 Hi, everyone. I'm Dr. Todd Harrison. I'm the Director of Strategic Science Initiatives at Planet, and if you're not familiar with Planet, we're going to talk about what we do today, and we image the whole world every day with a bunch of cubes sets that are about this big. So Planet is based out of San Francisco, and we have a mission to image the whole land rest of the Earth every day to make change visible, accessible, and actionable, and we do this through two constellations of satellites that we operate right now. The one that we're most known for are the cube sets, which we're going to focus on most today, and they are called the Doves, or the Super Doves, depending on what generation they are. And these are essentially always imaging the world every day as kind of a background scan of the planet. Any time they're flying over land, they're on and they're imaging, and they image in either four bands or eight bands of different colors, wavelengths, depending on what generation you're looking at. We also have another constellation of dishwasher-sized small sets called SkySat, and these are higher-resolution targeted satellites, so that means we actually have to specifically say, take a picture of this thing here at this time, and it will do it. So they work in tandem really nicely. We have the background scan at a slightly lower resolution, three to five meters, going on every day, and then we can use that to sort of inform where we want to take the higher-resolution 58 centimeter images with the bigger SkySat satellites. Right now, we have a little over 200 Doves in the constellation of the little cube sets, and we have 21 satellites in that dishwasher-sized SkySat constellation. We are constantly launching new satellites into the Dove constellation so that we can upgrade things like the tech, the software, when we want to improve the optics, maybe we want to add new bands, like our switch from four to eight, which we'll talk about in a little bit. So as these kind of age out technology-wise, they deorbit, they burn up in the atmosphere, and then we replace them with new ones. So we're not trying to build a gigantic mega-constellation. We're not launching thousands of satellites. We have about the number that we need to image the whole world every day, and then we just replenish that with better tech over time. Now, why do we want to image the world every day? Well, my whole background before coming to Planet was working on Mars missions, actually imaging Mars from orbit. And so I didn't really appreciate how cloudy the Earth is all the time until I came to work on Earth. And so this is an example of one month of coverage from two public satellites, Landsat, which is a satellite operated by NASA and the US Geological Survey. Landsat 8 and Landsat 9, so there's two satellites right now, currently are operational, and they've had a series of these satellites over the past most 50 years, imaging the whole world every 16 days or so. Now there's another satellite called Sentinel-2, which is operated by the European Space Agency, and this images the whole world about once every five to 10 days at 10 meter resolution. The Landsat's are 30 meter resolution, so these are a lot lower spatial resolution than the Dove satellites that we have. And when you are trying to image places, especially places that are already cloudy, like tropical rainforests or Southeast Asia during monsoon season, or from what I hear, the weather where you guys are right now is not that great. So trying to image the UK when it's cloudy and rainy, not the best if you can only image once every 10 days or 16 days, because you might go weeks or months without a clear atmosphere image. So this is an example of coverage from one month over Brazil, from Landsat and Sentinel combined. And if we compare that to one month of coverage from the Planetscope satellites and the SkySat satellites, you get a lot more chances to try to see through those clouds. So if you try to image the whole world every day, you get a lot more information about what's changing on the surface. So let's talk about the little cubesats that are in the title of the sock. In the past six years, we've had 14 different builds of these satellites. So like I mentioned before, we're just continuously iterating on them. And we leverage a lot of research and development from other industries. So these are mostly made out of commercial off the shelf parts so that we don't have to do a lot of sort of custom development of satellites, which has been sort of the way that NASA has done it traditionally. They will create these bespoke billion dollar satellites that are really, really carefully crafted to do a specific thing or answer a few specific science questions. But if we're using technology that's already out there for other things like drones or radiation hardened microprocessors or things that are being developed for other space applications, then it lets us save a lot of money because we don't have to redevelop. We don't have to read about the wheel each time. One of the cool things you might have noticed, these are renters of the actual satellites. There's art on the side panels. And so we do this through an artist in residence program where we have people come in either in person in the office and they get to actually be embedded with all the employees and they can create either visual art or written pieces, music, anything you can think of influenced by the satellites and the people and the mission and the imagery that we're getting. So if you look on our website, there's an art at Planet Section and you can see more about the artists that we've had before and some of the different ways that you can apply or try to get your art onto the satellites. You can design some of these side panels and there should be a way for you to do this at the camp if they haven't told you about this already. So every day we download about 30 terabytes worth of data, which is just a mind numbingly large amount. I would all get stored on the Google Cloud and this comes down to about