 Pleasure to be here. Let me thank you, Allie and Stefan and Kevin of fantastic hospitality and environment here. I've already benefited by, you know, learning more bids, but also algorithms from Stefan. So I hope we can work more together and productively. This idea of fire detection from space, it's not a new idea per se, but what we did, I think our group did was show, in fact, how strong the signals are, how easy it can be, and how it should be done. I would posit that all good planets in the galaxy have fire detecting satellites, and we should catch up with them. Anyway, so here's the idea is that we have a system called Fuego. It's an acronym for fire urgency estimator on geosynchronous orbit. It can sit out there 24-7, you know, particularly during, you can't see through clouds, but it looks in the mid-infrared range around 3.9 microns. There's a nice atmospheric window there. It can sit and it turns out that the signals are so strong, you know, that you get a million photons per second with a mirror this big out there at 25,000 miles. I mean, it's just huge signals. And I think the kind of skills or enthusiasm or maybe naivety I bring from astrophysics is we sit and we wait and wait night, you know, hours to get a thousand light particles from a supernova a billion light years away. We're so, oh wow, we got a thousand light particles. So looking, when we started to look at the signal noise levels here, we really were impressed with how strong fires are, how we might be able to slowly work our way up to the system. As you get it, probably as many of you know, satellites are, you know, kind of heavy metal of the space industry and take decades to plan. But in fact, there's been some breakthroughs in satellite deployment. They're very encouraging. Where smaller, cheaper satellites can be put onto buses. The Air Force has done this quite successfully, and we might even be using some of their data. So I think, I'm hoping, you know, within a decade, we can see progress towards a space-based system. Just a little, so Fuego, where the system we, Graham Fleming kindly supported us with, he helped fund Merrick Jakobowski, who was a good, fantastic graduate student, Maggie Kelley, one of our collaborators in the College of Natural Resources. And again, what we're looking at is to try to see if we can see fires earlier. The answer is we think we can reject false alarms enough. Again, that's a big question, because you have a lot of light. False alarms enough that we can see about a 10 square meter fire, you know, 10 feet by 10 feet, roughly, area of a sandbox size from outer space in every few minutes. I'm mean, when Ali raises the question, well, you know, will it help? I mean, can you, can you put it out enough? And there's kind of a magic window to our mind. You know, if a fire, fire is growing at a few hundred feet per second under Santa Ana, if you can get to it within half an hour, you can cover it maybe, you know, you could, with aerial tankers and ground crews, it's not turned into a mega fire. You know, you know, mega fires are really the evil part of the whole business here. They do most of the damage, most of the cost, most of the pain and suffering is called by these big fires. Hopefully we can get to the point where we can help, you know, with all the comp, also computational predictive tools, we can understand when a fire might be dangerous or not. And, you know, substantive, Cal Fire is a, is a, is a large, very effective organization, and we're trying to meet with them and talk with them and start to play a little bit bigger role in their day to day running of things. Now, we're not just this 10, 20 year away satellite, but we think we can be useful now in a, in a big data way. I mean, one might argue, like that's where the real low hanging food and fires over the next decade will be is in data management. So, uh, uh, well, let's see. So here's actually just a little back, back history. Uh, I, I help, I co-advice all Perlmutter with Richard Mulder. He was, he was, uh, started as a graduate student in our group when I, I was one of the leaders of Supernova Search. So we started out Supernova and it's, it's relevant. I mean, not only for Sol, but also the fact that finding point sources against the clutter background is something, you know, we've kind of done most of our life, although with primitive, much more primitive tools than Stefan has now, but we'd be looking at many, many images of the galaxy. Look for one, one or a few pixels that are bright, that seem to have some nice characteristics. And then if it looked like a Supernova, we turn other resources on it. That's kind of sounds like what happens with a fire from outer space. You look, you see, look for its reality, do tests, you know, look at, uh, hyperspectral imaging and time development and other characteristics. And eventually you could understand if this is going to be bad or a false alarm or it's going to be something real. We even have a patent filed on some of our algorithms, which apparently I'm told, actually, you'll run this in the talk. The Department of Defense has very powerful satellite resources up. And somebody says, oh, I mentioned our patent that we filed for, University of California filed for it. And they said, Oh, DoD knows all about that. Anyway, so that's just, again, here's, so here's what we've done in the Supernova world. We found, we had our first audit, we, we also, it turns out we had to automate telescopes, automate image processing and scanning and all this stuff. And we had the first