While longtime residents of the Great Lakes region are familiar with lake breezes, their opposite, land breezes can also impact local weather conditions. In the spring and summer, the density difference in the air over the warm land and the cool lake induces a breeze that blows cold air ashore. In the late fall and the early winter, however, that relationship is reversed. The land cools off quickly as polar air masses plunge southward from Canada, but the lakes retain their relative warmth for some time into the winter. This creates a breeze in the opposite direction as air from the cold land surface flows offshore onto the (comparatively) warm lake.

On the morning of 8 December 2025, land breezes from multiple shores of Lake Huron contributed to the development of snow over the lake through the creation of a convergence line in the middle of the lake which fostered enough vertical lift to induce snow. Let’s take a closer look at this case, starting with an animation of GOES-19 Channel 2, the high-resolution visible band. Start by looking at the temperatures: surface temperatures for the shoreline weather stations surrounding the lake are in the low 20s or upper teens Fahrenheit. The lake itself, however, is much warmer as can be seen with the 40 F buoy observation in the middle of the lake. However, even if the buoy were not present, you could still assume that the lake is warmer than the surrounding land as it hasn’t frozen yet and thus must be above 32 F. Thus, conditions that support a land breeze (strong lake-to-land temperature gradient, the absence of strong synoptic forcing) are present. That can also be seen in the surface wind observations. The shoreline weather stations all show westerly flow, even though stations in the center of the state are almost all calm. The weather station on Drummond Island, just to the east of the eastern tip of Michigan’s Upper peninsula, also shows flow oriented toward the lake with a substantially more northerly component than other stations in the area.

However, it’s hard to really see the convergent wind pattern with just a handful of stations on the edge of the lake. This is a sparsely populated region (especially on the Canadian side of the lake) and thus weather observations are rare. Fortunately, there was an ASCAT (Advanced Scatterometer) overpass at roughly this time. Scatterometers measure wind speed by measuring the surface roughness of large bodies of water; stronger winds mean rougher seas and the ASCAT retrieval algorithm can take this roughness and convert it to wind speed and direction. ASCAT instruments are mounted aboard both American and EUMETSAT polar orbiting satellites. While ASCAT winds don’t have the temporal continuity of a geostationary product, the fact that multiple polar orbiting satellites are available means that midlatitude locations are observed somewhat regularly with these systems.

In this particular case, the convergence pattern could hardly be clearer: flow from the lower peninsula of Michigan is largely westerly while flow from Ontario is largely northerly. Where they meet, a line of clouds has formed. These clouds are largely held in place by the opposing winds, and thus the cloud band doesn’t propagate ashore.

Of course, RGB images can help us better interpret what is happening in this environment. The following animation shows the Day Cloud Phase Distinction RGB for this scene. There are a few interesting things to see in this view. One is the clear demarcation of surface snow (green) from the ice clouds passing overhead (red). Of greater interest to this current discussion, however, is the change of color in the cloud band over the middle of the lake. Where the convergence pattern is weaker the clouds are not as vertical and thus they tend to be warmer, liquid clouds that show up as lavender. However, where the convergence pattern is more robust, the resulting clouds are deeper and changing phase into ice clouds. This appears as the yellow-green clouds in the animation.

Finally, with all this vertical growth taking place, is it precipitating? The Multi-Radar / Multi-Sensor System (MRMS) is a good tool for this task as it integrates both United States and Canadian radars into a unified radar mosaic product. These lake-impacted cloud events tend to be shallow, and the standard NWS radars may be too far away to clearly identify what is happening. However, MRMS clearly shows the development of precipitation. Given that these are ice clouds and the surface temperatures are well-below freezing, it is almost certain that this cloud band is producing snow.

View only this post Read Less

5-minute PACUS Sector GOES-18 (GOES-West) Shortwave Infrared images (above) displayed the pronounced thermal signature associated with an eruption of Kilauea on the Big Island of Hawai`i on 06 December 2025. According to the Hawaiian Volcano Observatory, this Episode 38 of the ongoing Halema`uma`u eruption (Episode 1 started on 23 December 2024) began at 1845 UTC — and there were 3 roughly equal sized 500 foot (150 meter) high lava fountains, with 2 from the north vents and 1 from the south vent. Such a triple fountain is an extremely rare event, and this is the first time during this current sequence of eruption episodes that it has been observed. Episode 38 then abruptly ended at 0652 UTC on 07 December, after 12.1 hours of continuous lava fountaining. It is notable that the formation of a dense volcanic cloud (cold, darker blue pixels) masked the GOES-18 thermal signature of Kilauea from 2001-2206 UTC.