four million images every single day. So this is something where you really need stuff like machine learning to go through all of this because you don't have enough grad students in the world to sit here and go through all this data. So we have so much data to go through so many things that we can try to gain insights from and we'll look at some examples of how this can be used in benefiting the world for science. So I mentioned before that our satellites have changed over time. The initial satellites that we had in the Doves were called the Dove pilots and these were launched from the International Space Station. These were just three-band satellites so they imaged in red, green and blue and for the optics or camera nerds in the audience, these had a bear pattern filter on them so the same kind of filter that's in your consumer digital camera or your cell phone to mimic what you would see if you were a human on the space station looking at the Earth. But this isn't always the best way to get the most information out of an image to actually understand what's happening on the surface of the Earth. So when we started launching our satellites on rockets that they were in a better orbit to image the Earth than you get from the space station because the space station doesn't cover the poles, we added a near infrared stripe to the filter. Near infrared is really useful for giving you information about vegetation health so things like looking at crops or looking at forests and so that was something that was really beneficial to add over time. And then we actually changed from a bear pattern filter to a butcher block filter so that we could get in the sort of short version, better spectral response, sharper images, better colors, better dynamic range. And we went there to first an iteration called DovR which was like the Dov Refresh and these we had four bands. We were still doing that red, green, blue and near infrared. And then with the newest satellites which are all the ones we're launching now going forward these super doves, we have eight bands. So we added a coastal blue band which can help you see through water in sort of shallow coastal regions as the name would suggest. We added a second green and a yellow band to help look at vegetation health. And we added a red edge band which is also really useful for looking at vegetation health. This is what those bands actually look like again for the optics nerds in the crowd. And we've been doing these changes a lot so that the data from the super doves and the older doves satellites are interoperable so they can be used at tandem with these public satellites like Landsat and Sentinel so that you can have a bunch of different data sets that you kind of fuse together to get all the benefits of the different satellites operating at the same time to get the most information about how our world is changing every day. But let's look at some pictures because that's the coolest part. So like I said before since you're trying to image the whole world every day you catch a lot of changes on really small time scales. I'm sure all of you heard about the Hangatanga-Hangapai eruption back in February of this year. We were actually monitoring this volcano for NASA for months before this. And so we got to watch the island grow slowly over the course of many, many weeks. And then before the big eruption that made the news the one that had the shockwave that you could see go all the way around the earth and I've seen some cool videos of that online. There was actually a slightly smaller eruption that caused this island to just grow by a factor of like two or three X in the span of a day. And so we caught that growth. And this is what the sort of last configuration of that island looked like before the big eruption. I'm sorry, the eruption was in January, February. And this is what it looked like after. So you can see that the entire central part of the island is just gone. This eruption was so large that took out the whole middle part of the island. And that volcano is still there but now the top of it is underneath the water so we can't see it anymore. This volcano actually only formed in the last five or 60 years at least in terms of being visible above the surface. Before it was just these two islands that are left that were here. And then this volcano solely grew in the middle of them and became this island. And this happens a lot in the Pacific. You'll see these kind of ephemeral islands that appear with big volcanic eruptions and then slowly get eroded away by wave action from the ocean. So this one was around for longer than most of these islands are but it's kind of cool to see geology in action. Cause a lot of stuff that we are used to as humans when we learn about geology doesn't really happen on human time scales. Certainly most islands don't appear within the span of a couple of days and then disappear in a similarly fast time scale. Now that bigger eruption had a lot of ripple effects. One of them was that it rocked an oil tanker near the coast of Peru and actually caused a massive oil spill. So here you can see some of the color variations on the water. That's actually the oil slicks that are sitting on top of the ocean here. This is a view from one of the SkySat satellites looking down into one of the volcanoes in Iceland that I'm not even going to try to pronounce. But with that near infrared band we can actually image heat basically. And so you can see the glowing lava inside of the caldera, the pole at the top of this volcano. And images like this just blow my mind the fact that we can see this kind of thing from space. We've also caught it in places like the eruptions in La Palma, Spain last fall. So this is again bringing in that near infrared band so you can actually see the glowing lava. And then you can kind of see in this other sort of greenish brown area the slightly older flows that aren't glowing hot anymore but are still from this most recent eruption. And then here if we zoom in this section where the lava is flowing into the water