Humans can only dream about such power…

We’ve all heard of the butterfly effect, how an insect flapping its wings in the Amazon could theoretically impact the weather at the Hokkai Bellybutton Festival. Chaos theory says that while such a relationship exists, we can never truly understand or predict it.

But on the large scale, connections between weather and climate systems are well studied. For example, teleconnections are the focus of studies examining how large-scale changes in one area can affect large-scale climate elsewhere (such as the hubbub around the current super El Niño). And then we have the story of Typhoon Halong, which formed southeast of Japan but ended up causing severe damage and the evacuation of over 1,500 people in Southwest Alaska.

Typhoon Halong for the Ride

Typhoon Halong formed in the warm waters north of Guam in October 2025. After a deke to Japan, it headed northeast towards Alaska.

A tropical cyclone’s primary energy source is the ocean’s surface. The water humidifies and warms the air, creating uplift that initiates convection. They are known as warm-core systems because the core is warm at higher altitudes.

As a cyclone moves over land or colder water, it may transition into an extra-tropical cyclone. Without access to warm ocean water, they draw their fuel from baroclinic instability, which relies on potential energy from strong temperature gradients between warm and cold air masses (fronts). For Halong, this happened on October 10, 2025, about 6 days after formation and 2 days after reaching peak strength. This was a transitional stage as it turned into a cold-core system.

Typhoon Pheonix

Usually typhoons weaken after they transition to extratropical storms. Halong had other plans, strengthening for roughly two days straight after the switch. Two things fed that. First, it crossed anomalously warm water, which pumped extra moisture and energy (heat) into the storm. Then it got wedged between another extratropical cyclone parked near Kamchatka, a landmass most people only know from Risk™, and a ridge to its east. A jet stream squeezing between those two systems steered, strengthened, and sped up Halong. This re-intensification rate ranks in the top 4% of warm-season Arctic Cyclones of Tropical origin.

When it reached Alaska, a wind-driven storm surge lifted houses off their foundations, wiped out water systems, and damaged airfields. More than 1,500 people had to be airlifted out of coastal areas and could not return until the Spring.

This rare storm caught the attention of Mingshi Yang and a team of researchers out of the University of Illinois Urbana-Champaign, the University of Alaska Fairbanks, the Naval Research Laboratory, and NSF's National Center for Atmospheric Research. They studied it using existing computer models run on archival data (“reanalysis”) and customized models that allowed them to isolate the impact of ocean surface temperatures. Their hunch was that a warming ocean may have amplified this storm. When the researchers ran their simulations with the ocean-warming signal stripped out, Halong's peak intensity weakened noticeably, confirming that the warm water was critical. They recently published these results in the Bulletin of the American Meteorological Society.

They also looked at historical data for evidence of an evolutionary trend in the frequency and paths of these storms. They split the record into two eras, 1940-1979 and 1980-2025, and mapped where these storms were born and how often tropical-origin systems decided to give Santa a visit.

More to Come

They found storms with tropical origins that make it all the way into the Arctic have increased 3-4 times in frequency (depending on the month studied), compared to 1940-1979. And the areas where they form have shifted from a tightly clustered zone south of the Bering Strait to a much more sprawling one that reaches deep into the subtropics. It is likely due to a stronger upper-level jet ferrying storms from lower latitudes north faster than it used to.

Combine that with warmer sea surface temps, and this type of thing may become more common. Halong is one of only 3 (ex)typhoons to have taken this journey this century. But all three of those were in the last 4 years. It’s another example of how, with weather systems changing so rapidly these days, global monitoring is important. What happens in one area can have major impacts on another.

And Now for Something Completely Different

Sesame Street recently aired a special on the topic of preparing for an upcoming big storm. Storm on Sesame Street can be viewed via PBSKids and Netflix with guests Post Malone and Michelle Buteau. It’s the first phase of a broader initiative by Sesame Workshop about extreme weather.

It got me thinking about weather resources for young children. Friend-of-the-newsletter Alan Gerard recently wrote eloquently about a book that was very influential in his life, Katy and the Big Snow.

My colleague (and frequent WWAT scientific reviewer) Beth Mills authored a book about hurricanes aimed at K-3rd graders. My daughter loved it.

Starting in the new year, AMS is launching a new initiative to create fun, free after-school weather activities for children of all ages, beginning with the younger ones. I’ll share updates when our first activities are ready.

We acknowledge additional contributions from Chris Nabholz, Greg Porter, and Beth Mills. We are grateful to Lockheed Martin for a grant supporting this newsletter.

Our archive of WWAT articles is here.

During the preparation of this work, the author(s) used Claude Sonnet 5 and Grammar.ly to copyedit text and Gemini 3.6 Flash to create a simplified version of the typhoon track map. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Weather with a Twist is published by the American Meteorological Society.

A waterspout is a tornado over a body of water. Check out their beauty:

There are two types of waterspouts: scary and non-scary. More scientifically, they are either associated with supercells or not. Supercell waterspouts are similar to tornadoes we see on land and are just as dangerous. Non-supercell waterspouts (a.k.a. fair-weather waterspouts) form beneath cumulus clouds during rapid growth. They are short-lived and typically have lighter winds (<EF0).

Birth of a Waterspout

In general, tornadoes form when rapidly rising air encounters wind shear (wind that changes direction and speed with altitude). It begins with a strong thunderstorm creating a column of rapidly rising air. As that column meets winds blowing in different directions at different heights, it starts to spin and takes on a life of its own. Supercell waterspouts form the same way.

Fair-weather waterspouts, by contrast, form from existing spin already present near the water's surface (similar to a dust devil) that is stretched upward as a developing cumulus cloud pulls air into it. This doesn't require the strong shear or a supercell that a tornado needs. And with the water surface being smooth, there is less friction to weaken the vortex and more uniform heating to support the spin. Also, it’s easier for its spin to pick up particles of water compared to dirt and debris, making waterspouts more visible. Thus, fair-weather waterspouts are more common than supercell waterspouts. They are actually quite common in some parts of the world, such as off the Florida Keys.

The Sailor’s Omen

Sailors are famously superstitious, perhaps as a psychological way to deal with a dangerous world over which they had little control. And they absolutely hated waterspouts. Old maritime lore is full of crews trying to destroy waterspouts by throwing knives or even firing broadsides at them. (Maybe the Navy was actually onto something…)

It sounds like pure superstition, but there's a sliver of physical plausibility buried in there. A fair-weather waterspout is a fragile thing. It usually dissipates within seconds of touching down because of surface friction. Perhaps a sufficiently violent disturbance near its base, especially in the early stages of formation, could weaken it? Whether cannon fire has ever actually done the job is impossible to say in hindsight. Either way, it was probably fun to try.

It’s Raining Fish

I first fell in love with waterspouts as a kid when I learned they are likely the source for the various it’s-raining-fish/frogs legends out there. There are plenty of documented reports of animal rain. But whether it’s due to waterspouts or other natural mechanisms is subject to debate.

Let’s take a deeper look at one such incident of fish falling during a thunderstorm over Texarkana, TX, on December 30, 2021.

I reviewed various records to determine whether we could find any evidence that a waterspout was the culprit. First, a strong thunderstorm did pass through the region that day. Also, there is also a source of fish upwind of the storm’s path - Wright Patman Lake. The falling fish were reported to be young white bass, which are indeed stocked in the lake and “very popular” among fishermen.

Let’s look at the radar data to see if there was any evidence of a waterspout and/or flying fish. We got help from Alan Gerard, a friend of the newsletter and author of the Balanced Weather Substack publication. Alan is a recently retired meteorologist with over 35 years of experience working for the National Weather Service.

He found radar evidence of possible storm rotation, of the type associated with weak tornadoes and waterspouts, above a northern finger of the lake.1 However, differential reflectivity and correlation coefficient do not show clear signs of any debris or other non-meteorological objects (i.e., flying fish). However, the radar beam was about 5,000 ft above ground level at that point2 and such debris would likely be lower. It would take a stronger tornado/waterspout to lift fish that high.

The Birds

However, there’s an alternative explanation. The Skeptoid podcast reports that ornithologists studied some of the unlucky fish and found signs of bird feeding and digestion. They suggest that birds had fed on the fish at the lake, then regurgitated them as they were stressed by the incoming storm. This is common with the types of birds seen flocking in the area before and after the storm. Though no one actually saw birds vomit fish at them. So that made me wonder if the birds fed afterward. Anyone who lives near a beach, as I do, knows how quickly birds grab free food.

Nevertheless, I’m probably 60% on the bird side and 40% on the waterspout side. We have decent evidence of a vortex over the lake and thousands of years of eyewitness reports of waterspouts tossing fish and other small marine critters onto boats and land. But on the bird side, we have direct evidence of feeding and a plausible mechanism (regurgitation to fly faster). Occam’s Razor is probably on the side of birds also, since it has less uncertainty (it explains how the fish got to altitude; AFAIK, no one has done the physics on waterspout fish elevation).

This is a fascinating challenge, even if bird vomit doesn’t quite measure up to the heights of The Great Debate. I’ll keep an eye on the news and look for other opportunities to investigate.

The scientific question isn't whether animal rain happens. That's well established. It's about the mechanism. There's plenty of documentation linking animal rain to waterspouts, but skeptics can always point to the lack of direct documentation (video, for instance) to fully close the loop. One Australian report, though, described fish falling encased in ice. In that case, it certainly wasn't a bird. Presumably, the fish was carried high enough to reach the cloud's freezing layer and picked up a coating of ice on the way down.

The debate is not new. This delightful 1946 letter to the editor of the journal Science reports on a 2,000+ year literature review on the subject. And this 1972 paper in Australia puts together almost 100 years of reports from down under, including many linked to tornadoes and waterspouts.

The author also supplied that bespoke illustration of raining fish in Australia. He finishes his article with a classic line: “We still have much to learn about these extraordinary occurrences.”

Our archive of WWAT articles is here.

During the preparation of this work, the author(s) used Claude Sonnet 4.6 to copyedit and fact-check. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Weather with a Twist is published by the American Meteorological Society.

About 99% of the liquid freshwater on Earth is found in groundwater. It supports about 2 billion people on the planet. During droughts, it’s especially important since it can augment low levels of surface water. But what happens when a surface drought reaches the depths below?

Out of Sight, Out of Mind

Groundwater drought is understudied compared to meteorological and surface water drought. Mostly because we simply don’t see it. Rivers and lakes dry up right in front of us, with immediate consequences. But deep down, only the mole people and CHUDs are eyeing dropping aquifer levels.

Underground water systems are complex, highly heterogeneous, and not easily accessible. To track their levels over time, scientists have resorted to using indirect measures such as their impact on the Earth’s gravitational field. And, heaven forbid, they even revert to computer models (a.k.a. educated guesswork). Both come with high levels of uncertainty due to the assumptions that go into their analysis. So, when possible, everyone prefers good old-fashioned direct measurements.

If only there were a vast industry of people who like to dig holes to find water...

Of course there is! The International Groundwater Resources Assessment Centre (IGRAC) has spent years consolidating data from tens of thousands of wells across North America and Europe into a single resource. The result is the Global Groundwater Monitoring Network (GGMN), a UN-backed, web-based platform that makes this data freely available worldwide.

Scientists at the University of Bristol (U.K.) and Centre for Climate Research Singapore used that dataset to map how meteorological drought leads to groundwater drought. They published their results recently in The Journal of Hydrometeorology.

Well…

After filtering the data from over 55,000 wells, they ended up with 3,274 that passed their data quality filters. They used the Standardized Precipitation Index (SPI) for meteorological drought, and its groundwater analog, the Standardized Groundwater Index (SGI). Both express how far conditions depart from a site's own long-term normal, allowing comparisons across different climates. They paired each groundwater drought with a corresponding meteorological drought (same location but started earlier) and then computed ratios for duration, intensity, and severity.

They analyzed the data in two stages. First, they looked at spatial differences between meteorological and groundwater drought. Then, they examined the properties (e.g., intensity) of a meteorological drought and how they changed as it moved underground.

A Slow Creep

Duration and severity of groundwater drought varied by more than an order of magnitude across sites, whereas the corresponding meteorological droughts varied much less. The ways these two types of droughts varied were also different. The correlations were weak across the board. This points to groundwater's response being influenced more by local factors than by the factors that led to the preceding meteorological drought. It’s location, location, location.

The median lag time between onset of meteorological drought and groundwater drought was almost 6 months. The depth of the water table was a key predictor in drought onset and duration. The shallower the depth, the more similar the groundwater drought was to meteorological drought.

So, in short, meteorological droughts lead to groundwater droughts, which tend to be less intense but last longer. As water drought conditions move into the Earth, their impacts slow down and spread out. Think of the water you pour onto the soil of a houseplant. It slows and spreads, with some of it eventually trickling out of the holes in the bottom.

Research like this helps water managers predict which regions can rely on groundwater as a drought buffer and which can't, helping plan water allocation strategies as climate change and human development make droughts more frequent.

