PyroCb

Wildfire smoke: from Alaska to Norway, via the Arctic and Atlantic Oceans

Scott Bachmeier | July 18, 2015

Note: even though this large-scale smoke event was not pyroCb-related, we feel that it is important enough to document in terms of the long-range transport of biomass burning smoke and the potential implications on weather and climate far from the fire source regions.

Meteosat-10 0.8 µm visible channel images (click to play animation)

EUMETSAT Meteosat-10 High Resolution Visible (0.8 µm) images (above; click to play animation; also available as an MP4 movie file) revealed the hazy signature of what appeared to be a ribbon of smoke aloft being transported eastward across the North Atlantic Ocean by the circulation of a large area of low pressure (surface | 500 hPa) on 17 July 2015. Early in the day, the smoke feature stretched from the east coast of Greenland to the central Atlantic Ocean; by the end of the day, the leading edge of the smoke had moved over the British Isles and was headed toward Scandinavia.

A portion of the smoke plume could be seen on Aqua MODIS and Suomi NPP VIIRS true-color Red/Green/Blue (RGB) images (below) as it was approaching the southern portion of Great Britain.

Aqua MODIS and Suomi NPP VIIRS true-color images [click to enlarge]

On the following morning, Meteosat-10 visible images (below; click to play animation) showed that the leading edge of the smoke ribbon was moving over southern Norway.

Meteosat-10 0.8 µm visible channel images (click to play animation)

The transport pathway of this smoke feature was rather interesting, as we shall explore with the following sets of images.

Suomi NPP VIIIRS 3.74 µm shortwave IR and 0.64 µm visible images on 06 July [click to enlarge]

The 2015 wildfire season in Alaska had been very active — as of 17 July, it was rated as the 4th worst in terms of total acreage burned. In early July, numerous wildfires burning across the interior of Alaska were producing a large amount of smoke, as can be seen in a comparison of of Suomi NPP VIIRS 3.74 µm shortwave IR and 0.64 µm visible channel images at 2131 and 2312 UTC on 06 July (above). The thermal signature of the wildfire “hot spots” showed up as yellow to red to black pixels on the 2 shortwave IR images, while the widespread smoke plumes from the fires are evident on the 2 visible images; even in the relatively short 101 minutes separating the two sets of VIIRS images, notable changes in fire activity could be seen.

Looking a bit farther to the north and west, a sequence of VIIRS 0.64 µm visible images centered over Cape Lisburne (station identifier PALU) in northwestern Alaska covering a 2-day period from 06 to 08 July (below) showed the initial transport of large amounts of smoke from the interior of Alaska northwestward over the Chukchi Sea between Alaska and Russia.

Suomi NPP VIIRS 0.64 µm visible channel images covering the 06-08 July period [click to enlarge]

Daily composites of Suomi NPP OMPS Aerosol Index covering the period of 04-17 July (below; courtesy of Colin Seftor; see his OMPS Blog post) showed the strong signal of this dense Alaskan smoke (denoted by the red arrows) as it moved from east to west over the far southern Arctic Ocean and along the far northern coast of Russia from 06-10 July. The Aerosol Index signal seemed to stall north of Scandinavia on 12-13 July, but then a small portion began to move toward Iceland and Greenland on 13-15 July around the periphery of a large upper-level low (500 hPa analyses). Finally, some of this smoke was then transported eastward across the Atlantic Ocean around the southern periphery of this upper-level low on 17 July, as was seen on the Meteosat-10 visible images at the beginning of this blog post.

Suomi NPP OMPS Aerosol Index images, covering the period 04-17 July [click to enlarge]

CALIOP lidar data from the CALIPSO satellite (below) showed the vertical distribution of the Alaskan smoke over and off the coast of northern Norway on 11 July. The signal of the smoke was located in the center portion of the images; while there appeared to be some smoke at various altitudes within the middle to upper troposphere, a significant amount of smoke was seen in the lower stratosphere in the 10-12 km altitude range.

