Innovative Tech for Natural Disasters
Good evening. My name is Cheryl Peach, director of Scripps Educational Alliances, here at Scripps Institution of Oceanography. It's a great pleasure to welcome you all to the Jeffrey B. Graham Perspectives on Ocean Sciences Speaker Series. It is my great pleasure this evening to welcome our speaker, Dr. Neal Driscoll.
Neal is a professor of geology and geophysics here in the Geosciences Research Division at Scripps UC San Diego. Neal does research on tectonic deformation and the evolution of landscapes and seascapes. His work primarily focuses on using the sedimentary record to understand the processes that have shaped the Earth. Neal is also the director of the ALERTCalifornia program, this is an early fire confirmation and situational awareness tool that consists of over 1,000 cameras across the state of California. These are state of the art pan tilt zoom cameras that are tools that firefighters use to help respond to fires.
Please join me in welcoming Neal for his talk on the ALERTCalifornia program Developing Technology to Stay Ahead of Natural Disasters. Prepare, respond, recover. Neal, thank you so much for being here.
Thank you Cheryl for those kind words. I'm mostly known here at the Scripps Institution of Oceanography as Cheryl's husband. What's the mission of this program, ALERTCalifornia? We're really trying to assess the fuels, quantify them, assess their moisture content, their health. We do this before events.
We also want to use our system to confirm early ignition, so we fight fire in the incipient phase. Finally, we want to understand how we recover how these burn scours, debris flows, erode, fill our riparian environments, which is a fancy word for fluvial or river environments. Finally, how do they revegetate? Do we go from chaparrals to grasses? How long does it take for the fuels to dry out? All of these questions remain and we're trying to get at them now to go into a little more detail, we want to quantify the pre-fire fuels, fire-inducing conditions and how we can use this for forest management and planning. This is where hopefully tonight I give you a little hope that we're going to move the needle. Secondly, we have sensors, these high definition pan tilt zoom, near infrared cameras, that provide early confirmation, actionable real time data, situational awareness for evacuations, for prepositioning assets where we think the fires are going to go.
Finally, as I said, we try to characterize the post-fire impact on sediment erosion, landscape evolution, landslides like in Montecito, that killed 22 people overnight. The whole game plan here is threefold. What do we do prior to events to better manage and plan? How do we work during events to respond and then after events, how do we recover? This research is carried on by a big team. I hate payroll day, but anyhow, these people are the backbone of ALERTCalifornia.
They're network engineers, computer engineers, field engineers. Here we lost one of our best last week unexpectedly, Bryan Hoban. I said, I'd be short and sweet.
Excellent worker and a good friend, I miss you. I'm back, death is so final, and he was 42 years old. It takes a village. We've built
this community where everyone pulls the wagon. There's trust, there's faith that you're going to do your job so they can do theirs. All the fire departments, all the utilities, wisps, counties, all work with us here at the University of California, San Diego to realize this platform. We're in a time of extreme climate. This was here in Derna, Libya back in early September beautiful city on the Mediterranean. To give you some bearings, Egypt is to the right.
This is what it looked like afterwards. Sixteen inches of rain, the dams broke, the fatalities are over 11,000 people. Eleven thousand over that devastating. We hear about these big storms and deluges like Montpelier, Vermont became a giant river.
We're getting extreme rainfall, extreme drought, heat waves. This is all extreme climate, air quality. Here I'm sure you all heard this summer was the hottest on record as released by NASA. Look where the heating is. A lot of that's in the polar regions north and south.
This heating it's beyond what we expected at this rate. If you look here on the average temperatures on the left, so over here what we can see is that 2023 is one of the hottest summers ever. Remember Florida, the water temperatures were so high, the reefs were bleaching out. I heard temperatures of up to 100 degrees C, that's phenomenal. Air quality this here is a map of air quality PM2.5 concentrations as we go from 2006-2020.
LA in 2020 had the worst air quality conditions of any county in the United States. PM2.5 is linked to the longevity of life. These fires emit huge amounts of particulate matter and CO_2, to give you a rule of thumb, 100 acre fire, 10 years ago was the largest fire a firefighter would ever have to fight. A 100 acre fire, now is considered medium. But 100 acre fire releases the same amount of CO_2 as seven million cars burning continuously for a year.