And Now for Something Completely Different

Our last issue was about clear air turbulence. Unbeknownst to us, a week later the Aviation Weather Center was to release a major upgrade to the Graphical Turbulence Guidance Nowcast - a name only a government bureaucrat could love. It’s designed to help pilots avoid turbulence by creating maps like these, updated every 15 minutes with ~3 km resolution:

The AWC smartly designed the system to be used with a pilot’s cell phone. Yet it’s still full of all sorts of fun popups and toggleable layers. It operates in an unholy mix of meteorological and pilot speak which can get quite technical.

The above report is from a pilot. Using a secret Ovaltine Decoder ring, it says that at 10:14 am local time (1514Z), a 737 MAX 8 roughly 10 nm west of a specific navigation aid, cruising at 20,000 ft, reported light to moderate “chop”. Chop specifically refers to rapid, rhythmic bumpiness without significant altitude change, as opposed to turbulence that sends food into the ceiling.

Personally, I enjoy turbulence. It's like a massage that quickly puts me to sleep. But I recognize that's not exactly a normal reaction. For many, it can be unsettling, and for pilots, it's a genuine operational hazard. Tools like this help them navigate around the worst of it, and they give anxious flyers a small but real measure of control over an experience that often feels entirely out of their hands. I’m really amazed pilots do this reporting manually, with no incentive other than to help each other out. It’s a great example of a professional community at its best.

We acknowledge additional contributions from Chris Nabholz and Beth Mills. We are grateful to Lockheed Martin for a grant supporting this newsletter.

Our archive of WWAT articles is here.

During the preparation of this work, the author(s) used Claude Sonnet 5 to copyedit text and Gemini 3.6 Flash to add the bottle of water to the mole person. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Weather with a Twist is published by the American Meteorological Society.

One of their many concerns is turbulence, and since there's no good way to detect it remotely, they rely on reports from actively flying aircraft. Some planes have turbulence detectors on board that send data directly to places like the AWC, but most do not. To fill that gap, pilots voluntarily send in verbal turbulence reports when they actually feel it. It's sky-truthing! However, these systems only cover areas where participating planes/pilots have recently flown. It’s not of much use to those flying less populated routes.

Clear Air Turbulence

Flight turbulence is mostly caused by shifting winds, especially vertical ones. This, mostly convective, turbulence near storm systems is relatively easy to detect via radar and satellite. But clear-air turbulence, the kind that happens far from any storm, is much sneakier since there’s no cloud or moisture for those tools to pick up, and can also hit in the middle of a smooth flight through clean, cloud-free air.1 This means there is little to no warning. As a result, CAT is responsible for 28% of turbulence-related injuries.

Current CAT detection options are all inadequate in some way. LIDAR works (shooting lasers from the noses of aircraft) but is expensive and power-hungry. Microwave radiometry and multi-sensor fusion approaches are in development but are still years from mass deployment. Numerical forecasting has improved but remains too computationally expensive to run at the resolution needed for real-time, aircraft-scale predictions. So what to do?

Infrasound

Instead of shooting lasers at the turbulence, what if we just listened for it?

Infrasound (a.k.a subsonic) is sound below the range of human hearing, roughly <20 Hz. Infrasound travels for extremely long distances with very little energy loss. It is already used to monitor other geophysical phenomena like volcanic eruptions. Like us, most animals cannot hear it, but homing pigeons may use infrasound to orient themselves before taking flight.

Severe thunderstorms also emit infrasound. Though the source mechanism is still debated, one theory is that pressure changes caused by vorticity and wind shear (like in a tornado) generate infrasound.2 Researchers at the University of Waterloo wondered whether CAT, with its own pressure fluctuations, also would produce detectable infrasound. They recently published a study on this in the Journal of Applied Meteorology and Climatology.

The Simulation

The team anchored their study to a real, well-documented event: an Emirates A380 (flight UAE216) that hit severe CAT over mountains in Wyoming, on 2 February 2020. The plane experienced vertical g-loads from 0.44 to 1.63 G over less than two minutes. One passenger was hurt with a fractured ankle, likely occurring while washing their hands in the lavatory.

The first step of their study was to simulate the conditions of that event. They used data from the HRRR weather model and (Space) Shuttle Radar Topography Mission (SRTM) terrain database at a horizontal resolution of just 111 meters (super small in the meteorological community). They were able to simulate the event at essentially the right time and place and deduced it was likely caused by a breaking gravity wave over a mountain.

With a physically realistic turbulence event in hand, the research team then applied an acoustic modeling framework originally designed for storms. Under the assumptions relevant to CAT (dry air, no moisture-related sources, negligible buoyancy, etc.), the dominant acoustic source reduces to one term driven purely by turbulent velocity fluctuations. This yielded seizure-inducing equations like thus:

That equation is their justification that any sound detected will be due to turbulence and nothing else atmospheric-related.

Next, they computed the resulting sound pressure at 13 virtual “microphones” arranged in a semicircle 40 km from the turbulence, then compared that signal to sound generated by a nearby, calmer “no-CAT” region as a control.

You Can Hear CATs (we think)

The CAT region produced dramatically stronger acoustic sources than the calm comparison region, on average more than two orders of magnitude larger. This translated into overall sound pressure levels more than 20 dB higher than the background.

Math and physics teachers will love the above graphs. Side-by-side, they look the same, but you gotta pay attention to the y-axis units. The turbulent air is producing pressure swings roughly 25 times larger than the quiet air beside it.

Finally, they ran a back-of-the-envelope thought experiment: if a detector sat 80 km from a similar CAT patch, the model suggests you’d get roughly 4 minutes of advance warning before an aircraft cruising at typical speed would reach it. That would give us 4 minutes to finish washing our hands, grab our seat, strap in, and enjoy the ride.

Caveats: The simulated turbulence occurred about 3 kilometers higher in altitude than where the aircraft actually reported the encounter. The authors suggest this is due to known sensitivities in modeling mountain-wave turbulence and not the infrasound modeling.

What’s Next

Hopefully microphones listening to the sky and warning us of hidden turbulence! But before we go from here to there, this solid simulation needs to be tested using real data. For example, the simulation doesn’t account for some real-world noise, like refraction and convection, between the source and the receiver. More sophisticated simulations and, ideally, real data from the ground (ground-truthing the air-truth) are the next steps.

And Now for Something Completely Different

In April, WWAT covered work on weather modification. Subsequently, a reader let us know about a book titled The Brothers Vonnegut, by Ginger Strand. I devoured it.

Last year, I happened to read Kurt Vonnegut’s book Cat’s Cradle, which has at its core a fictional substance called Ice-Nine. It was a form of solid water that would equally solidify all liquid water in contact with it. You can imagine the shenanigans an imagination like Kurt’s could devise with a MacGuffin like that. The book ends with a scene about its impact on weather. I went down a rabbit hole of trying to find out whether the ending was scientifically accurate (at the time more so than now).

Kurt’s brother, Dr. Bernard Vonnegut, was one of the pioneers of weather modification. He worked with Vincent Schaefer and Nobel laureate Irving Langmuir in the 1940’s to develop methods to stimulate rainfall from clouds. His particular invention, using silver iodide, is the main substance used today.

The book focuses on the development of their respective careers and how they influenced each other. They had a relatively strong bond as brothers go, with a dash of sibling rivalry beneath a lot of mutual respect. My only nitpick: the book ends suddenly, as almost all biographies do. (Why is that? Why do biographers eschew solid codas?) But overall, a great read.

We acknowledge additional contributions from Chad Kaufman, Greg Porter, and Chris Nabholz. We are grateful to Lockheed Martin for a grant supporting this newsletter.

Our archive of WWAT articles is here.

During the preparation of this work, the author(s) used Claude Opus 4.8 to copyedit text and help explain the pressure time series figure. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Weather with a Twist is published by the American Meteorological Society.

It also floats in the air, which makes it natural territory for meteorologists. Forecasting wildfire smoke has been around for a long time. But it's been the runt of the weather forecasting litter. Until recently, arguably less effort has been put into it than, say, studying the impact of weather on social media sentiment.1

Happily, that’s been changing. In the last decade or so, researchers (and NOAA) have invested much more in studying wildfire smoke forecasting. This is largely due to events like this:

I imagine tobacco executives, on the phone with each other late at night, telephone cable curling between their fingers, dreaming of days like that.

Increasing Fires, Increasing Smoke

It may be tempting to set up a huge NO SMOKING sign across the continental United States every summer… but wildfires are both normal and critical to a healthy ecology. The problem is that they are becoming larger.

And graphs like that lead to graphs like this:

After decades of improvement since the passing of the Clean Air Act in 1970, the average air quality over the United States has begun to decline once again. Anthropogenic pollutants are a contributing factor, but the main change over the last couple of decades has been the increase in wildfire emissions. In the last two decades, they have erased about a quarter of the air quality gains since 1970.2

Forecasting Smoke

Right now there are two main wildfire smoke forecasting tools. One is fire.airnow.gov, a collaboration of numerous government, academic, and commercial organizations that combine resources to produce a map of fires and smoke over CONUS. It also includes a daily forecast of overall air quality. (I use it for scheduling outdoor runs during fire season.)

The other is the High-Resolution Rapid Refresh (HRRR) model. HRRR is a high-resolution computer model released in 2014 by NOAA to provide 3km-resolution forecasts of key atmospheric variables (temperature, pressure, wind speed, precipitation, clouds, etc.) about once an hour or so. It’s a well-respected model. Many mobile weather apps use it as their source for near-term forecasts. In 2020, HRRR was upgraded to add data from satellites, ground stations, and additional atmospheric chemical simulations to create smoke forecasts <48 hours out.

2023 was the first year since the upgrade when CONUS experienced major wildfire smoke activity across the country (including the Northeast, where it is relatively rare). Researchers at the Johns Hopkins University Applied Physics Laboratory used data from that summer to evaluate the performance of the HRRR’s forecasts and published their results in the journal Weather and Forecasting.

The team pulled HRRR-Smoke forecasts for the entire continental US every six hours from June through August 2023, at lead times from 0 to 48 hours. They then compared those forecasts to three independent observing networks:

• AirNow.gov supplied an Air Quality Index (AQI) based on PM2.5 (particles <2.5µm in diameter) near the surface

• MODIS, the spectroradiometer flying on Aqua and Terra satellites, supplied aerosol optical depth, a measure of how much haze is stacked in the entire vertical column above a point. (a proxy for how smoky it looks outside.)

• AERONET, a global network of ground-based photometers that track the Sun, supplied a second independent check of optical depth.

OK on When, not-so-much on What

The HRRR did okay forecasting the timing of the smoke events. Tracking eight cities across the entire summer, the model predicted the onset of major smoke intrusions in places like Minneapolis, Buffalo, Newark, and Washington, D.C., to within roughly a day of when they actually arrived.

However, while the system was okay at predicting timing, it was not so great at predicting smoke intensity, largely because the model underpredicted aerosol optical depth (relative to MODIS) more than 90% of the time. The failure was worst over central New York on June 6, where smoke was so thick that MODIS's own satellite algorithm couldn't even return a reading.

The authors attribute the intensity failure to a structural limitation: HRRR carries a single tracer for smoke for the entire transport model. Smoke particles of different sizes, ages, and injection heights each disperse and settle differently. So we’re asking too much from a single tracer.

Goodbye to a Friend

Later this year (2026), the HRRR is planned to be put out to pasture, grazing for CPU cycles and making baby computer models.3 It is somewhat sad because it was an important piece of infrastructure and a big leap forward at the time. It will be replaced by the RRFS (Rapid Refresh Forecast System), NOAA's next-generation short-range weather model. It carries three aerosol tracers instead of one (smoke, fine dust, and coarse particulate matter), along with other general enhancements. This should theoretically improve smoke forecasts by offering a more complex tracing of the atmosphere. This paper empirically establishes a baseline methodology for HRRR performance that can be applied to the RRFS. We look forward to writing a WWAT article in a few years summarizing the results. 🤞

And Now for Something Completely Different

Many moons ago, in a creative writing class, we were doing poetry. I was finding the free-flowing forms challenging. Then the professor assigned us to write a terzanelle, which is highly structured. I took to it like a hipster to Wes Anderson and, in doing so, learned a valuable lesson: sometimes rules are fun.

Recently, I was reacquainting myself with the terzanelle and I found this web page, which I think describes in an entertaining way (forgive the background, which is either caused by or a symptom of a seizure). And I was happy to see that the author included a weather-themed poem as their “good example.”

Terzanelle in Thunderweather
by Lewis Turco

This is the moment when shadows gather
under the elms, the cornices and eaves.
This is the center of thunderweather.