CALIPSO CALIOP lidar data showing the smoke over northern Norway on 11 July [click to enlarge]

Possible PyroCb in Southeast Russia

Anna Sienko | July 13, 2015

On 12 July there was a possibility for pyroCb development in southeast Russia. By using satellite imagery from the new Japanese satellite Himawari-8 it was confirmed that these clouds formed near the fires but were not technically pyroCb. Himawari-8 detected the smoke plume and clouds around the fires, as well as the fire hot spots. Starting at 04:00 UTC on 12 July, the animation below shows visible (.64 μm) on the left and shortwave IR (3.9 μm) on the right (click image to play animation). In the shortwave IR images the darker black to red pixels indicate very hot IR brightness temperatures exhibited by the fire source region.

Himawari-8 0.64 μm visible (left) and 3.9 μm shortwave IR (right) images (click to play animation)

In addition, using Himawari-8 10.4 μm IR channel imagery the minimum cloud-top IR brightness temperatures could be found. The animation below, starting at 04:00 UTC on 12 July, shows that the possbile pyroCb reached around -55ºC (red color enhancement) at 8:00 UTC.

Himawari-8 10.4 μm IR images (click to play animation)

OMPS AI index image (courtesy of Colin Seftor) shows the transport of smoke. The image on 13 July shows a high AI value south of the fires discussed above. This is consistent with the animations above of a southward transport of this smoke.

OMPS Aerosol Index image on 13 July (click to enlarge)

Also looking at the CO mixing ratio image on 15 July (below, courtesy of Leonid Yurganov) shows a very high mixing ratio near the area where the possible pyroCb took place. According to the image the area around the wildfires has a CO mixing ratio of 220 ppbv (indicated by a dark red color), which is a very high mixing ratio.

CO Mixing Ratio on 15 July (click to enlarge)

Looking from above these smoke plumes are quite magnificent. The images (courtesy of Klaus) were taken approximately at 54 N 127 E at 08:00 UTC on 14 July. From the images the smoke plumes to be extending far up into the troposphere, possibly even parts of the lower stratosphere.

Smoke plumes in Asia on 14 July (click to enlarge)

To further investigate the transport of smoke from this fire CALIPSO was used. This LIDAR shows the height of the clouds from the wildfire. The first image below is the 532nm Total Attenuated Backscatter plot on 12 July from 04:06 UTC to 04:19 UTC. The smoke from this fire can be seen around 52 N indicated by a light grey color. This plot shows that the smoke is moving northeast. The next image is the Depolarization image the smoke is indicated by a red/pink color. The third image is 1064 nm Total Attenuated Backscatter plot, the smoke on this plot is indicated by a light grey color. The fourth image is the Attenuated Ratio plot between 1064 nm and 532 nm. The smoke is indicated by the teal and purple pixels. The final image is the Vertical Feature Mask. This plot shows the different features that are in the atmosphere, the smoke is attributed as a cloud on this plot and is indicated by a light blue color.

CALIPSO 532 nm Total Attenuated Backscatter on 12 July (click to enlarge)

CALIPSO Depolarization Ration on 12 July (click to enlarge)

CALIPSO 1064 nm Total Attenuated Backscatter on 12 July (click to enlarge)

CALIPSO Attenuated Color Ratio between 1064 nm and 532 nm on 12 July (click to enlarge image)

CALIPSO Vertical Feature Mask on 12 July (click to enlarge image)

Smoke over North America

Anna Sienko | July 11, 2015

In Canada and Alaska, 2015 has been the year for wildfires. As of 08 July, the number of acres burned in Alaska has been the second highest in recorded history. In Canada the acres burned already exceeds the annual 10-year average for an entire year (image below). The government has deployed 1,000 military personnel to help fight wildfires in Saskatchewan (information and image provided by Wildfire Today).