successful automated Supernova search back in 86. It was in the, this galaxy here. So this, this was the first Supernova totally found by automated means, again, which is like, will be like fires. Okay. And here's, we moved it to a deep Supernova search in Australia. It's always part of this work. The Anglo Australian Telescope is just some history. Again, the, the subtext of this talk is that there's been tremendous, tremendous development in, in image, image acquisition. Well, often because of Department of Defense initiatives, but the commercial sector has caught up industrial section and huge, huge arrays of things, micro-bulometer arrays. Now that would have been a fraction of a million dollars, now a fraction of a, you know, $1,000, for example. So there's just been factors of 10,000 price reductions and detectors at all wavelengths. So it's really exciting. Again, with the company, they're all coupled to computers and image processing and stuff. So I think that the sensor question is, there's been tremendous progress on that's going to help us. In fact, I mean, again, there's sort of a subtext. We came in to the Supernova business, we said, oh, we have microcomputers, PDP eights. Steffen, I know about those. We were, we were the hot kids on the block, because yeah, we were using a microvax. And, and, you know, we could afford microvax and a, you know, 400 by 400 CCD. We're the envy of the many astronomers doing this. And actually, there's a similar tale here. We were kind of outsiders, you know, physicists and stuff trying to go into astronomy. We met with a lot of, you know, well justified skepticism about our ability to do the field. We, you know, so I think so here we are kind of outside, you know, we're data people kind of coming into fire business. And but we're treading a lot more softly. I have enough scar tissue from the Supernova experience that, you know, we're working with people in the field and Cal, we'll be meeting Cal Fire people in Sacramento this month, I think. So that's, I think we can learn from that. Here's the, again, this is the instrument being mounted on the Anglo Australian telescope. And eventually it moved to the Canary Islands. We had bad weather in Australia. This is a two and a half meter Isaac Newton telescope there. That's where we found our first Supernova. That's actually, and then Saul took over the experiment at that time. He was a, he was a more skilled manager than I think and a better communicator. So that was good. So here's, here's the Nobel Prize. He kindly took, he was great. He took all the kind of founding members to, to Stockholm or the whole team to Stockholm. So I've never seen so much alcohol consumed in my life. But anyway, then well back to Fuego. So actually working with Merrick and Maggie and Scott, you know, the real, real experts on fires. We, we comb, we probed a number of data sources and found, you know, there were, even though there were backgrounds from these satellites. There's we have weather satellites, the, the Go satellite is taking images at a pretty good clip in providing us, it's a test bed and also a modus satellite. That was a low Earth orbit, but it was imaging fires pretty much. And we put all this data together and showed that we, you know, with some algorithms on time development and spatial characteristic, we could, we could repress the backgrounds enough that look like we'd have a credible case for fire detection from space. And again, all that, all things have gotten so much better since we wrote this paper. So that got refereed and published. Oh, and again, where we are in the sensitivity and, and, you know, we again, the world in the United States does have a lot of powerful satellites out there. But the Fuego satellite, we want to sit down here, be able to revisit a target every minute or so, also go much much, go down to about a megawatt fire, much much more sensitively, even goes are, which will be one of our nation's most powerful satellites. Well, we should be able to operate mainly because we've, we've optimized, optimized, optimized the satellite for fire detection, we should be able to go down 10 or 100 times more sensitive on that. Now, and we were working with goes our scientists, data scientists, Chris Schmidt is at the University of Wisconsin in Madison. He's a collaborator on this has really helped steer us and, and, and direct our efforts to a large degree. The the, the fire community, actually Elaine Prince was one of the leaders in this, in this movement at, she was at Madison. And they always had to fight against the weather people. You know, we got a hurricane coming, you know, they would tend the whole data flow, the whole workflows through the satellite, getting data out quickly and intermittently. It's just a tremendous nightmare because it's, it's optimized for hurricane prediction, I think, or, well, I might, I'm probably speaking a little bit beyond my knowledge, but it's been, Chris has struggled violently and successfully often to be able to get the data out of these satellites. And in fact, that's, that's getting better and better. Future satellites will have more and more capabilities in the Japanese satellites coming up. So I think actually the existing satellites, in fact, will get better and better and have these kind of capabilities. So here's a example of a fire tanker, aerial tanker in action. Actually, although not with this, this, not with this airplane, but there's also been recent progress in, in getting refurbished DC-10 super, you know, big, big airliners converted