and you can see the smoke that's coming off as that is being basically instantaneously cooled off by the water. The other reason to try to image the earth every day is that the earth changes very quickly. So this is an example from February of last year of one of the areas that was affected by wildfires in Australia on February 1st versus February 2nd. So when you have this immediate before and after imaging it's really easy to see what changes happened on the ground as a direct result of whatever event you're looking at whether it's a fire or a hurricane or a tornado, anything like that. And that can be really helpful for response efforts for mitigation when you're planning for future disasters especially with something like a wildfire. So it's really useful to have that fine time scale of information. This is another example. It shows how quickly the world can change thanks to just the pace of human activity. So you have to look very closely because the changes are small but you'll see like little dots pop up here and there. So these are some farms in India on April 25th of last year, 26th and 27th. So agricultural fields can change really quickly as well. So it's really good to have really fine temporal scale information to understand how these crops are being affected by weather, by watering conditions, fertilizer application, pest infestations, planting, crop rotation, any of that kind of stuff. This is an example just earlier this month of something called the Glacial Lake Outburst Flood in the Himalayas. And these are pretty catastrophic events that are happening more and more frequently due to climate change. And these are events where you have melting a part of a glacier kind of in the interior of the glacier to the point where it builds up pressure behind some kind of ice dam and then eventually that dam catastrophically breaks and releases a bunch of water and so you'll end up with floods and debris being deposited. Sometimes if you're in a mountainous area far away from any human population, it won't be necessarily that destructive but in the Himalayas, these can trigger really big downstream effects which can cause a lot of destruction to infrastructure and a lot of times a lot of people end up getting killed. So it's a really bad impact of climate change and monitoring these things help us understand we can look for the warning signs. Like can we see a lake so that here you can actually see the meltwater? Can we see these lakes forming with enough advanced notice so that we can evacuate people downstream? There's not really much we can do to stop this from happening short of stopping climate change at large, which is pretty, you know, we need to but it's pretty difficult to do at this timescale being able to save people when you see this happening. But we can try to plan to get people out of the way when we see the warning signs in advance. We can also with things like the Skysat satellites look at human activity on the ground. So here you can see some people and some of the vehicles from the Sound of Peace rally in Berlin from March of this year. These are some examples of wind turbines in Wales. So this can be really great if you're trying to look at say the construction of future turbines looking at the rate and where they're going or if you want to plan out where you're going to build future turbines or looking at the expansion of other types of renewable energy like solar farms. And because we're a bunch of fusioners of planet we monitor a lot of human activity on the ground and those tied to space. So this is the SpaceX launch site in Boca Chica, Texas when both Starship nine and 10 were on the pad. We've been trying so hard to catch one of these launches as they're happening, but just timing wise and angles of the satellites we haven't managed to catch one yet but you can believe if we do catch it it's going to be all over our social media. So stay tuned. We try every single time. This example of Blue Origin's landing pad in Texas. So they actually just had a launch yesterday of some civilians if you managed to catch the live stream of that. This is where they landed except this was in which last year. And then this is the Virgin Galactic spaceport of America in New Mexico. So this is where things like spaceship to have launched from in the past. They also do some launches from the Mojave Desert. So it depends on where they want to be going but this is the New Mexico location. So I work on the science side of Planet which is basically trying to get data out to researchers everywhere in the world to be able to use this information to actually act on that mission of Planet to try to help change the world. And through part of that effort we monitor all of the publications that we can find coming out of the science community using our data in some way. And we have over 2000 publications that have come out at this point. And so we've seen stuff applied in basically any area you could think of. Lots of machine learning because this is such a dense time stack of data. Lots of stuff tied to forestry, climate change in the Arctic, monitoring for illegal mining activities, looking at pollution levels in different places in the world, monitoring natural disasters like volcanic eruptions and earthquakes and hurricanes. We'll see some examples in the next few slides. But basically anything you could think of that you could maybe do with imagery from a satellite there's probably someone in the academic community that's dabbled in it in some way. So that's really exciting to see. So going back to the Himalayas this is another event that made a lot of splash in the media. So you might have heard about this last year. In February, there was a massive flash flood in the state of Uttarakhand in India that killed dozens of people and washed away to hydroelectric power stations. And the initial reports suggested that this was one of these glacial lake outburst floods like the example we saw before. But just by happenstance with the planet scope the dove constellation with that background