Canada are burned on a weekly basis through week of July 1, 2015 (click to enlarge)

As of 08 July the United States has had 30,017 fires burning 3,821,726 acres, Alaska has had 650 fires burning 3,208,107 acres, and Canada has had 4,672 fires burning 6,546,562 acres. The images below show the fire activity for Canada and Alaska at 2:45 UTC on 09 July (courtesy of Wildfire Today).

Canada Fires at 02:45 UTC on 09 July (click to enlarge)

Alaska Fires at 02:45 UTC on 09 July (click to enlarge)

From these wildfires, massive amounts of smoke have been produced. This smokes has been covering parts of Canada and the United States for most of the summer. OMPS Aerosol Index (AI) helps show the transport of this smoke. Below is the image of the OMPS AI on 04 July (courtesy of Colin Seftor and information from Rene). On this image are three different areas with very high AI values. The first area is around the Great Lakes at 47.72 N 88.3 W with an AI value of 14.3 at 18:46 UTC. This cluster was produce by numerous fires in Saskatcehwan. This smoke could possibly be from the three pyroCbs that produced at 21:00 UTC on 03 July. Farther north the next area at 52.27 N 92.5 W had an AI value of 16.6 at 18:48 UTC. This cluster again is most likely from the pyroCbs in Saskatachewan. The last cluster to the west at 53.84 N 106.07 W produced an AI value of 11.1 at 18:48 UTC. This cluster came from the fires in northern Alberta.

OMPS Aerosol Index image on 04 July (click to enlarge)

On 06 July the OMPS AI index image (courtesy of Colin Seftor) provides information about the smoke transport since 04 July. There are two areas of very high AI values. The first that covers part of Washington and Montana is from a fire in British Columbia that produced a pyroCb on 05 July. The other are of very high AI value is in the center of Canada. This value is from the multiple fires in Saskatchewan, some even producing pyroCbs.

OMPS Aerosol Index image on 06 July (click to enlarge)

The OMPS AI index images (courtesy of Colin Seftor) on 07 July show the transport of smoke across Canada and into the North Pole. The first image shows high AI values from fires in Saskatchewan and British Columbia. There are high AI values in Montana and North Dakota from the fires in British Columbia. The second image shows high AI values from fires in Alaska. This smoke is not moving east like it usually does, but is moving northwest into the North Pole.

OMPS Aerosol Index image on 07 July (click to enlarge)

Furthermore, the northwestward drift of the smoke from Alaska can be seen using VIIRS visible imagery.

Suomi NPP VIIRS True-Color visible image of Alaska Smoke o 07 July (click to animate)

On 08 July the OMPS AI index image (courtesy of Colin Seftor) showed the AI index in the shape of a wishbone. These high values of AI came from fires in Alberta and Saskatchewan. Furthermore, Suomi NPP VIIRS true-color visible, shortwave IR, and IR images showed the smoke adding a light brown hue to the convective clouds. This visible image correlates nicely with the high AI values shown in the first image. Also, the visible image shows smoke plumes originating form a cluster of hot spots just east of the Yukon/Northwest Territories. In addition, from the IR image the coldest cloud-top brightness temperatures were between -59ºC to -60ºC.

OMPS Aerosol Index image on 08 July (click to enlarge)

Suomi NPP VIIRS visible, shortwave IR, and IR at 20:54 UTC on 08 July (click to animate)

By 09 July the max AI values were in the northern part of the United States. OMPS AI index images (courtesy of Colin Seftor) show a band of smoke extending the northern part of the United States. This smoke is coming from fires in British Columbia, possibly from a fire that produced a pyroCb a few days later.

OMPS Aerosol Index image on 09 July (click to enlarge)

This smoke has provided some interesting features on visible and water vapor imagery. The images below show visible and water vapor imagery on 09 July at 17:45 UTC (courtesy of Darren Clabo). These images show this smoke band extending from Montana to the middle of Minnesota.