As early as 1856 UTC, the thermal signature exhibited a 3.9 µm brightness temperature of 137.88ºC (below) — which is the saturation temperature of GOES-18 ABI Band 7 detectors. This saturated Band 7 shortwave infrared brightness temperature was seen until 0651 UTC (the end of Episode 38).

During the ~2 hour period that the aforementioned dense volcanic cloud obscured the GOES-18 Shortwave Infrared thermal signature, there was a thermal signature seen in Next Generation Fire System (NGFS) imagery (below).

Later in the eruption, volcanic tephra and ash fall were observed in the town of Pahala as lower-altitude volcanic clouds moved southwest from Kilauea (below).

A larger-scale view of GOES-18 Ash RGB images created using Geo2Grid (below) showed the rapid eastward transport of 2 long plumes of volcanic cloud that were composed primarily of SO₂ (pale shades of green), in addition to the slower south-southwest transport of volcanic clouds that were composed of both SO₂ and a mixture of ash and SO₂ (shades of yellow). The dense volcanic cloud that emerged from the summit of Kilauea beginning around 2001 UTC, composed of both ash and SO₂, also moved eastward.

Plots of rawinsonde data from Hilo on the Big Island of Hawai`i (below) showed westerly winds above the 450 hPa pressure level, which transported the higher-altitude volcanic clouds to the east — and easterly to northeasterly winds within the 700-500 hPa pressure level, which carried lower-altitude volcanic clouds to the west and southwest.

GOES-18 True Color RGB images from the CSPP GeoSphere site (below) provided a closer view of the dense volcanic cloud that formed at 2001 UTC and moved eastward. In addition, an overshooting top was frequently seen directly over the eruption site.

View only this post Read Less

5-minute CONUS Sector GOES-19 (GOES-East) Nighttime Microphysics RGB + daytime True Color RGB images from the CSPP GeoSphere site (above) showed numerous power plant plumes (shades of red) embedded within a supercooled water droplet c  loud layer (shades of yellow) that covered much of Wisconsin on 02 December 2025.

The GOES-19 Cloud Top Phase derived product (below) confirmed that the cloud layer across the region was predominantly Supercooled (light green). The power plant plumes were classified as Uncertain (black).

An animation of GOES-19 Night Fog brightness temperature difference (BTD) + daytime Near-Infrared “Snow/Ice” (1.61 µm) images (below) also included plots of METAR surface reports — some of which showed light snow (denoted by the * symbol) in the the vicinity of the larger power plant plumes. These power plant plumes had the effect of eroding the supercooled water droplet cloud layer — via glaciation, initiated by the power plant emission of particles that acted as efficient ice nuclei, which then caused snow to fall from that portion of the cloud.

________________

One notable power plant plume in southern Wisconsin originated in west-central Dane County — and as this plume passed near/over the Madison airport (KMSN), the observer remarked that snow began at 1400 UTC and ended at 1451 UTC (above). The visibility at the 1453 UTC report time had improved to 8 statute miles — while the visibility at the Middleton airport, located just SW of KMSN, was only 4 statute miles at that time (Middleton is an AWOS site, which does not report precipitation type).

3 hours later, KMSN was reporting light snow with a visibility of 3 statute miles as the same power plant plume was passing near/over the airport (below) — while the visibility at Middleton was 2-1/2 statute miles (where it was also likely snowing). During the 1400-1900 UTC (0800-1300 CST) time period, intermittent reductions in visibility due to light snow were apparent in UW-AOS rooftop cameras facing north and northwest.

A closer view of the power plant plumes originating in Dane County is shown in a comparison of Suomi-NPP VIIRS True Color RGB and False Color RGB images valid at 1820 UTC, visualized using RealEarth (below). By that time, the leading edge of the Dane County plume had traveled northeastward over Columbia County; plumes from the Columbia Energy Center near Portage were also apparent.

________________

In northern Wisconsin, during the overnight hours a large power plant plume in Marathon County — likely the Weston Power Plant, located south of Wausau — was producing enough light snow to reduce the visibility to 1 mile at Wausau (KAUW), as seen in the GOES-19 0901 UTC Night Fog BTD image (below).

