1 00:00:00,000 --> 00:00:05,490 all right hello it's so exciting to be here at the grand opening of the earth 2 00:00:05,490 --> 00:00:10,920 at night book which our team I'm the science PI of the black marble project 3 00:00:10,920 --> 00:00:15,480 which is the data set that you see behind me and it's sort of that the 4 00:00:15,480 --> 00:00:19,380 theme of the book this year which has really has been exciting for our team to 5 00:00:19,380 --> 00:00:26,789 work on obviously this imagery is so sexy, um, it's beautiful to look at but it's 6 00:00:26,789 --> 00:00:31,949 also I'm hoping that this talk gives you some idea of the the depth of the 7 00:00:31,949 --> 00:00:36,000 scientific insights that it can also give us besides just being a bunch of 8 00:00:36,000 --> 00:00:42,860 pretty pictures but actually I'm going to start by looking at the Blue Marble 9 00:00:42,860 --> 00:00:48,180 which is the first slide that's going to come up here this is the next-generation 10 00:00:48,180 --> 00:00:54,449 Blue Marble it came out like less than a decade ago and this is a composite of 11 00:00:54,449 --> 00:00:59,609 the earth during the daytime and cloud free so we can see really interesting 12 00:00:59,609 --> 00:01:05,100 signals with this blue marble things like changes in natural land systems 13 00:01:05,100 --> 00:01:12,720 like the change in forest green up we can see changes in the ice caps etc but 14 00:01:12,720 --> 00:01:16,860 we what we can't really see with the Blue Marble or what you can barely 15 00:01:16,860 --> 00:01:23,189 perceive on this map is urban areas it's like a tiny little blip on the map 3% of 16 00:01:23,189 --> 00:01:30,810 the global land area and yet it has huge implications on planetary health so if 17 00:01:30,810 --> 00:01:36,689 you can flip the slide, urban areas even though they're only a tiny land area 18 00:01:36,689 --> 00:01:42,030 they produce 80 percent of the global GDP they house more than half of 19 00:01:42,030 --> 00:01:49,229 humanity and they also produce 76 percent of our carbon emissions so when 20 00:01:49,229 --> 00:01:52,710 you're thinking about what happens in this tiny little and land area it's 21 00:01:52,710 --> 00:01:56,820 actually going to be the site where the interventions of bringing about the UN 22 00:01:56,820 --> 00:02:02,149 sustainable development goals have to happen it's places like this that house 23 00:02:02,149 --> 00:02:07,380 millions of people who live in slums it's places like urban areas that are 24 00:02:07,380 --> 00:02:10,990 the drivers of planetary changes that we're seeing 25 00:02:10,990 --> 00:02:16,060 in the blue marble and so in some ways the black marble is like the yang to the 26 00:02:16,060 --> 00:02:24,070 blue marbles Yin it is it is the human stat to our natural land system stat of 27 00:02:24,070 --> 00:02:33,040 them of the MODIS so this is an imagery of black marble before the clouds are 28 00:02:33,040 --> 00:02:40,150 are taken away black marble is derived from a sensor aboard Suomi NPP called 29 00:02:40,150 --> 00:02:46,030 VIIRS, it's the VIIRS day/night band is what you're looking at and the that's 30 00:02:46,030 --> 00:02:50,440 been out since 2012 and since then our team has been working really hard to put 31 00:02:50,440 --> 00:02:54,910 out two products one is a top of atmosphere reflectance which looks like 32 00:02:54,910 --> 00:02:58,180 this it has all the clouds and everything intact and then the second 33 00:02:58,180 --> 00:03:01,810 which is coming out in the beginning of this year in just a couple of weeks is 34 00:03:01,810 --> 00:03:06,130 going to be that black marble composite that you see it actually it'll be daily 35 00:03:06,130 --> 00:03:10,270 but with the clouds removed with the moonlight removed with all of these 36 00:03:10,270 --> 00:03:17,260 aberrations that Mar your understanding of urban this is Cairo removed so that's 37 00:03:17,260 --> 00:03:25,090 a really exciting upcoming event