We're not going to solve this extreme climate problem until we solve these mega fires and the CO_2 release. You all heard we're going into what's going to be a super El Nino. You've all heard that. El Ninos usually bring wetter and warmer conditions. I say usually we emerged from three years of La Nina, and La Nina is supposed to be colder and drier. Well, I don't think those people up at Lake Tahoe and Mammoth thought it was drier.
Some of the predictors that we used to use that were robust are breaking down because we're in this new extreme climate. We need to move the needle. These were the largest fires in California and what I hope you noticed is 2021 and 2022 didn't make the cut. We looked at the Dixie fire, that was under a million, the August complex fire, these were in the Sierra Foothills, huge fires.
The money we pay to suppress these fires over here is in billions of dollars. This is the money to put the fires out. This isn't for new airplanes or for salary support.
This is money above and beyond Cal fires budget to suppress fires. We're always able to suppress fires. If we shift $1 to pre-events, start managing, start planning and I'm going to show you data that we might be in a better spot for these bad fires that burn super hot, that burn up into the canopy versus good fires which are colder and move more sluggishly. We want to drive the system in that way. One of the goals that we had when I talked to you last time, it was in December, was to map the Sierra.
To map all of the regions that are high fire threat, coastal ranges, transverse ranges, peninsula ranges, Northern California. Every area that was yellow and red on this map on the left is what we consider high fire threat. Then you can see over here on the right, the three phase program my team had so that we could go and map prior to the fuels. Here I'm really proud of our team. The blue is what we've mapped so far.
We mapped the largest dataset ever collected at one time with the Lidar. We mapped the entire Sierra. That's impressive. We also mapped large sections of Northern California.
We don't have as many cameras up there because the natives are growing stuff they don't want us to see. The greenish yellow is what's left to collect about 35,000 square miles. I'm working on funding to secure that through the California Resource Agencies because we want to characterize the fuels, the health of the fuels where we need to manage. Even modest management seems to be really important, as I hope I can convince you.
Together with the LiDAR data, and I'll tell you what LiDAR data is in a minute, it's light detection and ranging, but we want to collect imagery data, hyper and multi spectral data that tells us the health of the fuels, the water content. Combine there are force multiplier. We haven't gotten as far with the multi spectral data.
Only the purple has been done and the other phases are yet to be done. That's on our to do list. What is LiDAR? It's light detection and ranging.
Plane flies at constant elevation over the terrain and emits pulses. These pulses return from tree tops and from below the trees, from low vegetation in the ground. We can go in, and we can quantify the fuels that are in the canopy and on the forest floor. When we take this together with the multi spectral data, we can then tell the health of the forest. Here, this is Bixby bridge.
Everybody come on Big Sur. This is such a great drive when it's open. You can see here the detail of imaging the bridge. We only got to lime kiln because it's closed from there north, and I learned on this last trip up into Pacific Coastal Highway that a lime kiln is where they make the lime for vegetation and agriculture. I didn't know that. That Lime Kiln State Park
there where it's closed, is named after that. Bixby, the gentleman that is credited for named after this bridge or the bridge is named after him, had a huge lime kiln to help his neighbors. Here's the aircraft, there's the Sierra pilot and equipment. There's the LiDAR resolution. Eight pulses per square meter is what is recommended for quality level 1.
My team has been collecting 14 hits per square meter. So much greater resolution. Nominal spacing about a half a meter, and service altitude is about 2,000-2,500 meters. They want to stay constant elevation above the terrain.
They hated me for making them fly the Sierra. This isn't constant elevation Neal; and I go, yeah, I forgot. What do we get from these data? This here is Mount Whitney. This is the first return, everybody see Mount Whitney right there, the big peak, and you can see the forest.
This is the first return from the LiDAR. It hits these obstacles and it returns energy to the system light, and we can record it. We can get very high accurate maps. We can take the vegetation out and we can make what's called a bare earth model. Anybody that's been to Whitney portal, you can see the roads that transverse.
