While tropical beaches may be the first thing that comes to mind when thinking about Hawaii, it’s important to remember that the Big Island of Hawai’i has significant elevation. The highest point, Mauna Kea, is over 13,800 feet above sea level. Even though it’s the tropics, this is a high enough altitude that occasional snow can be seen during the winter months

Om 9 February 2026, conditions were ripe for a notable snowfall. Large levels of moisture coupled with an upper-level disturbance created an environment that could easily support snow at high elevations. Let’s start by looking at the moisture around Hawaii via the CIMSS MIMIC-TPW2 product. It’s clear that Hawaii lies right in the middle of the flow of an atmospheric river linking the continental United Staes to the Equator. With almost two inches of precipitable water available, the potential for a significant precipitation event is real.

Upper level support comes from an advancing trough, which can be seen in the 6.19 micron water vapor imagery from GOES-18 as the strong gradient in brightness temperatures and the general direction of flow from the southwest to the northeast. Thus, dynamic lifting is present in this very moist environment.

The Band 13 (infrared wind) imagery confirms the development of deep, moist convection. The tops of these clouds have brightness temperatures well below freezing, so snow is being formed here.

We can further confirm the cold nature of these clouds by looking at the day cloud phase distinction RGB, where the yellow clouds are indicative of thick, deep clouds in the ice phase.

We have said that it was moist. But how moist was it? The 1200 UTC sounding from Hilo (on the Big Island) can provide some insight here. This sounding was obtained from the University of Wyoming Radiosonde Archive and shows a deep layer of saturated air stretching from the surface to above the 400 mb level. The freezing level was at 640 mb (3900 m, or around 12,800 feet) and was well-below the maximum elevation of Mauna Kea.

The CIMSS Satellite blog, of course, is a strong proponent of using NUCAPS to help diagnose the thermodynamic conditions via satellite. However, this particular case exhibits some of the challenge of using satellite-based observations of temperature. The following image shows the gridded NUCAPS temperature at 700 mb. Note that the 700 mb temperature over the Big Island is right at freezing, and, givne that this is only 700 mb, the highest parts of the mountain would be well-below that critical temperature. However, the previous satellite images show that this is where the clouds are at their thickest, and precipitaiton is likely here. The colored dots representing profile quality are largely red, indicating that the NUCAPS retrieval results in this area are likely to be error-prone.

Given all of these conditions, the local weather service office issued a Winter Storm Warning for the high elevations of the Big Island. Such an event may be rare compared to the NWS offices in the upper midwest, but it’s not unheard of. Thanks to the fantastic archive at the Iowa Environmental Mesonet, it’s possible to track just how often a Winter Storm Warning is issued by a particular office. Some of these events have multiple warnings issued so it’s not trivial to connect the number of warnings to the number of significant snow events, but we can at least be sure that most years have at least one noteworthy snow event in Hawaii.

And, on a personal note, this post on Hawaiian weather is dedicated to NOAA NWS Honolulu meteorologist, native Hawaiian, and UW-Madison alumnus Will Ahue, who recently passed away. He was a friend to everyone he met, including those of us here at CIMSS. He will be deeply missed.

View only this post Read Less

As a low pressure system was located off the US East Coast on 07 February 2026, a so-called “Norlun Trough” (named after meteorologists Nogueira and Lundstedt) or “inverted trough” extended northwestward from the offshore low center to southern New England (above).

10 hours of 1-minute Mesoscale Domain Sector GOES-19 (GOES-East) Infrared images (below) showed the development of convective cloud bands (brighter shades of green) within the area of surface convergence in the vicinity of the Norlun trough. Surface convergence was also enhanced by an approaching arctic cold front (which eventually merged with the Norlun trough). These convective cloud bands helped to enhance snowfall rates.

5 hours of 1-minute GOES-19 Infrared images that included County outlines and County names (below) helped to identify the areas where the highest snowfall accumulations occurred (which as of 2002 UTC included 13.0″ in Essex County in far northeastern Massachusetts, 9.8″ in Washington County in far western Rhode Island and 8.0″ in Windham County in far eastern Connecticut).

A map of final Snowfall Totals on 07 February is shown below, depicting the highly-localized maxima across northeast Massachusetts, southeast Connecticut and southwest Rhode Island. These final snowfall totals included additional accumulation (after 20 UTC) from ocean effect snow.