The birds are quiet among these white leaves
where wind stutters, starts, then moves steadily
under the elms, the cornices, and eaves--

these are our voices speaking guardedly
about the sky, of the sheets of lightning
where wind stutters, starts, then moves steadily

into our lungs, across our lips, tightening
our throats. Our eyes are speaking in the dark
about the sky, of the sheets of lightening

that illuminate moments. In the stark
shades we inhibit, there are no words for
our throats. Our eyes are speaking in the dark

of things we cannot say, cannot ignore.
This is the moment when shadows gather,
shades we inhibit. There are no words, for
this is the center of thunderweather.

The point of the terzanelle is that the stanzas interweave, as you can see above. It reminds me of rolling shelf clouds like these.

We acknowledge additional contributions from Greg Porter and Chris Nabholz. We are grateful to Lockheed Martin for a grant supporting this newsletter.

Our archive of WWAT articles is here.

During the preparation of this work, the author(s) used Claude Opus 4.8 to copyedit text and create the fire acreage chart. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Weather with a Twist is published by the American Meteorological Society.

Current Status

We currently have about 6,600 subscribers. Which we’re very thankful for! We’ve done almost no marketing or advertising of WWAT. Almost all of that is from word-of-mouth from other newsletters. However, growth has definitely leveled off in the past six months.

But so has our productivity. It says something about American culture that the most popular idle chit-chat usually starts with: “How’s work?” Followed by: “Busy”. In fact, I decline to say “Busy” now. My inner emo-teenager1 refuses to conform, and now I reply with “my afternoon naps are now down to 20 minutes/day”.

Yet, as so often happens, reality rears its ugly head, and I have been distracted from my favorite part of the job (this!). But they’ve all been good things, actually. So complaining feels like…

Seriously, I hope to get back to an issue every 10 days. The dream is one per week… but we can’t achieve our dreams. Then they wouldn’t be dreams, would they?

No one has ever asked me, “How do you come up with the article topics?” So let me tell you. The awesome AMS Publications Department publishes over 12 research journals, which are combined at journals.ametsoc.org, a la the bad guys in Superman II. To help break them out, I start at journals.ametsoc.org/browse/.

Seriously, if you are a weather nerd then check out those links. You’d be surprised by how many of the articles are open access and free to read. If you’d like some less technical but still scientific reads, check out our new AMS Headlines portal.

The Future

I had a grand plan of organizing a cadre of authors who would take turns writing these newsletters. The hope was that different voices would have different truths to tell, and different memes to exploit. However, it turns out no one really wants to write mediocre copy about computer models, obscure scientific instruments, and explain how raindrops form for the 100th time.

But we do have a wonderful team of scientists, our Council of Doom, who review every issue. I am super thankful to all of them. They get almost nothing out of it, other than the satisfaction of blocking the worst of my misinformation from making it into the world.

One Last Favor - Do you like money?

Part of my job is doing education research. Right now, I'm working with the venerable Concord Consortium on an NSF-funded project to design a simulation that teaches the general public how to read uncertainty in hurricane forecast graphics.

We're in the very early stages of that project, which involves developing a survey to measure the simulation's impact. We'd love it if you'd help us pilot test it. We're offering a $25 gift card to the next 200 people who take the survey, which we estimate will take 10–15 minutes.

[survey link redacted from web site archive]

And, for fun, check out their Hurricane Simulator. An upgraded version will be one of the products of this project. It’s fun to play around with and can easily eat up a lot of your lunch break.

Thank you all for reading WWAT. It really is a joy to write.

Subscribe now

A colleague of mine ran across the concept of oceanicity as part of their general duties. They thought it was an interesting idea and shared it with me. And, following the theme of the last issue, I decided to steal it.

Said colleague also offered that they feel claustrophobic when away from the coast. I’m guessing that, to them, watching Dune was like watching Jaws for the rest of us.

Oceanicity describes the extent to which a nearby body of water shapes local climate and weather. Islands are probably the first thing we think about.

But the sea can also impact the climate hundreds of miles inland. For example, the Gulf Stream has a huge impact on the weather of northern and western Europe. Without it, London would be almost as cold as Moscow. Below is a map of mean surface temperatures. Notice how much warmer Europe is vs Canada at the same latitude.

It’s largely because of the Gulf Stream, which can be seen in the graphic below. It brings warmer water to the North Atlantic, acting as a hot water bottle for Europe. It’s oceanicity in pretty colors.

The Ingredients of Oceanicity

Ocean temperature is just one part of oceanicity. It mostly depends on being near a large body of water and having a steady wind blowing ashore. Then there is fetch, the uninterrupted stretch of water the wind travels over, picking up energy and water vapor along the way. Other factors matter too, and we will get to them. But distance and wind are the two parents in the front seat, bickering maybe, but directing the show.

The Conrad Index and Continentality

Oceancity has a counterpart known as continentality. That the latter is much better known, despite describing basically the same thing, shows the bias we humans have for land over water. Continentality can be illustrated via a measure known as the Conrad Index of Continentality. Victor Conrad was an Austrian-American meteorologist and seismologist who emigrated to the United States in 1938 to avoid persecution. He formalized the index in 1946 to quantify the oceanic-to-continental spectrum on a scale of 1-100.

The Spots with the Least and Most Oceanicity

Earth’s most landlocked point is known as the Continental Pole of Inaccessibility, and is located in northwestern China's Xinjiang province. It is positioned an equal distance away from the Arctic Ocean, the Arabian Sea, and the Yellow Sea.

It’s opposite, the point with the highest oceanicity, is likely Campbell Island/Motu Ihupuku, New Zealand. It has a Conrad Index of -5, which is amazing because it means it is more influenced by the ocean than an average spot on the ocean’s surface! This is likely due to influential currents combined with extreme wind. On average, it receives only 1.8 hours of sunshine per day and experiences a wind gust >60 mph about once every three days.

Ocean vs. Lakes

Oceanicity has the word ocean literally embedded in it. And the definition in the AMS Glossary of Meteorology refers strictly to the sea, but the influence of large bodies of water is not restricted to oceans. I live next to a big lake, large enough that you cannot see across it. And that has a tremendous impact on my weather. I can never come to terms with the official NWS readings for Chicago being recorded at an airport 24 miles away from downtown. They are consistently different from what most people in Chicago experience, yet O'Hare is our "official record." This creates all sorts of problems, especially for tourists and people who use automated weather apps that don’t know to account for the lake. I understand why the NWS doesn't want to change it (historical data), but I can still hold a grudge against whoever made that decision. If only they had been educated on oceanicity.

Oceanicity vs. Synchronicity

Water does not just shape climate. It shapes culture. Coastal people navigate by the water. Norwegians have a word, friluftsliv, for their entire outdoor-life philosophy, forged by a coastline so jagged it would take you over 50,000 miles to walk it. The British, parked on an island with maximum oceanicity, turned weather small talk into a national art form. Meanwhile, the people at the Continental Pole of Inaccessibility presumably talk about something else. Sand, maybe.

The most successful of all may be Corona beer. They created a brand associated with a chill-out beach vibe. It’s been so successful that Corona is now the most valuable beer brand in the world.

I wouldn’t mind a piece of that billion-dollar pie…

Coming Soon: Weather With a Twist of Lime.

Our archive of WWAT articles is here.

During the preparation of this work, the author(s) used Claude Sonnet 4.6 to copyedit and fact-check and Gemini 3.5 Flash to generate the oceanicity map which was then edited. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Weather with a Twist is published by the American Meteorological Society.

Unlike in some fields like economics, in science this is actually a problem. Science is based on facts and data (usually). The more accurate and precise the data, the better the analysis. And this brings up the wonderful topic of calibration.

Calibration is the process of comparing an instrument’s measurements against a known reference standard to determine its margin of error. It involves measuring the statistical difference between the instrument’s readings and the standard so that this difference can be applied to correct future data. For example, if you place a thermometer in that oven and notice that the displayed temperature runs 25°F too cold, you have calibrated the device; you can then mathematically add 25°F to the thermostat the next time you cook chicken-in-a-watermelon.

Calibrating Weather Radars

Calibrating weather radars can be a challenge. They measure the sky, and we don’t exactly have a reference sky handy. Radar engineers have many alternatives, but none of them work as well as we’d like. The best (light rain and Bragg scatter) can only be used in specific conditions or cannot be used at all (birdbath scans) by the National Weather Service’s WSR-88D2 radars, due to engineering limitations.

The "88Ds" are dual-polarization radars (after a significant upgrade to the original systems about a decade ago). They fire pulses oriented both horizontally and vertically, then compare the two echoes. The difference in strength between them is called differential reflectivity, a.k.a. Zdr. That fingerprint provides an approximate shape of the object, which meteorologists use to determine the type of precipitation (hail, snow, rain, etc.) or even whether it's precipitation at all (birds, debris, explosions, etc.).

A new paper, recently published in the Journal of Atmospheric and Oceanic Technology by J. Hu, P. Zhang, J. Krause, and A. Ryzhkov, proposes a method that measures snow in clouds, which would be more useful since there are often clouds hanging around somewhere in a radar’s field of view.

Measuring Falling Dinner Plates

Freshly grown ice crystals are flat, and as they drift down, they tend to align horizontally, like a sky full of tiny floating dinner plates. To the radar, they look extremely flat, with high and wildly variable Zdr.

But let those crystals fall for a while, and something nice happens. They collide, tangle, and clump into big, fluffy, low-density snowflakes. A fat clump like that tumbles as it falls instead of lying flat, so on average, it looks roughly round. It’s Zdr settles into a narrow, predictable band between 0 and about 0.25 dB. That’s the tell. Dry aggregated snow, “dry” meaning it hasn’t started melting yet, is the reference the radar needs.

Pray tell, where do we find such snow?

Near the top of a cloud, ofc. Specifically, in the band where temperatures sit around -10° to -20°C. This is the dendritic growth layer: the nursery where ice crystals grow fastest. The radar spots it as a stripe of enhanced signal. As crystals fall out of the nursery, they spend the next stretch of sky colliding and clumping into those fluffy aggregates. Lower down, in warmer storms, sits the melting layer, where snow turns to rain. It appears as a band of low correlation, the radar equivalent of TV static (younger folk will have no idea what I’m talking about), because half-melted particles are a chaotic jumble of shapes.

Dry aggregated snow lives in the sweet spot: below the nursery and above the melting line. The researchers’ main contribution is a new algorithm, QZdrCal, to verify that the layer is present and consists entirely of snow. In the past, similar calibration techniques always assumed the layer existed. So when it didn’t, it returned bad data.

When asked nicely, QZdrCal will test:

• Is there enough vertical room between the nursery and the melt for flakes to actually clump together? (It wants at least a kilometer.)

• Is there enough storm activity overhead to keep seeding fresh crystals down into that layer?

• Is reflectivity increasing on the way down, the telltale sign of flakes fattening up as they collect their friends?

• Is the melting-layer static deep enough to confirm that big aggregates really were arriving from above?

Plot Twist: Sometimes it’s cold enough to, you know, actually snow on the ground. There is no melting layer. The system automatically checks for that and, if it’s the case, then it measures the snow all the way down to the surface.

Only when the data passes the checklist does QZdrCal measure the median ZDR in that snow layer, compare it to the value such snow ought to have, and report the difference as the bias.

Testing the Calibration

Across a parade of test storms (a warm-season event over Milwaukee, a pure-snow night over Burlington, Vermont, and a messy front in transition over Delaware), the estimates came out stable and tight. The transition case especially so. As one storm morphed from having a melting layer to being frozen all the way down, the algorithm adjusted, and the bias estimate barely twitched. This is exactly what you want, since a given radar’s bias should not change much over the course of a few hours.

The team turned QZdrCal loose on 274 storms scattered across the country. In about one in five, the radar’s ZDR bias came in over 0.2 dB, a known value that lines up neatly with the radar network’s own monthly quality reports. More importantly, the estimates were consistent, with a typical scatter of around 0.06 dB. That is the kind of steadiness that makes an algorithm trustworthy enough to ship.

Zenlike Implementation

And ship it they did. QZdrCal is now built into the operational software that runs the entire U.S. radar network. Previous snow-based calibration routines computed a bias whether or not real aggregated snow was actually present, quietly averaging in a lot of nonsense. QZdrCal’s restraint, speaking up only when it genuinely sees the right snow, is precisely what makes it better. It’s almost zenlike, remaining quiet only when it has something to say.

And Now for Something Completely Different

The NASA emblem has a rich and varied history, famously highlighted by the wars over the meatball logo versus the worm logo. Today, the brand has become immensely popular, generating an estimated hundreds of millions of dollars in annual merchandise sales. Of which NASA gets zilch.

NASA is great and all. But, personally, I love the NOAA LOGO!

I see it on stickers on cars at NWS/NOAA offices all the time. And I want to steal them. Better yet, someone needs to make quality t-shirts out of this logo. It’s government IP, so it’s free (I think). And it can help build the NOAA brand at a time when morale is low. So how about it, apparel industry? RAYGUN I’m looking at you in particular… your weather t-shirt game is already on point. If you make one, I promise not to steal it.

