10-minute Full Disk scan GOES-18 (GOES-West) “Clean” Infrared Window (10.3 µm) images and “Red” Visible (0.64 µm) images with an overlay of the FDCA Fire Mask derived product (above) showed that a large wildfire north of Fort Nelson (CYYE) in far northeastern British Columbia produced a series of ~3 pyrocumulonimbus (pyroCb) clouds on 28 May 2025. The pyroCb clouds exhibited cloud-top 10.3 µm infrared brightness temperatures (IRBTs) in the -40s C (denoted by shades of blue to cyan) — a necessary condition to be classified as a pyroCb — beginning at 0010 UTC on 29 May. The coldest pyroCb cloud-top IRBT was -52.57ºC at 0320 UTC on 29 May (below).

According to rawinsonde data from Fort Nelson, British Columbia at 0000 UTC on 29 May (below), the coldest pyroCb IRBT corresponded to an altitude around 11.1 km.

This large wildfire burned very hot, exhibiting Shortwave Infrared 3.9 µm brightness temperature values of 137.88ºC — the saturation temperature of GOES-18 ABI Band 7 detectors — for 2.5 hours, from 2210 UTC on 28 May (2 hours prior to the formation of the first pyroCb) until 0040 UTC on 29 May (below).

This wildfire also made a significant run to the NNW in ~13 hours, as seen in a comparison of VIIRS Fire Radiative Power displayed using RealEarth (below).

The NNW run of the wildfire was also apparent in a 14-hour animation of Next Generation Fire System (NGFS) fire detection polygons (below).

===== 29 May Update =====

During the following daytime hours, GOES-19 (GOES-East) True Color RGB images (above) revealed a long ribbon of brownish-gray smoke-laden pyroCb ice cloud that was arcing eastward across the Northwest Territories then curving southward over Nunavut, Manitoba and Saskatchewan as the leading edge approached the Canada/US border.

View only this post Read Less

The slider below compares two different NGFS detections over the Canadian Plains: a VIIRS image (created using data from the direct broadcast antenna at CIMSS) and a GOES-18 image. The NGFS Microphysics looks slightly different between the two because of different view angles and slightly different channel characteristics between the VIIRS and GOES bands used. Note the far better spatial resolution with the VIIRS imagery however. The smoke plume is thick enough in both depictions that it has a presence in the RGB. This GOES-East true-color image (from the CSPP Geosphere site) shows the thickness and extent of the smoke plume later in the day on 28 May.

View only this post Read Less

Heavy rains developed overnight over American Samoa, as depicted by the animation above from the CSPP Geosphere site. Low clouds — white/cyan in the RGB — start to show vertical development (that is, they get more and more red) shortly after 1000 UTC with more vigorous vertical growth apparent after 1100 UTC. Pago Pago on the island of Tutuila has had a particularly wet month: 25″ of rainfall through 27 May, which is the second wettest May on record there (data from here). What clues were present to help anticipate this convective development? The animation below shows GOES-18 Clean Window infrared (Band 13, 10.3) imagery from 1000-1400 UTC, and it is overlain by GOES-18 estimates of Lifted Index (scaled from -5 to 4). American Samoa is at the northern edge of a pool or greater instability that is available given a suitable trigger.

Sometimes, LightningCast probability fields (available here) can give a suitable warning of convective development (and, subsequently, lightning) as discussed in this blog post (and elsewhere): that is, contours start to appear with convective initiation. Convective development on 28 May was quite rapid. Lightning occurred within about 20 minutes of the first contour appearing just south of Tutuila in the animation below. In other words, LightningCast in this case was helpful in alerting a forecaster that convection was developing — but that signal was also apparent in single-banded imagery!

At 1330 UTC (the middle of the night in American Samoa), a Flash Flood Warning was issued for American Samoa.

View only this post Read Less

Interleaved 1-minute Mesoscale Domain Sector and 5-minute CONUS Sector GOES-19 (GOES-East) Upper-level Water Vapor images (above) displayed the SpaceX Starship 9 rocket exhaust plume thermal signature after it was launched from the Starbase facility at Boca Chica Beach, Texas at 2336 UTC on 27 May 2025. Note the change in exhaust plume shape with time and atmospheric layer: at early altitudes of 20-50 km (where atmospheric gases within the Stratosphere had more density, and therefore higher ambient pressure), the Stage 1 Super Heavy booster plume was more linear — but after the 2339 UTC “hot stage separation” as the Stage 2 rocket reached higher altitudes of 70-100 km (where atmospheric gases within the Mesosphere and Thermosphere were much less dense, with lower ambient pressure) the plume was able to expand outward into more of a curved “boomerang” shape beginning at 2341 UTC.

The changing shape of the exhaust plume with time (and altitude) was also apparent in 1-minute GOES-19 Plume RGB images created using Geo2Grid (below).

A display of all 16 ABI spectral bands from GOES-19 (below) showed that a hot thermal signature of the Starship booster was apparent in Near-Infrared & Infrared bands 04-16 — especially 1 minute after launch (at 2337 UTC) near the coast. In addition, the rocket booster exhaust condensation cloud was seen in all 16 spectral bands, as it began to drift eastward away from the Starbase launch site.

A toggle between zoomed-in 16-panel GOES-19 images at 2336 and 2337 UTC (below) made it easier to see the distinct Super Heavy booster thermal signature in spectral bands 04-16 at 2337 UTC.

1-minute GOES-19 True Color RGB images from the CSPP GeoSphere site (below) displayed the rocket booster condensation cloud as it drifted eastward away from the Texas coast.

10-minute Full Disk scan GOES-18 (GOES-West) True Color RGB images (below) also provided a brief glimpse of the rocket condensation cloud as it cast long shadows onto to offshore waters.

View only this post Read Less