moment. Because 00:45:06.960 --> 00:45:10.119 I know satellite data is being used for loads 00:45:10.119 --> 00:45:13.420 of different things. even for mapping penguin 00:45:13.420 --> 00:45:15.480 colonies by looking at their shit you can do 00:45:15.480 --> 00:45:18.420 with satellite so what they're not being used 00:45:18.420 --> 00:45:26.920 for is kind of difficult to look at yeah I think 00:45:26.920 --> 00:45:30.960 a lot of places a lot more can be done I think 00:45:30.960 --> 00:45:35.440 for mapping these environmental claims for big 00:45:35.440 --> 00:45:38.340 organizations I think they can be used a lot 00:45:38.340 --> 00:45:41.139 more because I feel like for a lot of these cases 00:45:42.579 --> 00:45:46.559 companies will make a claim, for example, in 00:45:46.559 --> 00:45:49.320 the offsetting. They will share some sort of 00:45:49.320 --> 00:45:50.900 location with it or they will share some sort 00:45:50.900 --> 00:45:53.760 of report with it. But the data tends to be based 00:45:53.760 --> 00:45:56.739 on field work, which is nice, but which must 00:45:56.739 --> 00:45:58.920 be validated. We've seen it a lot where even 00:45:58.920 --> 00:46:01.760 companies we work with who do field work professionally 00:46:01.760 --> 00:46:06.099 and who do it for the project we asked them, 00:46:06.480 --> 00:46:08.619 give us data where there's just errors in there. 00:46:10.039 --> 00:46:14.039 That makes me wonder how much validation is being 00:46:14.039 --> 00:46:18.099 done to support these claims for, for example, 00:46:18.280 --> 00:46:21.039 the regulated carbon market. I know it's being 00:46:21.039 --> 00:46:23.699 done there, and I know it's kind of an open and 00:46:23.699 --> 00:46:26.599 short case, but I think more can be done there. 00:46:28.139 --> 00:46:30.519 Thanks, Lisa. We have about 10 minutes here left 00:46:30.519 --> 00:46:33.500 with Lisa Brughausen from Space for Good. If 00:46:33.500 --> 00:46:35.579 you have any questions, please type them in the 00:46:35.579 --> 00:46:39.190 chat. You've got 10 minutes to ask her. A question, 00:46:39.269 --> 00:46:42.489 we've got one here from SirSkippy. And SirSkippy 00:46:42.489 --> 00:46:46.909 asks, do you make the data that you collect available 00:46:46.909 --> 00:46:53.570 on OpenStreetMap for ECUs by others? No, we do 00:46:53.570 --> 00:46:56.289 not. And the data we collect, we don't have our 00:46:56.289 --> 00:46:58.650 own satellites. So the data we collect is the 00:46:58.650 --> 00:47:00.489 satellite data itself, the raw satellite data. 00:47:01.289 --> 00:47:04.070 And the data, the results, the information we 00:47:04.070 --> 00:47:08.269 generate We don't make available and I think 00:47:08.269 --> 00:47:10.329 so we also work for, for example, commercial 00:47:10.329 --> 00:47:14.409 entities to take their direct action. I think 00:47:14.409 --> 00:47:17.769 in most of those cases, it would not be beneficial 00:47:17.769 --> 00:47:19.829 to them to publish it. And then we also need 00:47:19.829 --> 00:47:21.949 to be cautious with that and rather focus on 00:47:21.949 --> 00:47:25.210 the direct action and on the mapping. Plus in 00:47:25.210 --> 00:47:27.269 OpenStreetMap, if we do, for example, deforestation 00:47:27.269 --> 00:47:30.329 detection. So like I said, what we do is we look 00:47:30.329 --> 00:47:33.900 at deforestation as there was a forest it's gone 00:47:33.900 --> 00:47:37.840 now there's no more forest if you look at most 00:47:37.840 --> 00:47:41.039 land cover maps they look at what is the land 00:47:41.039 --> 00:47:45.300 cover we the official land cover that's here 00:47:45.300 --> 00:47:48.119 the land use that's here so if you cut a forest 00:47:48.119 --> 00:47:51.480 and a new forest grows it is still the land cover 00:47:51.480 --> 00:47:55.619 class of forest and then there's no rapid change 00:47:55.619 --> 00:47:58.639 for change that we detect and then in the other 00:47:58.639 --> 00:48:00.360 land cover mapping we tend to have a bit more 00:48:00.360 --> 00:48:03.980 classes than are relevant for This road networks 00:48:03.980 --> 00:48:08.239 could make sense. But even then I think that 00:48:08.239 --> 00:48:10.440 working with the GPS tracker so people have locally 00:48:10.440 --> 00:48:14.880 are probably going to be more, uh, more directly 00:48:14.880 --> 00:48:17.099 applicable also because it's in private areas. 