We acknowledge additional contributions from Beth Mills and Chris Nabholz. We are grateful to Lockheed Martin for a grant supporting this newsletter.

Our archive of WWAT articles is here.

During the preparation of this work, the author(s) used Claude Opus 4.8 and Gemini 3.5 to copyedit, fact-check, and draw the cloud diagram. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Weather with a Twist is published by the American Meteorological Society.

I’m guilty of doing it myself. In a past life, I was looking for a rhythm in the fluctuations of a particular star’s brightness. I didn’t find what I was looking for. But I found something else that could be related. So I added quasi- to it, got published, and got a pat on the back from my professor. Then the world quickly forgot it. It was a quasi-success.

Quasi-Linear Convective Systems

Meteorology has its own quasi’s. It’s a favorite because the atmosphere rarely behaves in a perfectly pure way. It essentially means "seemingly" or "to a certain degree”. The official AMS Glossary of Meteorology includes 18 terms that start with quasi-.

But for most of us, the most interesting meteorological quasi- are Quasi-Linear Convective Systems (QLCSs)1. They are lines of thunderstorms with embedded rotating circulations that grow, fade, merge, and separate as the line progresses. While flooding and straight-line winds are the most common dangers, about 10–20% of all tornadoes form within them. They tend to be less intense than tornadoes born in supercells… but they are still tornadoes, with all that comes with that.

While a QLCS can form in any season, they tend to peak in April - July. Those that form overnight and during the cooler months (late fall through early spring), however, are more likely to include tornadoes. In contrast, supercell tornadoes are more likely in April-June. QLCS tornadoes tend to be found at the northern end of the QLCS and several hours after the system itself has formed. The large-scale conditions that favor QLCS formation are well studied, as are the internal dynamics within a QLCS as tornadoes develop. What has not been studied as thoroughly are the conditions immediately surrounding the QLCS after it has formed, but before tornadoes begin.

CAPE & Shear

Researchers at North Carolina State University and the NOAA/NWS Storm Prediction Center set out to look for differences in their environments that could help forecasters identify tornadic and non-tornadic storm formation in QLCSs. They recently published their study in the journal Weather and Forecasting.

Two of the most important ingredients for severe thunderstorm formation are instability and wind shear. Instability is measured by a quantity called Convective Available Potential Energy (CAPE). It is how much energy is available to raise a parcel of warm, moist air upward (a.k.a. buoyancy). The higher the CAPE, the more vigorously air will rise once it gets going, and the more explosive storm growth tends to be. Wind shear is the change in wind speed and direction with height. It gives storms their rotation and organizational structure (and produces a lot of the turbulence felt when flying in bad weather). Without shear, storm updrafts tend to be short-lived and disorganized.

High-shear, high-CAPE (HSHC) environments typically produce the supercell storms we know and love. But High-shear, low-CAPE (HSLC) environments, by contrast, present some of the most difficult operational forecasting challenges. They're associated with higher false alarm rates and lower probabilities of detection. The authors of the study suggest studying how shear and CAPE change immediately along the path of the QLCS may help us understand them more.

HRRRRRRRRRRRRRRRRRRRRs

The authors used data from the operational High-Resolution Rapid Refresh (HRRR) model for 79 QLCS events spanning 1,257 case hours between July 2018 and May 2023. The HRRR is a well-respected computer model for forecasting storms on short time scales in North America (hours to a few days). Its hourly update cycle and fine grid spacing make it the best available tool for the kind of analysis they were doing.

Cases were identified using the Storm Prediction Center’s severe weather archive, filtered to events in Louisiana, Alabama, Mississippi, Arkansas, Tennessee, and Missouri - because those areas are more sensitive to damage from QLCSs.2 To qualify, a QLCS had to maintain a quasi-linear reflectivity structure of at least 100 km for at least 3 hours.

To make sure they were studying the right kind of storms, the researchers required that a meaningful portion of the environment around each storm have low energy for storm formation, combined with strong wind shear. They set the bar for "low energy" a bit higher than some prior studies since the Southeast naturally tends to have more atmospheric instability than regions farther north.3

A key methodological contribution of the paper is the QLCS-relative coordinate system. The authors normalized the QLCSs by mapping each one along two axes: how far a point was from one end of the convective line to the other (expressed as a percentage, called the Line Length Percentile), and how far ahead of the line it was. This allowed them to combine all 79 systems into a common spatial framework regardless of each line's length, orientation, or geographic location. They then examined HRRR output conditions across this grid for each case hour throughout the full duration of every event.

Results

First, they found that CAPE varied significantly within the QLCS, largely due to the system’s own influence. Consequently, even if a QLCS begins in a low-CAPE environment, it may transition out of it over time, particularly at the southern end of the line. A QLCS can generate its own low-level potential vorticity (a measure of the circulation or spin of an air parcel) through latent heat release. Within the intense updrafts of the storm, water vapor condenses and releases heat, typically peaking in the mid-levels of the atmosphere. This enhances the low-level jet and wind shear just ahead of the storm, making it more tornado-friendly. However, they found tornadoes are still more likely to the north because, among other things,4 shear is more important than CAPE. The northern end may have less CAPE, but it has enough, and its shear advantage appears to be the deciding factor.

They also found that overnight hours are particularly dangerous. As the boundary layer (atmosphere just above the surface) stabilizes after sunset, low-level wind shear increases substantially. This happens because winds near and above the top of the boundary layer can still be strong, leading to greater contrast with height. And this effect is strongest at the northern end of the line, away from the moderating influence of the Gulf of Mexico.

The paper's core contribution is demonstrating that HSLC QLCS environments are not static. There are systematic gradients along the convective line and ahead of it, and the systems themselves modify their environments in ways that feed back on tornado production. For operational meteorologists, the practical punchline is that proximity soundings sampled well ahead of a QLCS may significantly underrepresent the shear and instability found near the convective line itself. More rapidly updating observational data and computer models would be helpful.

Share

And Now for Something Completely Different

Mike Trout is one of the best baseball players of all time. He’s also a fellow weather geek/nerd/enthusiast. Here he is calling into The Weather Channel (during the off-season!) with a report from a DC blizzard.

He proposed to his wife with an aircraft skywriting the proposal (awwww). But things didn’t go as planned:

“In summer 2016, Trout planned to propose to his longtime girlfriend, Jessica Cox, via skywriting. A plane would write the message in the air — “Will you marry me, Jess?” — while the two of them had lunch at the beach. However, Trout got the forecast wrong, and the planes had to be called off and rescheduled for a break in the weather.” - The Athletic/NYTimes

Nature doesn’t care if you are worth over $140 million. I think that story, officially, makes him one of us.

We acknowledge additional contributions from Beth Mills and Chris Nabholz. We are grateful to Lockheed Martin for a grant supporting this newsletter.

Our archive of WWAT articles is here.

During the preparation of this work, the author(s) used Claude Sonnet 4.3 to copyedit and fact-check. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Weather with a Twist is published by the American Meteorological Society.

Of course, the difference in scale from cooling a home to wholesale blocking out of the Sun is orders of magnitude. One requires a simple trip to Home Depot. The other, unlimited wealth and delusions of grandeur. Happily, there is a middle ground.

Cloud Seeding

As Jack Nicholson discovered in Chinatown, it’s all about the water.

Arid regions need more of it. When they find it, it usually comes from the ground. But imagine being in a desert and watching clouds float by. Looking up and seeing that sweet, sweet water pass overhead, taunting you like this:

What if we could somehow tap that floating water? For almost a century, researchers have studied how clouds could be “seeded” with silver iodide to induce ice crystal formation and growth, to trigger precipitation. Silver iodide has a crystal lattice structure that mimics that of ice, making it a highly efficient ice nucleus for supercooled water (water that is below freezing but not solid). Additionally, silver iodide is practically insoluble in water, ensuring that the particles don’t dissolve in the cloud.

The theory is solid, and it works in the laboratory. The problem is that it's been difficult to study in situ — that is, in the wild. The additional precipitation it creates is small relative to natural rainfall variability, so any signal gets lost in the noise.

The SNOWIE Project

So why not make a more unique signal that is easier to detect? That's exactly what the Seeded and Natural Orographic Wintertime Clouds: The Idaho Experiment (SNOWIE) set out to do. A coalition of academic, government, and private-sector researchers descended on the Payette River basin in west-central Idaho to finally answer the question of whether cloud seeding works — at least in the Idaho mountains.

They came up with a truly innovative solution. They flew an aircraft in a straight line, trailing silver iodide to trigger snowfall. At the same time, the wind blew the clouds perpendicularly across its path. This created a distinct zig-zag pattern in the clouds. So they looked for that exact same geometric shape in the radar data. Since nature almost never produces perfectly straight lines or sharp zig-zags in a storm, seeing that pattern would be smoking-gun proof that the snow was man-made.

A second technique they used was to follow the burst of precipitation as it fell to the ground. Regular snowfall would fall continuously, while seeding-generated snowfall would be temporary. Thus, they would see that enhanced snow descends in height amidst the continuous, regular snowfall.

In order to detect this, you need radars that are precise and can measure very close to the ground. This area of Idaho is also hilly and mountainous. So distant radar is blocked by both mountains and the curvature of the Earth. They needed something close and easily deployable. Friends of WWAT will know where this is leading…

The Doppler on Wheels mobile radar is a fan favorite here. We’ve reported on its studies of lake effect thundersnow, hurricanes, and tornadoes. Its mobility and extensive instrument suite allow meteorologists to get up close and personal with all sorts of weather phenomena.

There were 24 research flights over the 10-week campaign. Using data from the aircraft, DOWs, weather balloons, computer models, and other instruments, researchers identified and measured a precipitation pattern matching the cloud-seeding track. They published their results in an article in the Bulletin of the American Meteorological Society.

They reported distinctive zigzag patterns of enhanced radar reflectivity downwind of the seeding aircraft track, consistent with the dispersal of silver iodide and the subsequent nucleation and growth of ice particles. In situ measurements confirmed that inside the seeded regions, liquid water had been largely converted to ice, with particles and aggregates up to 4mm in diameter - much larger than what was found outside the seeded regions.

SNOWSCAPE

Inspired by this success, cloud seeding has once again become an active area of research. Subsequent studies have shown that a roughly 5–15% increase in precipitation can be attributed to seeding in a typical experiment. In the Mountain West, this is particularly significant. Snowpack has an outsized environmental and economic impact on the region and ultimately feeds the Colorado River, which supplies water to much of the desert Southwest.

Utah gets ~95% of its usable water from the snowpack. Not surprisingly, it is leading a large research consortium to develop a new, research-based framework for evaluating cloud seeding - Seeded and Natural Orographic Winter Storms CAtchment Processes and Evaluations (SNOWSCAPE). They just concluded a 2-month-long field campaign this winter to collect data in both seeded and non-seeded storms. Data is currently being analyzed, and we will report back when it is published.

SNOWSCAPE represents the most instrumented cloud seeding field campaign ever conducted in Utah. For the first time, we have the observational infrastructure to connect what’s happening within the clouds, at the aerosol and microphysical level, all the way through to the snowpack and streamflow response. We are building the scientific framework to better understand why and under what conditions cloud seeding works best. We are proud of the collaborative effort between Utah’s Division of Water Resources, the University of Utah, Utah State University, the National Center for Atmospheric Research, and Rainmaker Technology Corporation, which included over 50 participating students. - Jonathan Jennings, Meteorologist, Utah Department of Natural Resources, Division of Water Resources.

And Now for Something Completely Different

If you are of a certain age, you remember The Weather Channel's Local Forecast. It was a short segment that provided the forecast for your cable provider's broadcast area, airing roughly every 8 minutes between regular programming and following a consistent format, ostensibly to make automation easier.

The result was a consistent, relaxing heartbeat of weather during your day. Of course, as attention spans diminished and weather apps multiplied like wet gremlins, the need for such a product diminished, and it eventually was discontinued (sad_trombine.wav).

However, in a stroke of marketing genius, the Weather Channel recently brought it back! It is now known as Retro-Cast and can be viewed from any web browser.

I highly recommend enabling the sound, because the experience is not complete without the awesome ’90s smooth jazz soundtrack.

I went for a deeper dive, because this is the kind of thing the Internet was designed for, and discovered that a version of this has been around since 2020. The original WeatherStar 4000+ code was developed by Mike Battaglia, who released it as open source. Matt Walsh subsequently created a highly customizable version that works on web browsers. And he’s adding some cool new features as recently as last summer:

And, yes, the music is there.

We acknowledge additional contributions from Jonathan Jennings. We are grateful to Lockheed Martin for a grant supporting this newsletter.

Our archive of WWAT articles is here.

During the preparation of this work, the author(s) used Claude Sonnet 4.6 to copyedit text and some science background. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Weather with a Twist is published by the American Meteorological Society.

They went to college. They studied their maps. They carried their ones. It was close, but they passed their quals and made it through good ‘ol fashioned hard work and gumption.1 Eventually, they were hired as a real meteorologist, with their own grease pencil and green screen. They were proud.