Visible imagery on 09 July (click to enlarge)

Water vapor imagery on 09 July (click to enlarge)

Furthermore, this air seems to be trapped with a ribbon of very dry air associated with lowering of the dynamic tropopause. The reason the tropopause is lowering is due to a potential vorticity anomaly moving eastward across Minnesota. Furthermore, with this smoke being trapped in dry air it is common for it to cause stratospheric intrusions. By looking at the Total Column Ozone (shown in the image below) there are higher values in the ribbon of dry air. This is conclusive with the hypothesis that there are some stratospheric intrusions within this ribbon of dry air.

Total Column Ozone on 09 July (click to enlarge)

In addition, there are two vortices over Montana. The image below tweeted by the NWS office in Glasgow, MT shows the vortices. In addition, the GOES visible imagery below shows the two vortices on 10 July starting at 11:30 UTC.

Visible image over Montana on 10 July (click to enlarge)

GOES-15 visible imagery on 10 July (click to play animation)

On 10 July the OMPS AI index image (courtesy of Colin Seftor) showed the smoke that was in the northern part of the United States on 09 July moving east and off to the eastern coast.

OMPS Aerosol Index image on 10 July (click to enlarge)

Two PyroCbs in British Columbia

Anna Sienko | July 11, 2015

On 11 July there were a reported two pyroCbs in British Columbia. The first was at 56.4 N 123.9 W, producing a pyroCb around 00:30 UTC. GOES-15 detected the smoke plume and pyroCb cloud, as well as the fire hot spot. Starting at 22:30 UTC on 10 July, the animation below shows visible (.63 μm) on the left and shortwave IR (3.9 μm) on the right (click image to play animation). In the shortwave IR images the red pixels indicate very hot IR brightness temperatures exhibited by the fire source region.

GOES-15 0.63 μm visible (left) and 3.9 μm shortwave IR (right) images (click to play animation)

In addition, using GOES-15 10.7 μm IR channel the cloud-top IR brightness temperature could be found. The animation below, starting at 22:30 UTC on 10 July, shows the brightness temperatures of this pyroCb. The pyroCb reached -47.2ºC (lime green color enhancement) around 00:30 UTC on 11 July. From the Prince George sounding this puts the cloud top at 10.5 km.

GOES-15 10.7 μm IR images (click to play animation)

The second pyroCb occurred at 52.2 N 124 W around 05:00 UTC on 11 July. Since this pyroCb event happened during the evening hours visible imagery was not available. Again, GOES-15 detected the smoke plume (using 10.7 μm imagery) and pyroCb cloud, as well as the fire hot spot. Starting at 03:00 UTC on 11 July, the animation below shows IR imagery (10.7 μm) on the left and shortwave IR (3.9 μm) on the right (click image to play animation). In the shortwave IR images the red pixels indicate very hot IR brightness temperatures exhibited by the fire source region. Furthermore, in the 10.7 μm image the cloud-top brightness temperature was found to be -42.3 UTC around 05:00 UTC on 11 July. From the Kelowna sound this puts the cloud top height at 10.0 km.

GOES-15 10.7 μm visible (left) and 3.9 μm shortwave IR (right) images (click to play animation)

OMPS AI index image (courtesy of Colin Seftor) on 11 July shows the transport of smoke. The data from 11 July shows high AI values in central Canada. By using HYSPILT backward trajectory plot (below; courtesy of Rene) the smoke from the AI index image is from the pyroCb in British Columbia.

OMPS Aerosol Index image on 11 July (click to enlarge)

Backward Trajectory using HYSPLIT on 11 June (click to enlarge)

To further investigate the transport of smoke from this fire CALIPSO was used. This LIDAR shows the height of the clouds from the wildfire. The first image below is the 532nm Total Attenuated Backscatter plot on 11 July from 10:12 UTC to 10:24 UTC. The smoke from this fire can be seen around 61 N indicated by a light grey color. This plot shows that the smoke is moving northeast. The next image is the Depolarization image the smoke is indicated by a red/pink color. The third image is 1064 nm Total Attenuated Backscatter plot, the smoke on this plot is indicated by a light grey color. The fourth image is the Attenuated Ratio plot between 1064 nm and 532 nm. The smoke is indicated by the teal and purple pixels. The final image is the Vertical Feature Mask. This plot shows the different features that are in the atmosphere, the smoke is attributed as a cloud on this plot and is indicated by a light blue color.