We've done a really good job collecting awesome data. What I'd like to do is I'd like to hold our flag high. Scripps Institution of Oceanography is known for collecting large data sets of high quality for long periods of time that help us better understand the complex processes that shape our oceans and the land. We really have knocked the ball out of the park of creating datasets that the whole world uses to better understand the Earth.
Here, I can now take this, and what I've learned from firefighters, except for bad jokes, is that fire goes where water flows. Water flows down these canyons. Fires flow up them. Here, these bare earth models allow us to better understand fire behavior. This is one of the controls. Wind is another.
Vegetation health is another. We're trying to understand how all these work together to control fire behavior. Now, most of the time when we show Mount Whitney, we're looking west. When you look east, it doesn't look so recognizable, does it? Mount Whitney is right up there under like the mount. You can see it there in the pinnacle. We're looking to the west, and so here, excuse me, to the east and Mount Whitney is the northernmost peak.
But see all these valleys and conduits, the Caldor fire took advantage of these. Fires raced up the backside, the creek fire of the Sierra. The morphology is incredibly important. We can contour it, we can put elevation. We can put this in our fire behavior models.
We can also use it to determine what cameras can see in ignition because they might be blocked by a ridge line. It comes in very handy for that. Here, I can go in and identify each tree, and I can identify its morphology, and I can start doing forest analytics.
This is just breaking the trees into color by what type of tree it is. All of a sudden you can see the diversity here. Then we can go a step farther. We can now combine the LiDAR, which gives us the canopy, so 2015-2018, and we can merge it with the imagery data. Blue is good. We're doing
a normalized difference vegetation index. What this does is this allows us to look at how the fuels change during drought. The fuels, you can see them drying out in 2018, really dry fuels.
This is where car sparks can kick up a fire; lawn mower, and all of a sudden you're in a situation that the fuels are tinder dry and can really start getting a fire going. Here down the bottom here, you see the same thing. We had droughts earlier in 2016, but this immediately allows us to understand what parts of the forest and fuels need to be managed and what caution needs to be put in place. Here we can do biomass and carbon estimation. We're going to look at the altimetry of tree species, which is just a way of looking at how their morphology changes as they get older. We can put this with ground measurements and develop models to estimate forest production parameters.
We can use these in our models right away. I told you how the DEM helps us and helps us understand wildfire modeling. This here in areas like Montecito, where there were big landslides after fires and they got hit by atmospheric rivers, we can better understand areas that are susceptible to these landslides. This is the Oso landslide in Washington, I don't know if you remember this in 2014. The Oso slide over 420 people lost their lives in that slide.
The red here in this image shows where the material was vacated, and where sediment was deposited and you can see the river here. You can see that the river is filled with sediment. This here, then changes the oxygen content of the water and fish die offs. Here, can we save California's forests from going up in smoke? Here, this is the Mosquito fire, which was in Placer and Eldorado Counties. It was the largest fire in 2022. 76,000 acres, not a million acres, not 100,000 acres, but over here on the right, you'll notice where fire danger is lower and higher.
This area bounded by the yellow is lower. Prescribed burns modest management of the forest allowed the forest to here. Here's a pyrocumulus cloud from that fire. It just took off. This is from the R camera, and you can just see this pyrocumulus cloud just burst. Everything was so dry.
Here, Robert York, he's a forest ecology with UC Berkeley. He's using a drip torch to initiate prescribed burns in Blodgett in May. Some of his work has been very instrumental, and I just want to read this to you, "It's very clear that fire behavior changed and the treatments did accomplish that function of reducing fire severity, " York said. What it means is that on a bigger scale, in forest in California, if we can accomplish the level of fuel reduction and thinning that we did, then it's very likely that we can also accomplish reduced fire spread at that scale. Data is starting to emerge that's helping us understand forest management and prescribed burns.
I think this is what I was talking about with moving the needle. Here I want you to look at this area right here. This is the Oso Washington slide and it wasn't caused by fires it was caused by increased rainfall. But ALERTCalifornia is a multi-hazard platform. We can see these things.
