While Chicago O’Hare recently retook the crown of the country’s busiest airport in terms of number of flights, Atlanta’s Hartsfield-Jackson International Airport (ATL) continues to have the largest number of passengers. A great number of those ATL passengers were significantly inconvenienced on the evening of 6 March 2026 as a hailstorm passed directly over the airport just before 7:00 PM (0000 UTC on the 7th), one of the busiest times of the day for outbound flights. While the hail was not particularly large (no reports made it to the Storm Prediction Center’s Storm Reports for the day) it still forced a ground stop and inspections of dozens of aircraft to ensure that passengers were able to continue safely. Those inspections were hampered by the continued presence of lightning which kept crews indoors for their own safety. It took many hours for planes to be cleared for departure; hundreds of flights were cancelled and the ripples of this are still impacting flight operations the next day.

The atmosphere was somewhat primed for convective activity on this day. A NUCAPS profile near Atlanta from NOAA-20 at 1917 UTC (2:17 PM EST) as plotted by NASA SPoRT’s NUCAPS Savvy site is shown below. Note the presence of steep low-level lapse rates that are contributing to a decent level of surface-based CAPE (1486 J/kg).

The GOES-19 10.3 micron Band 13 view of the event is shown below. The hail event wasn’t the first storm to hit ATL that day, as a previous round of convection had already caused a ground stop earlier. The following loop begins just after the first round moved off to the east and captures the second round from initiation onward. The convection isn’t particularly deep, as can be seen in the relatively warm cloud top temperatures (you can also discern that in the NUCAPS profile above where the tropopause appears to begin around 280 mb or so).

Despite the somewhat low cloud tops, the presence of a vigorous updraft can be discerned with the presence of an enhanced V. We’ve highlighted below in the 2346 UTC frame; just move the slider back and forth to add or remove our subjective annotation.

CIMSS’s ProbSevere Product was also tagging this storm as one to watch, as shown in the following loop (the airport can be found at the right angle in the county boundaries in the center of the image).

Here’s what the ProbSevere metrics for this storm were. There’s an important caveat here: this was not a “severe” hail event, in that hail did not meet or exceed the 1 inch threshold that the National Weather Service has defined for hail to merit the severe designation. There do not appear to be any reports of damage to the planes, which is consistent with this “less than severe” designation. ProbSevere is calibrated to predict the probability of severe events. Therefore, while this 45% probability might seem small for an event that had such a significant impact, it is consistent with what actually took place.

The CIMSS LightningCast product was also monitoring this event. In addition to contour plots overlaid on satellite imagery, LightningCast also provides dashboards at key locations across the United States, including stadiums and airports. Here’s what the LightningCast dashboard for ATL looked like (red) alongside verification from the GOES-19 Geostationary Lightning Mapper. It was clear that lightning was going to hinder ground services.

One challenge with monitoring this event was that it occurred right around sunset, which meant that most of the RGB products used for convection monitoring were not reliable: it was too dark for the daytime products, but there was still too much reflectance in the 3.9 micron band for the nighttime products to be useful. Still, the panoply of other satellite products made it possible to monitor the evolution of this not severe, but still significant, event.

View only this post Read Less

5-minute CONUS Sector GOES-19 (GOES-East) Visible and Infrared images (above) showed a supercell thunderstorm that developed in northern Indiana and tracked northeast across far southern Lower Michigan on 06 March 2026. This thunderstorm formed in the vicinity of a warm front (surface analyses) that was moving northward across the region — surface plots indicated that air temperatures behind the warm front rose into the low-mid 70s F, with dew points in the 60s F. This supercell produced a least 3 tornadoes (one rated EF1, one rated EF2 and one rated EF3) that were responsible for 4 fatalities and numerous injuries (NWS ILX summary).

Note that surface observations at the Three Rivers Municipal Airport (KHAI) ceased after about 2030 UTC (below) — apparently due to either tornado damage or a tornado-related power outage.

The ProbSevere IntenseStormNet product (below) began to depict >50% probabilities of intense convection at 2011 UTC — the time when cloud-top infrared brightness temperatures were rapidly cooling, GLM Flash Extent Density was increasing, and tornadoes were being reported north and northeast of Edwardsburg (SPC Storm Reports).

Several hours later, 1-minute Mesoscale Domain Sector GOES-19 infrared images (below) showed a large cluster of supercell thunderstorms over northeast Oklahoma. One of the tornadoes was responsible for 2 fatalities near Beggs OK around 0119 UTC.

Around the time that the northernmost supercell began to produce tornadoes, it exhibited a notable Enhanced-V storm-top signature along with a subtle Above-Anvil Cirrus Plume in GOES-19 Infrared and Visible images, respectively (below).

1-minute GOES-19 Infrared images with an overlay of GLM Flash Points (below) depicted abundant lightning activity with these storms.

The coldest cloud-top infrared brightness temperature associated with these thunderstorms was -66ºC — according to a plot of rawinsonde data from Oklahoma City (below) that temperature roughly corresponded to a ~2 km overshoot of the Most Unstable (MU) air parcel’s Equilibrium Level (EL).

View only this post Read Less

In August 2024, the European Space Agency in partnership with EUMETSAT launched a pathfinder mission called Arctic Weather Satellite (AWS), designed to prove that high-quality passive microwave observations on a micro-satellite (about the size of a washing machine!) in polar low-earth orbit are possible. AWS has one instrument on board to do this, the Microwave Radiometer (MWR), that collects observations across 19 channels, with a pixel footprint size between 10 – 40 km, depending on the channel:

Notably, AWS MWR is the first weather satellite to include channels (4 of them) around 325 GHz, which is designed to add new information about ice clouds. All together, these channels provide information that can be used to determine vertical structure of temperature and humidity in all-sky conditions.