into aerial tankers, a company down in Albuquerque that's doing this. And that, that means, you know, you could get to a fire two or three times faster, dump a heavier payload and have a bigger impact. So they're kind of a natural synergy with our fire detection system. You know, we could, if we see something quickly, we'd want a quick response. So here's just, you know, it's really a global problem. This from Chris, this is from a talk Chris, Chris Smith gave, you know, some fires all around the world here, you know, various regions, you know, the usual deforestation fires in Brazil. He says actually, he went and gave talks about all the fires being burned in Brazil and actually he, he's scary, he can't go back there and talk about this anymore because he has, he's built up enough enemies in Brazil that he's not respected there or not appreciated there. Let me just, so let me just look at one example of where systems like ours might have been somewhat useful. This rimfire is an interesting example. Actually, here's a picture of it on the GOES satellite, which again, it saw it pretty early. It could have seen it pretty early, although we didn't get the data out until like a month later and the workflows were not made to this, but GOES could see the rimfire. This turned into a mega fire, third biggest, and I think mega fires are, you know, something more than about 50, 100,000 acres of fire and they do most of the damage. But it was very interesting. The started, I guess, by a campfire and it was reported by an aerial tanker, a cowfire tanker flying to another fire. He said, hey, hey guys, you know, call radio, and there's a fire over here and he flew on to his fire. And but again, you could imagine if you were coupled into a simulation system, you know, people could look at the moisture and the winds and everything. Somebody might have said, hey, put your water down on that other fire. But it didn't got out of control. Did a huge amount of damage. The suppression costs were 100 million. And then a question. Yeah, yeah, yeah, I think we could have seen it a little bit earlier than those guys. Hopefully, again, well, you know, then we had to have that I don't want to oversell our capabilities. I mean, I think we're probably an evolutionary step, not revolutionary, but you still have to take private taken, you know, minimum half an hour, 45 minutes to get an aerial tanker right there, except for the one that was flying right nearby. So again, the suppression costs were $100 million. And then this, this isn't doesn't seem to be well known. It became very apparent. It turns out that if you look at the real amount of dollars spent on fires, it's typically between well, two and 50 times the suppression cost. And, you know, somehow we all pay for that. And you know, it's a societal cost, you know, insurance, you know, goes into state budgets. And we don't know. But this is this you'll all still have another slide about where all that money disappears housing. I think this valley fire did a billion dollars worth of housing fire. And we're paying for that in our insurance costs. So again, Scott Stevens is a world expert on fires and gives this is in practice from his one of his talks. But in the 40s, we decided that we adopted what's called the smoky the bear policy that we'd start to put out fires. And I mean, that that's a arguably good thing to do. But also kind of tilts against natural what the natural role fires are in ecological systems. And that is that Scott has very interesting stories about actually how native Americans, indigenous peoples in California's would set fires. And, you know, at a lot higher rate than we see now. But when if we put out fires that the, you know, grass, underbrush, you know, forest fire, you know, forest pine needles, all this stuff will accumulate. And when when fires burn, they burn with a much more ferocious thing. I mean, this valley fire, people were amazed how fast it spread. Is they even with the moderate wind, it was going much faster than they thought, you know, because we've been in drought for so long, and fires have been had been suppressed in that area. The, for example, some of the fires within the national park system will let fires burn, as well known, apologize to experts in the field, but national parks will let fires burn. And when their fires go off, they're much, much easier to, you know, at least to contain or boundary. They do a lot less damage than a fire right outside the national park boundary. And Scott has lots of data on this from Yosemite. Again, but, you know, and actually, and the the Cal Fire does an amazing, amazing job. I mean, they're almost an A plus student. They're, they're putting out, you know, 95 to 98% of the fires very effectively. But it's a few percent that escape that are kind of problematic, which also, you know, which also links to the whole predictive what our needs for a data driven fire response system cries out for fires are not particularly there's, we don't have a lot of evidence that there's more acreage being burned. And this is way too noisy data to make any statement about increased fires. But for sure, the were the fire suppression costs are increasing, both because normal inflation also in California, we tend to put a lot of effort into protecting, you know, the urban, urban, wildland interface. Those are houses. And when fires he can do those the, there's a lot more money and put into putting these fires out. So we're spending several billion dollars a year on suppression about half of that's in