Look at that. Is that crazy? Over 420 people lost their lives in this slide, huge. You look the sediment where it was is in red, that's where we have a deficit where it's blue.
We have deposition, and look at it choked up all the rivers. The rivers on this side have no sediment in them or less. This is the problem is the sediments released during these slides get into our riparian environments and we have fish die off. What have we learned from the deadly Oso fire that happened in 2014? Look at this area, there's not many burn scours, but we're loading the area with continuous rainfall.
It rained for a long period, never really hard, but long enough that it weighted the soils and gave slip planes. Slope failure is something our group studies because we want to understand and at least not repopulate, areas that are prone to these failures. This is the Mckinney fire that burned in Klamath in July and then in August.
They had big rainfall here and what you notice is, look at this, these mud slurries in the Whitney Creek and others went down into the Klamath River and it caused a huge fish die off. Here, killing off thousands of fish. Really important to the indigenous people as a protein source, the Klamath River for recreation. All of a sudden a company, the attendant cascading threats we're seeing here. We see this with the creek before Camp Creek, before and after the Caldor fire, where you have carcinogens and ash flowing into the rivers, flowing into our reservoirs. Here the Sierra have taken quite a hit and that's why we wanted to start there this wildland urban interface, the WUI.
You look here at the North complex, biggest fire, Dixie second, Caldor. This actually crossed the Sierra, first time. One of these fires crossed the Sierra. Some people say it even crossed back, but I don't know what they base that on. Here, the Sierra have taken quite a beating and you look high severity acres burnt. This is a problem and hopefully some of the data we collect can help drive decisions.
We can also use this because we mapped the Sierra. We got a bonus. We got all the dams, because all the dams of the rivers flow to the west, and the dams are on the Western side of the Sierra. This here is the Orville Dam. We just mapped it this year. You can see the giant spillway.
Remember, not too long ago that giant spillway was almost gone. That was quite a threat. The rock here is what we call shift and its strength is very low and this could have been a catastrophe.
Here the US Army Corps uses our data also. They use it to do flow models. This was like Lake Tulare flooding. Remember the flooding of Lake Tulare? We had no snow and then we had so much snow, we were flooding everything. The Delta, the rivers.
The Army Corps used this is light our data with elevation shown by color. Here, they were able to use, this is Lake Isabella and the Kern River coming out of it and they were able to use our data to say that it was really important to open what's called the Kern River inter-tie. Rarely do they open this. It allows the Kern River to flow into the aqueduct and it stopped water from flowing north into Lake Tulare.
This was a big decision. Implications all around but it lowered the rate of water flowing up to Lake Tulare. Given all this, we've formed alert California. Basically, it's prepare, respond, and recover and these fires are deadly.
We've had a couple of years of reprieve, but don't kid yourself, look at other areas around the world. These fires are going to be with us for a while. There he is. [LAUGHTER] I really show this slide because I want to call special attention to Falco Kuester.
He's the co-PI of this project and has made it even more successful. Just a call out to Falco. Falco shows this picture to show me that even Condors are better looking than you, Neal [LAUGHTER]. Here we're on high alert. This is coming out in the university publication next week.
But we have data walls, you can see them pointing at a fire right here that we just captured on this data wall. We're spinning the cameras. We're detecting the fire's smoke within minutes or less and then we're using artificial intelligence to knock down the noise and watch standard fatigue. All of these agencies are using our platform and our AI and we're collaborating with CAL FIRE.
What's really cool is CAL FIRE is testing our AI. We have the subject matter expert. CAL FIRE, the biggest fire department in the world with expertise telling the AI yes that smoke or no it's not. How can I collaborate with anybody better? Here all of these people are involved and they move the cameras they have protected passwords so they can chase fires.
Here we have over 1,054 sensors up. These blue are the sensors and the direction they point is the direction of the cameras looking that you can see here. Alabama Hills is one of my favorite cameras. We've built these giant data walls for CAL FIRE at south ops and north ops so they can look at all the data that we're collecting.
