This website works best with a newer web browser such as Chrome, Firefox, Safari or Microsoft Edge. Internet Explorer is not supported by this website.

CIMSS Satellite Blog

MENU

Train derailment and fire in East Palestine, Ohio

By Scott Bachmeier •

GOES-16 Shortwave Infrared (3.9 µm) images, with railroads plotted in violet [click to play animated GIF | MP4]

At around 9 PM EST on 03 February 2023 (0200 UTC on 04 February) a 50-car train derailed in East Palestine in far eastern Ohio, with some of the rail cars carrying hazardous chemicals that caught fire (media report) — and GOES-16 (GOES-East) Shortwave Infrared (3.9 µm) images (above) showed the thermal signature of this long-lived fire, which was still evident 27 hours after the derailment (at 0006 UTC on 05 February). At times, thick cloud cover totally obscured the thermal signature; however, when relatively thin clouds were over the fire this thermal signature could still be seen. The peak 3.9 µm infrared brightness temperature was 17ºC, about an hour after the accident (at 0306 UTC).

===== 06 February Update =====

GOES-16 “Red” Visible (0.64 µm, top left), Shortwave Infrared (3.9 µm, top right), Cloud Top Phase (bottom left) and “Clean” Infrared Window (10.3 µm) / Cloud Top Temperature / Cloud Top Height (bottom right); railroads are plotted in violet [click to play animated GIF | MP4]

On the afternoon of 06 February, a controlled vent and burn of toxic chemicals (contained in 5 of the rail cars) was conducted (media report) — and 1-minute Mesoscale Domain Sector GOES-16 “Red” Visible (0.64 µm), Shortwave Infrared, Cloud Top Phase derived product, “Clean” Infrared Window (10.3 µm), Cloud Top Temperature derived product and Cloud Top Height derived product (above) showed signatures of the resulting black smoke plume that penetrated the top of a supercooled water droplet (light green in the Cloud Top Phase product) stratus cloud layer that was over the area at that time. The black smoke plume first emerged from the cloud top at 2139 UTC, casting a long shadow to the northeast at 2140 UTC. Beginning at 2149 UTC, the Clear Sky Mask derived product (animated GIF | MP4) mistakenly identified the black smoke as a hole in the stratus cloud deck — so with a “Clear” sky falsely indicated, the derived cloud products (Cloud Top Phase / Cloud Top Temperature / Cloud Top Height) were not created for those particular pixels. Note that the Cloud Top Height product was distributed in AWIPS at a reduced spatial resolution (4 km, compared to the native resolution of 2 km for infrared spectral bands used to create the product).

The black smoke cloud seen in Visible imagery also exhibited a signature in Dust RGB and Split Cloud Top Phase brightness temperature difference imagery (below) — which both leverage the 8.5 µm spectral band that is sensitive to differences in emissivity (in this case, the emissivity of the smoke particles differed from that of the supercooled water droplets along the top of the surrounding stratus cloud layer) .

GOES-16 “Red” Visible (0.64 µm) / Dust RGB / Split Cloud Top Phase brightness temperature difference (top left), Shortwave Infrared (3.9 µm, top right), Cloud Top Phase (bottom left) and “Clean” Infrared Window (10.3 µm, bottom right); railroads are plotted in violet [click to play animated GIF |MP4]

The infrared brightness temperature of the overshooting black smoke plume was 1-2ºC warmer than that of the adjacent underlying stratus cloud layer — and a plot of rawinsonde data from Pittsburgh, Pennsylvania (below) revealed the presence of a strong boundary layer temperature inversion, with air temperatures rising quickly with height. The presence of light winds at those low altitudes also prevented a rapid dispersion or significant advection of the above-stratus smoke plume.

Plot of rawinsonde data from Pittsburgh, Pennsylvania at 00 UTC on 07 February [click to enlarge]

GOES-16 True Color RGB images from the CSPP GeoSphere site (below) provided a better portrayal of the contrast between the dark black smoke plume and the surrounding stratus cloud deck.

GOES-16 True Color RGB images, with and without a map overlay [click to play MP4 animation]

Categories: Fire detection, GOES-16, Red-Green-Blue (RGB) images