California. Actually, and California does not have the most wildfires, but it has the most expensive half of the suppression costs are in California. Let's see, you know, and then I think some of you have seen this already. This is a graph from from a publication that Scott Scott was on with other other, other team, other collaborators. But they they looked at what what what fractional increase I apologize for the quality of this graph, what fractional increase in fires over the next few decades would result. It's kind of interesting, you see areas as the climate warms, this is within, they have several predictions based on several climate change models. But this is one typical one, but places like Humboldt will be somewhat closer to San San Diego LA area. I mean, see, maybe interestingly enough, San Diego LA can't get any worse. They're already maxed out in terms of there's climate change isn't going to make any difference. They're toast already, but more ways than one, but places like Humboldt, Northern California, we're going to see a lot more big fires if we keep ramping up on climate change. So that's a concern, I think for all of us. Oh, I had a little let me see if this will play. I don't need any here we go. Is that coming across? This is acute. This from whoops, that they have to detect the detector again. That can somebody move it off the screen? Okay. I apologize. I should test this. Now if I go full screen will disappear. Okay, well, you can get the idea. This is just a little sense. Stefan has friends in South Africa. We're really a global problem. It looks like, let's see. So play it again. This is a it's easily found on the NASA website. Did it disappear? Okay. But this just shows the seasonal variation of climates over a year. Summer in the North comes, we get up more up there. California wasn't so bad. I guess this is 2003 when this was. I guess we didn't have that bad a year of fires. You see Africa, South Africa, a lot in Brazil in summer. I guess that Philippines, Indonesia, but it's a cute, you know, it's kind of a scary, cute thing. Let's see now if I go back to, oh, look at that. I wonder if you run into the jungles like immediately sub Sahara, you have more, you know, yeah, scrub or more stuff. And then the jungles are too too wet to burn. Then you jump down to the Serengeti. But, you know, Indonesia has a huge number of fires, you know, Philippines, there are a lot, many man set human set for on purpose and stuff. But, you know, it is a global concern. See, so how do I see my cursor again? Oh, there I apologize. It's such a well run group and I'm clutching away here. Okay, so here's, this is interesting one wasn't, you know, probably just on the boundary of mega fire. But again, it was, this is an example of a fire over really, you know, droughty, dried out region. When it, when it, when it ran, it killed 95% of the trees. And you end up with this. And as Scott says, well, this is going to take a century, these places to come back. So that's again, kind of scary. So again, I meant, I've talked about the financial impacts that are growing. Acreage burn, not, not so much. Again, it's all on the few of these mega fires where things seem to happen. Again, we've talked also, you know, you'll get places when you'll have multiple events, multiple fires going on. You know, where do you, where do you put resources? Where's the action? And it's sort of become apparent that there's a benefit for society if we can have very high quality infrared data of a fire that's ongoing. And we know every pixel is telling us how much energy is in the, in the fire. And then I'll come to next to the, the data layers that go into the simulation programs. But if we, for example, if we really knew what, what was happening in a fire, and, and we had good pre-existing GIS data all prepared, then, then I think it would make this whole, whole issue better. This is a little too detailed. But again, we, we hope that we can, you know, with the detection and also the ongoing monitoring of fires and running simulations. At some point, this will take, I'm sure, a number of years to get us into a reality that we can be an assistance to the incident commanders. The, the incident commanders have, I've been told by one of my collaborators, you have to come to a fire. These, these guys are balancing all sorts of stuff, you know, where to put engine crews, and, you know, where to put airplanes, and, you know, they have a very, very, very hard job. And then for how to have them sit down, you know, our kind of life, look at a workstation, look at surface plots, run a few simulations, have a cup of coffee, talk to somebody, does this make sense? Kind of, you know, the kind of blessed academic life we live here is not in the life, not in the life of an incident commander. So you can imagine there are cultural, you know, we have to be very, we ought to be very respectful of what their current systems are, which largely work, and see if we can slowly migrate in and do something effective. So, oh, again, here's this, here's this one example of the San Diego fire I mentioned. See the fire suppression there, this is remarkably tiny cost, 2%, and the real cost to the rest of the community was huge, you know, unemployment insurance is 17%, about 50% was loss of property. So there's a huge multiplier on how much damage is done to civilization, and we all pay for that. Society pays for that, even though it's not clear how, but it comes out of a lot of stuff that we depend on. Again, I mentioned some characteristics of Fuego. One