for us, next slide, so this is not the first 38 00:03:25,090 --> 00:03:28,540 night lights data set that's been out there it's the first scientific night 39 00:03:28,540 --> 00:03:34,650 lights data set so before of years there was DMSPLOS which is a meteorological 40 00:03:34,650 --> 00:03:39,910 satellite that NOAA used to try to understand clouds actually, so that's 41 00:03:39,910 --> 00:03:43,870 been out since the 70s it's been digital since the 1990s this is the picture of 42 00:03:43,870 --> 00:03:49,900 Cairo from DMSP and you can see that it has a pretty coarse spatial resolution 43 00:03:49,900 --> 00:03:54,430 2.7 kilometers also with in Cairo you can't really see any differences between 44 00:03:54,430 --> 00:04:01,510 the center of the city and the suburbs but it has been out for a long time so 45 00:04:01,510 --> 00:04:07,840 it's good for that long-term record if you look now this is VIIRS it's almost 46 00:04:07,840 --> 00:04:12,070 like putting on glasses and compared to DMSP you can really make out the 47 00:04:12,070 --> 00:04:16,989 different heterogeneities within the urban area you also have a higher 48 00:04:16,989 --> 00:04:22,360 spatial resolution at 750 meters and maybe what's most important is there's 49 00:04:22,360 --> 00:04:25,750 calibration across time which means that from day to day we're getting 50 00:04:25,750 --> 00:04:29,970 a signal that's stable and that's in part due to a lot of the algorithmic 51 00:04:29,970 --> 00:04:34,540 algorithmic developments that we've put into place in the last year but this is 52 00:04:34,540 --> 00:04:38,080 going to be useful for a whole new set of research questions that DMSP 53 00:04:38,080 --> 00:04:41,340 couldn't even touch next 54 00:04:42,920 --> 00:04:45,400 so but of course it doesn't look like that pretty 55 00:04:45,400 --> 00:04:49,960 picture when it first comes out of the sensor it looks more like this here what 56 00:04:49,960 --> 00:04:54,400 you're seeing is different swaths the moonlight reflecting off of the desert 57 00:04:54,400 --> 00:04:59,830 isn't taken away and so the cities don't even pop out when moonlight is there so 58 00:04:59,830 --> 00:05:06,430 a big physics challenge for us was to try to even understand how to understand 59 00:05:06,430 --> 00:05:11,770 the physics of nighttime radiance reflectance etc if you go to the next 60 00:05:11,770 --> 00:05:14,980 slide I don't want to get into the weeds too much but I do think it's important 61 00:05:14,980 --> 00:05:19,450 to understand kind of what the algorithm is doing there are several sources of 62 00:05:19,450 --> 00:05:23,410 light in addition to urban light which the urban light is shown in red here you 63 00:05:23,410 --> 00:05:28,270 got the moon contribution you also have the reflectance off of the atmosphere 64 00:05:28,270 --> 00:05:32,290 and off of the ground for example from day to day I was 65 00:05:32,290 --> 00:05:37,570 looking at Texas once and it was in December and I didn't even think but in 66 00:05:37,570 --> 00:05:42,280 one day the radiant shot up really high and I was like what could this be it was 67 00:05:42,280 --> 00:05:46,360 because there is a freak snowstorm in Texas so of course the reflectance of 68 00:05:46,360 --> 00:05:50,620 the ground really matters for how streetlights reflect off of off of the 69 00:05:50,620 --> 00:05:55,060 ground in addition the vegetation occlusion of streetlights can really 70 00:05:55,060 --> 00:05:58,810 dampen the signal so we have to remove all of these artifacts in order to get a 71 00:05:58,810 --> 00:06:02,440 stable understanding of what what night lights are otherwise you think 72 00:06:02,440 --> 00:06:06,520 something's happening when something really isn't happening next 73 00:06:08,000 --> 00:06:08,640 right so 74 00:06:08,650 --> 00:06:14,110 this is the this is the global composite from 2012 it's gonna scroll over and 75 00:06:14,110 --> 00:06:19,990 show you 2016 we now have all the