Since the Moon was in the Waxing Gibbous phase (at 93% of Full), it provided ample nighttime illumination to reveal that the power plant plume originating near Wausau was causing enough light snow to fall from the supercooled cloud layer that a hole (or at least a significant cloud thinning) had developed — as seen in VIIRS Day/Night Band imagery from NOAA-21 and Suomi-NPP (below).

During the subsequent daytime hours, the Marathon County power plant plume was drifting far enough northeastward to produce light snow that reduced the visibility to 3 statute miles at Antigo (KAIG) (below).

_________________

As an aside, another interesting feature seen in the imagery was a packet of cloud-top waves associated with a bore that was moving inland (westward) across east-central Wisconsin during the morning hours (above). This bore had no influence on surface wind direction, indicating that the feature was not surface-based.

A plot of rawinsonde data from nearby Green Bay, Wisconsin at 1800 UTC (below) showed a pronounced temperature inversion from 948-900 hPa (1038-2058 ft) — within which the bore was likely being ducted as it propagated westward.

View only this post Read Less

When you hear about the lake effect, you almost always think about the Great Lakes, whose impacts on local weather and climate are well-known across the upper Midwest and into the northeast. It may, then, seem odd to consider a lake effect event in North Dakota. However, if conditions are right, smaller bodies of water can also create lake effect snow.

Lake Sakakawea is a reservoir that was formed in the 1950s when the two-mile long Garrison Dam was built across the Missouri River for flood control and hydroelectric purposes. While it’s the country’s third largest reservoir by volume, its long, narrow shape lends it the form of a very wide river more than a traditional lake.

With a polar air mass plunging southward from Canada, surface temperatures in central North Dakota were very cold on the afternoon of 1 December 2025. Even though it was daytime on the first day of December, surface temperatures were in the single digits and numerous stations had the potential to set new records for lowest daily maximum temperature. This was a sharp contrast to a little over a week earlier; on the 23rd of November, the high temperature at Garrison was 57 F.

Therefore, while the land surface was very cold, the lake was still relatively warm as its large thermal mass inhibited its heat loss. This made for a perfect set up for a lake effect event as the following visible animation from GOES-19 shows. Lake Sakakawea is the dark snake-like form in the middle of the image, and clouds can be streaming to the southeast from the southern portion of the lake. Note that the winds in the region are largely westerly, ensuring enough fetch along the lake to produce the effect.

Recent snows over the region had previously blanketed the area in white, meaning there was very little visible contrast between the clouds and the snow. Since these clouds are also very shallow, they have very little thermal contrast with the surrounding land, and thus the clouds are very difficult to discern using single channel infrared imagery as well. However, certain RGB products are well-suited to discriminating between snow and cloud, as can be seen in this view of the Day Snow-Fog product. Here, the lavender of the low clouds pops against the red of the snow-covered land.

It’s important to note that, as its name suggests, this product can only be used during the day. This is because of its dependence on multiple shortwave channels. At night, the 0.87 and 1.61 micron channels have no reflectance, and the interpretation of the 3.9 micron channel changes from reflectance to emission. These animations were recorded late in the day and wold soon become unreliable for further interpretation. A better product to use after sunset is the Night Microphysics RGB seen below. In this case the land has cooled even further with sunset; the temperature at Garrison on the northeastern edge of the lake dropped to 1 F at the time of this animation. This served to deepen the lake/land temperature gradient and further enhance the lake-induced dynamics.

Throughout this blog post, this has been called a lake effect snow event. However, there is no radar evidence that any precipitation formed from this. This animation from the Bismarck NEXRAD radar covers the area were we’d expect to see snow, but there’s nothing there beyond the expected ground clutter.

However, observers are reporting snow downwind of the lake. That’s visible in the station plots depicted in the GOES-19 animations at the start of this blog post. The asterisk in the station plot for Hazel, ND, indicates that snow was being observed at the same time that one of these lake-enhanced clouds was directly overhead.

So why are weather observers reporting snow when the radar says that it is clear? The answer is simple: lake effect snow events are shallow, and these highly localized lake effect events are even shallower. The radar beam gets higher above the ground the further away you get from the radar because of the tilt of the radar beam and the curvature of the earth. It’s approximately 60 miles from the NEXRAD in Bismarck to the shores of Lake Sakakawea, which means the beam is roughly 4500 feet above the ground at that distance. These shallow clouds easily slip beneath the lowest radar beam and produce snow that is undetected by the radar.

View only this post Read Less