We're collecting petabytes of data. Now here, what we can see here is I can go in. As a user, you can go to opsalertcalifornia.org and this page will pop up with all the cameras. You can roll over any one of these arrows and the camera will come up here. You can see this was last moved a minute ago, two minutes ago, two minutes, two minutes.
We have these top four boxes as the ones most recently moved. You also see up here that we have active cameras. Those are the cameras that have actively been moved. Before I get there, so I can go to this panel and just type in any letters and all of a sudden all the cameras with that name pop up, so I don't have to search through all of these cameras. Let's go to Bob Mountain, Sequoia, one of my favorite sites.
Here we are, the lookout. We're 30-50 miles away from everyone and this is the direction the cameras looking. This is the view shed and I started moving the camera so it would go over here,. You can see some rainfall coming out of these clouds. This here, really proud of the guys for this. This is here a star link receiver.
We've actually programmed our system and our data so it links with star link and we can backhaul it from anywhere in the world. We don't have to have microwave, we don't have to have connectivity to WiFi. We can go straight to the stars and back to our emergency command center. This was one of the biggest improvements this year. This is again Brian and he's chasing the rainbow.
Here just showing you that here with Beaumont. I can go in and do time lapse. I can look back three hours.
Imagine you're looking at something that looks like smoke. You can go back and do a time lapse and go, no, that's a dirt devil or that's dirt kicked up by a farmer. It's not real smoke. Then I can blow it up and I can look at this in detail.
I'm in the middle of nowhere having data come back to me at over 100 megabits per second. Even your teenager would like that. [LAUGHTER] I can also view it so I can check all these. I can look at the Irwin new starts, I can look here at US counties and states.
I put on all the counties, you can see the cameras that have this color proboscis. That means they've been moved purple in the last five minutes, 15 minutes, all the way up to three hours. Now here's what you can do.
You can say, I'm going to look at what this guy's looking at. What made that firefighter move the camera? You can say, well first he's going to say, this is in my area. These are the Cal fire units. They're three letter designated units and they combine counties. But here, the 21 Emergency Command Centers of Cal Fire, this is their breakout. Here, you can see this was moved in a minute, a few minutes.
Here I can go here now and look at one of these cameras and say, what are they looking at? This is Bell Canyon North, you can see here. We're looking over this area here. We're to the Southeast of Santa Anna.
Person that moved the camera, looking in this direction, I can blow this up and say, I don't see any smoke. What are they looking at? Well, I can go on and do this. I can go and say I know what they're looking at.
They were looking at some of these pocket fires. I can put my address and drop a pin and I can know exactly where these fires are with respect to me. My recommendation is this, get on this program. Ops alert California.org.
Understand where your house is. Understand where the Irwin fires are, the ignitions. This gives you, empowers you to make decisions and have information real time. How hard is it to go on the web and find out anything about a fire real time? It's hard. This cuts it down. You can make your base map, you can put your house and then you can put on ignitions.
You can put on what cameras have moved. You can layer it so that you're looking at data real time. Here, so we're front page of the New York Times and the front page of the LA Times on the same day. We were rooting for the home team, but we didn't manipulate this, it just happened. The New York Times guy wasn't so happy. But what I'm going to tell you is this, this isn't what's important.
He missed the meaning of our discussions. If you beat 911 by a minute, but you have situational awareness, the cameras can show you the smoke. Is it bent over? How fast is the fire growing? How am I going to change my response? How is my firefighter, my incident commander, my dispatcher, how are they going to respond differently based on the data they see? Beating 911 is good. But it's not the be all end all.
It's situational awareness, actionable, real time data. How did your response change? I have friends in Kern County and other counties that say Neal, five years ago, would send one battalion because we had to have eyes on the fire so that we knew how to deploy and what source of resources. I can look at your camera and I can tell in seconds what I'm going to do if I'm going to scale up or scale down. What a difference. Here, this little box here is fire and this was here on Mount Laguna and it wouldn't have been seen by anybody but Chief Slumpff who was in the dispatch.