way to 2019 and we have it daily you can't 76 00:06:19,990 --> 00:06:26,620 really see big changes here when it's like in this format so the next slide is 77 00:06:26,620 --> 00:06:32,160 going to show you a colored version where the areas of the 78 00:06:32,160 --> 00:06:35,820 world where there are new lights are shown in purple and where lights have 79 00:06:35,820 --> 00:06:42,210 declined are shown in orange and I'm so clear regional patterns start to stick 80 00:06:42,210 --> 00:06:48,150 out if you want to turn to the next slide you can see India Lower South East 81 00:06:48,150 --> 00:06:52,290 Asia having lots more additional lights this is largely because of 82 00:06:52,290 --> 00:06:58,350 electrification that has been happening there you can also see like Puerto Rico 83 00:06:58,350 --> 00:07:06,180 a huge loss of lights from a lot of the storms that had happened there so some 84 00:07:06,180 --> 00:07:09,330 of these regional patterns start to come out we we actually can't explain all of 85 00:07:09,330 --> 00:07:13,590 them we don't we don't know because we're not experts in each area but 86 00:07:13,590 --> 00:07:16,800 that's kind of what's exciting about this is it's like objective data that we 87 00:07:16,800 --> 00:07:20,730 can talk to regional experts and really start to understand what could these 88 00:07:20,730 --> 00:07:24,450 these different signals mean I mean ice and actually I think you can skip over 89 00:07:24,450 --> 00:07:31,200 this one this one's just a slow version right so I'm gonna just sort of talk 90 00:07:31,200 --> 00:07:35,970 about two different kind of areas of how we've used this data to look at new 91 00:07:35,970 --> 00:07:40,500 science questions the first is to understand how planetary changes are 92 00:07:40,500 --> 00:07:45,680 impacting urban areas and this has mostly been in the area of disasters 93 00:07:45,680 --> 00:07:52,560 this is an image of cyclone Edye which hit Mozambique this past year actually 94 00:07:52,560 --> 00:07:57,780 that area of the world was hit seven consecutive times this hurricane season 95 00:07:57,780 --> 00:08:02,100 and not unlike many other places in the world there's been escalating incidences 96 00:08:02,100 --> 00:08:09,150 of these storms and combined with the poor infrastructure and the low quality 97 00:08:09,150 --> 00:08:13,560 housing in Mozambique even though this was only a category two storm when it 98 00:08:13,560 --> 00:08:19,010 hit land hit landfall it killed over a thousand people so it had a huge impact 99 00:08:19,010 --> 00:08:26,730 next slide so while this storm was happening black marble was one of the 100 00:08:26,730 --> 00:08:30,660 only datasets that was available to really show the impact so we worked with 101 00:08:30,660 --> 00:08:36,660 the World Bank and other recovery teams on the ground to try to help them map 102 00:08:36,660 --> 00:08:40,680 and understand where the biggest outages had occurred and also where recovery was 103 00:08:40,680 --> 00:08:46,050 happening so this picture on the right is Mozambique before cyclone Edye on 104 00:08:46,050 --> 00:08:52,020 the left is 10 days after the the storm hit and you can see that the Manga 105 00:08:52,020 --> 00:08:59,430 region the Micucci region they were all very, very hard hit by this storm and so 106 00:08:59,430 --> 00:09:06,270 not just the before after but we can also progressively track how recovery 107 00:09:06,270 --> 00:09:12,290 efforts are doing which is a good way to keep and understand whether the the 108 00:09:12,290 --> 00:09:17,820 investments that are put into recovery are having an effect next slide it's not 109 00:09:17,820 --> 00:09:23,090 just disasters like hurricanes and cyclones that we can monitor also 110 00:09:23,090 --> 00:09:28,560 geopolitical conflict and disasters are another application where black marble 111 00:09:28,560 --> 00:09:33,840 has proven useful these areas of the world and you can keep going are really 