5-minute CONUS Sector GOES-19 (GOES-East) True Color RGB images created using Geo2Grid (above) showed aircraft glaciation trails drifting east-southeastward over southern Lake Michigan on 25 February 2026. As aircraft ascended/descended through a thin cloud layer composed of supercooled water droplets, additional cooling from wake turbulence (reference) — and/or particles from jet engine exhaust acting as ice condensation nuclei — caused the small supercooled water cloud droplets to transform into larger ice crystals (many of which then fell from the cloud layer as snow).

GOES-19 Day Cloud Phase Distinction RGB images (below) revealed that the cloud layer which the aircraft penetrated was composed of supercooled water droplets (shades of violet to purple) — while the ice crystal aircraft glaciation trails exhibited shades of green.

Cursor samples of the GOES-19 Cloud Top Temperature (CTT) derived product at 2 locations along the undisturbed supercooled water droplet cloud between aircraft glaciation trails (below) showed that while those CTT values were quite cold, they were still a few degrees warmer than the -40ºC temperature required to assure spontaneous freezing via homogeneous nucleation.

In a comparison of GOES-19 Shortwave Infrared brightness temperatures over an undisturbed portion of the supercooled water droplet cloud vs within an aircraft glaciation trail (below), enhanced solar reflection off the small spherical supercooled droplets produced a 3.9 µm temperature that was about 11ºC warmer than that of the ice crystals within a glaciation trail.

View only this post Read Less

The early morning hours of 21 February 2026 brought some intense tropical convection to the vicinity of the Samoan Islands. Radar coverage is lacking in this part of the world so satellites remain the best way to monitor storms in this region for potential safety hazards. A good first stop is checking the NUCAPS vertical profiles for a location to gauge the potential for convective instability. Since this post was written a few days after the event in question, our standard sources for NUCAPS profiles no longer have the relevant data, but the NOAA OSPO HEAP site allows us to go back 20 days to review the satellite-observed skew-T profiles. American Samoa is in the UTC –11 time zone (almost as far away from UTC as one can get) so this profile is around 1:20 AM local time. NUCAPS profile retrievals are applied to collocated infrared and microwave sounder observations. The microwave (MW) observations are able to see through the clouds but lack the vertical resolution of the infrared (IR) observations. The highest-quality profiles include both the MW and IR, but if the clouds are too thick, the retrieval algorithm will default to the MW only. The following plot shows both the MW-only (blue) and the MW+IR (maroon) profiles.

It’s interesting to see how the two different methodologies result in two very different levels of instability. The MW only retrieval has a CAPE of 3643 J/kg while the MW+IR retrieval brings it down all the way to 0. The biggest difference between the two appears to be differences in near-surface temperatures and the presence of a mid-level inversion between 600 and 800 hPa. However, those low-level temperatures appear to be an underestimate, especially for the MW+IR retrieval. At 1200 UTC, the surface temperature and dew point at the Pago Pago, AS, airport were 30/25 C respectively. That’s much warmer than NUCAPS estimated. If you append those values to the bottom of the sounding (the mean surface level pressure was 1005 hPa, and since the elevation of the airport in Pago Pago is just a few meters above sea level, the station pressure is basically the same as MSLP) you’ll see that the profile becomes much more positively buoyant. Here’s a hand-drawn illustration of that, with the new temperature and dew point added in red and the corresponding area of positive buoyancy shaded in orange. At the very least, it is likely that the CAPE value measured by the microwave-only retrieval is closer to the truth than the 0 value returned by the MW+IR retrieval.

LightningCast captured the potential for significant lightning from this storm. The following animation shows the LightningCast contours overlaid on top of the Band 13 infrared imagery from GOES-18. The occasional blue patches represent the GLM-observed flash density, helping to verify LightningCast’s predictions.

The deep reds of the Night Microphysics RGB also help indicate the presence of deep, thick convective clouds.

Together, the thermodynamic observations from the polar-orbiting satellite and the continuous, time resolved imagery from the geostationary perspective help offer a comprehensive portrait of the intensity of this event.

View only this post Read Less

5-minute CONUS Sector GOES-19 (GOES-East) Mid-level Water Vapor (6.9 µm) images (above) showed the evolution of a large and intense Nor’easter that produced as much as 37.9 inches of snowfall in Rhode Island (a preliminary State record) and wind gusts as high as 84 mph in New York (WPC Storm Summary) during the 22 February – 23 February 2026 period. The combination of high snowfall rates (up to 3 inches per hour) and strong winds produced widespread blizzard conditions.

A slightly closer view using 1-minute Mesoscale Domain Sector GOES-19 Mid-level Water Vapor images (below) included overlays of GLM Flash Extent Density and GLM Flash Points, along with plots of surface weather symbols and peak wind gusts (note that there was a brief outage of GOES-19 imagery on 23 February). Isolated inland GLM signatures suggested that some brief thundersnow may have occurred (0335 UTC | 0529 UTC | 1036 UTC | 1057 UTC) — in fact, regarding the 0529 UTC GLM signature, thundersnow was reported at nearby Teterboro, New Jersey (KTEB) and Manhattan, New York (KJRB) 3-9 minutes later (in their 5-minute METARs).

