The Outback on Fire

I’m not talking about a Subaru. I’m talking about the vast expanse of sparsely-populated Australia. We’ve already seen fires in the United States, Russia and the Canary Islands. Well, they have been happening down under, too. (Is there any part of this planet not currently experiencing a drought?)

Despite the risk of getting fire fatigue (“Another post about fires?” *yawn*), we’re going to look at these fires for two reasons. First, it gives me a chance to show off the “fire tornado” video clip that has been making the rounds on the Internet:

Second, VIIRS saw the fire that produced the “fire tornado” (and a whole bunch of other fires) and it gives me a chance to show off the newly christened “Fire Temperature RGB”.

First, let’s look at the boring (yet still valuable) way of detecting fires: identifying hot spots in a 3.9 µm image. Here’s what VIIRS channel M-13 (4.0 µm) saw over Australia on 19 September 2012:

VIIRS channel M-13 image of central Australia, taken 04:34 UTC 19 September 2012

VIIRS channel M-13 image of central Australia, taken 04:34 UTC 19 September 2012

Pixels hotter than 350 K show up as black in this image. Given this information, how many fires can you see? (Hint: click on the image, then on the “3200×1536” link below the banner to see the image at full resolution. And, no, wise guy – you don’t count all the black pixels outside the boundaries of the data.)

Here’s the “pseudo-true color” RGB composite (this time made of M-05 [0.67 µm, blue], M-07 [0.87 µm, green], and M-10 [1.61 µm, red]):

False-color RGB composite of VIIRS channels M-05, M-07 and M-10, taken 04:34 UTC 19 September 2012

False-color RGB composite of VIIRS channels M-05, M-07 and M-10, taken 04:34 UTC 19 September 2012

With this RGB composite, really hot fires show up as bright red pixels. More hot spots are visible in the M-13 image than the “pseudo-true color” image because M-13 is much more sensitive to the heat from fires than M-05, M-07 and M-10 are. M-10 only picks up the signal from the hottest (or biggest) fires. M-05 and M-07 don’t pick up the heat signal at all, because the radiation from the sun, reflected off the Earth’s surface, drowns it out (which is precisely why the hot spots look red). M-13 is also better at detecting fires because it works at night, unlike these three channels.

You can make the hot spots from the smaller/less hot (lower brightness temperature) fires more visible by replacing M-10 with M-11 (2.25 µm) as the red channel in the RGB composite. M-11 is more sensitive to hot spots than M-10. If you do that, you get this image:

False-color RGB composite of VIIRS channels M-05, M-07 and M-11, taken 04:34 UTC 19 September 2012

False-color RGB composite of VIIRS channels M-05, M-07 and M-11, taken 04:34 UTC 19 September 2012

Since the previous RGB composite is often referred to as “natural color”, maybe this one should be called the “natural fire color” RGB composite. Now, most of the hot spots (not just the hottest ones) show up as red.

It should be noted that the fire complex in the grid box bounded by the 24 °S and 26 °S latitude and 128 °E and 132 °E longitude lines is where the video of the fire tornado came from. That fire is currently burning close to Uluru (a.k.a. Ayers Rock), the site where the creator beings live, according to local legend. According to an Uluru-Kata Tjuta National Park newsletter from back in July, prescribed burns were taking place in and around the park, although it’s not clear if the fires seen by VIIRS now (in September) are part of the prescribed burns.

EUMETSAT recently held a workshop on RGB satellite products, where a new RGB composite was proposed for VIIRS: the “Fire Temperature RGB”, made from M-10 (1.61 µm, blue), M-11 (2.25 µm, green) and M-12 (3.70 µm, red). Here’s what that looks like:

False-color RGB composite of VIIRS channels M-10, M-11 and M-12, taken 04:34 UTC 19 September 2012

False-color RGB composite of VIIRS channels M-10, M-11 and M-12, taken 04:34 UTC 19 September 2012

In this composite, hot spots from fires show up as yellow, orange, bright red or white, depending on how hot they are. Liquid clouds show up as light blue. Ice clouds, which are missing from this scene, typically show up as dark green. The background surface shows up as a shade of purple. Burn scars, which show up as dark brown in the “natural color” and “natural fire color” composites, show up as more of a maroon color in the “fire temperature” composite. Coincidently, maroon is the “official color” of Queensland, although it looks like most of the maroon burn scars show up in the Northern Territory.