112 00:09:33,840 --> 00:09:39,390 hard to get on the ground data because because they're involved in conflict 113 00:09:39,390 --> 00:09:45,960 this this first picture is from 2012 and then 2016 and then changed on the far 114 00:09:45,960 --> 00:09:52,470 left and this is the Syria and you can see areas that were totally demolished 115 00:09:52,470 --> 00:09:58,950 by the conflict in Syria so this is like a spatial distribution of where battles 116 00:09:58,950 --> 00:10:03,990 took out the electric infrastructure in the country during the the war next and 117 00:10:03,990 --> 00:10:07,920 it's not just the spatial distribution that we get we also get the timing of 118 00:10:07,920 --> 00:10:13,980 these events so this is um Aleppo Syria and you can see that drop that just 119 00:10:13,980 --> 00:10:18,000 happened right there in 2012 that was the Battle of Aleppo which was a major 120 00:10:18,000 --> 00:10:24,180 battle in Syria and in Aleppo that really it forced hundreds of thousands 121 00:10:24,180 --> 00:10:29,280 of people to leave the city and so we can also see whether there whether 122 00:10:29,280 --> 00:10:33,180 there's a recert whether there's people coming back whether the infrastructure 123 00:10:33,180 --> 00:10:37,700 is um is repaired and as you can see in Aleppo there really hasn't been much 124 00:10:37,700 --> 00:10:43,440 recovery unfortunately next at the same time we 125 00:10:43,440 --> 00:10:50,640 can also see signals where many of these migrants are going so this is Zaatari 126 00:10:50,640 --> 00:10:54,520 refugee camp in Jordan houses 30% of the Syrian refugees in 127 00:10:54,520 --> 00:10:59,980 Gordon and you can see the increase in lighting over Zaatari as that camp grows 128 00:10:59,980 --> 00:11:05,290 and this is not just a people signal it's mostly actually that the increase 129 00:11:05,290 --> 00:11:11,200 in population in Zaatari was happening mostly in 2012-2013 this is the delayed 130 00:11:11,200 --> 00:11:14,590 signal for when infrastructure was provided for those people that are still 131 00:11:14,590 --> 00:11:21,550 living in the Zaatari refugee camp also an area where on the ground data has 132 00:11:21,550 --> 00:11:29,860 been hard to get so next so the next area that I want to talk about is how we 133 00:11:29,860 --> 00:11:33,900 use black marble to understand how urban areas are changing the planet not how 134 00:11:33,900 --> 00:11:43,620 the impacts within urban areas and so this is you can you can turn the slide 135 00:11:43,620 --> 00:11:49,600 so before we scrolled in and we saw that electrification signal in India you can 136 00:11:49,600 --> 00:11:57,850 see it again here in close contact so this is 2012 2016 and the change here so 137 00:11:57,850 --> 00:12:05,590 almost all of India is purple and that's because in 2012 over I think it was over 138 00:12:05,590 --> 00:12:11,020 a million I don't quite remember the number but several towns still didn't 139 00:12:11,020 --> 00:12:17,080 have electricity infrastructure in 2012 that was down to less than 20,000 towns 140 00:12:17,080 --> 00:12:20,470 that didn't have electrification so there was huge progress made between 141 00:12:20,470 --> 00:12:30,490 2012 and 2016 and that is reflected in the purple next a similar case is in 142 00:12:30,490 --> 00:12:36,070 Ivory Coast there was a lot of infrastructure investment there in both 143 00:12:36,070 --> 00:12:41,710 solar and hydro providing new electricity for that region these are a 144 00:12:41,710 --> 00:12:44,890 bunch of different news articles that came out during that time period 145 00:12:44,890 --> 00:12:50,290 tracking the power generation development next slide and then this is 146 00:12:50,290 --> 00:12:54,730 a black marble understanding of how electrification has happened in three 147 00:12:54,730 --> 00:12:58,840 cities in the Ivory Coast and where which which parts of the cities have 148 00:12:58,840 --> 00:13:04,240 benefited from that