This satellite downlinks its global, full-orbit data to a ground station in Svalbard. Since July 2025, ECMWF has been operationally assimilating AWS MWR into its forecast models with positive results. In addition, AWS has a direct broadcast capability in the L-band, allowing anyone with the right equipment to capture science data as it flies overhead. To collect this signal as the satellite passes overhead and make it useful, you need 5 things:

1. An antenna and feed capable of receiving the direct broadcast signal.
2. A demodulator configured to handle the received RF signal and turn it into a stream of packets.
3. An antenna control server that can routinely schedule reception of the satellite and capture the packets.
4. Software that can assemble, geolocate, and calibrate the raw packets to a Level 1 file.
5. Software that can read the Level 1 file and produce useful file formats for visualization.

On item 1, SSEC (along with our partners at the University of Alaska Fairbanks GINA) already operate a network of LEO direct broadcast antennas capable of L-band reception around the United States. Over the last year, we worked with the antenna vendor to develop a demodulator mode and schedule/reception configuration for AWS, covering items 2 and 3.

As of today, there two known direct broadcast antennas in the United States with all the hardware and software needed to routinely capture AWS direct broadcast – SSEC’s antenna in Madison, WI and GINA’s antenna at NOAA FCDAS in Fairbanks, AK.

In July 2025, EUMETSAT released their Level 0 / Level 1 processing package for AWS MWR. SSEC and GINA have since integrated this software into downstream processing servers, producing a real-time feed of AWS MWR Level 1 files. Then, on March 2nd, 2026, the CSPP team released Polar2Grid v3.2, which, among other things, adds the ability to read Arctic Weather Satellite MWR Level 1 files, and produce GeoTIFF and AWIPS tiled outputs. With that, all 5 items have been accomplished! For example, here’s a sample of Polar2Grid-generated images from the two systems of AWS MWR Band 19 (called AWS44 in the chart above), which is most sensitive to low level ice clouds and snow:

Work has yet to be done by various groups on developing useful colormaps to make these MWR Level 1 bands helpful for forecasters, and on the integration of AWS MWR into NOAA’s operational Level 2 microwave products (like MiRS). EUMETSAT plans on launching additional micro-satellites (up 6 operational satellites at a time, spread across 3 orbital planes), nearly identical to AWS, starting in 2029, creating a constellation called Sterna. Sterna satellites, along with the traditional NOAA JPSS and EUMETSAT MetOp constellations, would provide rapid-revisit passive microwave soundings across the whole globe to improve weather forecast accuracy.

More information about EUMETSAT’s future Sterna constellation can be found at: https://www.eumetsat.int/eps-sterna

Looking ahead, SSEC hopes to purchase the hardware needed to support AWS and other upcoming DB missions at its other 5 antenna sites. GINA is providing Alaska AWS DB data to NOAA/NESDIS for experimental inclusion in an improved Level 2 Snowfall Rate product. Work on this initial AWS direct broadcast development was partially supported by SSEC’s Polar Satellite Antenna Systems contract with the National Weather Service Office of Observations.

View only this post Read Less

The morning of 5 March 2026 saw widespread fog over the Midwestern United States. The following animation shows the Day Cloud Phase Distinction RGB product from GOES-19 with some surface observations overlaid on top. The surface observations show just how widespread the fog was. The standard meteorological chart code uses two horizontal parallel lines to denote mist and three such lines for fog. The visibility at a station (in miles) is shown by the number in the lower-right of each plot. In northwestern Illinois, for example, visibilities were between 2 and 3 miles, while in Chicago it was 1.5 miles and in Milwaukee it was as low as a quarter of a mile. Many of the lcoations across the map are also showing conditions at or very close to saturation, with numerous places have dew points at the same value as the temperature or just a degree or two apart.

The day cloud phase distinction product does well in discriminating between liquid (cyan) and ice (orange-yellow) clouds. A similar product, the Day Snow/Fog RGB, is another tool for identifying low clouds and fog from satellite. Here, teh frozen surface areas (lake ice or snow) show up as red with the flog as various shades of gray or yellow-gray.

An additional data source worth monitoring are the aircraft profiles. Many commercial aircraft provide observations of temperature and winds as they take off and land, while a small subset also provide water vapor. This smaller selection of temperature + water vapor observations (less than 10% of the total) is publicly available in real time, although there’s not a lot of places to easily access the data. Fortunately, a weather enthusiast has made a page to produce Skew-T plots of recent aircraft profiles, which can be found here. Here’s a plot of the observations from a flight into Chicago’s Midway Airport at 9:58 AM (1558 UTC) this morning. Note how the layer between the surface and the nocturnal inversion is fully saturated. Unlike radiosondes, where the disposable sensors often struggle to record full saturation, the aircraft-based sensors can be engineered to greater precision since they’ll be reused for thousands of profiles over their lifespan.

The low visibilities over the past day did have some adverse impacts on aviation. A plane scheduled to fly from Chicago O’Hare to Bloomington-Normal in central Illinois made it as far as the vicinity of its destination before having to divert all the way to Madison, Wisconsin, due to the low visibilities. The flight track, from FlightAware, is seen here:

View only this post Read Less