scanning we managed to capture two images taken 27 minutes apart as the landslide was happening. So the first image caught the time before both power stations were affected and the second image caught the time period after the first station was destroyed but before the second one was and that second image is the one on the right here. So I'm not sure if you can see my cursor but if you can, this is where the first power station was and this cloud that you see is actually the dust that's being lofted out of the valley as the flood and the landslide is moving through the valley. The second power station is down here. So this image is before the landslide made it that far. Since these images go through our web interfaces we have an interface called Planet Explorer which is kind of Google Earth-esque where you can go through and see the images basically as soon as they're available on the ground. We had scientists that were looking at these images within hours of the landslide and coordinating response to figure out what actually triggered it. And a lot of this was happening through Twitter which was amazing to see. I'm sure a lot of you were really active on Twitter and science Twitter is a big thing. Oh, sorry. So Professor Dan Sugar, he's at the University of Calgary and he's one of the world experts on landslides and he was kind of monitoring this in real time. Basically, I think he was just sitting there refreshing Planet Explorer waiting to see if new images were going to come down. And he was taking screenshots and annotating them and tagging other landslide scientists to try and analyze this within the span of minutes of getting this data on the ground. And these are just a few examples of the tweets going back and forth between different pretty prominent landslide scientists going through and saying, oh, hey, I thought this was a glacial lake outburst flood but now I see these images. I can see where it started. If this doesn't actually look like it's what it was. So can we figure out where the landslide started? Where did the slope fail? What actually caused this to begin with? And this is a whole new era for science really to be able to have so many people collaborating in almost real time to respond to an event like this. So after this group of scientists figured out where the start of this landslide occurred, they contacted us at Planet and they were like, hey, could you get one of the sky stat images of what we think is the source area so we can see how this landslide started. So we immediately targeted it, had to wait for the clouds to clear a little bit but we finally caught an image where we can clearly see where this landslide was triggered. And it's right here in this place marked rock fall site. So before the landslide, this area would have looked like the rest of this slope just snow covered glaciers, totally white. But you can see a really straight line here near where this rock fell off the slope. And essentially what happened is you had a glacier that was sitting in this cranny on the wall of this valley. The glacier detached, it fell almost two kilometers essentially straight down into the valley floor and then melted with the heat of friction when it hit the bottom. And it essentially just melted so quickly that it caused this massive flood that just dug up a bunch of dirt and dust as it moved its way down the valley floor and caused this flood that then moved down slope destroyed power stations and destroyed some of the other infrastructure and ended up killing a few dozen people. So having this kind of response really helped us understand the landslide risks in places like the Himalayas and high mountain Asia in general. And we started to go back after we got this image to look at historical imagery that we take in to see, could we tell that the slope was going to fail? And you could actually see the cracks starting to form here three years before the landslide actually happened. So now we know what kind of things you can look for in the imagery to try to predict these events. And again, with something of this scale there's not much we could do to stop it from happening but you could try to issue alerts to evacuate people down slope so that the very least while you'll probably have a lot of destruction to buildings and livestock and things like that down slope hopefully you can save a lot of lives. Another example of a really quick turnaround with response to a natural disaster was an earthquake in September of 2018 that occurred in Indonesia near a provincial capital city called Palau. And this was another unusual event in that it triggered some of the largest soil liquefaction blood flows that we've ever seen as humans. And this is a type of landslide where the ground is so saturated that when you start vibrating when you shake it with an earthquake it essentially just turns into liquid and flows all over the place. It ended up killing about 5,000 people and caused a ton of destruction to infrastructure all around the coastal regions and a little bit inland on this island. But since the images of this area were going out again through Planet Explorer in almost real time there were folks at the Earth Observatory in Singapore that were analyzing these images as they were coming down to try to figure out what landslides were caused by this shaking and where did they start? And the team that was analyzing this was able to correlate all of the locations of the landslides to places with a particular type of rice farming activity. And the rice farming in these areas had basically moved the water table so where the water underground meets the soil basically it moved it so shallow to the near the surface where in that shaking started it just liquefied all of the soil. And so this can help inform what types of farming practices you might want to have in areas that are prone to large earthquakes like this. See, so this is the before and after kind of wide angle view. If you look really closely you can see some changes along the coastline here like this dock disappears. You can