But soon reality hit them. Bureaucrats hated their new ideas, and middle managers feared their energy.

And, worse of all, the public they were trying to help cracked jokes and laughed in their face.

At first, they tried fighting the system. Beginning with the righteous power of their ideals…

When that went nowhere, they tried pleading…

But you know the tale. Of course, the system won. Sooner than was healthy, they transformed into… a jaded meteorologist.

They barely made it to work. “Wow!”’s became “meh”’s. Yet another severe thunderstorm warning? (rolls eyes). Each night shift, they felt like a vampire condemned to a world of V8.

At their deepest low, they searched within for anything to grab onto. And they found it - a tiny light still shining.

A nugget of an idea was forming. Not much more than a chicken nugget, actually. “This nugget is not made of chicken. And neither am I!”

They needed to recover their purpose. Through all the paper pushing and button pressing, they had lost track of why they went into this field - curiosity. But of what? Whose secrets were they really after?

They looked out the window and found their answer:

The distinguished cricket’s siren song. If a simple cricket can accurately detect the temperature and turn it into song, then surely our esteemed meteorologist can fight the powers and unlock the mysteries of nature. Some of them at least. Nature is the mystery, science is the tool, and they are the detective.

Renewed with hope and vigor, our hero is now an ex-jaded meteorologist. Inspired by science, they are ready to resume the march of progress. To face their peers, the public, that weird guy who sits in the corner office…

To once again, send one more forecast unto the breech.

And Now for Something Completely Different

There is a kernel of truth to this April Fool’s joke. Meteorology has always been a field shadowed by stress and anxiety. Every meteorologist I’ve met was inspired as a kid to study the weather. They do it for passion. No one enters it to get rich. But public expectations, institutional politics, relentless shift work, and the weight of decisions that carry real safety consequences all take their toll over time. And this past year, things have been especially tough.

If you know a meteorologist, buy them a coffee or a nice stiff drink or two. No joke - they probably need it, and they definitely deserve it.

Mother Nature is an artist. And sometimes she needs to scratch that creative itch. Consider, gravity waves:

This newsletter will be awash in beautiful imagery, because that is what gravity waves are. Sure, there is science and physics involved, and serious implications for our weather. But they're also art, writ large — Mother Nature painting on the canvas of Earth. And unlike most great art, hopefully we don’t have to wait for her death for it for people to appreciate it.

The Sky is Fluid

One of the big revolutions in meteorology came in the early 1900s, when Norwegian physicist Vilhelm Bjerknes had a deceptively simple insight: treat the atmosphere like a fluid. For a century, clever nerds had been building a new field in physics called hydrodynamics. Now it could be put to work predicting the weather.

Atmospheric gravity waves are ripples in the air that can be compared to ripples seen in water when you drop a stone in it. The water bobs up and down as the disturbance spreads outward. The atmosphere does the same thing. But it’s not due to dropping a rock on it, but rather something pushing up from beneath it.

The Setup

The key ingredient is a stable atmosphere —one that resists vertical mixing and pushes displaced air back from which it came. Stay in your lanes! It works best in layers when heavier, cooler air sits beneath lighter, warmer air on top.

Sometimes, the lower heavier air gets forced upward by the mountains, storm updrafts, etc. It’s now higher up in the atmosphere, surrounded by lighter air. And it wants to get back down to where the fun is. Gravity is there to provide the ride.

The Cannonball

So we have a stable air mass with heavier, cool air below and lighter, warm air aloft. What pushes the cool air upward?

Sometimes it is a land mass like a mountain. These mountain waves or lee waves are responsible for some of the most dramatic cloud formations you'll ever see: the smooth, lens-shaped lenticular clouds that stack up over peaks like flying saucers. Pilots know them well, because the powerful downdrafts beneath them can be dangerous.

Thunderstorms generate gravity waves too. As a storm's updraft punches into the stable air above, it acts like a piston, sending waves radiating outward in all directions. These waves can trigger new storms hundreds of miles from the original one. Behold one of my favorite radar videos (NWS Frederick).

It’s as if the thunderstorms are filming a Hadouken, a Japanese photo trend from 2013 where one person poses as if releasing an video game style energy blast, while others act like they’re being thrown back.

Or it can be caused by a frontal passage of colder air into a stable environment of less cool air beneath warm air. The denser, colder air acts like a plow moving through the atmosphere, wedging under the warmer air ahead and forcing it upward rapidly. That sudden vertical displacement sets off gravity waves, which radiate outward ahead of and behind the front. It is one reason why rough air and interesting cloud structures can show up long before a front actually arrives.

Kelvin-Helmholtz waves form when the upper layers of gravity waves move very quickly compared to the lower layers, causing them to crest. They are a relatively short-lived and rare phenomenon.

Implications for Weather

Gravity waves are one of nature’s great couriers of energy. They carry energy generated at low levels (from storms, terrain, or wind shear) upward into the middle and upper atmosphere, sometimes all the way into the stratosphere and mesosphere. When they eventually break, like ocean waves crashing on a beach, they deposit that energy and momentum, influencing winds at altitudes far above where the waves originated.

As mentioned, gravity waves created by thunderstorms can spin up new thunderstorms or amplify the strength of others nearby. There is one documented instance of a tornado caused by a gravity wave passing into an existing thunderstorm.

Gravity waves can be a source of clear-air turbulence in aircraft. They can develop in calm weather conditions surrounding the parent thunderstorm and may not be associated with any clouds. So a pilot may think they are maneuvering safely around a storm, only to suddenly encounter turbulence.

The waves come in a wide range of sizes. Some have horizontal wavelengths of just a few kilometers; others stretch for thousands of kilometers. Their periods range from minutes to hours. This enormous variety makes them tricky to observe and even trickier to represent in weather models. Most global forecast models still can’t resolve the smaller ones directly and have to parameterize their effects instead, which is an ongoing source of forecast uncertainty.

Undular Bore

There is a specific type of gravity wave known as an undular bore.1 An undular bore is a solitary wave or a series of waves usually triggered by the frontal passage effect. The bore part refers to its origin as a sudden disturbance that pushes into a stable layer, like a tidal bore in a river, while undular describes the train of wave crests that follow in its wake.

Undular bores are particularly associated with spectacular morning glory clouds in northern Australia and with nocturnal convection over the Great Plains, where they can trigger new thunderstorms in otherwise quiet air.

Fuqing Zhang

Fuqing Zhang (1969–2019) was a Distinguished Professor of Meteorology at Penn State whose work transformed our understanding of how gravity waves interact with severe weather. He was the first to rigorously examine how gravity waves are continuously emitted by unbalanced upper-level flow, and he demonstrated the critical role that gravity wave–convection feedback plays in wave amplification and storm development. He died unexpectedly at 49, leaving behind nearly 250 published papers, a generation of mentored scientists, and research questions his field is still working to answer.

Our archive of WWAT articles is here.

During the preparation of this work, the author(s) used Claude Sonnet 4.6 to copyedit text and provide background information on gravity waves. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Science is a creative act. And whether it’s serendipity or a purposeful sojourn into new territories, good scientists indeed look outside their fields for inspiration.

For some meteorologists, this includes looking into space. Most of the time at WWAT, we talk about looking back at Earth from space, through satellites. But sometimes we want to look from Earth into space to get a new perspective on an issue down home.

“Would you eat the Moon if it was made of ribs?” OK, maybe that’s not what’s keeping scientists up late at night.

The Earth through the Lenses of Other Planets

The physics and chemistry behind the greenhouse effect were discovered in 1824 by Joseph Fourier. However, at the time, many doubted it could have an impact on a planetary scale. The Earth is simply too large, and greenhouses are too small! What affects one could not possibly affect the other. To many, scientists seemed nuts to suggest otherwise.

In 1900, esteemed physicist Knut Ångström said bluntly, “Under no circumstances should the carbon dioxide be regarded as a factor of importance in the heat economy of the atmosphere.” As a leading authority in the field, his voice carried significant weight, and meteorological texts subsequently dismissed greenhouse gases as a trivial player in global temperatures.1

It took another famous scientist, Carl Sagan, to turn heads 60 years later by looking at a different planet: Venus. At the time, it was well known that Venus was significantly warmer than expected, with surface temperatures exceeding 850°F. While it is closer to the Sun than Earth, the distance alone couldn’t account for the heat. In fact, Venus is hotter than Mercury, despite the fact that Mercury practically violates the Sun’s personal space.

Our boy Carl calculated that the various greenhouse gases in the atmosphere created a runaway greenhouse effect. His theory was published in the journal Science and has since become widely accepted as fact. It also led people to think that if it could happen to Venus, it could happen here.

This is where astronomy as a science has real practical implications on Earth. We can perform experiments (both empirical and thought-based) on other worlds in ways we can’t on our own. We need another canvas to play with. We need to cross-pollinate.

From Venus to Titan

Meteorologists and astronomers at UCLA recognized some similarities between Earth’s atmosphere and Titan, a moon of Saturn. Titan is the second largest moon in the solar system, even larger than the planet Mercury (I don’t mean to keep picking on poor ‘ol Mercury, which is pretty neat in its own way.)

Among the similarities are a tropical climate, active hydrologic cycles, global circulation (Hadley cells), and convection. In fact, Titan may have all the components to form hurricanes. It has a warm surface liquid (methane instead of water), high atmospheric humidity2, deep convection (storms), and low-level vorticity (atmospheric spin).

But there are no hurricanes on Titan. At least, none that we’ve yet seen. It’s probably due to three main reasons:

• Rotation Period: Titan’s day is 16 days long. This lethargic pace slows down the atmospheric circulation patterns that would otherwise trap and concentrate heat near the equator. As a result, the temperature gradient from the equator to the pole is significantly shallower than on Earth. Because a sharp thermal gradient is a key driver of tropical cyclones, Titan’s atmosphere simply lacks the crank needed to start the engine.

• Methane instead of Water: The liquid of choice for the discerning drinker on Titan is methane, not H2O. While Titan has lakes, rivers, and rainfall similar to Earth, they are composed of liquid hydrocarbons. Its atmosphere contains a much higher relative concentration of methane than Earth’s atmosphere contains water vapor.
  

  For a tropical cyclone to form, there needs to be a significant disequilibrium between the surface liquid and the air above it. The air needs to be “hungry” enough to pull up more evaporating fuel. On Titan, the atmosphere is already so saturated with methane that this evaporation process is sluggish. Like someone overstuffed after a Thanksgiving meal, the “engine” of a cyclone never starts; it prefers to roll over and take a nap.

• Limited Sea Surface - Hurricanes are powered by heat from the ocean. Only 1-2% of Titan’s surface is covered with liquid (vs 71% of Earth). While Titan is the only other place in the solar system with stable surface liquids, there’s simply not very much of it. What does exist is mostly concentrated in three large lakes in the northern polar region.

Alien Laboratories

Scientists David C. Moore and Jonathan L. Mitchell explored this parallel in a recent article published in the Journal of the Atmospheric Sciences. They wondered what would happen if they took a well-known atmospheric model of the Earth and adjusted these 3 parameters. Would they be able to form tropical cyclones in those conditions? If so, perhaps we should be on the lookout for future hurricanes on Titan.

They used a suite of models based on the Isca 1.0 global climate/circulation framework. They adjusted the models to account for Titan’s slow rotation rate, the impact of methane in the atmosphere (compared to water), and the amount of land around the equator. Basically, in each iteration, they slowly turned the Earth into Titan.

The figure below combines three key output variables from the simulations.

In the above, the y-axis represents latitude, and the x-axis represents days of the year. The colors indicate the peak wind speed of a potential tropical cyclone (with red indicating stronger winds). The solid white contour lines outline areas where there is sufficient spin to organize a storm. The dashed white contour lines mark regions where dry air and wind shear are unfavorable for storm formation. Thus, favorable conditions are indicated by warm colors located inside the solid lines and outside the dashed lines.

At a glance, the data reveals a stark contrast: while Earth’s tropical storms can form almost anywhere (clustering most frequently near the equator) while Titan’s are far rarer. On Titan, the sweet spot shifts to the north pole, where even there, the likelihood of formation remains low and the window of opportunity brief.

The primary takeaway is that while Earth’s storms originate in the tropics, Titan’s would likely form at the poles. This shift occurs because planetary properties (rotation rate, humidity, and geography) alter where the necessary ingredients converge. The authors conclude that such a storm is unlikely but possible, and it should be monitored as we continue to send more probes to land on Titan.

NASA’s next such mission is Dragonfly, a nuclear-powered quadcopter drone designed to fly around Titan’s surface. Let’s repeat that for the people in the back. A NUCLEAR-POWERED QUADCOPTER DRONE. Enjoy your nightmares. Expected to land in 2034, it’s designed to monitor conditions for life. While it is primarily a chemistry and geology mission, it carries a dedicated meteorology package and several instruments to monitor Titan’s atmosphere.

