Wildfires in Northern California

October 9th, 2017 |

GOES-16 Shortwave Infrared (3.9 µm) images, with county outlines plotted in gray (dashed) and surface station identifiers plotted in white [click to play MP4 animation]

GOES-16 Shortwave Infrared (3.9 µm) images, with county outlines plotted in gray (dashed) and surface station identifiers plotted in white [click to play MP4 animation]

* GOES-16 data posted on this page are preliminary, non-operational and are undergoing testing *

GOES-16 Shortwave Infrared (3.9 µm) images (above) showed the “hot spot” signatures (black to yellow to red pixels) associated with numerous wildfires that began to burn in Northern California’s Napa County around 0442 UTC on 09 October 2017 (9:42 PM local time on 08 October). A strong easterly to northeasterly Diablo wind (gusts) along with dry fuels led to extreme fire behavior, with many of the fires quickly exhibiting very hot infrared brightness temperature values and growing in size at an explosive rate (reportedly burning 80,000 acres in 18 hours).

A comparison of nighttime GOES-16 Shortwave Infrared (3.9 µm) and Near-Infrared “Snow/Ice” (1.61 µm) images (below) offered another example of nocturnal fire signature identification — the bright glow of the fires showed up well on the 1-km resolution 1.61 µm imagery. Especially noteworthy was the very rapid southwestward run of the Tubbs Fire, which eventually moved just south of station identifier KSTS (Santa Rosa Sonoma County Airport; the city of Santa Rosa is located about 5 miles southeast of the airport. These Northern California fires have resulted in numerous fatalities, destroyed at least 3500 homes and businesses, and forced large-scale evacuations (media story).

GOES-16 Shortwave Infrared (3.9 µm, left) and Near-Infrared

GOES-16 Shortwave Infrared (3.9 µm, left) and Near-Infrared “Snow/Ice” (1.61 µm, right) images [click to play MP4 animation]

A toggle between 1007 UTC (3:07 AM local time) Suomi NPP VIIRS Shortwave Infrared (3.74 µm) and Day/Night Band (0.7 µm) images (below) provided a view of the fires at an even higher spatial resolution. Since the Moon was in the Waning Gibbous phase (at 82% of Full), it provided ample illumination to highlight the dense smoke plumes drifting west-southwestward over the adjacent offshore waters of the Pacific Ocean.

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

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

A closer VIIRS image comparison (with county outlines) is shown below.

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

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

A comparison of Suomi NPP VIIRS true-color and false-color Red/Green/Blue (RGB) images from RealEarth (below) helped to discriminate between smoke and cloud features offshore over the Pacific Ocean.

Suomi NPP VIIRS True-color and False-color RGB images [click to enlarge]

Suomi NPP VIIRS True-color and False-color RGB images [click to enlarge]

===== 10 October Update =====
Suomi NPP VIIRS true-color and false-color images [click to enlarge]

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

With the switch to southwesterly surface winds on 10 October, smoke plumes could be seen moving northeastward on RealEarth VIIRS true-color imagery, while the burn scars of a number of the larger fires became apparent on VIIRS false-color RGB imagery (above).

===== 11 October Update =====

Landsat-8 false-color RGB images, from 04 October (before the Tubbs Fire) and 11 October (after the Tubbs Fire) [click to enlarge]

Landsat-8 false-color RGB images, from 04 October (before the Tubbs Fire) and 11 October (after the Tubbs Fire) [click to enlarge]

A toggle (above)  between 30-meter resolution Landsat-8 false-color RGB images from 04 October (before the Tubbs Fire) and 11 October (after the Tubbs Fire) showed the size of the fire burn scar (shades of brown) which extended southwestward from the fire source region into Santa Rosa.

===== 12 October Update =====
Suomi NPP VIIRS true-color RGB images, with VIIRS-detected fire locations [click to enlarge]

Suomi NPP VIIRS true-color RGB images, with VIIRS-detected fire locations [click to enlarge]

A transition back to northerly winds on 12 October helped to transport the wildfire smoke far southward over the Pacific Ocean (above). Smoke was reducing surface visibility and adversely affecting air quality at locations such as San Francisco (below).

Time series plot of surface observations at San Francisco International Airport [click to enlarge]

Time series plot of surface observations at San Francisco International Airport [click to enlarge]

Suomi NPP VIIRS Aerosol Optical Depth values were very high — at or near 1.0 — within portions of the dense smoke plume (below).