see all the sediment that flowed into this river and then flowed into the bay. You can see a lot of changes in this area here like all of these buildings are just gone. Lots of changes along this area here. You had some little tsunamis that came onto the coastline. And during this earthquake the ground moved about 10 meters in the span of essentially a minute. And usually when you're talking about plate motion on the earth you're talking centimeters per year. So this moved so much that when these images came down we were worried that they hadn't rectified properly like they weren't showing up in the right place on the ground. So we reprocessed them and realized no the ground actually did move that much. So this is a really striking event. So since you can actually look at the images and say how far did any of these buildings move in the before image versus the after image you can make a map like this that shows you the displacement across the entire island. This is zoomed into just that specific area we were looking at near the dock. The dock would have been kind of in this area here. But this is the kind of analysis that would normally take weeks or months after an event like this. And this map was created within a matter of hours. We also saw a lot of folks analyzing changes around the world due to COVID lockdowns. I'm sure you saw in the news there were things like air quality getting better over large cities and some sort of true stories about dolphins reappearing in the canals around Venice. They attributed that to the water being cleaner. The videos of the dolphins were not real but the water was actually getting cleaner and we could see this from space. It was such a huge change. So this is a chart showing on the left the total suspended solids. So how much crud is floating in the water. In January and February before the COVID lockdown started in 2020 and then the right column is what the total suspended solids looked like in the early stages of lockdown. So March through April. You can see that this went down pretty significantly from like five grams per cubic meter to almost zero in a lot of places. And you can actually start seeing the structure of land underneath the water. That's normally not very visible in places around the canals of Venice because not only do you have so much total suspended stuff in the water but the water is also very turbid. It's very turned up from all the boats going through it. So this was a huge difference just from the reduction in boat traffic. I'm going to skip this example for time. This is an example here of some images taken just a few days apart of the collapse of the last what was the last intact Arctic ice sheet in Canada from July of 2020. There was a huge temperature increase in this area and a bunch of winds that came through and caused this ice sheet to break up. We managed to actually constrain the time period of when this happened with our images to a time span of 10 hours. These images are a little bit cherry picked because of the images the 10 hours before the collapse are really cloudy. So this is sort of the pretty versions you could clearly see what it looked like before and after. But we were able to see if you see these bright teal areas down here those are melt ponds. So you can actually see where the ice started to melt before the ice sheet broke apart. We also have folks looking at our data in the Arctic to look at how lakes are changing due to melting of permafrost. And these are areas where these lakes are actually releasing methane into the atmosphere which is obviously a powerful greenhouse gas. And so if we look at the dynamics of these lakes over time we can try to calculate how much of a contribution they're making to greenhouse gas emissions. And this is a really big problem across all of the Arctic as permafrost is melting because you have a lot of carbon that's locked up in that permafrost in the ground and now it's not so slowly being released into the air with the warming of the climate in the Arctic. So for this particular example they looked at lakes across huge swaths of Alaska and they're in Canada. So running short on time I'll go through this really quick. For anybody that is a researcher there are multiple access pathways where you can get free data access from planet. If you're based at a university so student, faculty, or staff anywhere in the world you can apply for access through our education research program. Anybody that is based in an ESA member state the European Space Agency including Canada or you're based in China you can get data through the ESA EarthNet program. If you're studying anything related to tropical forest loss or deforestation we have an agreement with the Norway International Climate and Forest Initiative or NICFI that gives you free access to data of the tropical forest regions all around the world. Any German researchers can access our data through an agreement that we have with DLR the German Space Agency through a program called the Rapid Eye Science Archive or VISA. And if there are any Americans that happen to be in attendance we also have an agreement with NASA through their commercial small stat data acquisition program which gives free access to our data to any researcher whose work is funded by any U.S. federal civilian agency. So groups like NASA and NOAA, USGS, USDA, Park Service, stuff like that. Or the National Science Foundation. So if you or anybody you know would be interested in applying to any of these programs you can just go to go.planet.com slash science and all your information about these programs and the details can be down there. We'll skip these. So I like to give everybody with this thought when you go off into your day after this particular talk what would you do if you could see daily change of anything anywhere in the world? And with that I think if we can do audio we'll have time for Q and A if you don't get a chance to ask a question now you can also just reach me online at tanyaplanet.com. Thank you very much Tanya. Round of applause. Thanks everyone.