Subscribe now

And Now for Something Completely Different

So, apparently, there was a cartoon in the 90’s called James Bond, Jr. Don’t ask me how they interpreted the sex and violence in the movies, but the cartoonish characters must have been an easy carryover. Happily for us, they produced an episode about a weather-controlling satellite with the brilliant title “Weather or Not”.

I would join this writing staff in a second. But until they invent a time machine, I’m slaving behind the Substack keyboard. Speaking of which, I really am infatuated by the episode title. I should have used it for this newsletter. What do you think?

We acknowledge additional contributions from Chris Nabholz. We are grateful to Lockheed Martin for a grant supporting this newsletter.

Our archive of WWAT articles is here.

During the preparation of this work, the author(s) used Gemini 3 Flash to copyedit text and help interpret Figure 15. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Weather with a Twist is published by the American Meteorological Society.

OK, if you aren’t a weather or science nerd, then you may have yawned and moved on. But then you wouldn’t be reading this far. Thus, I am pretty sure 99% of you agree with me - that’s one cool word.

Or rather, hot, because it involves fire. Specifically, thunderstorms that are made of fire. I’m surprised Stephen King hasn’t written a book about it yet.

From Firestorms to Thunderstorms

You probably already know about the terrible phenomenon known as firestorms. These occur when a fire becomes so large and hot that it generates its own weather. The heat causes air to rise at a tremendous rate, pulling in air near the surface and creating intense, hot winds. The circulation can also begin to rotate, producing vortices of fire known as firenadoes and their smaller cousins, fire whirls. Firestorms can occur naturally in large wildfires and, sadly and unnaturally, during heavy wartime bombing raids on cities.

If the firestorm is in the right conditions, then it can grow into something even greater - a full-blown thunderstorm with rain, hail, and lightning. Scientifically, they are known as pyrocumulonimbus, or even more scientifically, as pyroCB.

Our tale begins with a gentle firestorm rolling through a forest. If the forest is dry and dense, a little moist, and if the fire is moving slowly, it can pump out smoke as if it’s auditioning for a Cheech ’n Chong movie. This smoke is key to its evolution into a thunderstorm.

The fire’s intense heat drives a powerful updraft that lifts smoke particles high into the atmosphere. Those particles act as condensation and ice nuclei, helping water condense and freeze. This can enhance latent heat release and allow the cloud to glaciate rapidly, building into a deep, thunderstorm-like column. Within that column, collisions among ice and smoke-laden particles promote charge separation and lightning. Now we have the thunder for a thunderstorm.

PyroCbs can produce extreme updrafts and downdrafts dangerous to firefighting aircraft. It can also generate lightning to ignite new fires, hail, ember storms, and fire-generated tornadoes. As if that wasn’t enough, the smoke can sometimes rise into the lower stratosphere, where it can persist for over a year. That stratospheric smoke has been linked to changes in air circulation, ozone chemistry, and Earth’s radiative balance. PyroCBs are likely responsible for up to 25% of the black carbon and organic aerosols in the lower stratosphere.

Detecting pyroCBs

PyroCbs are identified primarily by their distinctive satellite signature. From space, they look like extremely tall, cold-topped thunderstorms growing directly out of intense wildfire plumes. The smoke embedded in the cloud creates unusually small ice particles at the top, which gives the storm a unique fingerprint when satellites view it at multiple infrared wavelengths.

Forecasters combine those cloud-top signals with satellite fire products that measure heat output and active fire locations to confirm that the storm is fire-driven. Detection is increasingly automated using algorithms that scan for these patterns, though analysts still review events to verify that what they’re seeing is truly a pyroCB and not just an ordinary thunderstorm with a smoky van driving through it.

To track wildfires, NASA maintains the MODIS Collection 6 Active Fire (MOD14) database, which records the location and radiative power of individual fire pixels detected by the Aqua and Terra satellites. Also, the Naval Research Laboratory maintains a pyroCB inventory using data from five different satellites covering the Earth. Finally, they included meteorological variables around the fires taken from the GFS database.1 Researchers used this data to look for trends and predictors in pyroCB behavior.

Analyzing pyroCB Patterns

One such team, led by Dr. Evan Krell, set out to better understand the atmospheric and fire conditions that favor pyroCb development and to test whether machine-learning models could detect them. They used those two datasets to train the models with data from 2013-2020. Then they tested the trained models on data from 2021 to see whether they could successfully identify pyroCb events. They published their methods and results in the journal Artificial Intelligence for the Earth Systems.

Most of their results concerned performance differences among different machine learning techniques. But there were some overall results worth reporting:

• The models performed well at detecting pyroCBs, but only when using both satellite and atmospheric data. Using either dataset alone was not as accurate. This supported the authors’ thesis that pyroCBs are a complex interaction between fire and a supportive atmospheric environment, so we need to consider both.

• Rare/unusual pyroCBs were not well captured. These include events where the atmospheric conditions were different than most of the others (ex: differences in wind shear). This isn’t really a surprise, given that machine learning needs prior data from which to learn. And the events being rare, there simply wasn’t much to train on. As the databases expand over time, the models may improve at detecting these less typical pyroCBs.

• Experts are still needed. The models statistically determined the best atmospheric variables used to predict pyroCBs in the 2013-2021 data. However, when tested with the 2021 data, the variables that experienced scientists flagged as important were more useful. In other words, machine learning worked best when guided by meteorological expertise.

• There is a consistent trade-off. Tweaking the model to reduce missed pyroCbs often increases false positives. This is a classic forecasting problem, but important because pyroCbs are high-impact yet infrequent. So small failures are quite visible and could lead to low confidence in future forecasts.

PyroCbs are rare events, but their frequency appears to be increasing as human-driven changes in climate and land use alter fire behavior and the environments that support these extreme storms. Research like this helps us better understand the mechanisms behind pyroCbs and, ideally, improve our ability to predict them.

Because most current weather models do not directly integrate detailed fire behavior data with atmospheric data, they are not well-suited to forecasting pyroCbs. The hope is that studies like this will inform more sophisticated, coupled fire–atmosphere models in the future.

And Now for Something Completely Different

We’ve talked in the past about changing climatology and the evolution of where tornadoes are now more likely to occur. NOAA has issued two interesting graphics to illustrate this shift.

The overall pattern remains largely the same, but the region of highest probability has shifted from the central Plains into the upper South. And check out this lucky area!

Below is a handy video that shows how the probabilities change daily over the course of a year. A bullseye literally forms over Oklahoma in May. If I were trying to convince someone not to sign with the OKC Thunder, I’d send them this video.

Data for these graphics comes from the NOAA NCEI Storm Events Database.

We acknowledge additional contributions from Dr. Milind Sharma and Chris Nabholz. We are grateful to Lockheed Martin for a grant supporting this newsletter.

Our archive of WWAT articles is here.

During the preparation of this work, the author(s) used ChatGPT-5.2 to copyedit text. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Weather with a Twist is published by the American Meteorological Society.

So why are meteorologists taking hits like Amare Stoudemire’s fire extinguisher?

Some folks are incensed that the storm tracked a little farther north than the early forecasts. Yet most people still saw enough snow and ice to justify the cost of the shovels and the French toast supplies they rushed out to buy. But a smaller, louder group never got to peel the plastic off their shiny new snow shovel and now finds itself sitting on a surplus of beer for the rest of the winter. Those are the angry ones. The beer probably contributing.

Let’s look at the science behind the storm, so we can decide who to properly blame. Because casting blame is the American Way - as was science, once upon a time.

Out of the Gate

The origins of the storm lay in a weak polar vortex, abundant Gulf moisture across the South, and a low-pressure system approaching Baja California. This system was unusual enough that the National Weather Service flew hurricane hunters into it to take key measurements, something they rarely do for Pacific storms at this time of year. And, sure enough, they found the low-pressure system to be abnormally strong.

As it moved across Mexico and the southwestern United States, forecasters sounded the alarm. Combining computer models with their own experience, they saw all the ingredients of a significant winter storm in a region that rarely sees them.

The low pressure was predicted to move across Texas and tug down the southern edge of a rare polar vortex that had already extended into the Great Plains and Midwest. This provided the cold air. At the same time, ample moisture was moving northward from the Gulf (of México or America, depending on your choice of headgear). Now we had the all the energy (the low-pressure system), cold (the polar vortex), and moisture (from the Gulf) needed for a fun winter storm.

The Forecasts

Let’s start with a map of watches, warnings, and advisories1 issued by the National Weather Service on Friday, January 23rd when the first frozen precip in Texas and Oklahoma began to fall.2

Now let’s look at the snowfall (l) and ice (r) reports from January 24-26, when the storm crossed the country.

Comparing the forecast (top) and the realized (bottom) maps shows that the forecast was reasonably accurate, though the storm did end up slightly further north than expected. In particular, areas of Georgia and eastern Tennessee received less winter precipitation, whereas areas of Illinois and Michigan received more. Most people were pretty happy with the forecasts they were given. But, especially in those southern areas, some were not.3

Trends are Your Friends

I recall watching the various weather channels on Thursday and seeing this northern trend appear in subsequent updates. I was staying with friends in East Texas and needed to drive to Houston the next day. We were already on the ice line, so a slight move north or south could make or break my trek. One of the presenters said that this type of northern movement of a winter storm forecast was common in them there parts.

I was able to make my 3-hour drive on time. And while I tried to keep my mind trained on more important things like bathroom breaks and the upcoming NFL Championship Games, the presenter’s comments kept creeping in. And it bothered me. If this is common, then why didn’t they account for it?

Causes of a Northward Shift

Let’s look at what caused that northward shift. First, as the polar vortex moved south it merged (i.e. phasing in scientific terms) with the low pressure system from the Pacific. This strengthened the system and pulled up warmer air from the Southeast. Sometimes this phasing doesn’t happen and the systems remain more independent, which would lead to a more southern track.

Next, as moisture from the Gulf feeds into the storm, condensation and freezing release latent heat, which subtly warms the middle layers of the atmosphere. That warming helps deepen the low pressure system, tightening its circulation and allowing it to strengthen. A stronger storm does not just get bigger, it also becomes more responsive to the larger atmospheric flow around it. Once that happens, small changes in where the temperature boundary sits can nudge the track farther north than early forecasts suggested.

Why northward? Midlatitude cyclones live off temperature contrasts where cold air sits to the north and warm air to the south. This contrast is their life force. As the storm intensifies, it aligns itself with the strongest part of that gradient and the nearby jet stream, which in southern winter setups is often positioned slightly poleward. The differences involved are small, but these storms are enormous. So even small changes are magnified and can translate into a noticeable northward wobble.

If it always happens, why don’t we account for it?

The process is well known to both experienced forecasters and also the computer models. However, to properly account for it we need specifics in the form of data. Recall that this system began off Baja California. Very little atmospheric data is routinely collected there. (All sorts of surprises routinely come out of that area of the Pacific for this reason.) There are no consistent balloon launches or ground stations reporting conditions there (hence why NOAA needed to fly a a hurricane hunter into it). Once the system entered the United States, more atmospheric data came in and the models/meteorologists updated their forecasts accordingly, moving it slightly northward.

Another example regards where the precipitation forms. If more of it forms on the front/right side of the storm, as is common in these types of storms, then it also can cause the storm to veer north. But we won’t know that until the first precipiation falls - and that won’t happen until the system meets the Gulf moisture (in this case over Texas and Oklahoma).

So even though we understand the mechanisms, we need real data before we can truly plan for it. And, to paraphrase Aaron Burr, we had to wait for it.

In the end, the storm was well forecasted for most. Of course, it wasn’t perfect. And, really, would you want the weather to be perfectly forecasted? We would lose the most common topic of small talk. And if, like me, you have locubrevisphobia, that would be a hell worse than opening an empty vault in front of 30-million people.

We acknowledge additional contributions by Alan Gerard.

Weather with a Twist is published by the American Meteorological Society.

* During the preparation of this work, the author(s) used ChatGPT-5.1/5.2 to copyedit and help with phrasing. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

2

Image legend:

We have trucks by Ford, backpacks by Dior, bespoke carpet by Luxeweave, and furniture by Ikea. The longest lines are for Wet Bulb coffee and if you squint you may catch Jim Cantore autographing an anemometer.

But you’ll also find great science! Like this little tid bit about space weather…

Hawaiian Mass Ejections

Space weather is about monitoring conditions around our planet that could affect us. This often involves activity originating from the Sun, including energetic particles, magnetic fields, and radiation. The most spectacular of these phenomena are coronal mass ejections (CMEs) - sudden and large eruptions of solar plasma into the solar system.

CMEs form when twisted or stretched magnetic field lines in the Sun’s corona suddenly reconnect and relax, releasing enormous magnetic energy. Much of that energy is converted into kinetic motion, flinging trapped plasma outward along the newly reconfigured magnetic fields, much like a stretched rubber band snapping and throwing whatever it was holding. If a CME happens to be directed toward Earth, it typically arrives a few days later and disturbs Earth’s magnetic field. This can lead to aurorae, affect satellite orbits and power systems, cause power outages on the ground, and all sorts of fun stuff.

