Time is ripe for fire detection satellite, say Berkeley scientists(newscenter.berkeley.edu)
newscenter.berkeley.edu
Time is ripe for fire detection satellite, say Berkeley scientists
http://newscenter.berkeley.edu/2013/10/22/time-is-ripe-for-fire-detection-satellite/
4 comments
You'd need a lot of UAVs to cover the whole west coast.
I'm not sure why you'd want to restrict coverage to daylight hours; camping fires are more likely to be lit at night so early spotting makes sense. Likewise I'm not sure you'd want to restrict observation to the summer months, since dry winters are common here and winter wildires are not unknown (eg there was a large one in CO in March of 2011). I don't know whether climate change is a major factor in wildfire incidence year over year, but drier conditions are the sort of effect predicted by theory.
But my main objection is about the size of the observation area. When you take in the # of UAVs, the launch/landing facilities, the fuel costs, the air traffic control overhead, and the data correlation requirements, the satellite doesn't sound so expensive after all.
I'm not sure why you'd want to restrict coverage to daylight hours; camping fires are more likely to be lit at night so early spotting makes sense. Likewise I'm not sure you'd want to restrict observation to the summer months, since dry winters are common here and winter wildires are not unknown (eg there was a large one in CO in March of 2011). I don't know whether climate change is a major factor in wildfire incidence year over year, but drier conditions are the sort of effect predicted by theory.
But my main objection is about the size of the observation area. When you take in the # of UAVs, the launch/landing facilities, the fuel costs, the air traffic control overhead, and the data correlation requirements, the satellite doesn't sound so expensive after all.
If you check out the abstract at: http://www.mdpi.com/2072-4292/5/10/5173 you'll see that it mentions a temporal resolution of a few minutes or so, for the entire west coast. A fleet of drones might be cheaper, but they wouldn't be able to cover that kind of area that often.
I worked doing something similar with using weather satellites to detect volcanic eruptions in the north pacific. This is totally feasible in the here and now, whereas I imagine public acceptance of massive surveillance drone use is a bit far off.
I worked doing something similar with using weather satellites to detect volcanic eruptions in the north pacific. This is totally feasible in the here and now, whereas I imagine public acceptance of massive surveillance drone use is a bit far off.
The other issue is repsonse times. Many of the responders are volunteers and the areas are inaccessible. The notion that there will be a 15 minute response time to something like the rim-fire is quite naive.
Second, the experience of the SD (Cedar) fires which I believe are still the worst in the state of CA [1] this system would likewise be irrelevant. That fire was well known for some time, but other decisions and resource constraints (including legal/policy directives) forced the fire to be triaged in a secondary rank.
I hate to be cynical, but this project really seems like a solution and a grant-money funnel looking for a problem and a pile of cash. The "several hundred millon dollar" pricetag would be better spend on prevention, education, and support of additiona fire-fighting infrastructure. Including preventive maintenace for things like controlled burns, etc.
The notion that we need 200-400 million dollars flying overhead to fight fires caused by trivial mis-behaviour is also astonishing. These fires are often set by people doing things like spilling over an alcohol stove, setting an emergency signal, or doing some other trivial but mis-guided task.
That all being said, the amount of CA forest land that has burned throughout the state in the past 10 years is absolutely insane. I'd encourage people to actively work towards solutions to the general problem.
[1] http://www.fire.ca.gov/communications/downloads/fact_sheets/...
Second, the experience of the SD (Cedar) fires which I believe are still the worst in the state of CA [1] this system would likewise be irrelevant. That fire was well known for some time, but other decisions and resource constraints (including legal/policy directives) forced the fire to be triaged in a secondary rank.
I hate to be cynical, but this project really seems like a solution and a grant-money funnel looking for a problem and a pile of cash. The "several hundred millon dollar" pricetag would be better spend on prevention, education, and support of additiona fire-fighting infrastructure. Including preventive maintenace for things like controlled burns, etc.
