Jakson Mapping Humanity From Space: NASA’s Black Marble with Dr. Miguel Román täydellinen litterointi

33 00:02:37.591 --> 00:02:38.072 Okay. 34 00:02:38.832 --> 00:02:49.038 And they were recruiting people fresh out of college like me saying, hey, calibrate this thing, which is, is the data coming out of this new instrument correct? 35 00:02:49.438 --> 00:02:53.780 Is the integrity, everything about this instrument working? 36 00:02:54.761 --> 00:02:58.923 And I trained as an engineer that switched over to science. 37 00:02:58.983 --> 00:03:07.828 So I was always kind of working around the integrity of the data, working with the instrument teams, working with the science teams and trying to be a bridge. 38 00:03:08.648 --> 00:03:14.333 And they asked me, is the instrument state the visible infrared imaging radiometer sweep? 39 00:03:15.334 --> 00:03:15.835 Figure it out. 40 00:03:16.696 --> 00:03:17.456 They gave me two years. 41 00:03:18.577 --> 00:03:23.342 So first thing you do is you ask the spacecraft team to look at the moon. 42 00:03:24.877 --> 00:03:25.978 Because the moon is very stable. 43 00:03:26.939 --> 00:03:32.724 So if you look at the moon every month and the moon starts looking dark, you got a problem with Uranus. 44 00:03:33.024 --> 00:03:33.425 Fair enough. 45 00:03:33.585 --> 00:03:33.805 Yeah. 46 00:03:33.985 --> 00:03:36.447 So we started doing that. 47 00:03:36.507 --> 00:03:38.689 And this is literally you just point the thing at the moon. 48 00:03:38.850 --> 00:03:39.050 Okay. 49 00:03:39.070 --> 00:03:39.250 Right. 50 00:03:40.351 --> 00:03:41.052 Very stable. 51 00:03:41.352 --> 00:03:45.055 Problem is there's a lot of variability even though it's stable. 52 00:03:45.916 --> 00:03:47.497 So we needed more data. 53 00:03:48.338 --> 00:03:53.064 So what is the next most boring stable place on Earth? 54 00:03:53.945 --> 00:03:58.411 Just look at any desert where there's not been any kind of source of civilization. 55 00:03:58.451 --> 00:03:59.712 So we pick the Saharan does it. 56 00:04:00.408 --> 00:04:05.891 And so this picture you showed, a section of the Sinai Desert next to Cairo. 57 00:04:06.131 --> 00:04:09.073 Up there's, you know, all these places, Jordan, Israel. 58 00:04:10.073 --> 00:04:20.319 And we're tracking how the data changes in a place where there's no clouds, where the moon goes up and down, up and down. 59 00:04:20.399 --> 00:04:22.560 And suddenly, so it doesn't reach a peak. 60 00:04:23.620 --> 00:04:27.222 And suddenly, boom, we have to look at it like three or four times. 61 00:04:27.322 --> 00:04:29.544 We're seeing peaks over like the bright lights. 62 00:04:30.464 --> 00:04:33.266 in all major metropolitan districts in this region. 63 00:04:34.187 --> 00:04:34.507 Why? 64 00:04:35.048 --> 00:04:45.956 That every research article that we had read since then told us that, yes, lights do increase gradually from year to year, but it takes time. 65 00:04:46.176 --> 00:04:49.258 It shouldn't be from one season to the next. 66 00:04:50.619 --> 00:04:52.581 And we thought the instrument was broken. 67

00:04:52.601 --> 00:04:53.562 Okay, interesting. 68 00:04:53.602 --> 00:04:54.062 Interesting. 69 00:04:54.560 --> 00:04:57.581 We thought, okay, the low gain stage is fine. 70 00:04:57.621 --> 00:05:01.423 The high gain stage that detects very small signals was broken. 71 00:05:01.443 --> 00:05:02.784 It wasn't broken. 72 00:05:02.844 --> 00:05:09.207 And we knew because we had corrected two data sets and the anomaly was following the lunar calendar. 73 00:05:09.827 --> 00:05:11.288 not the Gregori Academy. 74 00:05:11.728 --> 00:05:13.329 And who uses the lunar calendar? 75 00:05:13.450 --> 00:05:13.730 Right. 76 00:05:13.810 --> 00:05:14.190 Musloops. 77 00:05:14.470 --> 00:05:14.710 Uh-huh. 78 00:05:15.511 --> 00:05:24.797 And so when we looked at 2013 and 2014, the anomaly had perfectly matched against the last day of Ramadan, the AA. 79 00:05:25.257 --> 00:05:26.979 So we were mapping Ramadams from space. 80 00:05:27.739 --> 00:05:31.041 The instrument was fine, but it had given us a gift. 81 00:05:31.081 --> 00:05:37.946 They had given us a capability to monitor humanity from space and literally culture from space. 82 00:05:39.330 --> 00:05:42.272 And so the next thing is like, well, we're a U.S. agency. 83 00:05:42.292 --> 00:05:43.392 Let's go and get Christmas. 84 00:05:44.613 --> 00:05:45.774 Also found Christmas lights. 85 00:05:45.974 --> 00:05:46.334 Okay. 86 00:05:47.234 --> 00:05:52.957 And that brought me back to my home of Puerto Rico because we're very proud of Puerto Rico, if you don't know. 87 00:05:52.997 --> 00:05:55.499 But our Christmas are the best Christmas. 88 00:05:56.819 --> 00:05:57.880 Longest Christmas ever. 89 00:05:57.980 --> 00:05:58.200 Yeah. 90 00:05:58.240 --> 00:05:59.841 Get in the comments if you disagree. 91 00:05:59.881 --> 00:06:00.081 Yes. 92 00:06:01.422 --> 00:06:06.985 Let me say, you know, starts at Brad Fridays, ends in January 20th with the San Sebastian festivities. 93 00:06:08.462 --> 00:06:14.925 And I felt that this was an opportunity to connect with these communities. 94 00:06:15.885 --> 00:06:16.706 And that was amazing. 95 00:06:16.746 --> 00:06:26.010 I mean, the other thing is we knew that we needed to do data filtering elsewhere because snow, if you think about it, snow at night, very bright too. 96 00:06:26.770 --> 00:06:29.791 So, Puerto Rico doesn't have snow, but you still have Christmas. 97 00:06:30.171 --> 00:06:33.633 So, again, just like the deaths, find an easy case study 98 00:06:35.250 --> 00:06:36.010 turned out beautifully. 99 00:06:36.070 --> 00:06:38.751 So this map, there's a Puerto Rican map also. 100 00:06:39.271 --> 00:06:40.632 Right. 101 00:06:41.852 --> 00:06:47.314 And at the time, it began to sink in that we were looking at a goal map. 102 00:06:48.975 --> 00:06:52.356 But it was, like I told you, that wasn't our job. 103 00:06:53.156 --> 00:06:58.498 And that's where a lot of our science staff began to ask questions about, hey, 104 00:06:59.850 --> 00:07:06.853