We have expensive satellites in orbit to monitor the Sun for these CMEs, and they work well. However, the Sun is really bright - especially in space. So the satellites have to shade themselves from the Sun’s surface so they can see the CMEs as they are emitted from the corona. The closest the satellites can see is about 3 solar radii from the Sun before they would be damaged or otherwise made useless by the brightness.

But this is one situation where ground based observatories can do better. Berger, et al. (2026) reported on CME observations made with the NCAR Mauna Loa Solar Observatory (MLSO) in Hawai’i. The MLSO has a special tool that allows them to see as close as 1.05 solar radii from the Sun’s surface. This allows them to see a CME 55 minutes before they are detected by spacecraft - giving us more advanced notice of incoming mayhem. The observatory issues near real-time alerts to interested parties so they can prepare accordingly.

One catch: Hawaii sometimes experiences this dreaded thing known as night time, to the dismay of sunworshipping tourists. And it’s even cloudy sometimes. Egads! So the researchers use this success to propose a network of similar stations to be established globally, to get around these annoying facts of life on Earth.

Driving in Snow Squalls

Snow squalls are brief, intense, and often unpredictable, and they can be especially dangerous for drivers. In 2018, the National Weather Service began issuing snow squall warnings directly to cell phones. Because a cell phone alert is a serious intervention, forecasters aim to use them only when absolutely necessary. But in this case they were being sent out too often. So about five years later, the warning criteria were tightened and forecasters have been exploring other technologies to get the message out, such as remotely controlled roadside signs.

Bernhardt et al. (2026) tested the effectiveness of such signs, in both English and Spanish, by placing 216 participants in a Meta virtual driving simulator that led them into snow squall conditions. As participants drove through the simulation, they received warnings either via their cell phones or from roadside signs. The researchers found that drivers who reported more prior experience driving in winter weather paid less attention to both the alerts and the roadside signs, while those with less experience paid more attention to them.

Try it yourself. The Simulation software can be downloaded through the MetaQuest store.

Sou’easters

When I lived in New England, I loved nor’easters. I loved the cloudy, wet days of torrential rain and giant evergreens swaying back and forth in the wind. I loved the shaking house and rattling windows.

Nor’easters get their name from a strong northeast wind associated with them. Contrariwise, sou’easters are cold season coastal low pressure systems that bring strong southeasterly winds with them. As with nore’easters, they can cause flooding and wind damage. Yet they are not studied nearly as much as their nor’easter cousins. As in, maybe not at all.

Using storm tracks, wind and tidal data from the Gulf of Maine Research Institute, and meteorological data from past storms, Venarsky, et al. (2026) presented a post-2000 climatology of sou’easter events. They classified storms by where they formed, when they occurred, and what kinds of local damage reports they produced. Then they compared those categories using descriptive statistics and correlation tests to see which factors were related to more serious impacts.

The results are… vague. Sou’easters form under a wide range of synoptic setups and have many different origin stories. They come from many places and all across the year. The only meaningful divider was local storm report type, which best captured differences in meteorological and hydrologic impacts (flooding). Weak correlations existed between some impact variables, but nothing close to a clean one-to-one relationship.

For me, this raises a question: Are sou’easters a thing? There may be a reason the name never caught on in New England culture - maybe they just don’t exist. They’re less a single storm type than a family resemblance. And recognizing that variability may be more useful than trying to pin them to a narrow set of patterns. And that’s what makes studies like this useful.

So let’s let Wikipedia make the final decision. Are sou’easters a thing?

(sad trombone)

Subscribe now

We are grateful to Lockheed Martin for a grant supporting this newsletter.

Our archive of WWAT articles is here.

During the preparation of this work, the author(s) used ChatGPT-5.2 to copyedit text. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Weather with a Twist is published by the American Meteorological Society.

As in years past, I’ll post a couple of articles with quick-hit summaries of interesting research I saw in sessions. Then, for better or for worse, we’ll be back to our regularly scheduled programming next week.

Warnings and TORP Couplets

The continued improvement of tornado warnings has been one of the great success stories of the National Weather Service. That progress is the result of research into tornado formation, detection, and also the warning system itself. Despite these gains, uncertainty remains an inherent part of tornado detection, making systematic evaluation of warning performance especially important. Trey Holiday and coauthors conducted one such evaluation of the accuracy of over 2,000 tornado warnings issued by the National Weather Service in 2018.

Looking at archival radar data, they used a machine learning technique called TORP to estimate the probability of a tornado actually occurring in a warning area. TORP uses data from the WSR-88D radars to identify areas where wind is moving rapidly toward the radar on one side and away from it on the other, indicating rotation (a.k.a. “couplets”).

Recall that tornado warnings are intended to indicate that a tornado has been detected, not merely forecasted. Those detections are inherently imperfect: radar signatures are noisy, and visual confirmation can be ambiguous due to rain, debris, darkness, etc. As a result,forecasters always have to deal with false positives (when a warning is issued but no tornado occurs) and false negatives (when a tornado occurs without an issued warning).

The results shows that forecasters tend to prioritize avoiding false negatives over false positives, reflecting the much higher stakes of a missed tornado (death!) compared to an unnecessary warning (annoying beep on one’s phone!). They also found that false positives are less common in more densely populated areas, suggesting that forecasters may apply stricter warning criteria in areas with greater societal impacts. This pattern raises important questions about unconscious bias and equity in issuing warnings.

This study was supported by an NSF REU grant, a super-popular and effective program that provides undergraduate science students with research opportunities and mentorship. Such students often report some of the coolest results at these conferences.

Born of the Wind

Ground-based observations are just that: records of what’s going on at the surface. But the atmosphere is in 3D. Most of it is much above the ground. Satellites, radar, and other remote sensing provide some data about what’s happening up there, but there is nothing better than a good old-fashioned weather balloon taking direct measurements as it flies. It’s critical to our forecasting. And this is why weather balloons are always launching, and then also always falling:

Jake Spisak led a team from Winborne Systems and the Scripps Institution of Oceanography to present on a network of long-duration weather balloons. Usually the balloons go up, travel a bit, then fall after a couple of hours of glorious freedom. But their system goes up and stays up for thousands of hours, flying all over the world and into the most remote regions.

So far, they’ve launched over 5,800 flights totaling more than 1,000,000 hours of flight time (above). They presented calibration results comparing the accuracy and precision of their instruments with more traditional balloons (radiosondes) launched alongside some of them. The instruments were quite accurate, which is not surprising, as it is unlikely they would have published the work otherwise. What is surprising is the sheer scale of the data collection effort. By the end of 2026, they expect to have more than 400 balloons aloft at the same time, extending coverage of the Southern Hemisphere. And there are lots of atmospheric scientists salivating over that data.

The Return of… Megaflash

One of the first WWAT newsletter was about megaflashes - lightning flashes over 100km long. I’ve yet to see one myself, despite pleading and crying to the weather gods. Perhaps I need to move to Florida, where Nick Stewart, of the Midwest Weather Center, has been tracking lots of megaflashes detected near NASA’s launch facilities.

He combined data from the Geostationary Lightning Mapper (GLM) on NOAA’s Geostationary Operational Environmental Satellites (GOES) with a ground network of lightning detection devices in eastern Florida known as MERLIN. Lightning is both very common in Florida and also very dangerous for rocket launches (and cows). So Florida has one of the best detection networks in the world.

Stewart found many megaflashes that extended over the various launch complexes in the region. One such flash, on August 5 2025, lasted over 5 seconds and was over 173 km long! It struck the ground in seven different spots along that path. And it did travel right over Kennedy Space Center.

Tracking these flashes is important because they can occur from longer distances than typically associated with lightning. Right now, there is a rule that prohibits launches when there is lightning within 19km (10 nm) of the launchpad. Given the ramifications of a lightning strike during a launch, more consideration may need to be given to thunderstorms that are over 100km away. In 1987, a rocket and satellite was destroyed by lightning less than a minute after launching. Lightning risk may extend well beyond traditional safety thresholds, and that launch criteria may need to evolve as our understanding of long-range lightning improves.

More from AMS tomorrow…

We are grateful to Lockheed Martin for a grant supporting this newsletter.

Our archive of WWAT articles is here.

Share

During the preparation of this work, the author(s) used ChatGPT-5.2 to copyedit text. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Weather with a Twist is published by the American Meteorological Society.

In the United States and Canada, tornadoes are sorted by the amount of damage they cause. But what happens when there is nothing around to damage, like in a corn field? It raises the obvious question: if a tornado tears through a field and nothing is around to be damaged, did it exist?

Because I would prefer not to be beaten up in a dark alley at the next meteorology conference, this newsletter is not about litigating the best way to measure tornado strength. Instead, we’re going to look at a new method that aims to improve the current damage-based system.

The Enhanced Fujita Scale

But first, a look at that system. Measuring the direct wind speeds of a tornado is extremely difficult. Tornadoes are unpredictable and short-lived, making it rare for an anemometer to be in the right place at the right time, let alone survive the encounter. Radar can provide wind-speed estimates, but those measurements are limited and imprecise.

So in 1971, Dr. Theodore Fujita, a giant in the field of tornado research, proposed a scale to classify tornado intensity by relating observed damage to estimated wind speeds. The scale quickly became the standard for recording and comparing tornado strength. In 2007, it was updated (the Enhanced Fujita Scale) to incorporate improved engineering knowledge about construction quality and how different structures fail under extreme winds.

The scale assigns specific wind speeds to specific damage indicators. For example, the level of damage to a motel can be used to estimate a tornado’s wind speed from 54 mph (loss of roof shingles) to 217 mph (total destruction).

Canada uses its own version of the Enhanced Fujita Scale, with a modified set of damage indicators and degrees of damage that reflect differences in construction practices, tree species, and patterns of rural land use compared with the United States. This will be important later.

Existential Damage Indicators

But what if there is no damage to be indicated? First, we must thank our lucky stars. But the scientists among us still want that sweet tornado data. Much of the USA and Canada is farm and prairie, devoid of structures and trees. And tornadoes are pretty common in these areas - especially in Canada. So damage is hard to find.

If there is not much damage to be had, then the tornado is assigned a scale of EF-Unknown (USA) or EF0-Default (Canada). As you can guess, this leads to undercounting tornadoes and tornadic damage in these remote areas.

To address this, a research team led by Connell S. Miller at the Canadian Severe Storms Laboratory studied a method for estimating tornado damage using satellite imagery of cropland. They published their research in the Journal of Applied Meteorology and Climatology.

Their method focuses on the Normalized Difference Vegetation Index (NDVI), a measure of vegetation health commonly used in the environmental sciences. NDVI is calculated using measurements of red and near-infrared light reflected from a surface with plant life, such as cropland. Specifically, it is defined as the difference between near-infrared and red brightness divided by their sum, producing values that range from –1 to 1 (statistics nerds refer to this as normalization). Higher values indicate relatively more near-infrared reflectance, which is generally associated with healthier, denser vegetation. This works because chlorophyll, more present in healthy growing plants, absorbs red light for photosynthesis while reflecting much of the near-infrared light.

In Practice

While NDVI has been used to document tornado paths in the past, it has not been applied within a consistent damage-assessment framework. Here, the researchers develop and test such a system by combining Canadian tornado reports (using the Canadian EF scale), crop inventories, and satellite imagery from 2017–2023 to determine whether NDVI can be used to estimate crop damage and inform EF-scale ratings.

Their tornado data comes from the Northern Tornadoes Project, crop data from the Canadian government, and satellite data from Planet Laboratories, which operates over 150 polar-orbiting satellites (and just as impressively got the awesome planet.com domain name). Planet Laboratories’ constellation consists of tiny CubeSats, which are barely 5x5x6 inches.

The images were mosaicked to create a single continuous image covering the full tornado path. For each tornado, the researchers selected one pre-tornado and one post-tornado image mosaic. The change in NDVI is calculated from this data. Next, they filtered out background noise from factors such as harvesting, irrigation, soil differences, and natural plant stress.

The Verdict

In the above image, the blue polygons are cropland, which have been color-coded to reflect the drop in NDVI after the passage of a tornado. The colored overlay regions indicate tornado/straight-line wind speed estimates derived from NDVI. The pin pricks are areas where damage was estimated using eyes on the ground (ground truthing), used as a comparison check. They identified 23 EF2+ and 15 EF0-Default tornadoes that had sufficient data across all their measures to warrant this type of analysis.

Overall, the method performed well for moderate to strong tornadoes (EF2–EF4), but was largely ineffective for weaker events, with only 2 of the 15 EF0-Default tornadoes detected. The approach was not sensitive enough to consistently capture low-damage footprints. Performance also depended strongly on crop type and seasonality, with the clearest results occurring when tornadoes affected actively growing crops (as opposed to planting or harvesting time). In the future, new technologies such as drone surveys could be used to provide more precise data (through lower-level and more repeated flights), to increase the effectiveness of this approach.