00:48:17.119 --> 00:48:18.739 So they don't want the date. They don't want 00:48:18.739 --> 00:48:22.199 people on the roads. So no, we do not. The short 00:48:22.199 --> 00:48:24.960 answer. Thanks, Lisa. We got another one here 00:48:24.960 --> 00:48:30.340 from, um, uh, this one's from Charlie. Um, are 00:48:30.340 --> 00:48:32.599 you Charlie from, from Bellingham? And Charlie 00:48:32.599 --> 00:48:36.440 asks, I'm going to paraphrase here. Imagine you're 00:48:36.440 --> 00:48:38.079 listening to this and you go, wow, Space for 00:48:38.079 --> 00:48:40.980 Good sounds incredible. I think I have some skills 00:48:40.980 --> 00:48:43.099 that might make me a candidate. I'm going to 00:48:43.099 --> 00:48:45.639 look for postings on the Space for Good website. 00:48:45.739 --> 00:48:47.179 Maybe if they're hiring somebody, I'm going to 00:48:47.179 --> 00:48:49.739 apply. And the question to you, Lisa, is what 00:48:49.739 --> 00:48:52.400 kinds of skill sets do you look for in team members? 00:48:52.739 --> 00:48:55.639 Just in case somebody in the audience here wants 00:48:55.639 --> 00:48:59.079 to maybe give a shot at working with Space for 00:48:59.079 --> 00:49:03.599 Good. Yeah, thanks. So what we're looking for 00:49:03.599 --> 00:49:08.360 is really in -depth remote sensing and modeling 00:49:08.360 --> 00:49:10.559 skills. So if we look for a data analyst or data 00:49:10.559 --> 00:49:12.360 scientist and we look for people who really have 00:49:12.360 --> 00:49:14.900 experience in the data, but also have experience 00:49:14.900 --> 00:49:19.119 in one or two fields of interest. So we have, 00:49:19.199 --> 00:49:21.619 I myself have a mixed background with forestry. 00:49:22.000 --> 00:49:23.860 We have someone who has a background in doing 00:49:24.639 --> 00:49:27.159 carbon projects in the fields, getting that access 00:49:27.159 --> 00:49:29.320 with someone who's done a lot of work around 00:49:29.320 --> 00:49:32.599 fires in the past, agriculture. So knowing, and 00:49:32.599 --> 00:49:35.719 that's where I think most value lays in, combining 00:49:35.719 --> 00:49:39.099 the data sites with direct action in the field 00:49:39.099 --> 00:49:41.019 with something, someone who knows the background. 00:49:41.980 --> 00:49:44.639 If you don't know how a tree grows, then monitoring 00:49:44.639 --> 00:49:47.179 from above is going to be a lot harder and monitoring 00:49:47.179 --> 00:49:50.099 it when it's gone is going to be very difficult 00:49:50.099 --> 00:49:54.440 to assess. So that's ready for the data. They're 00:49:54.440 --> 00:49:56.099 the analytics, they're the scientist position. 00:49:57.599 --> 00:50:00.559 Then we have, if we work with, we have, for example, 00:50:00.579 --> 00:50:03.539 at the moment, one full stack developer that 00:50:03.539 --> 00:50:06.579 was looking for someone with cloud experience, 00:50:06.760 --> 00:50:09.420 scalability, et cetera, et cetera. But for us, 00:50:09.559 --> 00:50:11.780 the soft side is also very important where you've 00:50:11.780 --> 00:50:15.719 got a very cool team, if I may say so, and really 00:50:15.719 --> 00:50:17.599 working together as a team to build something 00:50:17.599 --> 00:50:20.820 that's better. And then, of course, around commercial, 00:50:20.940 --> 00:50:22.440 well, we've got a very small team around that, 00:50:22.500 --> 00:50:24.719 but that's also always good to have people who 00:50:24.719 --> 00:50:29.280 can try to get it to the final users and clients. 00:50:31.360 --> 00:50:32.820 Thanks very much. We have about eight minutes 00:50:32.820 --> 00:50:35.380 left here with Lisa. Get your questions in, folks. 