These are very bright lights, but it's like you've given us a new pair of glasses. 105 00:07:06.933 --> 00:07:15.436 And now we can see not just that activity, but we're also seeing these low light pieces of our geography at night. 106 00:07:16.096 --> 00:07:19.097 These places that before were being... 107 00:07:20.610 --> 00:07:28.553 You know, in a way, these bright lights were shadowing over the smaller light, rural communities, communities living outside the city. 108 00:07:29.653 --> 00:07:31.394 That's where the big reveal was. 109 00:07:32.054 --> 00:07:44.258 The fact that we can see all of humanity, not just the metropolitan areas that were occluding due to contamination of the lights, occluding these other pieces of society. 110 00:07:45.439 --> 00:07:47.240 And that's where the data came very proud. 111 00:07:47.701 --> 00:07:51.123 For a lot of people, it meant that you could track things like energy access. 112 00:07:51.203 --> 00:07:51.503 Right. 113 00:07:52.344 --> 00:07:53.725 You can track disasters. 114 00:07:54.926 --> 00:08:02.511 And at the time, we didn't know how powerful this tool was going to be for many different organizations. 115 00:08:02.752 --> 00:08:05.654 And we were just, I mean, we were building on our research to do that. 116 00:08:06.246 --> 00:08:08.867 Yeah, that is an incredible starting story. 117 00:08:10.608 --> 00:08:19.691 You were effectively just trying to calibrate this thing, and then you look to the side, and then there's just these random high-frequency events that are happening. 118 00:08:20.012 --> 00:08:20.732 That's pretty cool. 119 00:08:21.492 --> 00:08:25.855 And then from there, you know, this is sort of the first sojourn of the team. 120 00:08:25.975 --> 00:08:30.498 This is before even the team of Black Marvel was formed, right? 121 00:08:30.558 --> 00:08:32.499 We were just two at the time. 122 00:08:32.539 --> 00:08:32.759 Yeah. 123 00:08:32.779 --> 00:08:44.666 So how do you transition from like, okay, there's like random people doing calibration to now like there's, you get like a consortium of scientists and you and your colleagues who are now being recognized for the Golden Goose Awards. 124 00:08:45.266 --> 00:08:50.410 Like, how do you start convincing, you know, people that this is something that's worth... 125 00:08:51.130 --> 00:08:54.091 full-time scientists to work on. 126 00:08:54.791 --> 00:08:55.072 Yes. 127 00:08:56.212 --> 00:09:08.056 We were very blessed with a very multi-talented team that brought unique skills and different questions to our group. 128 00:09:08.777 --> 00:09:12.178 Because if it were for me, we would be just focusing on data filtering. 129 00:09:13.624 --> 00:09:23.268 To me, the most important job of a NASA scientist is to create highly pristine science measurements of the Earth system. 130 00:09:24.448 --> 00:09:34.212 And at the time, this has changed a lot, but at the time, we would just generate the gold standard product, throw it over the defense, and allow the science community to ask those questions. 131 00:09:34.932 --> 00:09:42.055 So in a way, years later, we then evolved as an agency, as an organization, and now... 132 00:09:42.935 --> 00:09:49.698 NASA scientists are doing a lot of work that crosses these boundaries, crosses the boundaries around nature, technology, and society. 133 00:09:51.078 --> 00:09:57.781 So in a way, we were pioneers in that new mindset of developing our science to action. 134 00:09:58.901 --> 00:10:02.182 And we brought urban scientists to our team very early. 135 00:10:02.342 --> 00:10:02.942 Oh, okay, cool. 136 00:10:02.962 --> 00:10:09.645 We brought people who had degrees from Yale in understanding urbanization as a process. 137 00:10:09.665 --> 00:10:09.745 Wow. 138 00:10:10.756 --> 00:10:22.649 And instead of giving them maps of land cover to show where are the cities and where are the non-urban areas, now they had something that helped describe the function of these places. 139 00:10:23.310 --> 00:10:28.195 Because you can see a building, if the lights are off, the building is still there during the day. 140 00:10:29.650 --> 00:10:40.134 And so this gave a lot of versatility to what kind of questions you can ask from an urban development, from an infrastructure point of view that we would never be able to do before. 141 00:10:40.274 --> 00:10:46.516 Unless you had a lot of money to kind of literally manually count buildings and how much energy they use and everything. 142 00:10:47.697 --> 00:10:51.258 And we understood that we had to be careful about what we were really seeing. 143 00:10:52.059 --> 00:10:53.079 We were seeing lights at night. 144 00:10:54.490 --> 00:10:57.312 And yes, it's been used as a proxy of many different things. 145 00:10:58.013 --> 00:11:01.055 But we were always very cautious in approaching that with integrity. 146 00:11:01.416 --> 00:11:08.901 Like we weren't overselling the capability and saying, this is a good way to replace estimates of total electricity. 147 00:11:08.921 --> 00:11:09.522 Right. 148 00:11:09.742 --> 00:11:09.882 No. 149 00:11:10.142 --> 00:11:10.383 Yeah. 150 00:11:10.603 --> 00:11:15.106 No, this is a piece that gives you better insights globally, geographically. 151 00:11:16.187 --> 00:11:22.212 And I think the part that we began to coalesce was when we hired someone who was entirely focused on applied science, 152 00:11:23.897 --> 00:11:33.926 And that's where the community became broader because we were doing things like light pollution, so monitoring dark skies, but we were also doing disaster response. 153 00:11:34.578 --> 00:11:45.684 So we were covering the entire gamut of energy conservation, wildlife preservation, and then emergency preparedness, resilience, and all these things. 154 00:11:46.264 --> 00:11:48.725 And by then, we were just a team of four or five people. 155 00:11:49.666 --> 00:11:51.587 Yeah, that's pretty incredible. 156 00:11:51.747 --> 00:11:57.450 And so we went

back to, you can't just run such an amazing capability with five people. 157 00:11:58.470 --> 00:12:03.793 And our NASA leaders pushed us to make it full and open to the world. 158 00:12:04.514 --> 00:12:08.156 That happened eight years after this discovery. 159 00:12:08.916 --> 00:12:11.358 And that, in fact, was our biggest contribution. 160 00:12:11.478 --> 00:12:14.920 We made, we really embraced open science. 161 00:12:15.901 --> 00:12:19.663 And that means that the data was not just in a web page. 162 00:12:20.884 --> 00:12:23.045 You know, you can find it, you can access it. 163 00:12:23.105 --> 00:12:26.407 You can reproduce it from start to, from scratch. 164 00:12:27.893 --> 00:12:31.881 And that's what it has attracted so many people from around the world. 165 00:12:32.943 --> 00:12:35.889 That before they were just like just a handful of teams. 166 00:12:37.242 --> 00:12:43.148 The majority of the research produced now is produced by outside organizations. 167 00:12:43.308 --> 00:12:53.799 NASA still published, but we are seen as the folks that preserve the integrity of the data. 168 00:12:53.939 --> 00:12:58.264 That still is why they come to NASA and ask us. 169 00:12:58.630 --> 00:13:05.779 Well, so speaking of, like, you know, preserving the data and presenting pristine data, that's a big passion of yours. 170 00:13:06.220 --> 00:13:10.345 So I wanted to ask you just how difficult it is, actually. 171 00:13:10.766 --> 00:13:15.131 If you don't mind, we like getting technical on this podcast because... 172 00:13:15.932 --> 00:13:17.914 You know, the name is From First Principles. 173 00:13:18.094 --> 00:13:22.779 So we try to develop the process of how science is done rather than just the answers. 174 00:13:23.059 --> 00:13:25.161 I think the answers are certainly really cool, right? 