To easily compare the different views of the fires (and make it obvious to everyone what the fires look like), here’s an animation, zoomed in on the lower left corner of each of the images above:

Animated loop of images of the fires in Australia as seen by VIIRS, 04:34 UTC 19 September 2012

Animated loop of images of the fires in Australia as seen by VIIRS, 04:34 UTC 19 September 2012

The yellow highlighted areas are where the active fires are.

Now that you’ve seen several different ways of displaying fire hot spots with VIIRS, which one do you like best?

VIIRS Captures a Glimpse of Hell

VIIRS has seen Hell and, luckily, it did not get scared. No, I’m not talking about Hell, Michigan, which is actually a nice place (and not as scary as their website would indicate). I’m talking about the Gates of Hell (or Door to Hell, depending on who you talk to) in Turkmenistan. You can see a single video of it here and, if that isn’t enough to get a sense of it, someone compiled a list of 296 videos of the Gates of Hell near Derweze/Darvaza, Turkmenistan.

Turkmenistan doesn’t have much – 80% of it is the Karakum Desert – but it does have a lot of oil and natural gas deposits. Back in 1971, the Soviet Union wanted to take advantage of these deposits, so they began drilling a gas well near the town of Derweze. Unfortunately, the drilling opened up a sinkhole that ate the drilling rig and caused the natural gas to leak out in large quantities. Oh, no! What to do now? Light it on fire!

The team of geologists thought that the best way to prevent the town from being suffocated by the toxic fumes was to ignite the gas, let it burn itself out in a few days, and return to see what the damage was. Guess what? That fire is still burning today – 41 years later!

This constantly burning crater is only 230 ft (70 m) across. So it may come as a surprise (to some people, at least) that VIIRS has no trouble seeing it. The highest-resolution channels on VIIRS have a spatial resolution of ~375 m at nadir. The fiery pit is so visible, the Day/Night Band (DNB), with ~740 m resolution, makes the Gates of Hell look like the biggest town in central Turkmenistan:

VIIRS Day/Night Band image of Turkmenistan, taken 22:26 UTC 13 September 2012

VIIRS Day/Night Band image of Turkmenistan, taken 22:26 UTC 13 September 2012

The red arrow points out the light source that is the Gates of Hell. One other thing to note from this image is all the lights in the Caspian Sea. Those are oil rigs, with the largest light source (the one closest to the center of the Caspian Sea) being the floating/sinking city of Neft Daşları (a.k.a Oily Rocks), which sounds like a pretty interesting/sad/weird place to work.

In case you think the lights are coming from the town of Derweze and not the actual Gates of Hell, here’s a zoomed in image from the DNB along with the M-12 (3.7 µm) brightness temperatures:

VIIRS Day/Night Band image of the Derweze "Gates of Hell", Turkmenistan, taken 22:26 UTC 13 September 2012

VIIRS Day/Night Band image of the Derweze "Gates of Hell", Turkmenistan, taken 22:26 UTC 13 September 2012

VIIRS channel I-04 image of the Derweze "Gates of Hell", Turkmenistan, taken 22:26 UTC 13 September 2012

VIIRS channel M-12 image of the Derweze "Gates of Hell", Turkmenistan, taken 22:26 UTC 13 September 2012. The color scale ranges from 210 K (white) to 300 K (black).

The Gates of Hell is the only light source that also shows up as a 345 K hot spot in channel M-12. Since this is a nighttime image, the signal in M-12 comes only from emission from the Earth (and clouds, etc.) without any contribution from solar reflection (as there would be during the day). What you see in the M-12 image is the temperature of the objects in the scene, just like a typical infrared (IR) satellite image, except with higher sensitivity to sub-pixel heat sources. The clouds show up as cold (bright, in this color table) above the warmer (darker) land surface. Sarygamysh Lake (and a few other smaller lakes) show up as really warm (dark) because the desert floor at night cools off much more than the water does.