He saw the fire. He called the San Diego Unit of Cal Fire said. You need to check out this camera. They deployed more battalions and right away this would have burned all night. Could have gone from like a 10 acre fire to 1,000 acre fire and who knows with wind conditions and stuff. Here he keeps his eye on the screens and he's looking at the AI.
Here's one thing I want to tell you about is this. Here you're in the firehouse, you're not going to look at 1,054 cameras. You're just not going to do it. You're going to fatigue. But if I tell you what cameras saw change that have been rotating every two minutes, taking six frames per rotation, then all of a sudden you're going to look at that camera.
The AI tells you what camera to look at. Now this image from New York Times went viral and my email blew up. All my friends said, wow Neal, the New York Times is really on the ball. They got your best angle [LAUGHTER] Most graduate students said, wow, he's wearing rainbow flip flops.
There's the different perspective of Californians and New Yorkers. [LAUGHTER] Here, this shows you all the fire ignitions. The yellow fire here is the camera I'm looking at. It tells you exactly where it is. These are other cameras, but I'm not looking at them now so they're black. The flames are black.
Can everybody see that little bit of smoke there? Did you buy that? Let's take a look. What was something that was hard and you might overlook with a human eye, was an ignition, it was smoke. This is where the artificial intelligence really helps us. Now we can say, and put it in context, is that a rug burn? Is that a car fire or is that truly something we have to respond to? All of a sudden, the firefighters, they know that they have a tool that reduces noise, reduces fatigue, and tells them where to look. What camera to look at? Here I think this is one of the best lines I've ever heard.
This is staff chief of Fire Intelligence at Cal Fire. The success of this project is the fires you never hear about. They're extinguished before they have a name. Cal Fire is the first and only firefighting agency in the world to have such a system in place.
Academia, collaborating with a fire agency, I have the best, the brightest firefighters confirming whether it's smoke. What a relationship. We don't have to listen to that guy.
Here you can't get better expertise. I can't tell you how proud I am of this collaboration we have with Cal Fire on many levels. We also have talented engineers that fly drones, and this drone movie here was from the Montecito slide region. These are all data, so we're looking now at lighter data with imagery data draped on top of it.
You can see where the slides are, we can evaluate where we've had first old growth on this slope. We're now trying to characterize the morphology. Montecito had just finished this bridge that was five years after the initial debris flows. The bridge was six months old, they got 20 inches of rain last January, no more bridge. Try twice, not three times. We've got to make really important decisions.
Insurance companies are not going to insure us anymore. We're going to run into big problems if we don't start really showing that we're moving the needle. I think the management and pre planning, control burns is a real important tool that we should start employing. But look at this, it looks like a picture, these data we collected 2.5 billion points. We have 200 and 300 points per square meter.
We're able to see features that are centimeters in, all asthmath. We're now defining the Earth that the scale processes shape it. Here in peak slide, we had a rapid deployment, it looked like it was going to cause fatalities and we were able to help out Kern County.
You can see the failure and these ridge lines. We were really concerned, it was raining, it was muddy. If this had let go, this would have been a disaster, so we have a rapid response team.
That rapid response team is going to be heading out on a hard two weeks, they're going up to Paradise Craggy. Has anybody been up here? Anybody here been to Paradise Craggy? Well, this is what my guys are going to be doing this week, so while you're having coffee tomorrow morning and the next morning, they're going to be lugging cable up this peak to this lookout. They truly love me [LAUGHTER] and they know I'm too old to do it. [LAUGHTER] Other things, we're using FLIR, so this is visible light. This is the Dixie fire, you can see where the fire is spotted. This is going to make our firefighters safer while they keep us safe, so they don't get trapped behind the lines that we can see where the fire is.
We can do this with the hyperspectral data too, it maps out the fire perimeter, really important new data. These are new technologies that are emerging. But the problem is this, I've just caught up with bandwidth necessary for the PTZ cameras. These features generate an order of magnitude more data.
I'm never catching up. I'm always behind with bandwidth, with back haul. I don't have enough juice to make the car run around the track on my game.