00:50:35.940 --> 00:50:38.599 We have a question from Dusty Wieners. By the 00:50:38.599 --> 00:50:42.119 way, Wieners, if you folks don't know. is the 00:50:42.119 --> 00:50:45.440 skin on your elbow is the weenus. I'm not sure 00:50:45.440 --> 00:50:48.059 if you are familiar with that term, but Dusty 00:50:48.059 --> 00:50:54.019 Weenus is asking the following question. They 00:50:54.019 --> 00:50:57.679 say, Dusty Weenus says, is there any way to verify 00:50:57.679 --> 00:51:00.159 direct carbon emissions from industrial sites 00:51:00.159 --> 00:51:06.340 from space? For example, A company claims a certain 00:51:06.340 --> 00:51:08.519 amount of scope one emissions, but the satellite 00:51:08.519 --> 00:51:11.539 data actually indicates a larger amount. Is that 00:51:11.539 --> 00:51:16.530 possible? For example, methane mapping. Thanks, 00:51:16.670 --> 00:51:19.329 Dusty or Venus, I'm not sure how to refer to 00:51:19.329 --> 00:51:22.090 you. For example, for methane, there's some satellites 00:51:22.090 --> 00:51:24.469 out there that can really map at different resolutions 00:51:24.469 --> 00:51:27.650 the direct methane emissions. It's always important 00:51:27.650 --> 00:51:30.429 to look at then the full scale of events. If 00:51:30.429 --> 00:51:32.630 they report CIRTOSCOPE -1 emissions, you need 00:51:32.630 --> 00:51:34.949 to have the correct overpass times, the correct 00:51:34.949 --> 00:51:36.829 models behind it. And there's always the issue 00:51:36.829 --> 00:51:38.909 that you look with the satellite at the full 00:51:38.909 --> 00:51:42.570 column. So at the entire atmosphere, basically. 00:51:42.880 --> 00:51:47.039 So there is a potential for a mix -up. However, 00:51:47.340 --> 00:51:49.039 these satellites have been used, for example, 00:51:49.159 --> 00:51:51.860 to identify when there have been methane leaks 00:51:51.860 --> 00:51:54.539 or large -scale methane emissions where they 00:51:54.539 --> 00:51:57.440 shouldn't have been due to any technical issues. 00:51:58.139 --> 00:52:01.559 So yes, it can be done for larger scale operations. 00:52:02.679 --> 00:52:06.480 The farmer next door with 20 cows doesn't exist 00:52:06.480 --> 00:52:08.320 anymore. But anyway, you're not going to be able 00:52:08.320 --> 00:52:12.000 to find that. But for the large factories, yes. 00:52:12.179 --> 00:52:17.300 And for the pipelines where there might be maybe 00:52:17.300 --> 00:52:21.500 slight issues with the maintenance as well. I'm 00:52:21.500 --> 00:52:24.719 being positive here. Thanks. Thanks Lisa. I think 00:52:24.719 --> 00:52:26.320 this is going to be the last question here. Somebody 00:52:26.320 --> 00:52:28.400 notices we're talking about, you know, work in 00:52:28.400 --> 00:52:31.400 the field and maybe working space for good. Somebody 00:52:31.400 --> 00:52:34.659 noticed that there's an astropreneur position 00:52:34.659 --> 00:52:36.320 on the website. Are you familiar with that one? 00:52:36.440 --> 00:52:38.599 I'm asking it because Chris is asking about this, 00:52:38.619 --> 00:52:40.679 but also I've never heard of that term. Do you 00:52:40.679 --> 00:52:42.619 know about the position, the vacancy? Can you 00:52:42.619 --> 00:52:47.099 talk about it? It might be a word we made up. 00:52:50.000 --> 00:52:55.059 More on the data side than that side. So let 00:52:55.059 --> 00:53:01.920 me quickly, where is that? It's in the, yeah, 00:53:02.000 --> 00:53:06.500 spaceforgo .com forward slash vacancies. Marketing 00:53:06.500 --> 00:53:10.179 and sales, market research and analysis, finance 00:53:10.179 --> 00:53:12.639 and investment. Oh yeah, talented and engaged 00:53:12.639 --> 00:53:16.500 entrepreneurs. So that's basically an open, always 00:53:16.500 --> 00:53:18.579 open vacancy where people who are interested 00:53:18.579 --> 00:53:21.380 can. let us know and we can see whether there 00:53:21.380 --> 00:53:23.900 is there is anything