The moon here was only ~10% full, so there wasn’t enough light reflecting off the few clouds in the scene for the DNB to detect them. In fact, with so little moonlight, everything is dark in the DNB. Everything, that is, except for the towns, villages and flaming craters of burning methane.

Fires in Paradise

Sometimes, it seems like the whole world is on fire. Siberia. The western United States (which has been burning for some time). And now, the Canary Islands. The Spanish islands have been under a drought, as has much of Spain. (As an indication of how dry it has been, one fire in mainland Spain was started by someone flicking a cigarette butt out of their car window in a traffic jam – a fire that ultimately led to two deaths.) Back in July, fires got started on Tenerife – a major resort destination – and earlier this month, fires began on La Palma and La Gomera. At least two firefighters have already died battling these fires.

For your reference, here is a VIIRS “true color” image (M-3 [0.488 µm], M-4 [0.555 µm], M-5 [0.672 µm]) of the Canary Islands, with the major islands labelled:

VIIRS true color RGB composite of channels M-3, M-4 and M-5, taken 14:01 UTC 5 August 2012

VIIRS true color RGB composite of channels M-3, M-4 and M-5, taken 14:01 UTC 5 August 2012

If you look closely at this image, from 5 August 2012, you can see smoke plumes coming off of La Palma and La Gomera. You can also see what looks like a von Kármán vortex street downwind of La Palma. That’s the west coast of Africa in the lower-right corner of the image.

As discussed previously, the true color RGB composite is better for viewing the smoke plume, but you can’t actually see the fire directly. So, here’s the M-5 (0.672 µm), M-7 (1.61 µm) and M-11 (2.25 µm) composite from the same time:

VIIRS RGB composite of channels M-5, M-7 and M-11, taken 14:01 UTC 5 August 2012

VIIRS RGB composite of channels M-5, M-7 and M-11, taken 14:01 UTC 5 August 2012

It’s easy to see where the fires are actively burning with this composite. Let’s zoom in to make it even more obvious:

VIIRS false color RGB composite of channels M-5, M-7 and M-11, taken 14:01 UTC 5 August 2012

VIIRS false color RGB composite of channels M-5, M-7 and M-11, taken 14:01 UTC 5 August 2012

All the bright red pixels indicate where the fire is actively burning. You can also see the burn scar on Tenerife (not as easily as in Siberia) where the M-5, M-7, M-11 RGB composite shows the fire was back in July:

VIIRS false color RGB composite of  channels M-5, M-7 and M-11, taken 14:38 UTC 18 July 2012

VIIRS false color RGB composite of channels M-5, M-7 and M-11, taken 14:38 UTC 18 July 2012

La Gomera has been the hardest hit island, where thousands of people had to be evacuated, and approximately 10% of Garajonay National Park has burned. Garajonay National Park is home to one of the last remaining laurisilva forests, which has been around for 11 million years. That lush vegetation burned hot, and channel I-04 (3.7 µm) reached saturation as that area went up in flames:

VIIRS channel I-04 image of fires in the Canary Islands, taken 14:01 UTC 5 August 2012

VIIRS channel I-04 image of fires in the Canary Islands, taken 14:01 UTC 5 August 2012

The two white pixels on La Gomera are where I-04 reached saturation and “fold-over” due to the heat from the fire. M-13 (4.0 µm), which is a dual-gain band designed to not saturate, reached a brightness temperature of 451 K over La Gomera, compared with a saturation brightness temperature of 367 K for channel I-04.

The fires also showed up in the Day/Night Band that night:

VIIRS Day/Night Band image of the Canary Islands, taken 02:25 UTC 6 August 2012

VIIRS Day/Night Band image of the Canary Islands, taken 02:25 UTC 6 August 2012

The red arrows point out the fires on La Palma and La Gomera. The fire on La Gomera covers a significant percentage of the island. The yellow arrow points to Lanzarote, which, for some reason, is not part of IDL’s map. On the night this image was taken, the moon was approximately 84% full, so you can see a number of clouds as well the city lights from the major resort areas of the Canary Islands. The biggest visible city in Africa is El Aaiún, the disputed capital of Western Sahara.