175 00:13:25.201 --> 00:13:33.688 The fact that you can see Muslim cities of Cairo and Amman fluctuating in brightness during Ramadan and it like coincides with Eid. 176 00:13:33.728 --> 00:13:34.869 But I know where you're getting with this. 177 00:13:34.889 --> 00:13:36.091 Do you know what it takes to get? 178 00:13:36.271 --> 00:13:36.711 Exactly. 179 00:13:36.791 --> 00:13:37.512 It's not a lot. 180 00:13:37.812 --> 00:13:38.553 A lot of work. 181 00:13:38.613 --> 00:13:44.378 A lot of work goes into convincing yourself that you're not being fooled because the easiest person to fool is yourself. 182 00:13:44.778 --> 00:13:50.163 So what I wanted to ask you with this, this is a figure from, I think it's the first paper. 183 00:13:50.183 --> 00:13:52.065 Algorithm theoretical basic document. 184 00:13:52.145 --> 00:13:52.605 Exactly. 185 00:13:52.826 --> 00:14:00.693 Which is NASA's speech for no one had written the theory on understanding how lights interact with the sky at night since 2018. 186 00:14:00.913 --> 00:14:03.195 You have to create a theory from scratch. 187 00:14:03.435 --> 00:14:04.196 Exactly. 188 00:14:04.416 --> 00:14:07.941 So I wanted to ask you, like, so there's moonlight, obviously, in this figure. 189 00:14:07.961 --> 00:14:09.442 There's air glow? 190 00:14:09.622 --> 00:14:13.387 Yeah, air glow is just the atmosphere glowing? 191 00:14:13.868 --> 00:14:18.373 Yes, from precipitation electrons from the hemiosphere. 192 00:14:18.713 --> 00:14:19.113 Oh, wow. 193 00:14:19.133 --> 00:14:20.394 So interacting with charged particles. 194 00:14:20.414 --> 00:14:22.816 Yes, hydroxies and atmospheric chemistry happening. 195 00:14:23.217 --> 00:14:26.039 So every time there's a storm event, you see aurora. 196 00:14:26.659 --> 00:14:28.000 And that becomes noise for us. 197 00:14:28.140 --> 00:14:35.346 But hemeophysicists love to go get those maps to do early warnings for potential impacts to infrastructure like a power grid. 198 00:14:35.846 --> 00:14:36.567 It's the same data set. 199 00:14:36.887 --> 00:14:37.268 Oh, it's the... 200 00:14:37.448 --> 00:14:38.729 Wait, so hold on. 201 00:14:38.829 --> 00:14:41.291 So let's start from just the satellite, okay? 202 00:14:41.811 --> 00:14:44.073 What are the wavelengths that it's sensitive to? 203 00:14:44.516 --> 00:14:46.640 So there's only one band. 204 00:14:46.680 --> 00:14:47.782 There's only one band. 205 00:14:47.822 --> 00:14:50.987 There's only one panchromatic one from 0.5 to 900 nanometers. 206 00:14:52.795 --> 00:14:57.298 And it captures all the infrared, visible and infrared light in that. 207 00:14:57.598 --> 00:14:58.578 So it's just brightness. 208 00:14:58.678 --> 00:14:59.519 It's just brightness. 209 00:14:59.839 --> 00:15:01.340 Okay, that's crazy, actually. 210 00:15:01.640 --> 00:15:02.241 Yes. 211 00:15:02.581 --> 00:15:17.029 And because had it been designed by someone with expertise in night-time lights, it would have been done completely different, including the LED pump 420, which a lot of the lights that we see now are transitioning to LEDs. 212 00:15:17.269 --> 00:15:18.030 We can't see that. 213 00:15:18.390 --> 00:15:18.630 Why? 214 00:15:18.850 --> 00:15:20.331 Because it was not built to do that. 215 00:15:20.431 --> 00:15:22.853 It was built to measure clouds at night for weather. 216 00:15:23.113 --> 00:15:23.934 It's just a little sound. 217 00:15:23.954 --> 00:15:29.237 Wait, so wait, no, now I'm jumping the gun because like, so the last paper that I wanted to, we're just going for it. 218 00:15:29.537 --> 00:15:31.678 So the last, the last figure I have is this one. 219 00:15:31.698 --> 00:15:33.880 This is the very recent nature paper. 220 00:15:34.420 --> 00:15:38.244 And there you see Europe dimming, but that's because of the... 221 00:15:38.564 --> 00:15:40.546 It's not dimming because Europe is poor. 222 00:15:40.646 --> 00:15:42.809 It's not dimming because there's some power. 223 00:15:43.049 --> 00:15:47.433 It's dimming because all of Europe decided to switch two LED lights in three years. 224 00:15:47.613 --> 00:15:47.773 Right. 225 00:15:47.793 --> 00:15:49.375 So the

brightness could remain the same. 226 00:15:49.475 --> 00:15:49.675 Yes. 227 00:15:50.016 --> 00:15:50.176 Right. 228 00:15:50.236 --> 00:15:51.217 It's not that the brightness is the minute. 229 00:15:51.237 --> 00:15:52.178 The brightness is getting worse. 230 00:15:52.478 --> 00:15:53.139 Yeah, yeah, of course. 231 00:15:53.159 --> 00:15:56.701 But in our data, it looks like it's going down because we don't see the blue. 232 00:15:56.961 --> 00:15:58.062 Yeah, that was unclear to me. 233 00:15:58.262 --> 00:16:04.807 So before we did this interview, we did like a deep dive on five papers that you were co-author on. 234 00:16:05.227 --> 00:16:05.908 And that was the last one. 235 00:16:05.928 --> 00:16:10.151 And it was kind of unclear to me why the brightness would go. 236 00:16:10.211 --> 00:16:15.075 I mean, the explanation was in the paper that, you know, there's policy changes and there's different... 237 00:16:15.675 --> 00:16:42.984 types of lights but now it's very clear that it's just a single band and the thing is outside the band and it is not only a single band it's a single band with a narrow aperture in wavelength that we made trades at the time we needed to be more in the infrared because i i sacrificed the blue but i gained infrared life to monitor fires and flaring okay fair enough so that's always the 238 00:16:45.803 --> 00:16:47.725 What do you give off and what do you gain? 239 00:16:48.385 --> 00:16:56.112 And at NASA, we've had to do that many, many times to help address the needs and priorities of broader science communities. 240 00:16:56.393 --> 00:17:04.780 You know, when we built this same instrument was built to measure clouds, to measure ocean color, to measure fires, to measure forests. 241 00:17:05.501 --> 00:17:10.946 So by the time you build those requirements, you're talking about a $7.40 million instrument. 242 00:17:12.355 --> 00:17:15.616 And so at the end, you add, oh, and I want to add blue LEDs. 243 00:17:15.736 --> 00:17:17.217 Okay, that's going to be another $2 million. 244 00:17:17.237 --> 00:17:18.918 Like, at some point, you got to stop. 245 00:17:19.478 --> 00:17:26.041 You know, so that's... You can't just... You can't... What's the technical reason why the band has to be, like, a certain width? 246 00:17:26.161 --> 00:17:29.322 Why can't it be a little... Like, why is it more expensive so much? 247 00:17:29.402 --> 00:17:31.203 It's true of my requirements. 248 00:17:31.303 --> 00:17:32.883 It's not just this one decision. 249 00:17:33.064 --> 00:17:35.905 It's decisions going back to the requirements of... 250 00:17:36.825 --> 00:17:41.588 a large swath of stakeholders across the federal community. 251 00:17:41.608 --> 00:17:42.448 Yeah, that makes sense. 252 00:17:42.468 --> 00:17:45.290 The OD, NOAA, and NASA, they all have different customers. 253 00:17:45.810 --> 00:17:46.970 They all have different stakeholders. 254 00:17:47.030 --> 00:17:51.553 You've got folks who need to support weather.