The notion that we need 200-400 million dollars flying overhead to fight fires caused by trivial mis-behaviour is also astonishing. These fires are often set by people doing things like spilling over an alcohol stove, setting an emergency signal, or doing some other trivial but mis-guided task.
That all being said, the amount of CA forest land that has burned throughout the state in the past 10 years is absolutely insane. I'd encourage people to actively work towards solutions to the general problem.
[1] http://www.fire.ca.gov/communications/downloads/fact_sheets/...
We are probably close to the point of diminishing returns on education and prevention, given people's limited attention spans, the time dedicated to reading warning messages, already prominent notices when you go to campsites, and limitations on human intellect/awareness/abilities. "According to the Ad Council, Smokey Bear and his message are recognized by 95% of adults and 77% of children."[1]
The cost of fighting the Rim Fire was $127 million dollars[2], and that doesn't include some much smaller cost to repair the damages it caused (cited to be in the tens of millions by "experts"). The Cedar Fire cost $32 million in firefighting plus 15 lives and 2800 structures. The Zaca Fire cost $117 million to fight. From my perspective it's just a mathematical question whether it would be cheaper to put a satellite in the air vs. other mitigation strategies, and what the cost would be over the lifespan of the satellite for accelerated response vs. typical response. If we plug in the expected value of $200 million in additional funding to awareness I think we will be dissatisfied with the results.
1: https://en.wikipedia.org/wiki/Smokey_Bear
2: https://en.wikipedia.org/wiki/Rim_Fire
3: http://interwork.sdsu.edu/fire/resources/fire_facts.htm
4: https://en.wikipedia.org/wiki/Zaca_Fire
The cost of fighting the Rim Fire was $127 million dollars[2], and that doesn't include some much smaller cost to repair the damages it caused (cited to be in the tens of millions by "experts"). The Cedar Fire cost $32 million in firefighting plus 15 lives and 2800 structures. The Zaca Fire cost $117 million to fight. From my perspective it's just a mathematical question whether it would be cheaper to put a satellite in the air vs. other mitigation strategies, and what the cost would be over the lifespan of the satellite for accelerated response vs. typical response. If we plug in the expected value of $200 million in additional funding to awareness I think we will be dissatisfied with the results.
1: https://en.wikipedia.org/wiki/Smokey_Bear
2: https://en.wikipedia.org/wiki/Rim_Fire
3: http://interwork.sdsu.edu/fire/resources/fire_facts.htm
4: https://en.wikipedia.org/wiki/Zaca_Fire
I think the analytical approach is a good one, but I'm not sure the results are as simple as you imply, or have the result you may be led to believe. The damages of the Rim and Zaca fires were large, but the question is could the system actually prevent the ignition of the fire and decrease the costs of containment. In both of those cases, the answer to the former is <no> and to the latter is <unlikely>.
_________________
The Yosemite Rim Fire was man-made and had outsize damages because of inaccesibility and bad policy. Viz:
The fire was caused by a hunter's illegal fire that went out of control
The blaze was difficult to fight because of inaccessible terrain and erratic winds, forcing firefighters to be reactive
Also contributing to the fire was a pre-1980s policy of suppressing small natural fires. The lack of those fires created nearly a century's worth of fuel to burn, resulting in a massive forest fire killing virtually all plant life in its path.
A widespread heat wave and drought conditions helped to spread the fire and make it difficult to combat.
_________
It takes hours if not days to get men and supplies into some of these locations. On the flip side, the grow very, very quickly. Lets look at the Rim Fire:
Only 40 acres when it was discovered, it grew to 10,000 acres within 36 hours and 100,000 acres after four days.
The premise that a fire is like an ICBM launch that can be shot down before it explodes via star-wars is flawed. Its the sort of thing that people in Ivory towers believe to make themselves feel better. Which is fine, but its not clear at all that such a mindset correlates with reality.