Finally, here’s the “pseudo-true color” composite of VIIRS channels I-01 (0.64 µm), I-02 (0.87 µm) and I-03 (1.61 µm) from 13:42 UTC 6 August 2012. This is a full granule at the native resolution of the Imagery bands with no re-mapping, showing the rich detail of VIIRS high-resolution imagery, including more interesting cloud vortices:

VIIRS false color RGB composite of channels I-01, I-02 and I-03, taken 13:42 UTC 6 August 2012

VIIRS false color RGB composite of channels I-01, I-02 and I-03, taken 13:42 UTC 6 August 2012

Make sure to click on the image, then on the “6400×1536” link to see it in its full glory.

Fires near the “Coldest City on Earth”

Raise your hand if you’ve only ever heard of Yakutsk because of the board game “Risk”. (If you raised your hand, you might want to look around and make sure that no-one saw you raise your hand for no reason.)  Yakutsk is actually the capital city of the Sakha Republic (a.k.a. Yakutia), which, according to Wikipedia, is the largest sub-national governing body in the world (only slightly smaller than India in terms of land area). Over 260,000 people live in Yakutsk, which has been called the “Coldest City on Earth” (with 950,000 total in Yakutia) even though, according to this article, it doesn’t sound very pleasant in the winter (or summer, for that matter). In January, the average temperature is -42 °C (-45 °F), and it isn’t very far from Oymyakon, where the lowest temperature ever recorded in a permanently inhabited location was observed (-71.2 °C or -96.2 °F). In the summer, it can make it up to +35 °C (95 °F) and legends tell of reindeer dying from choking on all the insects that cloud the air.

This summer, large areas of Siberia (including Yakutia) have been on fire. Some pictures from MODIS have already been circulating around the internet (e.g. here and here). And someone beat me to posting VIIRS images already. To make it easier to judge the size of the fires that are visible in the VIIRS Day/Night Band (DNB) image in the last link, here is a close-up with latitude and longitude lines added:

VIIRS DNB image of fires in Siberia, taken 16:25 UTC 4 August 2012

VIIRS DNB image of fires in Siberia, taken 16:25 UTC 4 August 2012

At this latitude, longitude lines are ~55 km apart. The latitude lines are ~111 km apart. So, you can see that these fires cover quite a large area. Unfortunately, you can’t see Yakutsk, which is underneath the clouds (and possibly smoke) at about 62° N, 130° E.

For comparison, here is the M-13 (4.05 µm) image from the same time. The primary purpose of M-13 is to detect wildfires. Notice how all of the hot spots (black spots) line up with all of the light sources that the DNB saw:

VIIRS channel M-13 brightness temperature image taken 16:25 UTC 4 August 2012

VIIRS channel M-13 brightness temperature image taken 16:25 UTC 4 August 2012

The visible image from earlier that day showed just how much smoke was produced by all of these fires:

Visible image of fires in Siberia from VIIRS channel M-5, taken 02:38 UTC 4 August 2012

Visible image of fires in Siberia from VIIRS channel M-5, taken 02:38 UTC 4 August 2012

Except for a few clouds near the edges of the scene, that is pretty much all smoke.

A few days later, the burn areas were easily visible with many fires still active, although not producing nearly as much smoke. RGB composites can really highlight what is going on with these fires, so let’s look at a few.

