Sir Ivan Fire pyroCumulonimbus in New South Wales, Australia

February 12th, 2017

Himawari-8 0.64 µm Visible (top), 3.9 µm Shortwave Infrared (middle) and 10.4 µm Longwave Infrared Window (bottom) images [click to play animation]

Himawari-8 0.64 µm Visible (top), 3.9 µm Shortwave Infrared (middle) and 10.4 µm Longwave Infrared Window (bottom) images [click to play animation]

Himawari-8 Visible (0.64 µm), Shortwave Infrared (3.9 µm) and Longwave Infrared Window (10.4 µm) images (above / MP4 ; zoomed-in over fire source region: GIF / MP4) showed wildfires burning in New South Wales, Australia on 12 February 2017. The larger Sir Ivan Fire near Dunedoo produced a pyroCumulonimbus (pyroCb) cloud, which first cooled below the -40ºC Longwave Infrared brightness temperature “pyroCb threshold” at 0530 UTC (-47ºC) and quickly reached its minimum temperature of -56.6ºC at 0540 UTC. An animation of Himawari-8 true-color images is available here (courtesey of Dan Lindsey, RAMMB/CIRA).

Consecutive true-color images from Suomi NPP VIIRS (0402 UTC) and Aqua MODIS (0405 UTC) viewed using RealEarth (below) showed the large smoke plume about 1.5 hours prior to pyroCb development.

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

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

A high fire danger was well-anticipated across this portion of Australia:

Some ground-based photos of the pyroCb cloud:


Oil well fire in Utah

January 6th, 2017

GOES-15 Visible (0.63 µm) images, with hourly surface reports [click to play animation]

GOES-15 Visible (0.63 µm) images, with hourly surface reports [click to play animation]

GOES-15 (GOES-West) Visible (0.63 µm) images (above) showed a small, short-lived black cloud that formed south/southwest of Vernal (station identifier KVEL) in northeastern Utah on 06 January 2017. This feature was the result of a fire at an oil well site (media report | well location) that apparently started around 11:30 am local time (1830 UTC); the black cloud from the burning oil tanks — which was first apparent on the 1930 UTC visible image — stood out well against the snow-covered ground. The initial northwestward transport of the smoke plume was consistent with lower-tropospheric winds in Grand Junction, Colorado rawinsonde data at 07 January/00 UTC, which showed southeasterly winds as high as 784 hPa (2185 meters or 7169 feet above ground level). The sounding profile also showed that this height was the top of a well-defined temperature inversion, which acted as a cap to prevent the smoke from reaching higher altitudes (photo).

GOES-13 (GOES-East) Visible (0.63 µm) images (below) also displayed the dark smoke plume. The viewing angles from the 2 satellites were similar (~53 degrees from GOES-15 vs ~57 degrees from GOES-13), but the time sampling was slightly better from GOES-15 (due to the extra “SUB-CONUS” scan images at :11 and :41 minutes nearly every hour). Image frequency will be even better with the GOES-R series of satellites (beginning with GOES-16), with routine scans every 5 minutes; the visible image spatial resolution will also be improved (to 0.5 km, vs 1.0 km with the current GOES).

GOES-13 Visible (0.63 µm) images, with hourly surface reports [click to play animation]

GOES-13 Visible (0.63 µm) images, with hourly surface reports [click to play animation]

MODIS Visible (0.645 µm), Shortwave Infrared (3.7 µm) and Infrared Window (11.0 µm) images from a 2036 UTC overpass of the Aqua satellite (below) showed the black smoke cloud in the Visible, but there was no evidence of a fire “hot spot” in the Shortwave Infrared (the media report indicated that the fire was extinguished about 2 hours after it started, which would have been around or just before the time of the MODIS images). On the Infrared Window image, the smoke plume actually did exhibit a slightly colder (darker blue color enhancement) signature, which is unusual since conventional fire and wildfire smoke is normally transparent to thermal radiation.

Aqua MODIS Visible (0.645 µm) and Shortwave Infrared (3.7 µm) images at 2036 UTC [click to enlarge]

Aqua MODIS Visible (0.645 µm) and Shortwave Infrared (3.7 µm) images at 2036 UTC [click to enlarge]

A view of the 250-meter resolution Aqua MODIS true-color Red/Green/Blue (RGB) image from the MODIS Today site is shown below.