Lets look at the Zaca Fire:
______________
The Zaca Fire was a wildfire which began burning northeast of Buellton, California, in Santa Barbara County, California. The fire started on July 4, 2007, and by August 31, it had burned over 240,207 acres (972.083 km2), making it California's second largest fire in recorded history after the Cedar Fire of 2003.[1][2] The fire was 100% contained on September 2, 2007.[3] It was declared controlled on October 29, 2007.[4]
The fire was started as a result of sparks from a grinding machine on private property which was being used to repair a water pipe.
______________
Again, man made fire. People doing stupid things. Entirely preventable. On the flip side, would the Bird save any money by early warning? Highly unlikely. There was no delay in getting at this other than it being called in. So, you would not save any material amount on this.
The order of magnitude of cost savings would appear to be perhaps 1/10 of the actual cost, not 100/100. So, in this case, perhaps you save $12MM on the Rim Fire ans $11MM on the Zaca fire. So, for these two events you save $20MM. At a cost of 200-400 million, you would need to incure 10x to 20x of these large scale fires. But this case would be devastating, if ever incurred. So using it as a rational backdrop seems off. If we are really expecting that, we should adopt strategies of prevention and fuel mitigation as much if not more-so than early warning.
_________________
The Yosemite Rim Fire was man-made and had outsize damages because of inaccesibility and bad policy. Viz:
The fire was caused by a hunter's illegal fire that went out of control
The blaze was difficult to fight because of inaccessible terrain and erratic winds, forcing firefighters to be reactive
Also contributing to the fire was a pre-1980s policy of suppressing small natural fires. The lack of those fires created nearly a century's worth of fuel to burn, resulting in a massive forest fire killing virtually all plant life in its path.
A widespread heat wave and drought conditions helped to spread the fire and make it difficult to combat.
_________
It takes hours if not days to get men and supplies into some of these locations. On the flip side, the grow very, very quickly. Lets look at the Rim Fire:
Only 40 acres when it was discovered, it grew to 10,000 acres within 36 hours and 100,000 acres after four days.
The premise that a fire is like an ICBM launch that can be shot down before it explodes via star-wars is flawed. Its the sort of thing that people in Ivory towers believe to make themselves feel better. Which is fine, but its not clear at all that such a mindset correlates with reality.
Lets look at the Zaca Fire:
______________
The Zaca Fire was a wildfire which began burning northeast of Buellton, California, in Santa Barbara County, California. The fire started on July 4, 2007, and by August 31, it had burned over 240,207 acres (972.083 km2), making it California's second largest fire in recorded history after the Cedar Fire of 2003.[1][2] The fire was 100% contained on September 2, 2007.[3] It was declared controlled on October 29, 2007.[4]
The fire was started as a result of sparks from a grinding machine on private property which was being used to repair a water pipe.
______________
Again, man made fire. People doing stupid things. Entirely preventable. On the flip side, would the Bird save any money by early warning? Highly unlikely. There was no delay in getting at this other than it being called in. So, you would not save any material amount on this.
The order of magnitude of cost savings would appear to be perhaps 1/10 of the actual cost, not 100/100. So, in this case, perhaps you save $12MM on the Rim Fire ans $11MM on the Zaca fire. So, for these two events you save $20MM. At a cost of 200-400 million, you would need to incure 10x to 20x of these large scale fires. But this case would be devastating, if ever incurred. So using it as a rational backdrop seems off. If we are really expecting that, we should adopt strategies of prevention and fuel mitigation as much if not more-so than early warning.
You've convinced me that the satellite would probably not in the end recoup it's cost. It makes intuitive sense that the most dangerous and costly fires are those close to civilization that can be reported quickly. I was mentally working with a probability of something like 35%, but it probably should have been something less than 5%, without running any actual analysis.
I wasn't trying to imply that the total figures were the likely savings, which is why I explicitly noted it as the cost of an accelerated response vs. a typical response, whatever that may be.
I wasn't trying to imply that the total figures were the likely savings, which is why I explicitly noted it as the cost of an accelerated response vs. a typical response, whatever that may be.
Your dismissal of an earlier response to the rim fire sort of depends on how long it took to get to 40 acres. I imagine it wasn't that long, but it matters.