investment so what's really nice about this data is you know 149 00:13:04,240 --> 00:13:06,700 you know that the investment is going there but it's 150 00:13:06,700 --> 00:13:11,380 it's a way for people to objectively track how successful that has been at 151 00:13:11,380 --> 00:13:16,900 the end user level so one of the key sustainable development goals I think 152 00:13:16,900 --> 00:13:24,190 it's SDG 7 is to ensure modern energy services for all and so this is the key 153 00:13:24,190 --> 00:13:28,180 way that we're gonna have to be able to track that across places that have very 154 00:13:28,240 --> 00:13:32,380 different levels of statistical data collection systems 155 00:13:32,960 --> 00:13:33,500 next 156 00:13:35,200 --> 00:13:36,680 ok so I wanted 157 00:13:36,680 --> 00:13:41,800 to end by talking about a couple of future directions for our team a lot of 158 00:13:41,800 --> 00:13:46,000 times in the remote sensing literature nightlights has been used as a proxy for 159 00:13:46,000 --> 00:13:51,340 everything it's a proxy for GDP, it's a proxy for where people are, people use it 160 00:13:51,340 --> 00:13:56,680 to map where urban areas are but we want to move a step beyond that instead of 161 00:13:56,680 --> 00:14:00,880 thinking of it as a proxy we want to look at places where infrastructure 162 00:14:00,880 --> 00:14:05,350 where lights might diverge from those other things for example where does 163 00:14:05,350 --> 00:14:09,430 population grow but lights doesn't where does lights grow but population 164 00:14:09,430 --> 00:14:13,510 doesn't where has land change happened without infrastructure development and 165 00:14:13,510 --> 00:14:17,110 what are these different signals where those signals diverge what are those 166 00:14:17,110 --> 00:14:21,310 mean so this ternary diagram is something 167 00:14:21,310 --> 00:14:25,540 we've developed where we're looking at a global population data set, a global land 168 00:14:25,540 --> 00:14:28,960 data set and a global infrastructure data set which is nightlights at the 169 00:14:28,960 --> 00:14:32,920 same time and seeing what kind of classes are pulled out and what kind of 170 00:14:32,920 --> 00:14:37,060 new information about urbanization is pulled out by using these three data 171 00:14:37,060 --> 00:14:44,530 sets in tandem, orthogonally, as opposed to as a proxy of one another, ok next so 172 00:14:44,530 --> 00:14:50,230 this is the ternary diagram that we've developed in this corner this is a when 173 00:14:50,230 --> 00:14:54,010 population increases but the other two don't we call that a population dominant 174 00:14:54,010 --> 00:14:58,360 class a land dominant class and an infrastructure dominant class we have in 175 00:14:58,360 --> 00:15:02,920 the center we have our most common conceptualization of urbanization that's 176 00:15:02,920 --> 00:15:07,240 where they all happen at once and then these other ones are what we're 177 00:15:07,240 --> 00:15:12,370 referring to as skewed urbanization where two of the three growth signals 178 00:15:12,370 --> 00:15:14,440 happen next 179 00:15:15,560 --> 00:15:18,740 and so you can pull out different kinds of 180 00:15:18,750 --> 00:15:23,220 development from that from taking these three together and these different types 181 00:15:23,220 --> 00:15:28,020 of development have really different impacts on natural land systems in the 182 00:15:28,020 --> 00:15:33,020 environment so for example vertical growth in cities is a really different 183 00:15:33,020 --> 00:15:38,400 type of urbanization than growth in informal settlements so though these 184 00:15:38,400 --> 00:15:42,090 have very different social implications and also very different like energy and 185 00:15:42,090 --> 00:15:44,880 resource implications okay now so I'm just going to show you 186 00:15:44,880 --> 00:15:48,450 some examples of what this classification pulls out so this first 187 00:15:48,450 --> 00:15:56,130 one is this is the center