255 00:17:52.073 --> 00:17:56.936 And so at the time, the requirement was we need to be able to see 256 00:17:58.341 --> 00:18:20.613 low fog and cloud in alaska and these are the bands that allows us to optimize again that christmas lights was not a factor yeah yeah and it's about it's about weather at the end of the day it's all about this is our nation's weather workforce yeah and so the priority is let's try the five-day forecast if there's signs that comes out of that we'll revisit it right okay 257 00:18:21.473 --> 00:18:29.622 And so, okay, so back to the figure, you've got moonlight that penetrates our atmosphere, hits the ground, goes back. 258 00:18:30.042 --> 00:18:36.609 Sometimes it goes back into the ground and then comes back because there's like just photons are doing weird things. 259 00:18:37.089 --> 00:18:40.533 So there's the air glow from the atmosphere. 260 00:18:41.073 --> 00:18:42.555 Then there's clouds. 261 00:18:42.835 --> 00:18:46.939 There's all sorts of like random crap that photons are hitting. 262 00:18:46.979 --> 00:18:47.219 Yes. 263 00:18:47.420 --> 00:18:48.301 Before it hits your sensor. 264 00:18:48.321 --> 00:18:51.344 Even before we did this, we didn't even know how difficult it was. 265 00:18:51.524 --> 00:18:52.505 So here's another problem. 266 00:18:53.730 --> 00:18:56.232 The lights from buildings also vary by angle. 267 00:18:57.193 --> 00:18:58.474 So it matters a lot. 268 00:18:58.774 --> 00:19:00.736 And that's where we accounted for that in this paper. 269 00:19:01.136 --> 00:19:01.536 Oh, wow. 270 00:19:01.556 --> 00:19:01.796 Okay. 271 00:19:01.916 --> 00:19:10.763 If you're in New York City, Times Square, and your spacecraft is looking obliquely at 45 degrees, you're not seeing the lights. 272 00:19:10.783 --> 00:19:11.264 The billboard. 273 00:19:11.304 --> 00:19:11.984 The billboard. 274 00:19:12.325 --> 00:19:12.705 Ah, okay. 275 00:19:12.745 --> 00:19:18.349 So you have to be looking from really across the light of sight. 276 00:19:18.409 --> 00:19:19.710 So what we call Nader. 277 00:19:20.691 --> 00:19:25.735 And so we realized that this instrument had a lot of structure in the data. 278 00:19:26.055 --> 00:19:35.983 If you understand the directional signature, the multi-angular signature, because you have natural variation, but you also have anthropogenic variation going on in the mirror. 279 00:19:36.283 --> 00:19:36.663 Wow. 280 00:19:36.763 --> 00:19:46.130 So the biggest breakthrough between this paper and this paper was the fact that we understood that the world, the light in the world, also vary a lot angrily. 281 00:19:46.550 --> 00:19:46.971 Wow. 282 00:19:47.071 --> 00:19:48.712 And we didn't account for that. 283 00:19:48.972 --> 00:19:49.392 Right. 284 00:19:49.412 --> 00:19:51.094 Here you have angle from the moon. 285 00:19:51.174 --> 00:19:51.374 Yes. 286 00:19:51.554 --> 00:19:52.535 You've accounted for that. 287 00:19:52.575 --> 00:19:52.795 Yes. 288 00:19:53.115 --> 00:19:54.156 That we carry for that. 289 00:19:54.256 --> 00:19:57.737 And we basically use atmospheric correction routines from daytime. 290 00:19:57.817 --> 00:20:00.078 And we just basically plop the song on the moon. 291 00:20:00.139 --> 00:20:00.679 That's all we do. 292 00:20:00.699 --> 00:20:00.979 Ah, okay. 293 00:20:01.059 --> 00:20:01.879 Yeah, fair enough. 294 00:20:01.939 --> 00:20:06.582 Yes, because the inherent optical properties of the atmosphere is really the same. 295 00:20:06.642 --> 00:20:08.583 You have aerosols, you have molecules. 296 00:20:09.383 --> 00:20:10.784 So you can make some approximations. 297 00:20:10.885 --> 00:20:13.607 And those approximations have hold now for 12 years. 298 00:20:14.248 --> 00:20:18.612 The biggest state of the art now is trying to figure out the angular structure of artificial life. 299 00:20:19.293 --> 00:20:25.218 And that is very fascinating because you could relate those different modes. 300 00:20:25.899 --> 00:20:30.243 So even though we don't have colors, we have a lot of understanding of the architecture of the lights. 301 00:20:31.264 --> 00:20:34.145 And that is how we started looking at the volatility of the lights. 302 00:20:35.045 --> 00:20:41.607 Because the fact that it's dark at this angle and it's bright at this angle, what does it mean? 303 00:20:42.068 --> 00:20:42.728 I have something. 304 00:20:43.248 --> 00:20:44.048 I have building height. 305 00:20:44.448 --> 00:20:44.708 Right. 306 00:20:45.149 --> 00:20:49.070 So it's literally I can extract building heights from the lights there. 307 00:20:49.370 --> 00:20:49.630 Yeah. 308 00:20:50.030 --> 00:20:51.511 Yeah, yeah, yeah. 309 00:20:51.871 --> 00:20:52.831 That's pretty cool. 310 00:20:53.031 --> 00:20:59.013 And if I can track building lights over a 40-year period, I know how your country is growing. 311 00:21:00.201 --> 00:21:02.984 China urbanized in 20 years. 312 00:21:03.544 --> 00:21:05.766 They built up and they ran out of space. 313 00:21:06.607 --> 00:21:07.067 Who's next? 314 00:21:07.588 --> 00:21:07.848 India. 315 00:21:08.569 --> 00:21:12.592 So India is going to be the next civilization that is still building horizontally. 316 00:21:12.713 --> 00:21:14.814 And now they're building up. 317 00:21:15.255 --> 00:21:17.997 Eleanor Stoke, she's our expert in urbanization. 318 00:21:18.037 --> 00:21:20.540 She came out of Karen Seelos, grew back at Yale. 319 00:21:21.321 --> 00:21:22.101 They're the experts. 320 00:21:22.121 --> 00:21:36.598 They're the ones informing mayors on how to build cities sustainably, where you don't end up in a situation where you're building so fast that you're not gaining benefits from living in cities that are more accessible. 321 00:21:37.479 --> 00:21:39.181 Those are the questions this field of 780s. 322 00:21:40.302 --> 00:21:48.047 Hopefully Los Angeles also takes note because I would love to have more housing in Los Angeles. 323 00:21:48.328 --> 00:21:59.134 So in order to do

that, you need to understand the makeup of transportation, commercial, housing, understanding how those functions change over time. 324 00:21:59.835 --> 00:22:03.117 For the first time, we have data sets that can help you understand that variation. 325 00:22:04.498 --> 00:22:09.642 And you can now understand the volatility of those sectors as a result of disasters. 326 00:22:11.023 --> 00:22:13.505 So that's another factor that's very unique to this. 327 00:22:13.725 --> 00:22:14.265 Yeah, yeah. 328 00:22:14.345 --> 00:22:18.428 Actually, so one of the things that I wanted to ask you about was the big Puerto Rico paper. 329 00:22:18.889 --> 00:22:25.214 So first of all, this must have been personally a very big deal for you, right? 330 00:22:28.076 --> 00:22:30.518 How'd you come across doing this? 331 00:22:31.643 --> 00:22:38.285 Yeah, so I remember the day I was at Goddard, had my monitor, 2016, September 20th. 332 00:22:39.745 --> 00:22:47.548 The first major outage, like, of an entire area was... 333 00:22:48.744 --> 00:22:53.710 a power outage that happened at the Aguirre power plant in Guayama that day. 334 00:22:55.051 --> 00:22:59.696 And with this as early Black Marble, we were still manually downloading data. 335 00:22:59.917 --> 00:23:01.318 No, we weren't in the cloud. 336 00:23:02.059 --> 00:23:04.362 And I put Puerto Rico on just like I do every day. 337 00:23:04.422 --> 00:23:07.325 I drink my cafecito and I'm going to look at Puerto Rico. 338 00:23:07.345 --> 00:23:08.947 And I have to look three times. 339 00:23:10.488 --> 00:23:13.711 And my hard job, because it had disappeared. 340 00:23:13.991 --> 00:23:14.231 Yeah. 341 00:23:14.732 --> 00:23:16.333 And this is just after the hurricane? 342 00:23:16.493 --> 00:23:17.554 Before the hurricane. 343 00:23:17.734 --> 00:23:18.915 A year before the hurricane. 344 00:23:18.935 --> 00:23:20.096 Oh, before the hurricane. 345 00:23:20.616 --> 00:23:22.557 That was kind of a warning sign. 346 00:23:22.918 --> 00:23:23.478 Interesting. 347 00:23:23.498 --> 00:23:27.481 That the electricity system of Puerto Rico was already very vulnerable. 348 00:23:27.641 --> 00:23:29.443 And sending us an early signal. 349 00:23:30.063 --> 00:23:40.384 of what became the largest power artist in Puerto Rico in U.S. history at the time, which is Hurricane Irma and Maria that followed each other within two weeks. 350 00:23:42.494 --> 00:23:46.398 And the project changed forever after that. 351 00:23:46.418 --> 00:23:49.662 We knew what was going on. 352 00:23:50.423 --> 00:23:52.926 We weren't, you know, our head didn't skip at that time. 353 00:23:53.387 --> 00:23:59.694 And rather than having a concern, it was, you know, we had a sense of urgency as an agency. 354 00:24:00.695 --> 00:24:01.756 Not just our team. 355 00:24:01.856 --> 00:24:08.319 I mean, I remember being part of NASA's disaster response coordination team. 356 00:24:09.120 --> 00:24:16.384 We were in meetings at the White House with our partners and other federal agencies helping understand the situation in Puerto Rico from us. 357 00:24:17.724 --> 00:24:24.088 And everything we had done prior to this was curiosity-driven. 