Also, there are thousands of wildfires each year. There is plenty of opportunity to scrape payback out a little at a time, no need for dozens of major incidents.
(I don't have strong feelings about whether such a satellite makes sense, but with early detection capability, maybe it makes sense to increase the capability to drop suppressant on short notice, and so on.)
Also, there are thousands of wildfires each year. There is plenty of opportunity to scrape payback out a little at a time, no need for dozens of major incidents.
(I don't have strong feelings about whether such a satellite makes sense, but with early detection capability, maybe it makes sense to increase the capability to drop suppressant on short notice, and so on.)
Fair points, but I think the important takeaway is that $100mm per fire in <savings> is incredibly flawed. The flipside is if you save $1-2mm/per fire, you'd almost be encouraging fires (100-200?) in order to see a positive ROI.
As a practical matter, many of these fires cannot be put out once started, they can only be contained. (You don't just drive in a truck and shoot water out of a hose). The methods of containment are not fast, and they don't work all that well in areas like the Yosemite backcountry. There are other political issues involved. For example, the inability to use machinery in a Wilderness Area. Without taking a position on the logic of that restriction, consider the practical effects. Removal of fuel, lack of fire roads, walk-in only access, etc.
Empirically, its common for fires that start in very populated areas to still be incredibly damaging. Early warning in these areas does not seem it will materially alter the situation on the ground. Take for example the Powerhouse fire earlier this year in LA. [1] That was visible to the naked eye almost immediately. But the are is a dry, high-desert tinderbox. Ground zero was immediately next to the local San Francisquito <fire station>, yer it took 11 days to contain, and burned 30,000 acres despite its immediate alarm and proximity to roads and fire-fighting gear.
The Yosemite back-country is far more inaccessible and has far more fuel below tree level that the area around the powerhouse fire. I think the notion that once started it could have been contained in under 24 hours to be almost implausible. And once it reached 40 acres, we saw what happened next. Just as a sanity check, Assuming it was 20 miles in to start, it would take 2 days to walk to the area from the nearest road. You will not be dropping in smoke-jumpers like navy seals into this terrain in the middle of the night. Its steep, rugged, and when its not is densley forested. The nearest meadow may be in the next canyon.
In any event, it would be interesting to thing about what we would do with such a system if it were in place. Let's say that we got if for FREE. Then think about the next 200 million of infrastructre that would be required to actually put out the fires more quickly. My guess is that doing this logistical exercise would be more the bottleneck than the actual alert time. The prevelance of cell phones, sat-phones, epirb beacons, and Spot devices greatly increases the ability to comuunicate <SOS>. They do not make it very quick to get help, however. The benhcmark is 24-48 hours at a minimum if you are in a remote area, even with a Helo in many cases.
[1] http://www.theatlantic.com/infocus/2013/06/the-terrible-beau...
[2] The fire station is on the road in the middle of this image of the fire:
https://d2aoguav525b8o.cloudfront.net/sites/default/files/re...
Empirically, its common for fires that start in very populated areas to still be incredibly damaging. Early warning in these areas does not seem it will materially alter the situation on the ground. Take for example the Powerhouse fire earlier this year in LA. [1] That was visible to the naked eye almost immediately. But the are is a dry, high-desert tinderbox. Ground zero was immediately next to the local San Francisquito <fire station>, yer it took 11 days to contain, and burned 30,000 acres despite its immediate alarm and proximity to roads and fire-fighting gear.
The Yosemite back-country is far more inaccessible and has far more fuel below tree level that the area around the powerhouse fire. I think the notion that once started it could have been contained in under 24 hours to be almost implausible. And once it reached 40 acres, we saw what happened next. Just as a sanity check, Assuming it was 20 miles in to start, it would take 2 days to walk to the area from the nearest road. You will not be dropping in smoke-jumpers like navy seals into this terrain in the middle of the night. Its steep, rugged, and when its not is densley forested. The nearest meadow may be in the next canyon.