You should already be familiar with the “true color” image (M-3, 0.488 µm [blue], M-4, 0.555 µm [green] and M-5, 0.672 µm [red]):

True color image from VIIRS channels M3, M4 and M5 of fires in Siberia, taken 03:22 UTC 7 August 2012

True color image from VIIRS channels M3, M4 and M5 of fires in Siberia, taken 03:22 UTC 7 August 2012

And the “pseudo-true color” image made by combining the first three I-bands (I-01, 0.64 µm [blue], I-02, 0.865 µm [green] and I-03, 1.61 µm [red]):

False color (or "pseudo-true color") image of fires in Siberia from VIIRS channels I-01, I-02 and I03, taken 03:22 UTC 7 August 2012

False color (or "pseudo-true color") image of fires in Siberia from VIIRS channels I-01, I-02 and I03, taken 03:22 UTC 7 August 2012

The “pseudo-true color” image may be referred to as “natural color” depending on who you talk to. It should be noted that these last two images were kept at the native resolution of VIIRS with no re-mapping or re-sizing the image. There is only cropping to keep the file sizes manageable.

As discussed before, the pseudo-true color composite has the advantage of easily distinguishing ice and snow from liquid clouds, and it is really sensitive to vegetation. Plus, scattering by molecules in the atmosphere is greatly reduced, so you don’t have to do any atmospheric correction to produce a nice image. There is also the advantage that it uses I-bands, which have twice the resolution of the M-bands. But, that advantage was almost always neutralized by the fact that the images would have to be compressed to create a reasonable file size so that it would fit on this blog. If you click on the images above, then on the full-resolution link below the banner, you can easily compare the true resolution between the M-band image and the I-band image.

You can see here that the burn scars (all the dark brown areas) show up really well in the pseudo-true color image. (Some of the lighter or reddish brown areas are mountain ranges.) You might also notice that the active fires are still producing smoke, which shows up a lot better in the true color image. Some of the burn scars cover an area close to 60 km across.

As luck would have it (or, more accurately, the planning ahead by the scientists and engineers who designed VIIRS), channels M-5 (0.672 µm), M-7 (0.865 µm) and M-10 (1.61 µm) are very similar to the first three I-bands, so we can easily produce an M-band “pseudo-true color” image:

"Pseudo-true color" composite of VIIRS channels M-5, M-7 and M-10 of fires in Siberia, taken 03:22 UTC 7 August 2012

"Pseudo-true color" composite of VIIRS channels M-5, M-7 and M-10 of fires in Siberia, taken 03:22 UTC 7 August 2012

For reference, the location of Yakutsk has been identified. Also, if you’re curious, the big river that curves from the left-middle of the image to the top-center is the Lena River. It is up to 10 km wide in parts, particularly north of Yakutsk. Its second largest tributary, the Aldan River, is also easily visible as it meanders through a lot of the burn areas.

If you replace M-10 with M-11 (2.25 µm) as the red channel, you get this image:

False color RGB composite of VIIRS channels M-5, M-7 and M-11, taken 03:22 UTC 7 August 2012

False color RGB composite of VIIRS channels M-5, M-7 and M-11, taken 03:22 UTC 7 August 2012

Here, the green is darker due to the lower reflectivity of the surface in M-11 compared with M-10. The advantage of this RGB composite it that, if you zoom in, you can actually see where the fires are still active, as those pixels show up bright red. (If the fire is hot enough, you’ll get red pixels in the “pseudo-true color” composite also, but M-11 is more responsive to heat from fires than M-10, so you can see lower temperature fires this way.) You can also see the faint bluish smoke plumes originating from the areas that are actively burning.

If you go in the other direction and use only the shortest wavelengths, the surface becomes difficult to see, but the smoke stands out more. Here is the RGB composite of M-1 (0.412 µm [blue]), M-2 (0.445 µm [green]) and M-3 (0.488 µm [red]):

False color RGB composite of VIIRS channels M-1, M-2 and M-3, taken 03:22 UTC 7 August 2012

False color RGB composite of VIIRS channels M-1, M-2 and M-3, taken 03:22 UTC 7 August 2012

Here, the wavelengths of these channels range from the violet to the blue portion of the visible spectrum. At these shorter wavelengths, scattering in the atmosphere becomes much more important and the solar radiation has a tough time making it all the way to the surface. All the smoke and haze increases the scattering, so it is difficult to pick out features on the surface. That same scattering, though, really highlights the smoke plumes, which are difficult to see in the other false color composites.  Since the scattering by the stuff in this image doesn’t vary much between these three channels, you get an image without much color to it.