358 00:24:25.068 --> 00:24:29.171 It was understanding the night, understanding how to protect the night, 359 00:24:30.015 --> 00:24:33.503 you know, the natural patrimony that is to be able to have dark skies. 360 00:24:35.128 --> 00:24:37.850 Puerto Rico changed that. 361 00:24:38.691 --> 00:24:41.714 We started asking different questions because of this. 362 00:24:41.814 --> 00:24:48.900 And it wasn't not enough to just coming up with a power outage map and putting it in the front cover of Washington Post. 363 00:24:49.221 --> 00:24:50.202 We used to do that a lot. 364 00:24:50.382 --> 00:24:50.622 Yeah. 365 00:24:51.062 --> 00:24:53.144 But after this, it wasn't enough. 366 00:24:53.444 --> 00:25:00.551 We had to stay and monitor which communities were still without power for how long. 367 00:25:02.092 --> 00:25:06.636 And the entire project became a focus on recovery and resilience. 368 00:25:08.277 --> 00:25:16.984 And, yeah, so we were, I remember being on calls with the Army Northern Command in Puerto Rico, and they were moving diesel power generators around the island. 369 00:25:18.465 --> 00:25:21.547 And I got a call from a captain, like, sir, like, we're going to move. 370 00:25:22.508 --> 00:25:24.730 We want to go to this town, Barranquitas. 371 00:25:26.071 --> 00:25:31.375 And I was like, yeah, they've been without power for 60 days based on the DOE report and our night time. 372 00:25:33.050 --> 00:25:39.575 And we were part of the search capacity force, the federal-wide search capacity force that had not been activated since Katrina. 373 00:25:40.476 --> 00:25:40.856 Wow. 374 00:25:42.037 --> 00:25:49.182 Nowadays, there were a lot of FEMA folks in Puerto Rico eating MREs for weeks. 375 00:25:49.883 --> 00:25:53.866 And that taught us something very important because we were here, we were in D.C. 376 00:25:56.107 --> 00:26:02.392 It taught us that you can look at all this stuff from space, but if you can't connect with those communities and people with that knowledge... 377 00:26:04.258 --> 00:26:15.507 And that's where Black Marvel evolved, thanks to Puerto Rico, because we began to realize that you can't do Earth action, Earth science to action, without involving the local expertise. 378 00:26:17.668 --> 00:26:22.372 And, you know, at the time, like I've grown now, you know, since then, so it's been nine years. 379 00:26:22.412 --> 00:26:24.013 Yeah. 380 00:26:24.193 --> 00:26:28.577 It really, that personal connection at the time was very hard. 381 00:26:29.409 --> 00:26:36.072 But now I've sort of seen how much we've learned from these events. 382 00:26:37.013 --> 00:26:39.574 And the world has learned. 383 00:26:39.654 --> 00:26:48.118 The world has become more resilient in a way because we know what happens when you have these

large-scale events. 384 00:26:48.879 --> 00:26:51.580 And how can that impact centralized electricity systems? 385 00:26:52.835 --> 00:27:00.342 And you go to Puerto Rico, everybody has, you know, domestically, you know, every home now has batteries and solar power. 386 00:27:01.103 --> 00:27:01.423 Nice. 387 00:27:02.104 --> 00:27:02.364 Right? 388 00:27:02.484 --> 00:27:08.169 And there was both local investment, federal investment going into restoring that. 389 00:27:08.310 --> 00:27:10.371 And you can see that in the data, too. 390 00:27:10.772 --> 00:27:11.032 Right. 391 00:27:11.192 --> 00:27:11.513 Really? 392 00:27:11.653 --> 00:27:11.813 Yeah. 393 00:27:12.033 --> 00:27:12.694 You can see that. 394 00:27:12.714 --> 00:27:17.138 You see, like, the coffee-grown mountain region of Puerto Rico, Alcuntas. 395 00:27:18.411 --> 00:27:20.412 They're now fully electrified with solar. 396 00:27:21.192 --> 00:27:23.193 And every time there's an hour, they're still on. 397 00:27:23.273 --> 00:27:23.853 They're still on. 398 00:27:24.174 --> 00:27:26.054 And you can see that in our anomalies course. 399 00:27:26.955 --> 00:27:27.755 Years later. 400 00:27:27.935 --> 00:27:28.715 Eight years later. 401 00:27:29.116 --> 00:27:33.418 Yeah, so that's like significant changes at the ground level. 402 00:27:33.958 --> 00:27:36.819 Because of... 403 00:27:36.919 --> 00:27:39.520 I mean, this is like a classic golden goose story, right? 404 00:27:39.540 --> 00:27:43.122 You're chasing random holiday lights. 405 00:27:43.342 --> 00:27:48.204 And it turns into this incredible story where communities are changed forever. 406 00:27:48.756 --> 00:27:53.160 You know, now that I'm 40, I love salsa music more than I'm used to when I was a kid. 407 00:27:53.360 --> 00:27:53.660 Okay. 408 00:27:53.981 --> 00:27:56.263 And there's this amazing singer-songwriter. 409 00:27:56.323 --> 00:27:57.744 His name is Ismael Libera. 410 00:27:59.746 --> 00:28:03.809 And now every time I go to Goddard, I have, you know, Amazon Music on. 411 00:28:03.849 --> 00:28:04.690 I'm listening to Maero. 412 00:28:06.321 --> 00:28:11.463 And he talks about this little chantita off of San Juan called La Perla. 413 00:28:11.824 --> 00:28:17.046 If you ever seen the music video of Despacito, that's La Perla, right? 414 00:28:17.326 --> 00:28:18.647 Oh, okay. 415 00:28:18.727 --> 00:28:19.167 All right. 416 00:28:19.247 --> 00:28:21.148 Sorry, I have seen the video, yes. 417 00:28:22.268 --> 00:28:29.651 And they talk about this town outside the walls, the old walls of Old San Juan. 418 00:28:30.892 --> 00:28:32.953 He talks about this community with dignities. 419 00:28:34.000 --> 00:28:42.605 You know, and it's funny, like it took me, I had to come work at NASA to understand some of the things this man was saying. 420 00:28:43.706 --> 00:28:51.491 That a lot of what we have learned from Black Marble are just basic things about protecting a community, right? 421 00:28:52.972 --> 00:28:57.895 Not everything that's important in life is the thing that shines the brightest. 422 00:28:58.867 --> 00:29:00.087 It's hard for me to translate. 423 00:29:00.107 --> 00:29:00.567 Yeah, yeah. 424 00:29:00.607 --> 00:29:00.947 Okay. 425 00:29:01.588 --> 00:29:02.088 I'm not sure. 426 00:29:02.248 --> 00:29:06.169 People can get hit really hard and they get up. 427 00:29:06.249 --> 00:29:07.669 That's a very American thing. 428 00:29:08.089 --> 00:29:11.690 You see it everywhere in hurricane-torn regions, in tornado-torn regions. 429 00:29:12.110 --> 00:29:14.330 People just feel bad, better. 430 00:29:14.650 --> 00:29:15.631 They just get up. 431 00:29:16.991 --> 00:29:21.312 And again, the credibility comes by understanding that community. 432 00:29:22.620 --> 00:29:23.500 At a personal level. 433 00:29:23.580 --> 00:29:25.601 And we just came back from Puerto Rico last week. 434 00:29:25.701 --> 00:29:30.582 We opened an earth information center at the local science museum. 435 00:29:30.782 --> 00:29:31.442 Okay, nice. 436 00:29:31.582 --> 00:29:40.384 And it was a small event, so we had 50 registrations every hour. 437 00:29:40.444 --> 00:29:42.465 So we did 300 people, right? 438 00:29:44.265 --> 00:29:52.287 People from over the island wanting to learn about Black Marble, wanting to learn about how can NASA data help their communities. 439 00:29:54.233 --> 00:30:03.159 I would have never thought, you know, we would be able to reach communities from Puerto Rico learning about what we're doing at NASA. 440 00:30:03.259 --> 00:30:03.699 It's just. 441 00:30:04.680 --> 00:30:06.140 That's that is that is really cool. 442 00:30:06.160 --> 00:30:09.643 And that must have been so meaningful for you to go back to that community. 443 00:30:10.143 --> 00:30:14.966 And going back to a community where you see the outages of places that you know. 444 00:30:16.556 --> 00:30:17.780 Streets that you have not. 445 00:30:17.940 --> 00:30:18.562 Right, right. 446 00:30:19.745 --> 00:30:20.407 Yeah, that hits you. 447 00:30:21.490 --> 00:30:22.493 That's pretty cool. 448 00:30:23.094 --> 00:30:23.816 I did want to... 449 00:30:25.827 --> 00:30:35.675 One of the innovations, I think, at least from what I read in this paper, that was kind of like, it sounded miraculous to me when I first read it. 450 00:30:35.775 --> 00:30:46.263 It's this figure where it shows the resolution of the Black Marble Standard product, which is like 500 meters, so half a kilometer by half a kilometer. 451 00:30:46.843 --> 00:30:47.564 Which, okay, fine. 452 00:30:47.584 --> 00:30:49.226 That's believable to me, right? 453 00:30:49.326 --> 00:30:50.347 That it's like nighttime. 454 00:30:50.367 --> 00:30:51.569 You don't get a lot of photons. 455 00:30:51.949 --> 00:30:53.951 Yeah, but how did we get there? 456 00:30:54.011 --> 00:31:00.639 But then it's completely unclear to me what data magic happened to get you to 30 meter resolution. 457 00:31:00.659 --> 00:31:01.260 I can explain.