CALIPSO 532 nm Total Attenuated Backscatter on 11 July (click to enlarge)

CALIPSO Depolarization Ration on 11 July (click to enlarge)

CALIPSO 1064 nm Total Attenuated Backscatter on 11 July (click to enlarge)

CALIPSO Attenuated Color Ratio between 1064 nm and 532 nm on 11 July (click to enlarge image)

CALIPSO Vertical Feature Mask on 11 July (click to enlarge image)

PyroCb in British Columbia

Anna Sienko | July 5, 2015

On 05 July there was possibility that a fire in southern British Columbia (located at 50º N, 123.5º W) produced a small pyroCb cloud; upon further analysis it was determined that this fire did indeed produce a pyroCb because the satellite-detected cloud-top IR brightness temperature reached the -40º C or colder threshold. GOES-15 3.9 µm shortwave IR images shown below detected the large “hot spot” signatures of these fires, as seen starting at 08:00 UTC on 05 July (click to play animation). The black to red pixels indicate very hot IR brightness temperatures exhibited by the fire source regions. Usually GOES visible imagery can be used to see the associated smoke plumes, but since the time of the pyroCb formation was during nighttime hours visible imagery was unavailable.

3.9 μm shortwave IR images (click to play animation)

In addition, using GOES-15 10.7 μm IR channel imagery the cloud-top IR brightness temperatures could be sampled; the animation below, starting at 08:00 UTC on 05 July, showed that the IR brightness temperature of this pyroCb cloud reached a minimum of -39ºC (lime green color enhancement) around 09:30 UTC.

GOES-15 10.7 μm IR images (click to play animation)

A side-by-side comparison of GOES-15 3.9 µm and 10.7 µm imagery (below; click to play animation) incorporates the 2 extra images per hour (at :11 and :41) of the “SUB-CONUS” sector, making it a bit easier to follow the initial formation and fast southeastward motion of the pyroCb cloud which originated from the southernmost of the 2 largest and hottest fires.

GOES-15 3.9 µm shortwave IR (left) and 10.7 µm IR (right) images (click to play animation)

In a comparison of Suomi NPP VIIRS 3.74 µm shortwave IR, 0.7 µm Day/Night Band (DNB), and 11.45 µm IR channel images at 1017 UTC (below), the actively burning fires (yellow to red to black pixels on the shortwave IR image) also exhibited a bright glow on the DNB image; in addition, due to ample illumination from a Full Moon, a large smoke cloud could be seen spreading southwestward away from the southernmost large fire (demonstrating the “visible image at night” capability of the DNB). At the time of the VIIRS images the pyroCb cloud feature had drifted about 100 miles south-southeastward away from the fire source region and was located near the US/Canada border — yet still had a minimum IR brightness temperature of -54ºC (darker orange color enhancement). This example helps to highlight the advantage of using polar-orbiting satellite imagery to help locate and track pyroCb features: due to much higher spatial resolution (375-meter with VIIRS, vs. 4-km with GOES) a colder and more accurate cloud-top IR brightness temperature could be detected, and due to minimal parallax viewing error the cloud feature could be more precisely located (VIIRS vs GOES-15 IR comparison).

Suomi NPP VIIRS 3.74 µm shortwave IR, 0.7 µm Day/Night Band, and 11.45 µm IR channel images (click to enlarge)

On 05 July OMPS Aerosol Index (AI) images were useful to see the transport of the smoke. The AI images (below; courtesy of Colin Seftor) confirmed that smoke from the British Columbia fires was being transported southeastward.