With much of Colorado and, really, much of the western U.S. having burned already this year, it’s easy to know what the people of Siberia are going through. Fortunately, none of the fires have really threatened any towns. And, another plus: I bet those clouds of mosquitoes don’t like the dry weather that has caused all of these fires.

Wild Week of Wildfires, Part II

Last time on “Wild Week of Wildfires“, we looked at the Little Bear Fire and High Park Fire, two lightning-ignited fires burning out west that were so hot they caused saturation in the two 3.7 µm channels on VIIRS (I-04 and M-12). There was mention of the Duck Lake Fire, a lightning-ignited fire in northern Michigan, which VIIRS also saw, and I couldn’t resist showing some more images.

On 9 June 2012, the same day the High Park Fire exploded (figuratively speaking), the Duck Lake Fire finally reached 100% containment after burning over 21,000 acres. The next day (10 June 2012), Suomi NPP passed over the Upper Peninsula of Michigan, and it was actually a clear day. (This joke comes courtesy of 20+ years experience of living in Michigan.) Even with 100% containment, the hot spot of the fire was still clearly visible in VIIRS channel I-04 (3.7 µm) that afternoon:

Channel I-04 image of the Duck Lake Fire from VIIRS, taken 18:18 UTC 10 June 2012

Channel I-04 image of the Duck Lake Fire from VIIRS, taken 18:18 UTC 10 June 2012

The highest brightness temperature in the burn area in this channel at this time was    ~331 K. As we saw before with the Lower North Fork Fire, the high resolution false color composite of channels I-01, I-02 and I-03 is useful in highlighting the burn area:

False color RGB composite of VIIRS channels I-01 (blue), I-02 (green) and I-03 (red), taken 18:18 UTC 10 June 2012

False color RGB composite of VIIRS channels I-01 (blue), I-02 (green) and I-03 (red), taken 18:18 UTC 10 June 2012

Notice the large, brown area that coincides with the hot spot in the I-04 image. The combination of wavelengths used in this composite (0.64 µm [blue], 0.865 µm [green] and 1.61 µm [red]) is quite sensitive to the amount (and health) of the vegetation.

You might have also noticed several other interesting features in the image that show up better when you zoom in:

False color composite of VIIRS channels I-01, I-02, and I-03 from 18:18 UTC 10 June 2012

False color composite of VIIRS channels I-01, I-02, and I-03 from 18:18 UTC 10 June 2012

The Upper Peninsula of Michigan was based on mining for most of its history, and several large mines and quarries still exist, which VIIRS can easily see.

If you didn’t know any better, you might confuse the iron mine southwest of Marquette, Michigan with a frozen lake, or miraculously un-melted snow leftover from winter, since that is just what snow and ice look like in this kind of RGB composite. Compare that with the true color view of the same area:

True color RGB composite of VIIRS channels M-3, M-4 and M-5, taken 18:18 UTC 10 June 2012

True color RGB composite of VIIRS channels M-3, M-4 and M-5, taken 18:18 UTC 10 June 2012

In this case, the iron mine stands out as a bright red. Why?

The true color composite uses wavelengths at 0.48 µm (blue), 0.55 µm (green) and 0.67 µm (red). The red channel in the true color composite is actually in the red portion of the visible spectrum. The blue channel in the false color composite (0.64 µm) is also in the red portion of the visible spectrum.

This example shows that the iron oxide (rust) produced at the iron mine is highly reflective in the red portion of the visible spectrum. That’s what gives it the characteristic rust color. Iron oxide is not nearly as reflective at shorter or longer wavelengths, so it shows up blue when red wavelengths are used as the blue channel (as in the false color composite) and red when they are used as the red channel (as in the true color composite).

Let this be a lesson to anyone who uses the false color composite as part of a snow and ice detection algorithm. Snow and ice are not the only things to show up that color. You may be looking at a really large iron mine.