thing that comes up here is that we've, we've had, I think, a productive dialogue with some of the people in our, United States Air Force, who have a satellite system called SHERP, that actually not, not, not flying now, but there might be, there might be something like it up there now, and I don't have any security clearances, but these, these satellites, in fact, do provide close to the kind of sensitivities we would need that are up there already. We're hoping there's some way we can work with them that we can use some of our algorithms, some of Stefan's algorithms, to, on the, on the existing satellites might be processing behind, you know, behind, on their computers behind the wall, but we can get out alerts from them. In fact, there is a system, a very good gentleman named Everett Hinckley from the US Forest Service, has in fact worked with a number of the military and classified satellites in space, and they're, they're letting, they're sending out alerts effectively. Again, a lot of his job was to suppress the, the, the backgrounds again, which I think, you know, I hope the machine learning tools we've talked about can be a, have real significance there. Let's see. Oh, and again, drones, we, we got in the newspaper over drones. Turns out that, I mean, drones are, is, you've come to learn our, our amazing data tools. You can sweep up huge amounts of data, and you know, there, there are significant FA worries and firefighters worries. There was, I think there was one fire, a drone came in and almost, the guy, it freaked out a bulldozer operator and he almost tipped the bulldozer operator, almost tipped the bulldozer over. You know, there've been near misses on drones coming in and cow fires, helicopters have had near misses and airplanes. And so again, the, the whole atmosphere of fire is so, so much, there's life threatening and property damage and all this stuff going on. And having an ammeter come and fly a drone to look around is really, is not a good solution. So I think that's given drones a, you know, in some ways that justifiably, you know, create a lot of skepticism about the use. But in fact, there are ways they can be used immediately. For example, the mop up operation after a fire, you could, you can fly a drone with one of our infrared detectors. I mean, we're, we're working with the teams that have these drones and infrared detectors. And you can, they can see and mop ups are costly and they're not dangerous, but you have to go and find all the hot spots. This is actually such, such low hanging fruit for a drone and a, and a detector data system. You can see one inch embers easily from the drone at a thousand feet. I mean, these systems are so sensitive. Okay. And I mentioned, you know, in as much as Fuego will be, you know, five years, 10 years before we can even consider building our own satellite. We're looking at all these things in between. And again, we, we're working with a drone company and a, and a, and a good company that integrates data systems with the drones. And we know, and, and actually this team has made significant progress on infinite duration drones that I mean, Google and Facebook and are, you know, a lot of IT companies are planning, you know, long term, the balloons or drones, which we, you know, we've talked about, but this, this company has something light and cheap and ready to fly now. And they, they, I think it's conceivable that within a year and a half, they can have a drone that could fly at 80,000 feet forever, you know, solar, and there's some algorithms, there's solar and recharging and they have some other patented technology that can keep these things up there. And one like that could sit and look at all of LA County, for example. You could have one drone over every, you know, one over San Diego County, one over LA County, etc., with probably, you know, one foot, one foot pixels. That's, that's, that's all very, very possible now. So I think that's something we can do very quickly. You know, data will play a huge role in it. And, you know, maybe I hope, hope we can form some working collaboration with bids on this. I mentioned the history. We got into this from the supernova side and the Oakland fire. But at the, again, at the time, the infrared detectors in 89 would have been a billion dollars just for the camera and probably, you know, another factor of two or three on top of that for the satellite. But things have gotten a lot cheaper. Both existing satellites out there seem to be willing to share their data. Well, this T for SAC, this, this is a, it's, it's an interagency coordinating committee. Turns out that there, there are half a dozen agencies, at least, you know, federal and state all, all involved about fires. They never talked to each other. So this is a group actually we've become part of and it's helped us get a lot more understanding of how, how we, we might be possible useful in this, in this business. But anyway, so again, we learned of the, of some of these drone collaborators who we're working with right now. And that's, that's, that's been fabulous. Also, and through T for SAC, we learned of the San Diego wildfire tires and a good system called HP RIN. Well, they have cameras, cameras on, on fire towers that are watching can cover a huge amount of San Diego looking for smoke. And it turns out they're mainly used now for, for looking for confirming fires. If somebody says there's a fire, they'll get a call in and they'll look at made try, try to triangulate with two tires and fire two towers in terms of fire