OMPS Aerosol Index image on 05 July (click to enlarge)

To further investigate the transport of smoke from this fire CALIPSO was used. This LIDAR shows the height of the clouds from the wildfire. The first image below is the 532nm Total Attenuated Backscatter plot on 05 July from 09:22 UTC to 09:36 UTC. The smoke from this fire can be seen around 52 N extending to about 48 N indicated by a white/light grey color. This plot shows that the smoke is moving southeast. Around 48 N the smoke increases in altitude as the LIDAR moves southeast insinuating that smoke is following that trajectory. The next image is the Depolarization image the smoke is indicated by a red/pink color. The third image is 1064 nm Total Attenuated Backscatter plot, the smoke on this plot is indicated by a light grey color. The fourth image is the Attenuated Ratio plot between 1064 nm and 532 nm. The smoke is indicated by the teal and purple pixels. The final image is the Vertical Feature Mask. This plot shows the different features that are in the atmosphere, the smoke is attributed as a cloud on this plot and is indicated by a light blue color.

CALIPSO 532 nm Total Attenuated Backscatter on 05 July (click to enlarge)

CALIPSO Depolarization Ration on 05 July (click to enlarge)

CALIPSO 1064 nm Total Attenuated Backscatter on 05 July (click to enlarge)

CALIPSO Attenuated Color Ratio between 1064 nm and 532 nm on 05 July (click to enlarge image)

CALIPSO Vertical Feature Mask on 05 July (click to enlarge image)

PyroCb in Alberta

Anna Sienko | July 4, 2015

On 03 July a pyroCb was produced at 23:30 in Alberta (58.2 N 118.3 W). GOES-15 detected the smoke plume and pyroCb cloud, as well as the fire hot spot. Starting at 22:00 UTC on 03 July, the animation below shows visible (.63 μm) on the left and shortwave IR (3.9 μm) on the right (click image to play animation). In the shortwave IR images the red pixels indicate very hot IR brightness temperatures exhibited by the fire source region.

GOES-15 0.63 μm visible (left) and 3.9 μm shortwave IR (right) images (click to play animation)

In addition, using GOES-15 10.7 μm IR channel the cloud-top IR brightness temperature could be found. The animation below, starting at 22:00 UTC on 03 July, shows the brightness temperatures of this pyroCb. The pyroCb reached -44.2ºC (lime green color enhancement) around 00:00 UTC on 4 July.

GOES-15 10.7 μm IR images (click to play animation)

To further investigate the transport of smoke from this fire CALIPSO was used. This LIDAR shows the height of the clouds from the wildfire. The first image below is the 532nm Total Attenuated Backscatter plot on 04 July from 19:46 UTC to 19:59 UTC. The smoke from this fire can be seen around 58 N indicated by a white/light grey color. In addition, the transport of smoke southeast is shown on the plot around 44 N. The next image is the Depolarization image the smoke is indicated by a red/pink color. The third image is 1064 nm Total Attenuated Backscatter plot, the smoke on this plot is indicated by a light grey color. The fourth image is the Attenuated Ratio plot between 1064 nm and 532 nm. The smoke is indicated by the teal and purple pixels. The fifth image is the Vertical Feature Mask. This plot shows the different features that are in the atmosphere, the smoke is attributed as a cloud on this plot and is indicated by a light blue color. The last image shows the subtype of the aerosols that have been detected by the LIDAR. This shows that the aerosols that the LIDAR has detected are smoke (indicated by the black pixels) around 58 N and southeastward around 44 N. This LIDAR data helped to conclude the direction of the transport of the smoke from this fire.

CALIPSO 532 nm Total Attenuated Backscatter on 04 July (click to enlarge)

CALIPSO Depolarization Ration on 04 July (click to enlarge)

CALIPSO 1064 nm Total Attenuated Backscatter on 04 July (click to enlarge)

CALIPSO Attenuated Color Ratio between 1064 nm and 532 nm on 04 July (click to enlarge image)

CALIPSO Vertical Feature Mask on 04 July (click to enlarge image)

CALIPSO Aerosol Subtype plot on 04 July (click to enlarge image)