Suomi NPP VIIRS true-color RGB image and Aerosol Optical Depth product [click to enlarge]

Suomi NPP VIIRS true-color RGB image and Aerosol Optical Depth product [click to enlarge]

Wildfire burning in Greenland

August 4th, 2017 |
GOES-16 Visible (0.64 µm) and Shortwave Infrared (3.9 µm) images [click to play animation]

GOES-16 Visible (0.64 µm) and Shortwave Infrared (3.9 µm) images [click to play animation]

* GOES-16 data posted on this page are preliminary, non-operational and are undergoing testing *

GOES-16 “Red” Visible (0.64 µm) and Shortwave Infrared (3.9 µm) images (above; a zoomed-in version is available here) displayed a subtle hazy signature of a smoke plume along with an intermittent “hot spot” (darker black pixels) associated with  a small fire — located near the center of the cyan circle — that was burning close to the southwest coast of Greenland on 01 August 2017. The approximate latitude/longitude coordinates of the fire were 67.87º N / 51.48º W, a location about halfway between Ilulissat (station identifier BGJN) and Kangerlussuaq (station identifier BGSF) and about halfway between the western edge of the Greenland Ice Sheet and the west coast .

Closer views using daily composites of 250-meter resolution Terra and Aqua MODIS true-color Red/Green/Blue (RGB) images (from 30 July to 04 August), sourced from RealEarth (below) indicated that the fire may have started close to 1540 UTC on 31 July — when a small white smoke and/or cloud feature (just north of the cursor) was seen at the fire source location on the Terra image (overpass time). The Aqua overpass time was around 1600 UTC.

Daily composites of Terra MODIS true-color RGB images, from 30 July to 04 August [click to enlarge]

Daily composites of Terra MODIS true-color RGB images, from 30 July to 04 August [click to enlarge]

Daily composites of Aqua MODIS true-color RGB images, from 30 July to 04 August [click to enlarge]

Daily composites of Aqua MODIS true-color RGB images, from 30 July to 04 August [click to enlarge]

Similar daily composite RGB images from Suomi NPP VIIRS (31 July to 04 August) are shown below. Note that the initial fire signature was not seen on the 31 May VIIRS image, due to the earlier overpass time  (1513 UTC) of the Suomi NPP satellite.

Daily composites Suomi NPP VIIRS true-color RGB images,.from 31 July to 04 August [click to enlarge]

Daily composites of Suomi NPP VIIRS true-color RGB images,.from 31 July to 04 August [click to enlarge]

On 03 August, a 1507 UTC overpass of the Landsat-8 satellite provided a 30-meter resolution Operational Land Imager (OLI) false-color RGB image of the fire (below). This was the same day that a pilot took photos of the fire, as reported on the Wildfire Today site.

Landsat-8 false-color RGB image [click to enlarge]

Landsat-8 OLI false-color RGB image [click to enlarge]

A comparison of one “before” (27 July) and two “after” (03 and 05 August) Landsat-8 OLI false-color RGB images (below) showed differences in smoke plume transport as the wind direction changed.

Landsat-8 false-color images on 27 July, 03 August and 05 August [click to enlarge]

Landsat-8 OLI false-color images on 27 July, 03 August and 05 August [click to enlarge]

It is possible that this “natural fire” is similar to the Smoking Hills type of spontaneous combustion that has been observed in the Canadian Arctic (thanks to Ray Hoff, retired UMBC Professor of Physics, for that tip).

Credit to Mark Ruminski (NOAA/NESDIS) for first bringing this interesting event to our attention.

===== 09 August Update =====

The animations of daily Terra and Aqua true-color RGB images (below) have been extended to 09 August and 08 August, respectively.

Daily composites of Terra MODIS true-color RGB images, from 30 July to 09 August [click to enlarge]

Daily composites of Terra MODIS true-color RGB images, from 30 July to 09 August [click to enlarge]

Daily composites of Aqua MODIS true-color RGB images, from 30 July to 08 August [click to enlarge]

Daily composites of Aqua MODIS true-color RGB images, from 30 July to 08 August [click to enlarge]

Suomi NPP VIIRS true-color RGB images from 04-09 August (below) include VIIRS-detected fire locations plotted in red. The 09 August image showed that smoke from the fire had drifted west-southwestward over the adjacent offshore waters of Davis Strait.

Daily composites of Suomi NPP VIIRS true-color RGB images, from 04-09 August, with fire detection points plotted in red [click to enlarge]

Daily composites of Suomi NPP VIIRS true-color RGB images, from 04-09 August, with fire detection points plotted in red [click to enlarge]

===== 12 August Update =====

Landsat-8 OLI false-color images on 03, 05 and 12 August [click to enlarge]

Landsat-8 OLI false-color images on 03, 05 and 12 August [click to enlarge]

Another overpass of Landsat-8 on 12 August provided a glimpse of the fire burn scar, which appeared as a darker hue of reddish-brown. Note that the fire had burned eastward to the coast, during a day when stronger westerly winds prevailed.

Related sites:

NASA Earth Observatory

NPR

ESA Space in Images

AGU EOS

 

Hail damage swath in South Dakota and Minnesota

July 4th, 2017 |

SPC storm report plots, from 12 UTC on 21 June to 12 UTC on 22 June 2017 [click to go to SPC storm reports list]

SPC storm report plots, from 12 UTC on 21 June to 12 UTC on 22 June 2017 [click to go to SPC storm reports list]

* GOES-16 data posted on this page are preliminary, non-operational and are undergoing testing *

As seen on the map of SPC storm reports from 21 June 2017 (above), nighttime thunderstorms (during the pre-dawn hours of 22 June) produced a swath of hail (as large as 2.0 inches in diameter) that damaged emerging crops at some locations across eastern South Dakota and southwestern Minnesota (NWS Aberdeen summary).