I want to leave you close with this planes are sexy, dropping fire retardant, helicopters and water. We all gravitate towards these, but what a firefighter wants most, if you ask a firefighter, they want to know where they are with accuracy and they want to be able to communicate back to the ECC. This is the future right here, this is a Starlink receiver. They can get over 200 megabits per second, they can communicate where the fire is, understand from intel, where we have all the data coming in. Where is the perimeter, are you safe, where do you move to? This is something that every firefighter would tell you. If you ask them, and I've been at meetings, I say, what is it that we can do that will make your job safer? They don't say planes, they say communication and knowing where we are.
They think that's very important. Here, sorry, I broke up a little bit in the beginning, but it's still raw. Governor Newsom is very supportive and the state. California resources, the utilities, the firefighters, Cal fire, all the contract county fire. We're providing early confirmation of ignition, we're quantifying the fuel loads, which I think is really important.
I think this is where we can make advances. We've got to attack this problem. Crawling baby steps show results, have data drive decisions.
Constraints on fire behavior, defining forest health, and guiding forest management. Here I thank you all for coming, I want to see data drive decisions. I thank you very much for your time. [APPLAUSE] [APPLAUSE] Thank you very much for your presentation.
Did you do a similar presentation about four years ago, five years ago on the fire? I've done so many presentations. Well here, I've probably given more perspectives, talks than anybody else. I've talked about tectonics and formation of La Joya. I've talked about the offshore faults, the San Andreas, and why it hasn't ruptured in 275 years. But this talk is showing data.
We've never shown the data before or implications. Yes, I have talked on this before. I don't think I pirated too many slides, so that was time for you to take a cat nap.
[LAUGHTER] My real question is has other regions in the world asked for this type of technology in their regions? Like last year Canada was on fire and so was Europe. Yeah, Greece. Is this technology been asked for in other places in the Earth? I'm working with Hawaiian Electric right now, Canada, the EU, Spain.
Yes, people have taken notice worldwide of this technology and want to have systems themselves. What I say to them is this. I have to keep my focus on California during fire season.
When we emerge from and it's a lesser state of threat, then my teams can come and help you build out a network. But my focus is on the safety of California. [APPLAUSE] If you've got eyes on a fire, but you only have a single camera on it, how do you go about arranging the location of that fire from the camera? That's a good question. When I have 1,054 cameras, the AI tells me immediately what cameras can see that ignition based on the digital elevation model. If the cameras are blocked by ridges, the AI doesn't pick them up. Immediately the AI says, I think this is ignition, I think this is smoke, and these are the 10 cameras in and around that region that can potentially see it.
Here we can do that. We've come up with a really neat process where we have the digital elevation model. We can then project the camera on that. Then by matching the elevation and the camera, we can get a lat-long and we can push that out.
But we like to have triangulation and as many cameras as possible. Does that answer your question? I can't see you, I can't see if you're nodding your head. It has. Sorry. Thank you. [LAUGHTER] [OVERLAPPING] Next question. You mentioned that the firefighters say that the number one thing they want is better comms, but comms is also good for residents like ourselves.
I understand that in Maui, a lot of the residents didn't receive evacuation notices because the comms went down. I didn't see many comms on the list of partners. I saw you, PG&E, SDG&E, SCE. You mentioned WISP, the wireless Internet service providers.
Can you talk about what work is being done with landline and wireless providers to make sure residents get evacuation notices? If you had looked down that list a little further, we have probably about 30 WISPs that work with us, that develop communication either through microwave or fiber optic or WiFi. We've developed an approach and we've shared with them with the Starlink. My field team was instrumental in getting that up and working, and it wasn't easy and we're still fighting with it on certain mountaintops like Tobias.
We believe that network is the most important thing. We're working really hard to increase the bandwidth connectivity of the network because sensors will come and go. We want to have the bandwidth. As I told you, for some of these multi-spectral, some of the fleer forward-looking infrared cameras, we want to have the ability to bring that data back so we can interpret it and look at it. Now, petabytes of data are huge volumes of data and we use on the upper campus Falco's Lab, we have racks of servers storing this data and collecting it. We're working really hard to increase our connectivity.