458 00:31:01.340 --> 00:31:03.122 Yeah, so that's why I need that. 459 00:31:03.202 --> 00:31:04.083 This came out of a need. 460 00:31:04.103 --> 00:31:04.324 Yeah. 461 00:31:04.384 --> 00:31:04.584 Yeah. 462 00:31:06.646 --> 00:31:12.628 When we were talking to FEMA and local communities, they're saying, we love this. 463 00:31:12.648 --> 00:31:15.990 You're telling us where the lights have gone down. 464 00:31:16.590 --> 00:31:19.031 But your data set is so coarse. 465 00:31:19.331 --> 00:31:21.532 I need to get down to the house level. 466 00:31:22.052 --> 00:31:23.272 And I'm like, I can't help you. 467 00:31:23.972 --> 00:31:24.673 I can't help you. 468 00:31:24.713 --> 00:31:26.553 I only have this weather salad. 469 00:31:26.573 --> 00:31:26.773 Right? 470 00:31:28.074 --> 00:31:33.296 So, but then one of our team members say, but we have Landsat. 471 00:31:34.417 --> 00:31:40.620 And Landsat, another USGS NASA mission, can look at buildings at the hectare scale, right? 472 00:31:40.860 --> 00:31:42.861 You can see streetlights and streets. 473 00:31:43.461 --> 00:31:46.662 You don't know where those streets have streetlights or not. 474 00:31:48.023 --> 00:31:49.844 Why didn't we merge those two data sets? 475 00:31:50.570 --> 00:31:55.791 The Landsat Archive, which is like the crown jewel of like our science. 476 00:31:55.871 --> 00:31:56.952 Yeah, it's amazing. 477 00:31:56.972 --> 00:31:57.972 50 years going in. 478 00:31:59.372 --> 00:32:02.133 And we're like, well, we know we can count on Landsat being good. 479 00:32:02.393 --> 00:32:05.373 Yeah, because there's other guys like you and Landsat team. 480 00:32:05.413 --> 00:32:10.155 Yeah, other big Ls at Landsat who know like, oh, they can get down the quality of the data down to point. 481 00:32:10.175 --> 00:32:12.235 Yeah, their whole thing is quality of data. 482 00:32:12.675 --> 00:32:19.917 So we know that if we can get in previous area for Landsat, meaning anywhere that there's a map, 483 00:32:21.298 --> 00:32:24.781 and you make this assumption, the more cement, the more legs. 484 00:32:26.002 --> 00:32:26.262 Okay. 485 00:32:26.863 --> 00:32:28.164 I think it's a fair assumption. 486 00:32:28.264 --> 00:32:28.965 Fair assumption. 487 00:32:29.746 --> 00:32:37.233 Then now you can create a fused image between, in that pixel, anywhere where there's cement, 488 00:32:39.093 --> 00:32:43.218 They attribute the likes of that value over that area over that region. 489 00:32:43.478 --> 00:32:48.064 This came out of a need because we don't have a Black Marble satellite at 10 meter resolution. 490 00:32:48.625 --> 00:32:52.770 We do now, thanks to NASA's Commercial Satellite Data Apposition Program. 491 00:32:53.906 --> 00:33:04.433 So I can now, this is the awesome thing about industry being part of NASA, I can now ask a company to take a picture at night and buy the data for research. 492 00:33:05.173 --> 00:33:12.278 And they go through an award and then they make their satellite system available to researchers. 493 00:33:13.643 --> 00:33:18.251 So we have done, as of last year, there was a major power outage in Mississippi. 494 00:33:18.932 --> 00:33:22.298 And our disaster response system was, I want high-resolution images. 495 00:33:22.559 --> 00:33:26.526 And we detected power outages with real, real nighttime nights. 496 00:33:27.602 --> 00:33:36.244 But between now and then, we were basically creating a simulated map. 497 00:33:36.344 --> 00:33:36.624 Yeah. 498 00:33:36.824 --> 00:33:40.285 And that map is not inside Traso. 499 00:33:41.245 --> 00:33:41.865 What does that mean? 500 00:33:42.205 --> 00:33:45.926 It means you can visualize it, but you don't know what the values mean. 501 00:33:46.406 --> 00:33:49.126 It's just for situational awareness only. 502 00:33:50.267 --> 00:33:52.127 Nothing replaces a physical measurement. 503 00:33:53.585 --> 00:33:58.007 So that's why these maps, you can see them in museum displays. 504 00:33:58.027 --> 00:33:58.588 You can see them. 505 00:33:58.968 --> 00:34:03.270 But analysis, you still use the fire in a mirror. 506 00:34:03.350 --> 00:34:05.571 It needs something of high integrity. 507 00:34:05.731 --> 00:34:11.894 And you cannot use that assumption in previous area lights and assume that that's going to apply everywhere. 508 00:34:12.134 --> 00:34:12.414 Right. 509 00:34:12.434 --> 00:34:13.435 That's not the case. 510 00:34:13.595 --> 00:34:17.257 Across the world, different countries use lights differently. 511 00:34:17.717 --> 00:34:19.758 That assumption only holds visually. 512 00:34:21.392 --> 00:34:24.193 And I, you know, from an asset point of view, it's like, then why are you doing it? 513 00:34:24.793 --> 00:34:26.293 I'm doing it because our users need. 514 00:34:27.194 --> 00:34:29.794 Because a lot of our users want a 60% solution. 515 00:34:30.194 --> 00:34:31.115 And it's good enough. 516 00:34:31.135 --> 00:34:33.375 The decisions they're making are like Keenan-Awe. 517 00:34:33.895 --> 00:34:39.497 They can't wait three months for a science paper, you know, three years to make a decision. 518 00:34:40.137 --> 00:34:43.238 So we were tailoring a solution. 519 00:34:44.275 --> 00:34:54.945 And that we card a lot of technical skill, not just in building the algorithms, but figuring out the data pipelines to do this because there's a lot of high volume in there. 520 00:34:55.445 --> 00:34:55.725 Yeah. 521 00:34:56.366 --> 00:34:57.647 Okay, that makes a lot of sense, though. 522 00:34:57.687 --> 00:35:00.630 So it's with that assumption and then you kind of do like a... 523 00:35:03.687 --> 00:35:10.850 you know, the Bayesian kind of analysis of like, okay, given this ground truth, what's the probability? 524 00:35:10.910 --> 00:35:12.510 Well, we typically go downscaling. 525 00:35:12.810 --> 00:35:15.791 And in the sciences, this is very common. 526 00:35:15.871 --> 00:35:19.252 You measure flux, different kind of fluxes. 527 00:35:20.573 -->

00:35:23.194 And you have a tower collecting flux data. 528 00:35:23.854 --> 00:35:28.976 And then NASA satellites come in, give you a big perspective, and then you downscale it against that anchor. 529 00:35:29.856 --> 00:35:34.262 And you can come up with maps of like ozone, nitrogen oxide, everything. 530 00:35:34.302 --> 00:35:37.046 Like there is, it's just geostatistics. 531 00:35:38.267 --> 00:35:41.872 In this case, you do that, but we always say buyer beware. 532 00:35:43.124 --> 00:35:44.906 It is useful as a tool. 533 00:35:45.506 --> 00:35:50.411 If you do analysis, you need to get ground-based data and colocated against the lights. 534 00:35:50.531 --> 00:35:51.792 And we're also doing that. 535 00:35:52.653 --> 00:36:05.184 I just came from Cornell University, one of our partners, and they took me to an experimental agriculture site, made me go on a John Deere tractor, turn on the lights, and they detected the lights from space. 536 00:36:05.664 --> 00:36:06.345 No way. 537 00:36:06.645 --> 00:36:07.166 I can show you. 538 00:36:07.946 --> 00:36:08.507 Yes, wait. 539 00:36:09.578 --> 00:36:12.120 So imagine... That's so few photons. 540 00:36:12.180 --> 00:36:13.221 That doesn't make it... What? 541 00:36:14.021 --> 00:36:14.382 Wow. 542 00:36:14.642 --> 00:36:18.865 Because, again, the sensitivity is not related to the actual support. 543 00:36:19.966 --> 00:36:28.672 A fire, a small fire, can still be detected, but the scale at which it's reported is broader because you're running a CTVR8. 