out there uh that always 00:53:23.900 --> 00:53:27.059 depends a bit on projects on current needs on 00:53:27.059 --> 00:53:29.840 skills whether we can fit something in or not 00:53:29.840 --> 00:53:33.820 um lately usually not i have to admit because 00:53:33.820 --> 00:53:36.360 we're with quite quite a strong team at the moment 00:53:36.360 --> 00:53:39.460 uh but quite frequently we can find figure out 00:53:39.460 --> 00:53:44.139 something so um yeah entrepreneur combination 00:53:44.139 --> 00:53:47.739 of uh astro and entrepreneur so also looking 00:53:47.739 --> 00:53:49.900 at people who want to We do work as well with 00:53:49.900 --> 00:53:52.860 a shell of freelancers who take specific projects 00:53:52.860 --> 00:53:56.500 or who we support in giving specific projects 00:53:56.500 --> 00:53:59.320 when there are specific skill sets we don't have 00:53:59.320 --> 00:54:01.539 in the team, but we need in certain projects 00:54:01.539 --> 00:54:05.059 or that just want to be somehow connected because 00:54:05.059 --> 00:54:08.860 we are a lot of fun. Great. Thank you so much 00:54:08.860 --> 00:54:10.579 for that. And we've got this last question here, 00:54:10.699 --> 00:54:12.900 folks, again, from SIRS -KP, has the change in 00:54:12.900 --> 00:54:16.219 US government funding priorities affected Space 00:54:16.219 --> 00:54:17.739 for good at all. We get that question about again 00:54:17.739 --> 00:54:21.159 all the time. We're passing it on to you Lisa 00:54:21.159 --> 00:54:23.780 space for good. Yeah, you don't have to ask the 00:54:23.780 --> 00:54:30.079 book Affected us. I mean emotionally financially 00:54:30.079 --> 00:54:32.780 directly or indirectly. I think it's affecting 00:54:32.780 --> 00:54:36.860 affecting the entire world At the moment, it 00:54:36.860 --> 00:54:39.579 has not affected any of the current of the projects 00:54:39.579 --> 00:54:41.519 that we do. And that's also where we try to have 00:54:41.519 --> 00:54:44.320 a mix between commercial projects, projects with 00:54:44.320 --> 00:54:47.280 NGOs and governmental projects. We do know that 00:54:47.280 --> 00:54:49.099 a lot of our partners have been affected quite 00:54:49.099 --> 00:54:52.619 strongly and we do see quite some risk here because 00:54:52.619 --> 00:54:56.199 the budgets are also affecting the data sets. 00:54:56.360 --> 00:54:58.159 And then we're looking at the climatological 00:54:58.159 --> 00:55:00.219 data sets, which we use if we make predictive 00:55:00.219 --> 00:55:03.260 models. It's important to have those available 00:55:03.260 --> 00:55:06.039 globally rather than only in certain regions 00:55:06.039 --> 00:55:09.719 because the climate is kind of a global thingy. 00:55:11.199 --> 00:55:15.219 So there we do expect it to affect us on the 00:55:15.219 --> 00:55:18.199 data set quite a bit and on the projects not 00:55:18.199 --> 00:55:22.000 on the active projects we have but on the yeah 00:55:22.000 --> 00:55:26.159 future ones unfortunately I think that's yeah 00:55:26.159 --> 00:55:31.650 that's great well On that depressing note, we're 00:55:31.650 --> 00:55:34.030 going to call it. Thank you so much, Lisa. Brooke 00:55:34.030 --> 00:55:37.429 Hellsend from Space for Good. Thanks for your 00:55:37.429 --> 00:55:39.369 wonderful presentation. I learned a lot. I know 00:55:39.369 --> 00:55:40.949 folks here learned a lot as well. Thank you very 00:55:40.949 --> 00:55:43.269 much, Lisa, for spending this hour with us. Thank 00:55:43.269 --> 00:55:45.289 you. And have a good evening and day, everyone. 00:55:45.829 --> 00:55:48.670 Thank you for listening to the Stage Talk. If 00:55:48.670 --> 00:55:51.670 you'd like to catch a Stage Talk live where you 00:55:51.670 --> 00:55:54.349 can ask the guest questions, join the Bell and 00:55:54.349 --> 00:56:02.639 Cat Discord server by visiting www. The music 00:56:02.639 --> 00:56:06.119 you've heard is titled Dawn by Newer Self and 00:56:06.119 --> 00:56:07.800 is courtesy of Artlist.