Aqua MODIS true-color image at 2036 UTC [click to enlarge]

Aqua MODIS true-color image at 2036 UTC [click to enlarge]

Gatlinburg, Tennessee wildfire

November 29th, 2016

GOES-13 Shortwave Infrared (3.9 µm) images, with METAR surface reports [click to play animation]

GOES-13 Shortwave Infrared (3.9 µm) images, with METAR surface reports [click to play animation]

Wildfires had been burning in the Great Smoky Mountains for a few weeks (see previous blog posts) as extreme to exceptional drought persisted over the region. However, on 28 November 2016 weather conditions became conducive to extreme fire behavior — and this allowed the Chimney Tops 2 Fire south of Gatlinburg, Tennessee to race rapidly northward (fire perimeter map), driven by strong southerly winds gusting to at least 30-40 knots (as were recorded in Knoxville KTYS, located about 25 miles northwest of Gatlinburg). Widespread evacuations were necessary, and at least 13 fatalities were reported.  4-km resolution GOES-13 Shortwave Infrared (3.9 µm) images (above) showed the development of a fire “hot spot” (the cluster of pixels at the center of the images exhibiting a black to yellow color enhancement) during the day, before clouds moved overhead to mask the fire hot spot signature. The warmest infrared brightness temperature seen during this time period was 326.8 K (brighter yellow pixels) on the 1700 UTC image.

Even though cloud cover was increasing, a detailed view of the fire hot spot was provided by an AWIPS II image of 375-meter resolution Suomi NPP VIIRS Shortwave Infrared (3.74 µm) data at 1815 UTC on 28 November (below). An AWIPS I version of this image is available here. Due to the cloudiness, no discernible hot spot appeared on the lower-resolution 1815 UTC GOES-13 Shortwave Infrared image.

Suomi NPP VIIRS Shortwave Infrared (3.74 µm) image [click to enlarge]

Suomi NPP VIIRS Shortwave Infrared (3.74 µm) image [click to enlarge]

Props to NWS meteorologist Carl Jones for spotting this somewhat unexpected result: the glow of the fire was evident on the following nighttime Suomi NPP VIIRS Day/Night Band (0.7 µm) image, even though there was a thick layer of clouds over the fire itself:


An AWIPS II image comparison of VIIRS Infrared Window (11.45 µm), Shortwave Infrared (3.74 µm) and Day/Night Band (0.7 µm) data at 0816 UTC on 29 November is shown below. Cloud-top Infrared Window brightness temperatures were in the -40 to -55º C range over the fire region (such air temperatures were foundd within the 9.5-10.5 km altitude range on the Nashville sounding when the cloud band was over central Tennessee at 00 UTC). While no fire hot spot signature was evident on the Shortwave Infrared image (due to masking by the clouds), the very distinct bright glow of the fire (which appeared rather large in size, due to scattering of light by the water and ice particles present in the various cloud layers) was seen on the Day/Night Band image. AWIPS I versions of these images are available here.

Suomi NPP VIIRS Infrared Window (11.45 µm), Shortwave Infrared (3.74 µm) and Day/Night Band (0.7 µm) images [click to enlarge]

Suomi NPP VIIRS Infrared Window (11.45 µm), Shortwave Infrared (3.74 µm) and Day/Night Band (0.7 µm) images [click to enlarge]

Additional information is available on the Wildfire Today site (post 1 | post 2 | post 3 | post 4 | post 5).

The Loma fire in California

September 27th, 2016

Suomi NPP VIIRS imagery from the Day/Night Band visible (0.70 µm) and Infrared (3.74 µm) at 0936 UTC on 27 September, and terrain [Click to enlarge]

Suomi NPP VIIRS imagery from the Day/Night Band visible (0.70 µm) and Infrared (3.74 µm) at 0936 UTC on 27 September, and terrain [Click to enlarge]

Between 0900 and 1000 UTC on 27 September (2 AM and 3 AM PDT) Suomi NPP overflew the Loma fire that is burning in the high terrain between Santa Clara and Santa Cruz counties in northern California. (News Article 1; News Article 2) The toggle above shows the glow of the fire in the nighttime visible imagery from the Day/Night Band on the VIIRS instrument. This glow is along the border of the two counties, well removed from the glow of nearby cities. The fire hot spot as detected by the 3.7 micron channel is apparent as well. Smoke from the fire is difficult to detect in this low-light scene (the waxing quarter moon was below the horizon at the time of the image, shedding no light on the scene).

GOES-15 can provide 3.9 µm imagery roughly 4 times per hour (when GOES-R is launched, shortwave infrared imagery will be produced every 5 minutes over the continental United States) allowing a better indication of how the fire is evolving with time. The animation below, from 0500 through 1530 UTC, shows a cooling trend in the warmest pixels (hottest pixels are colored red in the animation, then yellow, then black), which is expected as winds that drive the fire relax at night. There is notable motion in the navigation of this image. GOES-15 is operating with only 1 Star Tracker (vs. the usual 3), resulting in less-precise image navigation.

GOES-15 shortwave infrared (3.9 µm) imagery from 0500 UTC through 1500 UTC on 27 September [Click to animate]

GOES-15 shortwave infrared (3.9 µm) imagery from 0500 UTC through 1500 UTC on 27 September [Click to animate]

Visible Imagery from GOES-15 after sunrise on 27 September shows a long smoke plume moving southeastward from the fire source.

GOES-15 visible (0.62 µm) imagery from 1430 to 1530 UTC on 27 September [Click to enlarge]

GOES-15 visible (0.62 µm) imagery from 1430 to 1530 UTC on 27 September [Click to enlarge]