544 00:36:29.532 --> 00:36:38.058 So you can detect a 20,000-lumen light if you know how to shine it based on the geometry. 545 00:36:38.079 --> 00:36:38.219 Right. 546 00:36:39.095 --> 00:36:42.338 So John Deere has lights this way, this way, that way. 547 00:36:42.838 --> 00:36:43.279 We got it. 548 00:36:43.439 --> 00:36:45.080 We had a Honda Sirica next door. 549 00:36:45.120 --> 00:36:45.600 Couldn't detect. 550 00:36:45.620 --> 00:36:46.381 Couldn't detect that. 551 00:36:46.441 --> 00:36:46.621 Okay. 552 00:36:46.641 --> 00:36:48.463 But the John Deere had so many conical lights. 553 00:36:48.503 --> 00:36:49.964 Yeah. 554 00:36:50.024 --> 00:36:50.524 Yeah. 555 00:36:50.865 --> 00:36:51.585 The signal on it. 556 00:36:51.765 --> 00:36:52.106 Mm-hmm. 557 00:36:52.726 --> 00:36:53.887 Why does that mean to me? 558 00:36:55.088 --> 00:37:06.978 Well, if you care about electrifying Africa, where the nighttime signal is three times smaller than the moon-condaminated signal, that's 10% of the human population right there. 559 00:37:08.232 --> 00:37:17.880 So from a sustainability planning point of view, if you think about 2100 demographically, Africa is going to be a large, large contingent of human population. 560 00:37:18.842 --> 00:37:28.506 So if you want to measure those trends and how that's going to impact global economies at scale, you need to look at the darkest place in the world that is developing right now. 561 00:37:28.886 --> 00:37:29.106 Yeah. 562 00:37:29.486 --> 00:37:35.928 And that 10%, so you're saying because Africa, I'm guessing, is like near the equator, moon contamination is worse? 563 00:37:36.069 --> 00:37:36.529 Is that the idea? 564 00:37:36.549 --> 00:37:39.710 No, Africa is still the last mile in terms of energy. 565 00:37:40.190 --> 00:37:44.452 I'm saying that Africa is still in, there are countries that are still not electrified. 566 00:37:44.672 --> 00:37:45.212 Right, okay. 567 00:37:45.292 --> 00:37:46.472 And so like 10% of the... 568 00:37:48.111 --> 00:37:50.495 Global population still has no electricity. 569 00:37:50.655 --> 00:37:51.236 Oh, okay. 570 00:37:51.477 --> 00:37:57.586 So if you want to detect those regions, your data has to be of high integrity because those are very low signals. 571 00:37:59.510 --> 00:37:59.730 Mm-hmm. 572 00:38:00.031 --> 00:38:10.061 And so that's why NASA expends so much effort in validating these results and understanding where's the threshold at which you can make science. 573 00:38:10.121 --> 00:38:17.809 Because if the World Bank is going to go to Africa and electrify cotivora, which they have... 574 00:38:18.910 --> 00:38:27.313 They can then use Black Marble to assess the effectiveness of those electrification programs over years and say, we made an investment of $200 million. 575 00:38:27.933 --> 00:38:31.394 We use Black Marble and now we're going to look and it still looks dark. 576 00:38:31.714 --> 00:38:32.614 We got a problem. 577 00:38:32.794 --> 00:38:33.035 Right. 578 00:38:33.215 --> 00:38:33.435 Right. 579 00:38:33.595 --> 00:38:34.715 Oh, it looks bright. 580 00:38:35.455 --> 00:38:36.816 Great. 581 00:38:36.916 --> 00:38:40.757 And then they begin to do assessments against, okay, why did it work in this city? 582 00:38:41.697 --> 00:38:42.578 And why didn't it work? 583 00:38:42.998 --> 00:38:47.279 And so you can start making really good decisions about the future investments of technology. 584 00:38:47.657 --> 00:38:51.899 countries that are very, you know, that don't have access to resources. 585 00:38:51.919 --> 00:38:53.860 We can do that too as a federal government. 586 00:38:54.780 --> 00:39:09.668 If I can tell you where are the places that are most frequently hit by disasters, that are constantly observing volatility, you can concentrate support, expertise, capital in those regions. 587 00:39:09.788 --> 00:39:12.789 That's what Earth Observations offer you. 588 00:39:13.289 --> 00:39:14.350 We don't make the decisions. 589 00:39:15.143 --> 00:39:20.607 But we're giving you a very strong formation by which you can make decisions to help everyone. 590 00:39:20.907 --> 00:39:21.147 Yeah. 591 00:39:21.267 --> 00:39:24.950 And if the data can be trusted, then... That's the big if. 592 00:39:25.010 --> 00:39:25.310 Yeah. 593 00:39:26.030 --> 00:39:27.051 That's the big if. 594 00:39:28.252 --> 00:39:32.154 Because that requires a lot of expertise and institutional muscle memory. 595

00:39:32.194 --> 00:39:36.617 You can't just calibrate an instrument like this and backfill it with somebody else. 596 00:39:37.078 --> 00:39:41.521 You need a team of people that does that. 597 00:39:41.761 --> 00:39:43.102 That's where NASA steps in. 598 00:39:43.912 --> 00:39:52.611 Because we made a commitment that data systems, data and integrity is a capability that must be sustained for our niche. 599 00:39:54.046 --> 00:40:00.008 Just in the same way the U.S. government has a vested interest in the GPS network. 600 00:40:00.028 --> 00:40:01.368 Like, why? 601 00:40:01.428 --> 00:40:02.149 Because we cared. 602 00:40:02.489 --> 00:40:04.669 We couldn't find this place if it weren't for GPS. 603 00:40:04.689 --> 00:40:08.771 Well, it takes a lot of space hardware to run the GPS network. 604 00:40:09.631 --> 00:40:15.873 And many different federal agencies own the capabilities around position, navigation, and timing. 605 00:40:17.667 --> 00:40:26.491 about understanding changes in the atmosphere and how that impacts radio occultation signals. 606 00:40:26.531 --> 00:40:28.392 Like everyone is involved in this. 607 00:40:29.292 --> 00:40:37.136 Black Marble does nighttime lights, but we are not the only agency involved in asking energy-related questions or disaster-related questions. 608 00:40:38.702 --> 00:40:43.765 But we've been given the distinct privilege to ask the weird questions. 609 00:40:43.785 --> 00:40:44.025 Yeah. 610 00:40:45.026 --> 00:40:46.967 Think differently about these questions. 611 00:40:47.047 --> 00:40:50.929 And I think that's really the proudest thing about winning this award. 612 00:40:51.929 --> 00:40:57.812 That there were folks at NASA that believed in us, even though our leaders were crazy at the time. 613 00:40:57.832 --> 00:40:58.112 Yeah. 614 00:40:59.173 --> 00:41:05.637 Yeah, so, I mean, I wanted to ask you, you know, now we're about to go to the Library of Congress for the ceremony. 615 00:41:06.218 --> 00:41:11.161 What does winning this award mean for you, for your team? 616 00:41:11.802 --> 00:41:15.804 You know, you've been working on this for, what, like 10, almost 15 years now? 617 00:41:16.545 --> 00:41:20.628 15 years since that, the Ramadan photo, at least. 618 00:41:21.668 --> 00:41:22.709 It's been a long time. 619 00:41:22.729 --> 00:41:23.189 Yeah. 620 00:41:24.050 --> 00:41:27.114 I'm sure you didn't start expecting to win an award. 621 00:41:27.134 --> 00:41:28.055 I don't think anyone does. 622 00:41:28.856 --> 00:41:32.061 But what does it mean that you're getting recognized like this? 623 00:41:33.783 --> 00:41:35.145 I think it validates... 624 00:41:36.683 --> 00:41:49.949 Something that our leaders at NASA have been saying for a long time, which is we have given, we have produced a very amazing capabilities, not just black marble, but earth science in general, the fleet of NASA satellites. 625 00:41:51.169 --> 00:42:03.755 And we're going through this evolution because now one of our