In any event, it would be interesting to thing about what we would do with such a system if it were in place. Let's say that we got if for FREE. Then think about the next 200 million of infrastructre that would be required to actually put out the fires more quickly. My guess is that doing this logistical exercise would be more the bottleneck than the actual alert time. The prevelance of cell phones, sat-phones, epirb beacons, and Spot devices greatly increases the ability to comuunicate <SOS>. They do not make it very quick to get help, however. The benhcmark is 24-48 hours at a minimum if you are in a remote area, even with a Helo in many cases.
[1] http://www.theatlantic.com/infocus/2013/06/the-terrible-beau...
[2] The fire station is on the road in the middle of this image of the fire:
https://d2aoguav525b8o.cloudfront.net/sites/default/files/re...
> during the summer months and daylight hours.
That's one advantage of a statellite: it's always looking.
Also, I suspect the UAVs would be more expensive than one might expect, once you factor in the manpower necessary to operate them (pilots and maintainence).
That's one advantage of a statellite: it's always looking.
Also, I suspect the UAVs would be more expensive than one might expect, once you factor in the manpower necessary to operate them (pilots and maintainence).
I would imagine that forest fires are less likely to be started in the winters and at night, unless they're caused by humans i.e. campfires.
I agree re:winters, but am not sure about nights. I imagine nighttime storms are not infrequent and so you could get a reasonable number of lightning strikes at night, causing fires.
Though, the UVA—satellite tradeoff also requires thinking about how soon you want to see a fire after it's been started. E.g., is 12 hours acceptable?
Though, the UVA—satellite tradeoff also requires thinking about how soon you want to see a fire after it's been started. E.g., is 12 hours acceptable?
I heard Kerry Emanuel (the climatologist) talk a few years back. He explained it like this: one can gather weather data with balloons and dropsondes, or with satellites. The first method is cheap and yields fine-grained data, the second one is sexy.
This. Cheap, solar powered, slow moving UAVs or something like that would be ideal for this.
Also, for those HNers that haven't learned about California forests:
Part of reason we have massive fires today is because in the early 1900s we didn't understand that forests in California depended on fire for regrowth. Loggers didn't like big fires burning down their precious trees, so we suppressed fire with great prejudice[1] - even though the Native Americans used low-intensity fire frequently to manage the forest and we knew that! [2]
So, today, we're still dealing with "insane fires" that we basically set up a century ago when we let much fuel accumulate in the forest. It's still going to take some time (see: decades) before all this excess fuel sorts itself out.
[1] http://ucanr.edu/sites/forestry/files/78651.pdf page 5 [2] ibid page 2
Also, for those HNers that haven't learned about California forests:
Part of reason we have massive fires today is because in the early 1900s we didn't understand that forests in California depended on fire for regrowth. Loggers didn't like big fires burning down their precious trees, so we suppressed fire with great prejudice[1] - even though the Native Americans used low-intensity fire frequently to manage the forest and we knew that! [2]
So, today, we're still dealing with "insane fires" that we basically set up a century ago when we let much fuel accumulate in the forest. It's still going to take some time (see: decades) before all this excess fuel sorts itself out.
[1] http://ucanr.edu/sites/forestry/files/78651.pdf page 5 [2] ibid page 2
Agree. At least with UAVs you can constantly upgrade the equipment to keep reducing costs further. You'd just need one solid fuselage and then you can keep improving the imaging sensors, engine and energy source.
Maybe if we declared a War Against Forest Fires, we can de-weaponize military UAVs and re-appropriate them for fire detection.
Maybe if we declared a War Against Forest Fires, we can de-weaponize military UAVs and re-appropriate them for fire detection.
UAVs are probably more likely to crash and cause a forest fire than a satellite.
Lol their acronym is FUEGO for the project (Spanish for fire). I swear these people think up an acronym first and work backwards...
A very similar system has been developed and deployed already (10 years ago), but not using a dedicated satellite: http://www.jpl.nasa.gov/news/news.php?release=2003-113
It's too bad the investigators are unaware of prior work.
