A fire on Sunday 9 March in suburban Milwaukee (Milwaukee Journal Sentinel articles 1, 2, 3; I read it about it in the paper: Page 1; Page 2) burned about 100 acres just north of Capitol Drive as shown in the 10 March 2025 image above from the Journal Sentinel. What kind of alert did NGFS provide for this small fire? The slow animation below shows the NGFS Microphysics RGB from routine CONUS scanning (that is, every 5 minutes; mesoscale sectors producing imagery every minute were positioned elsewhere on the 9th). The first NGFS alert was at 1846 UTC, a bit after the 1:39 PM 911 call that alerted first responders.

NGFS imagery includes Google Maps that can inform a user about the location where the burn is detected. The slider below reveals the pixel size relative to the burning region, the marsh that burned, and the roads and businesses near the fire. (Here’s another google maps view — direct link — that includes the Ford Dealership mentioned in the article, and a circle that highlights the intersection at the bottom of the burn-scar image shown up top). The area of the infrared pixel used to detect this fire is about twice the size of the area consumed; you might conclude that very little of the pixel at any one time was burning, and the detection of a fire of this small size is remarkable.

View only this post Read Less

On 20 February 2025, high winds at Pago Pago International Airport (located at 14.3^oS, 170.7^oW) meant Hawaiian Airlines Flight 465 (an A330 with 160 passengers on board) was forced to return to Honolulu (News Article; FlightAware screenshot) without its scheduled landing in Pago Pago. American Samoa at the time was within the South Pacific Convergence Zone, and both flood and high wind warnings (issued by the National Weather Service office at the airport) were in place (Here is the High Wind Warning Text, and here is the Forecast Discussion valid at that time). What did satellite imagery show during this time?

The animation above shows MIMIC TPW estimates of Total Precipitable Water derived from Microwave sounders. The abundant moisture of the SPCZ is shown are red/dark purple over Samoa. February saw 25″ of rain fall in Pago Pago.

CIMSS Turbulence (online here, with a training video available) shows fields created via machine-learning algorithms that are trained on aircraft observations of Eddy Dissipation Rate (EDR, i.e., turbulence), and the fields displayed below show probability of Moderate-or-Greater (MOG) turbulence. The animation shows MOG fields from GOES-West from 0000-0930 UTC on 21 February 2025 (HA465 typically lands at 9 PM Samoa Standard Time, which is 0800 UTC on the following day). As noted above, Pago Pago is near 14^oS, 171^oW, near the center of the box drawn on the image. The satellite imagery and GFS model fields used to compute MOG Probability show large values around the Samoan Islands. This predicts upper-level turbulence, not the turbulence that might be occurring closer to the surface.

JMA‘s Himawari-9 also views American Samoa. CIMSS Turbulence fields computed using data from that satellite, below, tell a similar tale.

GOES-18 Clean Window infrared imagery, below, shows widespread convection associated with the SPCZ over the Samoan Islands. Clear-air Total Precipitable Water (TPW), a level-2 GOES-R Product shows little information because of the widespread cloudiness (this is when microwave-based TPW estimates, like MIMIC shown above, are most important) although values (showing very moist conditions) occasionally appear in breaks in the cloudiness.

Upper-level Water Vapor infrared imagery, below, similarly shows widespread convection over the Samoan Islands. You can also infer large scale upper-level divergence: Equator-ward motion over the northern part of the domain and poleward motion in the southern part of the domain.

Soundings from Pago Pago (upper air station 91765, from here) on 20 and 21 February are shown below. The very moist airmass is noteable (60+ mm of Total Precipitable Water), and so are the strong winds: 40-50 knots below 800 mb!

Scatterometry on 20 and 21 February shown below (from the ‘manati’ website) show strong low-level winds over/around the islands.

---

The SSEC/CIMSS Tropical Website includes wind analyses over the southeast Pacific. What did those show during this time? Low-level convergence and upper-level divergence, shown below, are consistent with the development of very strong convection within the moist atmosphere.

Observed winds (observed by tracking cloud motions or water vapor gradients) at low levels show the broad convergence over the area, and broad swaths of low-level winds at 20-25 knots. Cloudiness over the Samoan Islands, however, prevented observations of derived motion winds there. Upper-level winds for the same time are concentrated over the cloudband that overlies the Samoan Islands. It’s pretty easy to see very strong diffluence over Samoa.