leaders says we're part of a broader ecosystem where, you know, there are many different actors. 626 00:42:04.880 --> 00:42:09.382 And the thing about being part of the ecosystem is you no longer control the ecosystem. 627 00:42:10.123 --> 00:42:20.928 And it's a huge success story that our nation made a big bet around investing in Earth observations. 628 00:42:22.573 --> 00:42:28.754 Back in the 90s and 80s, you know, and different administrations came in and said, let's do a big bet. 629 00:42:28.794 --> 00:42:30.655 Let's understand Earth as a system. 630 00:42:31.175 --> 00:42:34.556 Let's build large observatories the size of school buses. 631 00:42:35.536 --> 00:42:36.336 Let's deploy them. 632 00:42:36.356 --> 00:42:48.619 Let's try to get all the... And now we have all this data pried into the next chapter of U.S. Our next big bet, which is AI. 633 00:42:50.307 --> 00:42:53.850 and, you know, digital modernization, all this. 634 00:42:54.410 --> 00:43:08.782 We have, we're one of the few countries in the world that has what matters in order to be able to have predictive understandings of the future, which is gold, highly pristinely maintained, highly integral product. 635 00:43:08.802 --> 00:43:10.443 Because with AI, you know how it works. 636 00:43:10.503 --> 00:43:11.424 Garbage in, garbage out. 637 00:43:11.444 --> 00:43:12.385 Yeah, yeah, 100%. 638 00:43:12.405 --> 00:43:13.886 But we spent- Worst garbage art. 639 00:43:14.066 --> 00:43:14.567 Exactly. 640 00:43:14.967 --> 00:43:18.390 But with, but the fact that we have produced and we had- 641 00:43:19.050 --> 00:43:23.351 the foresight to know that maybe AI wasn't a thing back then, but it is now. 642 00:43:25.111 --> 00:43:33.673 And we're building metrics, allowing people to create a protective understanding of risks. 643 00:43:35.013 --> 00:43:38.533 And I wish we had that 15 years ago, but we didn't, but we had it now. 644 00:43:39.774 --> 00:43:47.575 I think that's what makes this award so special because it's really timed at a moment where our agency... 645 00:43:48.757 --> 00:43:53.021 has shown and produced a lot of public value for our nation. 646 00:43:54.302 --> 00:43:54.782 Certainly. 647 00:43:55.202 --> 00:44:05.371 A lot of times, scientists don't get involved in that element of like, hey, how much, you know, we do the justifications, we conduct reviews, we do all that due diligence. 648 00:44:06.591 --> 00:44:18.234 But to be able to represent our division in demonstrating the many thousands of times NASA Earth Science generates public value, it's quite the privilege. 649 00:44:18.674 --> 00:44:20.475 Because we're not the only guys. 650 00:44:21.255 --> 00:44:31.497 There are people who can monitor air quality at fine resolution, who can protect firefighters from the lines of fire using NASA satellites. 651 00:44:32.117 --> 00:44:33.918 There's so much going on. 652 00:44:35.133 --> 00:44:35.974 And they picked us. 653 00:44:36.475 --> 00:44:49.212 But let me tell you, for every Black Marble, there's 100 other NASA Earth scientists out there who are just doing crazy stuff, thinking differently and generating results that are directly impacting people's lives. 654 00:44:50.098 --> 00:44:55.966 Yeah, well, we are huge fans of NASA at the podcast, and I think all of our viewers are huge fans. 655 00:44:56.346 --> 00:45:06.019 So I think you're doing a great service for doing this kind of work, and AAAS and the Golden Goose Awards are just such a great way to recognize something like this. 656 00:45:06.139 --> 00:45:06.459 I hope... 657 00:45:07.080 --> 00:45:09.101 You know, we continue funding this kind of stuff. 658 00:45:09.181 --> 00:45:11.823 We continue doing this kind of stuff. 659 00:45:12.184 --> 00:45:14.065 It's an incredible privilege to talk to you. 660 00:45:14.145 --> 00:45:18.748 Congratulations on the award to not just you, but also your co-recipients. 661 00:45:20.349 --> 00:45:21.930 Yeah, this was really fun. 662 00:45:22.090 --> 00:45:22.831 I learned a lot. 663 00:45:24.252 --> 00:45:26.893 I hope you don't mind the technical questions, but I don't think you did. 664 00:45:26.953 --> 00:45:28.254 I think you're an engineer at heart. 665 00:45:28.394 --> 00:45:30.196 You've enthused me at last for sure. 666 00:45:30.916 --> 00:45:31.236 All right. 667 00:45:31.316 --> 00:45:33.598 Well, yeah, I'll be seeing you later today. 668 00:45:33.618 --> 00:45:34.438 It's going to be a fun time. 669 00:45:34.458 --> 00:45:34.718 Thank you. 670 00:45:35.118 --> 00:45:35.759 Thank you again. 671 00:45:35.779 --> 00:45:36.999 Thank you for all the work you do. 672 00:45:37.119 --> 00:45:43.763 I mean, it means a lot to us to share how we do our work with integrity. 673 00:45:45.190 --> 00:45:46.992 because it takes years. 674 00:45:47.273 --> 00:46:01.049 Yeah, and I think... And for you to deep dive on the algorithm, it means a lot because it shows that you're doing the due diligence that's required to build that credibility with your audience. 675 00:46:01.229 --> 00:46:02.071 Yeah, thank you. 676 00:46:02.091 --> 00:46:02.731 That's... 677 00:46:03.412 --> 00:46:06.713 A lot of the time folks, most of the folks tell me, I don't know what's going to read that. 678 00:46:06.753 --> 00:46:07.773 It's like people read it. 679 00:46:07.813 --> 00:46:10.173 No, I mean, you'll be surprised. 680 00:46:10.234 --> 00:46:19.576 You'll be surprised at how many people resonate to the methods of how you get the data, you know, because as scientists, I think, you know, I'm just going to go on a little tired. 681 00:46:19.616 --> 00:46:30.218 But like as scientists, convincing the public that what we're saying is not made up is, I think, as important as communicating what it is that we found. 682 00:46:30.978 --> 00:46:38.642 I think that's a very important point because you're giving us the opportunity to show that we're not using a black box. 683 00:46:38.702 --> 00:46:40.703 Yeah, it's not made up. 684 00:46:41.303 --> 00:46:44.485 It's not a neural network that somehow works. 685 00:46:44.625 --> 00:46:46.306 This is actual physics. 686 00:46:46.506 --> 00:46:49.408 Yeah, this is just physics being applied in a very reasonable way. 687 00:46:49.828 --> 00:46:51.469 There's assumptions that are open. 688 00:46:51.889 --> 00:46:53.410 You can see what the assumptions are. 689 00:46:53.810 --> 00:46:59.573 And most of them, I mean, it's almost destiny, but a lot of the physical processes and principles that... 690 00:47:00.333 --> 00:47:02.315 The team we learn as grad students. 691 00:47:02.835 --> 00:47:10.760 We learn from the real pioneers of Earth observation that build those legacy systems that are now being retired. 692 00:47:11.341 --> 00:47:16.744 And so that legacy, I mean, we couldn't take full credit card. 693 00:47:16.764 --> 00:47:21.488 It weren't for those who were standing on the shoulders from. 694 00:47:22.168 --> 00:47:26.451 People going back to the 80s, the 70s, would build the first radiator transfer model. 695 00:47:26.531 --> 00:47:31.074 I mean, the science is just, if you really want to start, you have to go all the way to pipiagoras. 696 00:47:31.134 --> 00:47:32.494 You're like, I'm freaking away. 697 00:47:33.535 --> 00:47:39.739 Well, maybe we'll have you on the podcast for a much more lengthy conversation and we can start with at least radiative transfer. 698 00:47:40.980 --> 00:47:44.642 But it was a pleasure talking to you and congratulations again.