Nearly 2 weeks later, on 04 July, the hail damage swath was still apparent on GOES-16 imagery. In a comparison of “Blue” Visible (0.47 µm), “Red” Visible (0.64 µm) and Near-Infrared “Vegetation” (0.86 µm ) images (below), the northwest-to-southeast oriented hail damage swath was best seen on the 0.64 µm imagery (in part due to its higher spatial resolution, which is 0.5 km at satellite sub-point); healthy vegetation is more reflective at 0.86 µm, so the crop-damaged hail swath appears slightly darker in those images.

GOES-16

GOES-16 “Blue” Visible (0.47 µm, top), “Red” Visible (0.64 µm, middle) and Near-Infrared “Vegetation” (0.86 µm, bottom) images [click to play animation]

A signature of the hail damage swath was also seen in Near-Infrared “Snow/Ice” (1.61 µm) and Shortwave Infrared (3.9 µm) images (below). The hail damage swath warmed more quickly on the 3.9 µm imagery — exhibiting a darker black appearance with time — compared to the adjacent fields of healthy crops.

GOES-16

GOES-16 “Red” Visible (0.64 µm, top), Snow/Ice (1.61 µm, middle) and Shortwave Infrared (3.9 µm, bottom) images [click to play animation]

Why was the hail damage swath also seen on the 1.61 µm “Snow/Ice” (Band 5) imagery? A look at the Spectral Response Functions for GOES-16 ABI  bands 3, 4, 5 and 6 — plotted with the reflectance of asphalt, dirt, grass and snow (below) — show that the 1.61 µm Band 5 happens to cover a portion of the radiation spectrum where there is a minor peak in grass relectance (denoted by the green plot).

Spectral Response Functions for GOES-16 ABI Bands 3, 4, 5 and 6, along with the reflectance of asphalt, dirt, grass and snow [click to enlarge]

Spectral Response Functions for GOES-16 ABI Bands 3, 4, 5 and 6, along with the reflectance of asphalt, dirt, grass and snow [click to enlarge]

======================================================

Aqua MODIS Land Surface Temperature product {click to enlarge]

Aqua MODIS Land Surface Temperature product {click to enlarge]

Regarding the warmer temperatures seen on GOES-16 Shortwave Infrared images, the 1-km resolution Aqua MODIS Land Surface Temperature product at 1738 UTC (above) revealed a 10º F difference between the warmer hail damage swath (which appeared to be about 100 miles in length) and adjacent fields of undamaged crops. A similar result was noted on 03 July by NWS Aberdeen (below).

A comparison of before (21 June) and after (02 July) Aqua MODIS true-color Red/Green/Blue (RGB) images from the SSEC MODIS Direct Broadcast site (below) clearly shows the hail damage path.

Aqua MODIS true-color RGB images, before (21 June) and after (02 July) the hail event [click to enlarge]

Aqua MODIS true-color RGB images, before (21 June) and after (02 July) the hail event [click to enlarge]

On 05 July a closer view of the hail scar was seen using a Suomi NPP VIIRS true-color RGB image from RealEarth (below).

Suomi NPP VIIRS true-color RGB image [click to enlarge]

Suomi NPP VIIRS true-color RGB image [click to enlarge]

Incidentally, on 02 July the Sentinel-2A satellite provided 10-meter resolution true-color imagery of the hail swath:

===== 07 July Update =====

The hail damage swath was also evident on a 30-meter resolution Landsat-8 false-color RGB image from 07 July:

Landsat-8 false-color RGB image [click to enlarge]

Landsat-8 false-color RGB image [click to enlarge]

Landsat-8 false-color RGB image, zoomed in on Castlewood, South Dakota [click to enlarge]

Landsat-8 false-color RGB image, zoomed in on Castlewood, South Dakota [click to enlarge]

Other examples of satellite-observed hail damage swaths can be seen here and here.

 

Mud Creek landslide along the California coast

May 22nd, 2017 |

As seen in the Tweet above from NWS San Francisco Bay Area, a major landslide occurred along the California coast in the Big Sur area (at Mud Creek) during the nighttime hours on 20 May 2017. A large portion of coastal Highway 1 was closed by the massive amount of debris.

A timely overpass of the Landsat-8 satellite at 1840 UTC on 22 May (along with the cooperation of a gap in cloudiness) provided 30-meter resolution false-color Red/Green/Blue (RGB) imagery (source) which showed the landslide debris extending off the coast and into the adjacent nearshore water of the Pacific Ocean (below). Before/after photos of the landslide site can be seen here.

Landsat-8 false-color images [click to enlarge]

Landsat-8 false-color images [click to enlarge]

text