there and we're saying, wait, how come you're not looking for fires with these? That's, that's, that's what you're doing is well, you know, we don't do that. So actually they're, they're, you know, they're, they didn't have enough resources to develop a automated fire detection algorithm. And I have, I have students working with now and and hopefully we can make some progress with, with Steffen on this. So and here's some of the team and we have people burst in GIS. This Bill Cruz is a really good GISer that mentioned the, this company Fireball is in Reno and Drone America is there. Work with the University of California, San Diego Super Computer Center. That's Ilkay Al-Tentis and mentioned Drone America. A woman who used to work at VMware and is a cloud computer specialist. She has, they call them workflows. She has our workflows working on their system already with data aggregation and data simulation starting to work. Again, so I think our long-term goals are to eventually get there, but we don't have enough credibility to justify the satellite yet. But I think over the next five to 10 years we'll develop that baseline of information that we could could develop the satellite. And again, we want to, there's, we want to take advantage of a pretty large amount of data that exists out there that this, these military satellites who we wanted, we will soon have some access to that. And also working with the drones and the infrared dating systems that turns out the forester has a pretty good system called an IROPS which once a night will fly out over all the fires or a number of fires in the Western United States. It'll go and survey them all. It's based out of, I think, Boise, Idaho. And but that's not enough of information. I think what comes out as a perimeter, really what you want to know is how bright every pixel is at high resolution. And so we're trying to hopefully support and build on the good work of an IROPS and the Forest Service. And again, the drones. And we have a system, Tim Ball, a fireballer, a collaborator, has airplanes now with very high resolution imagers. And hopefully that can supplement and IROPS and make our simulations a lot more effective. This, again, we have a lot of activities, you know, again, trying to probe various places where our use of data and algorithms could be helpful. Well, here's, again, this is an example of a workflow diagram. You know, let me, I think this is a public knowledge in, but there was one of our collaborators was in a fire situation. And their GIS data, their input data, was going into the fire simulation program. And it looked like it was going to be a pretty fast burn across this. So they, they, I think they alerted some folks, but then it found out this one area turned out, they thought was a grassy field, turned out it was a reservoir. So it'd be very, very little chance of any fire damage or anything like that. But they had taken that input data from a federally maintained GIS site. Again, just shows the kind of issues in, you know, data validation, you know, qualification of this. And that, that's why we really, there's really work at almost every, every, every piece, I also work in science education and high schools and, and middle schools and almost every, every, every stone you and turn, and overturn the system that has lots of really wiggly things on it. I think there's, you know, like many things in society, same could be said that there's a lot of room for improvement in how we handle data with fires. And again, I mentioned we're going to use the existing fire products that try to really get the best GIS data. Then there's kind of, there's pretty static stuff like contours won't change, but winds will be changing. You know, even fires will generate their own winds. We have some algorithms who want to test on fire-generated winds. And, and kind of make this thing detector-independent enough you know, particular computer-based system independent so it can run in the cloud and be of use, so, wherever. So that's, and, oh yeah, here are some of our, this is from Bill Cruz, kind of one of our main data guys. This is what we want to do within the data management. We want to be cloud deployable. We want everybody, like a firefighter, to be able to look down and see, uh oh, the fire's coming this way. I mean, not, not only an instant commander, but most of the firefighters have cell, have smartphones and you know, an airplane flying above can be a local hot spot for communication. There is software out there for doing that. So we can, you know, these problems like in Colorado a year or two ago, you know, fire came down the canyon and wiped out a whole bunch of firefighters in a real tragedy. But that's something probably, again with modern detectors, modern communications, that could have been averted. See, again, the system, UCSD, that's another talk in itself has a really wonderful system called Wi-FiR, which tries to, which again aggregates data, handles the workflows, collects data from the towers and pushes data out to the incident commanders and the firefighters and the community for evacuation plans. So they, I think that's the, we're, we're, you know, building, we're kind of writing on their coattails and trying to support Wi-FiR to the best review possible. But that's a superb system that I think is going to get better and better. And they kind of view us that the, you know, we bring some of the algorithms and the sensors, like I mentioned that, where our hopes to be able to find smoke from their watch tower system would