It's too bad the investigators are unaware of prior work.
I had the same thought, but apparently the resolution is 'only' 1000 meters [1]. Additionally, it revolves the Earth in ~2 days - making early warning kinda hard.
As for similar sensor packages - some 'Satellite-Derived Sea Surface Temperatures' [2] are collected hourly and could get around the time issue. Resolution ranges from 1 to 8km. Then there are things up there from the NRO, DoD, etc which likely aren't as friendly with data sharing...
[1] http://en.wikipedia.org/wiki/MODIS
[2] http://www.opc.ncep.noaa.gov/sst/newSST/GOES_SST.shtml
As for similar sensor packages - some 'Satellite-Derived Sea Surface Temperatures' [2] are collected hourly and could get around the time issue. Resolution ranges from 1 to 8km. Then there are things up there from the NRO, DoD, etc which likely aren't as friendly with data sharing...
[1] http://en.wikipedia.org/wiki/MODIS
[2] http://www.opc.ncep.noaa.gov/sst/newSST/GOES_SST.shtml
Its resolution is lower, but OTOH, it is an operational system ;-)
http://sensorweb.nasa.gov/new/validationReport/IEEE-SMC-2005... (section 4)
http://sensorweb.nasa.gov/new/validationReport/IEEE-SMC-2005... (section 4)
Agreed.
I wish the article was more clear what advances the Berkeley folks have come up with. They claim a theoretical 3 meter resolution compared to the 'next-gen' sensor package (VIIRS [1]) at 750 meters.
[1] http://npp.gsfc.nasa.gov/viirs.html http://collaboration.cmc.ec.gc.ca/science/rpn/SEM/dossiers/2...
I wish the article was more clear what advances the Berkeley folks have come up with. They claim a theoretical 3 meter resolution compared to the 'next-gen' sensor package (VIIRS [1]) at 750 meters.
[1] http://npp.gsfc.nasa.gov/viirs.html http://collaboration.cmc.ec.gc.ca/science/rpn/SEM/dossiers/2...
Is the resolution 3m or can they detect a 3m x 3m fire? That kind of resolution seems unlikely from such a high orbit.
I had the same question. From "Table 1" of the linked article, they have 72m X 72m pixels, and they are trying to detect a 3m X 3m fire (burning at 1100K) within that pixel.
To bring up a related matter, it seems like they are not proposing to do the detection on-board (see section 2.2.2, where they talk about a database of past images). If they are not, they have to transfer all the data to ground. The data rate can be very demanding (and the high orbit makes it more so). Accurate pointing will be another issue (high orbit, small pixels). They have a longer way to go than the press release indicates.
Notwithstanding all this griping, it's an interesting concept.
To bring up a related matter, it seems like they are not proposing to do the detection on-board (see section 2.2.2, where they talk about a database of past images). If they are not, they have to transfer all the data to ground. The data rate can be very demanding (and the high orbit makes it more so). Accurate pointing will be another issue (high orbit, small pixels). They have a longer way to go than the press release indicates.
Notwithstanding all this griping, it's an interesting concept.
I rather doubt that. FTA:
“The point is, satellites like Landsat and GOES provide great information after a fire starts; they can focus and monitor a fire by looking at smoke plumes, fire spread, hot spots at the edges, etc.,” Kelly said. “FUEGO is designed for early detection of smaller fires. Right now, we lose a lot of time because fires are already big by the time we see them.”
“The point is, satellites like Landsat and GOES provide great information after a fire starts; they can focus and monitor a fire by looking at smoke plumes, fire spread, hot spots at the edges, etc.,” Kelly said. “FUEGO is designed for early detection of smaller fires. Right now, we lose a lot of time because fires are already big by the time we see them.”
Are not fires natural? Sequoia trees need fires to reproduce. Highly efficient fire-fighting methods creates timing bomb, as more and more material accumulates.
Just a thought, I could be totally off base here. I'm just not seeing the inherent advantage of a satellite, especially given the cost.