okay then this first one this Center triangle right 188 00:15:56,130 --> 00:16:00,200 here that's concurrent change so this is what we think of this is Noida India 189 00:16:00,200 --> 00:16:05,760 before in 2012 and 2017 basically a whole city grew up out of nowhere this 190 00:16:05,760 --> 00:16:10,590 is a very common signal to see in India where population growth happens 191 00:16:10,590 --> 00:16:13,710 infrastructure growth happens and land change happens all at once and people 192 00:16:13,710 --> 00:16:17,220 who have visited this area of the world understand this is what people usually 193 00:16:17,220 --> 00:16:20,040 mean when they say urbanization next 194 00:16:21,420 --> 00:16:24,500 these are three of the independent types 195 00:16:24,510 --> 00:16:29,610 so this is actually where I live in Washington DC and NOMA this is the 196 00:16:29,610 --> 00:16:34,470 previously developed area that started to grow vertically and so this is we saw 197 00:16:34,470 --> 00:16:38,250 an increase in population here but no change in land and no change in 198 00:16:38,250 --> 00:16:42,240 infrastructure because already there were streetlights already those places 199 00:16:42,240 --> 00:16:46,320 were paved so they didn't get pulled out in an impervious signal but what was 200 00:16:46,320 --> 00:16:48,930 really changing was a vertical change and so that was pulled out in the 201 00:16:48,930 --> 00:16:54,660 population signal this is in Uttar Pradesh this is a electrification of a 202 00:16:54,660 --> 00:16:59,220 road these red you can barely see it but these red circles are showing the shadow 203 00:16:59,220 --> 00:17:04,949 of the street lamps that were picked up by the Google images so this is just an 204 00:17:04,949 --> 00:17:08,850 electrification change no population change and no land change associated and 205 00:17:08,850 --> 00:17:14,970 then this is just a land change where one undeveloped land was changed into 206 00:17:14,970 --> 00:17:20,430 like a garden farming so these are three different signals that are pulled out as 207 00:17:20,430 --> 00:17:24,900 urbanization depending on what what data set you're using but they're really 208 00:17:24,900 --> 00:17:28,560 different processes and they have really different implications on sustainability 209 00:17:28,560 --> 00:17:33,040 and next and then finally oh sorry and this is 210 00:17:33,040 --> 00:17:38,440 also from the electrification this is the electrification one where I was 211 00:17:38,440 --> 00:17:43,660 showing you the roads so we looked at this class across all of India and we 212 00:17:43,660 --> 00:17:48,160 compare that to the statistical data from surveys and we found that actually 213 00:17:48,160 --> 00:17:51,940 it mapped pretty well the electrification signals so if you pull 214 00:17:51,940 --> 00:17:56,310 out just that that class it's a it's a pretty good indication of 215 00:17:56,310 --> 00:18:01,960 electrification in terms of estimating the population in those pixels okay next 216 00:18:01,960 --> 00:18:09,940 and so finally the three classes we haven't talked about next are these 217 00:18:09,940 --> 00:18:16,120 three so this class is where population and infrastructure are changing but no 218 00:18:16,120 --> 00:18:21,370 land change this is Broadmoor neighborhood in Detroit so here we 219 00:18:21,370 --> 00:18:26,680 actually saw a decrease in population a decrease in lights but no change in land 220 00:18:26,680 --> 00:18:31,720 and that's because when Detroit went into the Great Recession lots of people 221 00:18:31,720 --> 00:18:36,550 moved out of this area the streetlights weren't even kept on because they lost 222 00:18:36,550 --> 00:18:40,120 so much population but there's very little change in land you can see a few 223 00:18:40,120 --> 00:18:44,050 houses here and there that have changed but it's not really picked up in a lot 224 00:18:44,050 --> 00:18:49,630 of our land data so it was picked up in the other