I think that it's something that not just firefighters want, we want that increased connectivity. It's asymmetric too. Downloads are faster than uploads, and when we're in the field we're not downloading as much as we're uploading. That changes things too. We're working really hard on it and as I say, one of our biggest partners that helped develop a lot of the platform is Digital Path, a worldwide known WISP, GoLinks, Conifer Valley.
All of these WISPs are working with us to increase the bandwidth and the resilience of the signal. Does that get to your question? Yeah, thank you You're welcome. Hi, I'm glad you talked about data-driven decisions, how to close your talk. I love data. I looked at trends on two of your charts and I drew the opposite conclusion of what I think you were trying to say. The first was what we spend per year in California to fight fires. By the way, right now works out to about $25 per person per year, which to me sounds like a bargain.
If you were to take that chart and adjust by population and by inflation, I suspect that number per capita per year in constant dollars would be almost flat over the last 40 years. My second observation is you talked about acres burned in California going back the last 10 or 15 years. If we looked at acres burned in the United States, going back 100 years, in the 1930s, we had five times as many acres burned per year in the US as we have today.
To me, that means we've got a lot less fires and the trend line seems to be getting much better rather than worse. You can look at timescales and very different timescales. For the last 20 years, fires have been getting bigger from the Cedar Fire, which was 270,000 acres, through the Witch in 2007, through the Mendocino Complex.
We've never seen, we don't have data that shows us we have million acre fires. Okay. That data doesn't exist here in California. This is the first time with the August Complex Fire and the Dixie Fire that we had two fires that were a million acres.
I'd love to talk to you more, but we're talking two different timescales. I don't want to normalize cost of suppression by population because a lot of these fires that I showed you in the Sierra, the population isn't very high. Like Paradise, the Creek. These aren't burning in LA or San Francisco they had 1,906, they burned in the earthquake. I would say to you that I don't think scaling by population will yield as flat line as you think because these are fires in very low populated regions that are some of the largest. You look at in Siskiyou, huge fire.
Here, a lot of the biggest fires, severe fires, the slide I showed are in the Sierra, and the population in the Sierra is nowhere near as the urban areas. We've got a question. [OVERLAPPING] We want to talk to you more about the August data are fun. I'm just saying we can agree to disagree, but let's do it together. [LAUGHTER] We've got a young lady back here who has a question for you, Neal, back here in the corner close to the doors. Good. [LAUGHTER] How long did it take you to build this? Oh hang on My fault.
Over here. I'm just going to answer to the noise [LAUGHTER]. Okay, here you go. How long did it take to build this organization? This was a 10-year overnight success. [LAUGHTER] That's how old the questioner is. [LAUGHTER] Hopefully, we'll see you on my team.
[LAUGHTER] Mine's not actually a cognitive question. It's just more of a comment. Back in 1964, I was a firefighter for the United States Forest Service Station in Mount Laguna. I would have loved to have had this system back then because we used to have to go up to the highest point, sit on the tops of our trucks, and then if we see smoke pinpointed between two or three other fire towers to try to figure out where it was and unusually hike in to go suppress the fire before it got too large, but would have been nice to have a system like that back then. That was a long time ago. [APPLAUSE] Thank you so much for your service as a firefighter.
I'm going to say this, I think we're going to end on that really good question. [LAUGHTER] Thanks everybody and please come talk to us and oh, one thing that's really important is my team has kept this in a university environment. The University of California, San Diego. All of the data are open source. Grab some, and let's talk. [LAUGHTER] Here it's going to be stored in perpetuity, and you can have access to it.
People are using it for wildlife, they're using it for tree mortality, for snow retreat. These cameras are a step in the right direction of giving us actionable real time data, situational awareness. We have towers on Laguna that would have helped you out very much. Oh, yeah. Oh, yeah. [LAUGHTER] Think about it.
When you're doing it and triangulating and you're in the lookouts, it's very stressful because that depth is very hard to get. This depth, much easier. How you push that out as you know it's very difficult. Well, thank you so much, Neal Driscoll. [APPLAUSE] Thank you all for coming.
[APPLAUSE] [MUSIC]
2023-12-06 04:02