View only this post Read Less

A recent question from the Weather Service Office in Pago Pago, Amercian Samoa, brought up the issue of parallax. GOES-18 is located at a longitude of 137 degrees west, while American Samoa is at 170 degrees west. This means there’s around 33 degrees of longitude between the two. By comparison, this is similar to viewing Billings, Montana, from GOES-19 (East) or Denver, Colorado from GOES-18 (West). This results in a significant amount of parallax. This CIMSS Satellite Blog post from 2006 (twenty years ago!) gives a great overview of parallax and its issues. However, we’ll briefly summarize parallax here: the farther away from nadir the satellite is viewing, the more oblique the scan angle. For deep clouds, this means the satellite is more likely to see the side of a cloud rather than the top as the cloud gets closer to the edge of the satellite’s field of view. This also means that the position of the cloud will be incorrectly placed. Check out the following cartoon:

The red house is in the cloud’s shadow, while the blue house is in clear sky. However, because the satellite is viewing the cloud at an angle, the position of the cloud is mapped to a different location than where it actually is. From the satellite’s point of view, the red house is actually in clear sky while the blue house is beneath the cloud. Imagine the surprise of the Blue House citizens when they’re told that it’s cloudy overhead!

American Samoa is far enough west that it is within the field of view of Japan’s Himawari-9 geostationary satellite, too. That satellite is located at a longitude of 141 degrees east, putting it about 49 degrees of longitude away from American Samoa (similar to viewing Seattle from GOES East). Given the deep convection that is frequently found in the tropical Pacific, it’s interesting to compare the two perspectives. The following image pair shows American Samoa, the independent nation of Samoa, and the tropical Pacific at the same time using the True Color RGB from both Himawari-9 (left) and GOES-18 (right). You can drag the slider left and right to see how the change in satellite changes the apparent positions of the clouds.

There are a couple of interesting things to note in this set of images. One of the most glaring (literally!) changes is the difference in sun glint. This occurs when the solar angle over a particular location in the ocean is the same as the satellite viewing angle of that location. The ocean acts as a mirror and reflects the sun’s light into the satellite’s imager. The geometry is just right that GOES-18 detects a significant amount of sun glint at this time. However, because Himawari-9 is in a different position, it doesn’t detect any glint even though it’s imaging the same location at the same time. The angles just don’t work to produce the same result.

The other interesting thing to note is the different impacts of parallax depending on cloud height. Low clouds, like the developing cumulus in the lower right of the images, are largely unaffected by parallax since they’re so close to the ground. High clouds, by contrast, see significant displacement between the two satellite views. This is easily seen in the animation below, in which the small cumulus clouds in the upper right show much less of a spatial change than the upper level cirrus streaks seen elsewhere in the image.

Of course, these location differences are also seen in the infrared imagery as well. In fact, the differences can appear to be even larger in the infrared than in the visible. This is because clouds tend to be white whether you’re seeing their tops or their sides and so it can be difficult to discern what part of the cloud a satellite is seeing.. However, because the temperature of deep convection changes dramatically from bottom to top, the IR signatures will look quite different from the two perspectives, as you can see with the slider below. Again, these are images from the same time and location, just differing from the angle at which they were taken.

The next effect of parallax is to make clouds appear farther from the sub-satellite point than they actually are. This effect is negligible for shallow clouds or for locations that are near the sub-satellite point. But for locations that are on the edge of a satellite’s field of view, like in higher-latitude regions or those, like American Samoa, which are tropical but far away from the sub-satellite point, deep convection can experience significant parallax.

2 blog posts that mention the effect of GOES-18 parallax for deep convection (that produced heavy rainfall) over American Samoa are here and here.

View only this post Read Less

As a strong arctic cold front moved southward across the Gulf of Mexico toward southern Mexico on 04-05 February 2026, the cold front fractured as it moved inland across Mexico’s Isthmus of Tehuantepec — the cold air was then channeled southward through Chivela Pass and emerged as a Tehuano (or “Tehuantepecer“) gap wind that eventually fanned outward across the Gulf of Tehuantepec and adjacent Pacific Ocean. 10-minute Full Disk scan GOES-19 (GOES-East) Near-Infrared images (above) showed the hazy plume of dust that was being transported offshore — along with a narrow arc cloud that marked the edges of this Tehuano flow.

The pulse of Tehuano winds emerging southward across the Gulf of Tehuantepec and the adjacent Pacific Ocean was seen in ASCAT winds (source) from Metop-B and Metop-C (below).

The highest Metop-B wind speed was 40 kts (below).

At the leading (southern) edge of the Tehuano flow, a ship reported NE winds gusting to 35 kts at 1800 UTC (below).

The broad plume of dust lofted by Tehuano winds was apparent in True Color RGB images (source) from both GOES-18 and GOES-19 (below).

A relatively narrow smoke plume was seen near the middle of the broad dust plume — and a closer look using 5-minute CONUS Sector GOES-19 GeoColor RGB images with Next Generation Fire System (NGFS) Fire Detection polygons (below) showed the larger/hotter wildfire that was responsible for producing this smoke plume.

Just south of the Pacific coast of Mexico, wind-driven significant wave height values derived from SWOT were as high as 9.55 ft at 2321 UTC (below).

‘

View only this post Read Less