5-minute GOES-19 daytime True Color RGB + Nighttime Microphysics RGB images from the CSPP GeoSphere site (below) displayed the Nor’easter during the 22-24 February period.

View only this post Read Less

While much of the weather world on 23 February 2026 has turned its eyes to the east coast of the United States due to the presence of a strong bomb cyclone and an associated blizzard in the most densely-populated region of the country, the effects of this event are being felt much further out. One of the more unexpected impacts is the presence of lake effect snow in the northern part of central Indiana, nearly a hundred miles from the shore of Lake Michigan.

The impact of the Great Lakes on snowfall totals is well-known. Generally westerly flow across the Great Lakes causes higher snow totals on the eastern edges of those lakes to have notably higher annual snowfall totals. This is because the (relatively) warm and moist lakes inject sensible and latent heat into air that passes over the lakes. Those air masses then increased surface friction and terrain once they reach the shore, factors which cause low-level convergence and orographic lift which combine to help increase the low-level snowfall formation process. A map is available here from our friends at NOAA’s Great Lakes Environmental Research Laboratory; note how the downstream side of each of the Great Lakes hosts a local maximum in annual snowfall.

But what happens when the wind pattern isn’t the standard west-to-east flow that we see throughout much of the year?

The 700 hPa map from 1200 UTC on the 23rd gives an easily accessible overview of the synoptic state of the atmosphere on this day. The obvious weather maker is the cyclone off of the east coast of the continental United States, with a broad ridge across the western US. Over the Great Lakes, flow is almost due northerly.

The northerly flow across the Great Lakes is forcing lake effect snows in places that don’t always receive them. In particular, the fetch across Lake Michigan runs nearly the entire north-south extent of the lake. This puts a substantial plume of lake effect snow into northern and north-central Indiana. We’ll start our analysis with a true color view from the GOES-19 ABI as shown by SSEC’s Real Earth. The telltale signs of lake effect snow are present as narrow parallel bands of cloudiness forming over the lake, extending over the shore and penetrating into the land. But instead of the east-southeast orientation that is typical for Lake Michigan, the bands are almost straight north-south.

The Day Microphysics RGB can provide some insight to the makeup of the clouds producing these snows. The central band of cloudiness extending southward from Lake Michigan shows a lavender center where the clouds are geometrically thicker and a yellowish-green edge where they are thinner. Both of thse colors are associated with liquid clouds in this RGB, with the pinker colors representing clouds with larger water droplets. This is consistent with lake effect snow, as these clouds are much more shallow and closer to the surface than synoptically-forced clouds and thus are likely to be too warm to be purely ice clouds.

The Cloud Phase RGB similarly depicts a salmon-colored band where the clouds are liquid and have larger water particles, both of which are consistent with lake effect snowfall.

The Day Cloud Phase Distinction RGB also provides a unique perspective on the evolution of this event. Here, snow covered grown shows up s a bright green color. The wide swath of snow from last week’s event is clearly visible stretching from eastern Nebraska into north-central Wisconsin, for example. However, you can also see the surface snow cover in Indiana between the individual clouds of the lake-induced convection. The green and orange clouds assoicated with the thick ice clouds of the east coast cyclone are also visible on the far eastern edge of the image.

Finally, here is the 0.64 micron (red) visible satellite product with the Warsaw, Indiana NWS NEXRAD radar overlaid on top ot if, along with the surface weather conditions. The finer horizontal resolution of the 0.64 micron imagery really shows how small individual convective elements in a lake effect event can be. Persistent northerly winds ensures a constant flow of moisture and snow into the central part of Indiana. Based on this view, there’s clear lake effect influence as far south as Kokomo, meaning that town is hardly a tropical getaway today.

However, lake effect events are notoriously shallow. It is likely that the lake effect snow is propagating even further to the south, but after it reaches a certain distance away from the radar, the angled radar beam is overshooting the top of the snow rendering it invisible to that radar. If you also include the radar from the more southerly Indianapolis office, you can see individual lake effect precipitation echos penetrating as far south as the northern suburbs of Indianapolis, over 120 miles away from the Lake Michigan shore.

Lake effect events in northern Indiana aren’t altogether rare. South Bend, Indiana, and Joliet, Illinois are two cities at similar latitudes and similar distances away from Lake Michigan. However, South Bend’s position downstream of the lake contributes to its much higher snowfall of nearly 65 inches a year compared to Joliet’s annual snowfall of only 16 inches. However, the inland penetration of today’s event is certainly unusual, thanks to a stronger than normal flow over Lake Michigan that is almost perfectly oriented to maximize the influence that this lake can have on the regional weather.

View only this post Read Less