The NDVI method for estimating tornado damage won’t replace ground-based survey estimates, which continue to be the gold standard. But they can be used when such surveys are not available due to a lack of damage indicators, remote locations, or other reasons.

And Now for Something Completely Different

We publish this newsletter about every other week. But sometimes it’s not enough. That shaking feeling isn’t low blood sugar or the pending arrival of your in-laws. It’s your need for a meteorological fix. And we’ve got just the IV drip for you:

My bosses at the American Meteorological Society, a non-profit group that supports weather, water, and climate science worldwide, recently launched a new website that's your one-stop shop for all your meteorology news and updates.

There’s a lot here, from the latest news to profiles of scientists to columns by people like moi (though much more professional, polished, and sensible). If you want to keep up with the world of weather, water, and climate research, bookmark the site and enjoy losing a few hours each week learning how the atmosphere and ocean work on this strange planet of ours.

We are grateful to Lockheed Martin for a grant supporting this newsletter. We acknowledge additional contributions from Chris Nabholz.

Our archive of WWAT articles is here.

During the preparation of this work, the author(s) used ChatGPT-5.2 to copyedit text and create the mugging cartoon. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Weather with a Twist is published by the American Meteorological Society.

The Sirocco is a hot, dusty wind originating in the Sahara Desert that blows north across the Mediterranean Sea into Southern Europe. Some people love it, most hate it, but it is vital to the region’s ecology and leads to a fascinating meteorological phenomenon - red rain.

It begins with the formation or passage of a cyclonic low-pressure system in the western Mediterranean Sea, which rotates counterclockwise. As the system moves eastward, winds ahead of it blow from the south. These winds lift Saharan dust and carry it across the Mediterranean and into southern Europe. Strong upward motion along the system’s leading edge helps loft the dust high into the atmosphere, allowing it to travel long distances before settling out.

The air over the Sahara is typically warm and dry. As it moves northward across the Mediterranean Sea, it picks up moisture. By the time it reaches Europe, it often arrives as a warm, humid breeze. Over time, much of the dust settles out of the atmosphere under gravity. In some cases, however, the warm air encounters cooler air over the continent, triggering uplift, cloud formation, and rainfall that can bring the dust down in precipitation.

Regular Red Rain

Dust in the atmosphere serves as a condensation nucleus for water vapor, allowing it to condense at lower humidity than it would in clean air. As rising air cools, water collects on these dust particles, forming droplets. They grow through further condensation and collisions with other droplets until they become heavy enough to fall as rain.

As it falls, the rain appears no different from ordinary rainfall (though the sky itself can appear red due to increased atmospheric dust). But because the dissolved dust is rich in iron oxides (rust) and other minerals, rainwater that collects on the ground can appear reddish or brown. After it dries, it leaves behind a thin red film on cars, buildings, and vegetation. Sometimes this is referred to as blood rain, which can occur anywhere downwind of dusty regions.

A Bad Rap

In general, people don’t like the Sirocco and its red rain. The wind itself raises temperatures, limits nighttime cooling (due to increased humidity), worsens air quality, and coats cars, buildings, and laundry with fine reddish grit.

The Sirocco can trigger headaches, respiratory irritation, and allergies. And while rain is usually welcomed in these dry regions, sirocco rain tends to leave grime and stains rather than relief - unless your response to injuries is to rub some dirt on it.

Historically, this has led to the sirocco being blamed for everything from tuberculosis to having a case of the Mondays. As far back as 1815, a doctor in the Royal College of Surgeons published a paper associating the Sirocco with poor health and a general “depression of animal spirits.” Additionally, after blowing for seven days straight, he throws his arms up in the air and notes, “It gives a degree of lassitude, both to the body and mind, that renders them absolutely incapable of performing their usual functions.”

Peter Gabriel wrote his song Red Rain for a planned movie about villagers punished for their sins with a deluge of red rain. The movie never came to pass. But it left us one heck of a song.

Unearned Grief

Poor Sirocco. Poor, misunderstood Sirocco. No one gets you as I do. Look at me. I see you.

You may think of the Sahara Desert as a barren wasteland. And while it has very little water or plant life, its dust is surprisingly fertile, consisting mainly of fine mineral particles produced by wind erosion of soils and sediment. These particles contain nutrients such as iron and phosphorus, which are essential for healthy ecosystems downwind.

As the Sirocco blows this dust northward, it serves as a giant fertilizer for the Mediterranean Sea and southern Europe. The nutrients support plants/farms on land and plankton in the sea. It may also play an important role in climate change. Historically, increased mineral dust in the atmosphere has been associated with lower global temperatures.

Saharan dust affects far more than southern Europe. It can rise to high altitudes and travel thousands of kilometers, reaching as far as Siberia. More broadly, Saharan dust, including plumes blown westward over the Atlantic, accounts for roughly 55% of all airborne dust worldwide.

A Cultural and Natural Crossroads

This area of the world has always been known as a cultural crossroads - where African, Middle Eastern, and European influences mix into a stew that gives us flamenco, The Odyssey, and the practice of bashing dead octopi against rocks until they foam.

It is also a natural crossroads where two continental landmasses with different climates meet with a large body of water between them to magnify storms and moderate coastal climates. This creates the classic Mediterranean climate pattern of hot, dry summers and cool, wet winters, which is rare globally and tightly tied to the basin’s shape and latitude. Only about 2–3 percent of Earth’s land surface has a Mediterranean climate. It occurs in just five regions worldwide: the Mediterranean Basin, coastal California, central Chile, southwestern Australia, and the Western Cape of South Africa. No other major climate type is so fragmented and geographically limited.

For millenia, people around the Mediterranean have lived with the Sirocco and built stories, beliefs, and habits around it. The red rain has been interpreted as a warning, a curse, or something simply really, really annoying. Today we understand the physics behind it and why it’s important to ecology. Yet its cultural imprint remains, shaped by generations of people adapting to a wind that arrives from elsewhere and leaves its mark behind.

For more WWAT articles about wind:

• Deep in the Helm

• Flying Dirt

Weather with a Twist is published by the American Meteorological Society.

* During the preparation of this work, the author(s) used ChatGPT-5.1/5.2 to copyedit. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

But for most radars, there is a spot they can never see - straight up. Meteorologists colloquially call this the cone of silence.

The WSR-88D radars operated by the National Weather Service cannot physically tilt higher than 19.5°. Thus, they cannot see what is happening directly above them. Fortunately, most NWS radars sit within range of neighboring radars, allowing meteorologists to refer to those views when needed. It’s their version of asking a friend, “Do I have something in my hair?”

Vogelbadvergnügen

However, not all weather radars are alike. The DWD (Deutscher Wetterdienst - German Weather Service) operates a network of radars that can indeed look straight up. In fact, they do it all the time. It ain’t no thang (admires fingernails).

When a radar is pointed straight up, it is colloquially called a birdbath scan. The name ostensibly comes from the way it resembles a birdbath collecting water. I suspect it is also a gentler way of saying that it becomes something of an (ahem)… target for the birds.

The DWD radars routinely do birdbath scans during normal operation (nominally around every 5 minutes). Their radars operate on a C-band frequency, which is shorter than the S-band the NWS WSR-88Ds use. This shorter wavelength makes them more sensitive to various sources of noise, so they have to calibrate more often. The NWS radars also do routine calibrations, but they use other techniques.

Calibrating Hamburgers

Recall from a recent WWAT that most raindrops flatten like hamburgers as they fall. Differential Reflectivity (Zdr) measures the difference between the vertical and horizontal dimensions of that hamburger shape, and that tells us a lot about the precipitation.

Now imagine standing beneath the drop and looking straight up. The hamburger would appear as a perfect circle. Because of that, a radar looking straight up at the same raindrop should measure a Zdr of 0 dB. If the radar instead reports a different value, such as +0.2 dB, we know the system has a +0.2 dB bias. In that case, we can subtract 0.2 dB from future measurements from that radar. Sing with me now: ♫ “This is how we calibrate the radar using real-world data...” ♫

DWD recently upgraded their processing system so that these birdbath scans capture not just the usual reflectivity and velocity values, but the entire Doppler spectrum. This is the same type of information cloud radars rely on to study microphysics. Like a Pokemon leveling up, weather radars that were built to watch rain at a distance can now do much more.

Cloud Radar

Birdbath scans are the natural realm of radars designed to study clouds. These systems are highly sensitive, operate at a very high frequency, and are pointed straight up - all the time. My neck hurts just thinking about it.

As you can imagine, there are not many of these cloud radars around. And they are fixed in location and field-of-view. So they are at the mercy of the weather gods when it comes to studying passing clouds. However, the German DWD network has a bunch of C-band radars covering the landscape like bathroom lines at Oktoberfest.

The DWD’s C-band radars offer something cloud radars cannot: they keep working even in heavy rain or hail. They can peer straight into the core of a thunderstorm, track hailstones as they form aloft, follow melting snow as it descends, and watch rimed particles speed up as they grow heavier.

Two Examples

Researchers at the Meteorological Observatory Hohenpeißenberg (DWD) and Meteorological Institute, Ludwig-Maximilians-Universität München (LMU) recently published a paper in the Bulletin of the AMS studying how to use these radar’s calibration scans to study precipitation and cloud physics.

The researchers identified a place in Germany, Lindenberg Observatory, which had an operational cloud radar, which was near a DWD C-band radar in Prötzel, Germany. They used both to monitor a slow, soaking rainstorm as it moved across the area, and compared results.

The cloud radar at Lindenberg Observatory provided a detailed, second-by-second picture of reflectivity and fall speeds, which could be compared to C-band birdbath data from the radar in Protzel.

Despite updating only every five minutes and being far less sensitive, the C-band profiles was able to detect some of the same key features in the cloud radar data:

• Cloud tops rising and falling through the event

• A crisp, recognizable melting layer

• Strong vertical motion patterns

• Distinct zones where ice particles began to grow rapidly

Where things got truly compelling was in the Doppler spectra. At 7 km altitude, the spectra showed slow-falling tiny ice crystals. Deeper in the cloud, these merged into larger snowflakes, falling faster. Around 3.5 km, a second “slow” velocity mode appeared, a signature scientists often associate with secondary ice production, where ice particles shatter or splinter during collision or riming.

Another test event occurred on 29 June 2021 near Munich during an intense storm with heavy precipitation. A local cloud radar tracked the storm well until it was directly overhead, at which point the radar became saturated and could no longer detect anything within a cone-shaped region more than one kilometer above. When the storm moved over a C-band radar operating in birdbath mode about half an hour later, however, that radar was able to resolve the core of the storm with no attenuation.

Additionally, what it saw in the core was kind of amazing. It tracked 2–3 cm hailstones falling at about 10 m/s, directly following their descent from the upper parts of the storm as they continued to grow. (Ground observations confirmed the size of the hailstones.) This is the kind of hailstorm tracking operational meteorologists have long hoped for because they can support near real-time estimates of hail size.

At this point, it seems we could make a drinking game out of novel uses for meteorological radars. They were built for one purpose: tracking precipitation, and they do that well. Yet dozens of unforeseen applications have emerged from the same data. From tracking birds, wildfires, precipitation amounts, meteors, even highway traffic - weather radars are the perfect holiday gift that keeps on giving.

And Now for Something Completely Different

The Twisters 31 reference got me thinking more about the 2024 movie.2 So I dug into how well it performed. Apparently, it was a minor hit in the US but didn’t do as well overseas, which is increasingly important for Hollywood. It earned $372 million worldwide, but 72% was from North America. Usually, it’s about 50% for a Hollywood tentpole. These maps may explain why:

Tornadoes are such a major part of meteorological cultural lore (and research) that I kind of assumed everyone cared about them. But that was ethnocentrism at work. Tornadoes aren’t really a thing in most places. But if you have ever hidden in the basement or bathroom during a tornado warning, you probably were at least intrigued by the movie. This relationship between the movie’s revenue and tornadic activity exists even at the state level. In the USA, per capita revenue in the tornado alley major markets mostly outpaced revenue from outside the region.

I wonder if this will impact future Hollywood movie choices. There haven’t been a bunch of weather-themed Hollywood movies (which blows my mind, TBH). But consider the heatmap for revenue from the Frozen movies:

So I guess, globally, ice beats thunderstorms if it comes with sibling rivalry and catchy songs.

We are grateful to Lockheed Martin for a grant supporting this newsletter. We acknowledge additional contributions from Chris Nabholz.

Our archive of WWAT articles is here.

* During the preparation of this work, the author(s) used ChatGPT-5.1/5.2 to copyedit text and edit the DWD radar image. The original photo showed the radar dish at approx. 45 degree tilt. Gemini and Bing were used for the box office heat maps. WWAT’s author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Weather with a Twist is published by the American Meteorological Society.