be, might be a somewhat added asset to the system. So here's, again, Bill made this, this picture of the data flows you know, including the real-time sensor input, data services, the old, the GIS stuff, which is up to date. Then we'll have cloud services that should be independent of, of a lot of the stuff. And then we want to be able to get, eventually, we're not there and I think we're, you know, at least a couple of years away, be able to get, send something out to the incident commander after we know the things are working our simulations work, saying, well, you know, put assets here. They want something very simple and quick. They don't have time to look over these things. And I think, hopefully our, our, our software stacks will be mature enough to help, help to undertake this. And this is more, more than you need to know. But again, Bill, we've gotten into some, some of the, all the, all the possible inputs. I thought, I thought this was cute. The LiDAR products. So the Bill Cruz, our collaborator and he's in small company and on the peninsula. But I think working with the Save the Redwoods, he's mapped every redwood tree I think in Marin County. Just, I mean, this, the cloud is very big and, and spacious and can give us exquisite knowledge. So I mean, imagine, you know, if you know where every tree is in, in a, in a, I don't, again, I think it's a lot of work to dig that pool, pull the points out of, out of the cloud to do that. Takes him, you know, it takes a lot of cycles to generate a tree. But I think, you know, with GPUs and advanced systems, you know, if we know where every tree is, where the wind is, how hot it is, how likely it is, we're going to see huge, I think, substantial advances in the simulations. There's a, there's a number of fire simulation programs. One of the best tested and probably pretty reliable is a program called Fires, Farsight comes by Mark Finney from, I think, from U.S. Forest Service. But he just, here's some of the parameters that go into, the data layers that go into the, to the simulation programs. You know, things you might expect the elevation slope, the fuel model, the crown, all these things about trees and forests that you need to know that will predict what's going to happen, including the moisture. There's things, moisture and the local wind and et cetera that are also puts into this. But it's, it's probably done everything right. It's never been, it's never been, it's used a fair bit, but the validation is still not very good because he doesn't get good enough data input. I mean, he, and we say we want to get, you know, resolution of meters, at least from our airborne UAV systems, meters to send in them. He says, oh, that'd be such a blessing. But right at the present state, state, I think that computationally we're ahead of the sensor and the data input with Farsight. I mean, I think it's probably a very good program but it needs, right now in some ways it's garbage in, garbage out sort of situation. I think that's going to improve a lot over the next few years. Again, there's, there's been, oh, there's, this is some of the algorithms on that Chris Schmidt is deployed for GOES, for example, we're working with him. So, you know, I think, so I view our, guess hard, but about the end of the, pretty, we want to get this GIS, probably particularly for San Diego, get the GIS data together for them, vet it, make sure it's in great shape. And then the next fire, when we have a little bit of funding, we're going to fly, fly and get exquisite. We can get 100, it's, it's kind of sad, you know, with, with a few thousand of us, about $5,000 a day to rent the airplane and the cameras. Live for, you know, for doing one fire, we can have 10 or 100 times the best data ever collected on a fire, which tells me how, how vulnerable this field is for improvement. So then, then, you know, we'll be doing this data experiment. We don't want to talk to, we don't want to bother incident commanders. We'll map the fire, run, run through our simulation programs, you know, as the fire goes on, you know, and, and, you know, see how good our, our predictions are, see if our, our predictions for the fire actually match what happened. That would, that would be a first. So that's, that's kind of, I hope it's sort of a continuing, a conservative goal of who we're going to go. So anyway, I think, again, you know, that's, it's an area that's being kind of unplumbed, you know, when you see the, how the silos of government funding has gone, and you know, we go to, we go to government agencies and they say, oh, it's a great idea, but we're on continuing resolution or sequester. You know, we have, we have $1,000 for this widget. You know, we get laughed at. So, but I think, you know, we're, we're still trying. We have two NASA grants in right now and it might be an SBIR we'll be applying for in a few more months. And, we've been in conversation with some venture capitalists who maybe on the drone side see this as a low-hanging fruit. One, by the way, one of our, our drone collaborators has just made some progress with a, with a power line inspection. That in fact, there's, there's cash flowing in that, that regime already. So I think that, that tells us that the, there are, besides the wine industry, there are viable uses of the, of the data management with drones. So, again, I'm, I'm delighted to be here. I've just appreciated the quality of people, quality of people in the interest here. And I hope that, you know, we can find a way to work together and advance society. Thank you.