two this is a really important 225 00:18:49,630 --> 00:18:53,320 land class that's pulled out by this this is the growth of informal 226 00:18:53,320 --> 00:18:58,600 settlements in in Mumbai so here you could see that there was a change in 227 00:18:58,600 --> 00:19:01,450 population and a change an infrastructure but the land data set 228 00:19:01,450 --> 00:19:06,250 really didn't capture because the pixels were were large and there's not a ton of 229 00:19:06,250 --> 00:19:12,700 I mean it's it's an infill land change and then here this is the final one this 230 00:19:12,700 --> 00:19:17,680 is like a sports facility in California there was a huge change in lights 231 00:19:17,680 --> 00:19:22,180 because they put a sports field there was a big change in land because they 232 00:19:22,180 --> 00:19:26,350 also created this parking lot and this new development but no change in 233 00:19:26,350 --> 00:19:30,310 population no one's living there so that kind of commercial development is also 234 00:19:30,310 --> 00:19:34,090 pulled out by this urbanization signal which is really really interesting 235 00:19:34,090 --> 00:19:38,770 because then we're getting it to land use not just lot land cover right okay 236 00:19:38,770 --> 00:19:43,950 so next and so finally one other next step is 237 00:19:43,950 --> 00:19:49,320 the fusion of our black marble data with different spatial resolution data so 238 00:19:49,320 --> 00:19:54,570 this is black marble in its native form it's 500 meters this is overlaid on an 239 00:19:54,570 --> 00:20:00,690 Open Street Map of University of Houston okay next one so when we downscale that 240 00:20:00,690 --> 00:20:04,860 based on Landsat we get kind of a 30 meter which we've been calling the black 241 00:20:04,860 --> 00:20:10,590 marble HD product and this is where we start to guess for each pixel what are 242 00:20:10,590 --> 00:20:14,130 the uses within each pixel that we're using the light how do we distribute the 243 00:20:14,130 --> 00:20:18,750 light within each pixel based on our Landsat there's a paper out about how we 244 00:20:18,750 --> 00:20:24,840 do this but it starts to get us to more exact targets especially for those cases 245 00:20:24,840 --> 00:20:28,620 where we're looking at disaster impacts of like where the outages might be 246 00:20:28,620 --> 00:20:32,580 occurring and what are the uses that might be affected and then this last 247 00:20:32,580 --> 00:20:37,290 image is something nobody's seen before and this is the 2 meter version with 248 00:20:37,290 --> 00:20:43,200 digital Globes data so you can see how you can start to get really specific 249 00:20:43,200 --> 00:20:47,460 about how lights are potentially distributed within each pixel and you 250 00:20:47,460 --> 00:20:50,850 can have a very high-resolution understanding of the energy 251 00:20:50,850 --> 00:20:56,730 infrastructure in a city I think over here it's gonna be a full-scale version 252 00:20:56,730 --> 00:21:03,240 of Houston based on that data so a little bit of eye candy to leave you 253 00:21:03,240 --> 00:21:10,620 with next so yeah I just wanted to wrap up by saying I think that black marble 254 00:21:10,620 --> 00:21:13,799 is really exciting because it's making the case for how we start to bring 255 00:21:13,799 --> 00:21:18,960 together this human system side with the natural system side that NASA has been 256 00:21:18,960 --> 00:21:23,340 so focused on and really we can't understand one without the other because 257 00:21:23,340 --> 00:21:27,450 now that we're in the Anthropocene it's the human systems that are driving a lot 258 00:21:27,450 --> 00:21:30,660 of the natural system changes so we really have to understand the processes 259 00:21:30,660 --> 00:21:36,210 within urban areas to understand why their impacts are the way they are so 260 00:21:36,210 --> 00:21:41,700 this is a one step towards completing that picture thanks so much for 261 00:21:41,700 --> 00:21:43,820 listening