Deep-layer mean flow on 20-21 February 2025 was towards the east-southeast. Convective development was moving in that direction during the 36h shown in the animation (Click here to see an animation of 850-mb vorticity).

Shear values, shown below for 0000 and 1200 UTC on 21 February, suggest a favorable environment for Tropical Cyclone spin-up is present just south of the Samoan Islands.

What do all these fields mean put together? Are they enough to cancel the flight at the outset? Not really (in my opinion, not that anyone asked me back on the 20th), but they certainly support the presence of the strong winds at the airport that led to two landing attempts and forced the plane to return to its starting point.

---

What did LightningCast probabilities (as computed with the latest CSPP Geo software beta release) show? The animation below shows GOES-18 LightingCast probabilities at 10-minute timesteps from 0700 through 0830 UTC on 21 February 2025. Convection that will likely produce GLM Observations in the next 60 minutes is over Independent Samoa, but not over American Samoa.

Values computed with Himawari-9 data, shown below at 0730 UTC on 21 February 2025, are very similar to those from GOES-18.

View only this post Read Less

In a sequence of daytime True Color RGB + Nighttime Microphysics RGB images from GOES-16 (GOES-East) and GOES-18 (GOES-West) (above), signatures of the SpaceX Starship Test Flight 8 launch were seen — the Super Heavy Stage 1 rocket booster condensation cloud cast a long shadow across the water before sunset, as it became sheared by high-altitude winds over the Starbase launch site near the South Texas coast.

1-minute Mesoscale Domain Sector GOES-16 daytime True Color RGB + Nighttime Microphysics RGB images (source) are shown below.

A 16-panel display of 1-minute GOES-16 images (above) showed reflectance or thermal signatures from all 16 ABI spectral bands. It is notable that a thermal signature of the Super Heavy Stage 1 rocket booster was apparent in Near-Infrared and Infrared spectral bands 04-16 at 2333 UTC — and at 2334 UTC in bands 08/09/10/16, thermal signatures of (1) the Starship Stage 2 rocket (just after separation from the Stage 1 booster) and (2) the boostback burn of the Stage 1 booster rocket as it began its return to the Starbase launch site.

A toggle between GOES-16 Upper-level Water Vapor (6.2 µm) and Shortwave Infrared (3.9 µm) images at 2336 UTC (below) showed thermal signatures of the Starship Stage 2 rocket as it was traveling east across the Gulf of Mexico. At that time the rocket had reached altitudes of 70-80 km (where air within the Mesosphere was much less dense, with low ambient pressure) — so the rocket exhaust plume was able to expand outward into more of a curved “boomerang” shape in the Water Vapor image.

View only this post Read Less

CSPP Geosphere imagery on 5 March, above, (direct link to imagery), shows the development of a smoke plume from a fire at the north end of San Jose Island. When did the Next Generation Fire System first identify the fire that caused the smoke plume? The NGFS Alerts Dashboard, below, shows a detection over Aransas County.

What did those CONUS (every five minute) detections show?The animation below shows NGFS Microphysics, Fire Temperature RGB and Band 7 (Shortwave infrared, 3.9 µm) imagery from GOES-16. The NGFS detection occurred at 1646 UTC. The fire develops quickly and the Fire Temperature RGB and Band 7 imagery show a very obvious fire signal by 1701 UTC.

---

NGFS detections also occur with VIIRS imagery that has much greater spatial resolution, as fine as 375m for some infrared (and visible) channels. The NOAA-21 overpass shows a smoke plume in the GeoColor imagery below, and three separate fire detections in these Real Earth screenshots (from here).

A zoomed-in view of the NGFS detections and the NGFS Microphysics is shown below. Of particular note is that VIIRS data resolves three separate fires.

What did GOES show at the same time? That’s shown below in a toggle between GeoColor and the NGFS Microphysics RGB. The advantages of the higher-resolution VIIRS imagery shown above is obvious. VIIRS coverage is somewhat limited over CONUS; it is much more plentiful over Alaska where GOES pixel sizes have become large because of the distance to Alaska from the sub-satellite point of GOES-18.

This detection on 5 March was over a relatively uninhabited region, where satellite detection will likely give the earliest alert for responders. On 4 March, when much of south Texas was within a critical fire weather outlook from SPC, a fire developed within the pixels that included at the NWS Corpus Christi office! The NGFS detections for that (brief) grassland fire are shown below.

This example on 4 March demonstrates why it’s important for a weather office to have windows!

View only this post Read Less