A whole bunch of unusual, rare, and wild weather is happening simultaneously all over North America

One-sentence summary

Between extreme warmth, tornadoes in odd places, raging fires in Texas, and a blizzard in California, there is no shortage of wild, rare weather to close February.

Tornadoes in Michigan in February?

In and of themselves, tornadoes in Michigan are not uncommon. Tornadoes in winter are rare. Only six previously documented tornadoes have occurred in Michigan in December, January, or February since 1950. Add another two to that after last night.

Storm reports from Tuesday show several tornado reports in red dots in Ohio, Michigan, and Illinois. (NOAA SPC)

According to the Storm Prediction Center, two tornadoes were reported yesterday, one near Marshall and another near Grand Blanc, MI. The other winter tornadoes included one in January 1996 near Richland, MI (just outside Battle Creek and Kalamazoo), one west of Detroit in December 2015, and then four on the last day of February in 2017 near the Indiana border.

So, winter tornadoes are rare to begin with, yet having one as far northeast as Grand Blanc, which is near Flint seems completely out of place in the wintertime. This comes on the heels of Wisconsin recording their first February tornado on record earlier this month. January tornadoes have occurred in Wisconsin, with Janesville on the receiving end of an especially bad one in 1967.

Front page of the Janesville, WI Daily Gazette from January 25, 1967, the day after a tornado ravaged the area. (Newspapers.com)

Tornado statistics are tough to do a whole lot with authoritatively. Some tornadoes are very brief and never reported. Reports have increased dramatically since records started being kept, so it’s entirely possible we’re missing some reports from back in the day. But however you want to slice it, tornadoes this far north during this time of year are straight up rare. And to have more than one distinct event in a very far north part of the country this close in proximity to one another further underscores how unusual this is.

Record warmth over a tremendous swath of real estate

On the day of the Janesville tornado noted above, they hit 60 degrees before dropping to about 25 the next morning. Interestingly, if you look at Cold Spring, WI just northeast of Janesville, they pulled off one heck of a change yesterday to today. Temperatures were over 71 degrees yesterday and dropped to below 13 degrees this morning.

Dozens and dozens of records were smashed this week, including multiple monthly temperature records. Monthly records fell in Quebec, Ontario, Missouri, Illinois, Wisconsin, Michigan, Minnesota, and Arkansas, according to Maximiliano Herrera, who keeps an exhaustive record of these things.

Dozens of daily record temperatures occurred yesterday and this morning across much of the Midwest, Mid-Continent, and Great Lakes. (Coolwx.com)

Not only are some of these monthly records being replaced, they’re being replaced by several degrees and being attained in multiple, disparate spots in these states. Even Mexico has seen incredible extent of record warmth. The breadth of the warm weather we’ve seen at times in the last few years boggles the minds of many meteorologists, including those like ourselves that tend to view things from a grounded, level-headed perspective.

Texas Panhandle burning

Over the last few days, a number of fires have broken out in parts of the Texas Panhandle. Fire behavior went off the rails yesterday, as strong winds ahead of a cold front and strong winds and dry air behind the cold front allowed for extreme growth and spread.

A annotated map of the extent of various Texas Panhandle fires as of Wednesday afternoon. (Wildfire.gov)

We’ve seen bad fires in Texas before, especially in the Panhandle. This in and of itself is not necessarily uncommon. March and April tend to be the big Texas fire months historically, so this is certainly at least a bit early. Texas’s largest wildfire (going back to the 1980s) was the East Amarillo Complex, which occurred in March of 2006. The Smokehouse Creek complex is now the second largest on record in Texas, thanks in part to how quickly it grew yesterday. What is intriguing to me is how this happened despite that part of Texas being mostly out of drought. Compare the Drought Monitor maps below showing current conditions to the conditions preceding the 2006 East Amarillo complex. There are notable differences here.

An image comparison of the current drought status in the Panhandle and the drought status preceding the largest wildfire in Texas history, the East Amarillo complex. (U.S. Drought Monitor)

Whatever the case, the burn scar is clearly visible this morning on satellite imagery, a depressing coda to a pretty terrible situation for the Panhandle.

The burn scar from wildfires in the Texas Panhandle is visible on satellite imagery today. (College of DuPage)

California blizzard warnings for the Sierra

When we think of the Sierra Nevada in California, snow certainly comes to mind. But not often do Blizzard Warnings get issued in that part of the world. The storm that will impact them over the coming days has prompted some pretty severe warnings about travel and conditions within the mountains. Upwards of 5 to 10 feet of snow could fall, likely shifting the Sierra from deficit to surplus on the season in terms of snowpack.

Timeline of the upcoming storm in the Sierra that will dump 5 to 10 feet of snow in some spots. (NWS Sacramento)

The last such blizzard type event literally occurred a year ago. This storm looks even more powerful.

Additional rain will occur for parts of California as well, including a chance of thunderstorms for places including San Francisco.

Between rain and the water contained in the snowpack that’s going to build this weekend, a whole lot of water is headed to California again. (Pivotal Weather)

The bottom line is that the weather is pretty wild right now. While some of it is typical for a spring-type setup, a lot of abnormal, unusual, and rare things are occurring simultaneously, a theme we’ve grappled with on more than one occasion recently.

Seasonal hurricane commentary is beginning to buzz: What we actually know right now

One-sentence summary

With early hurricane season commentary beginning to emerge, we take a look at what is known here in late February.

Tropical aggression

Yesterday, AccuWeather, the private forecasting company from State College, PA published an article expressing concern for the upcoming hurricane season, using words like “super charged” and “blockbuster.” I’m not writing today to dunk on AccuWeather. I personally know a number of folks that work there, several of whom I would consider friends of mine. But I want folks to understand both the realities and limitations of hurricane forecasting in February.

About a month ago, we posted an article in which we shared some commentary about the upcoming hurricane season. In it, I wrote that “the combination of warm water and a weakening El Niño probably suggests an active hurricane season ahead in 2024.” I also noted that you’d probably be seeing some aggressive seasonal outlooks. Consider that article from AccuWeather to be a first or second volley. It’s important to note a couple things though. As is often the case, it’s the buzzwords that grab people’s attention and leads to accusations of hype in some circles. But the content of their actual article is fundamentally sound! There is one element I take issue with, and that’s their description of the Texas coast being at higher risk than usual for the upcoming season.

Forecasting landfall patterns consistently this far ahead of time is really, really hard. I’ve never found anyone that has done it successfully and at a consistent level. Anyone can claim they nailed a seasonal call this far out, but they’ve probably also had 50 false alarms to go with their handful of victories. Colorado State University, which releases a highly anticipated outlook each season has actually put their verification metrics online, and they did well this past season. But while they offer a landfall probability table by state, they are intentionally broad in their discussion, because specifically narrowing down a section of, say, a state’s coastline being at higher risk than another is difficult. If Texas is at higher risk, does that mean Louisiana is not? What of Mexico? At what point does “higher risk” kick in?

I don’t want to fault anyone or say I am against seasonal forecasts. I am not. But we have to be extremely careful how specific we get in February. The peak of hurricane season is still about six and a half months away.

So where do we stand now?

If you read last month’s post, the general takeaway was that there was no reason to forecast anything other than an active hurricane season in 2024. How active and how bad was TBD. Has anything changed since then? Not really.

First, a look at sea surface temperatures (SSTs) shows that the Atlantic remains obscenely warm. Obscenely warm.

Sea surface temperatures in the Atlantic Ocean, especially the Main Development Region west of Africa are significantly above normal. A lot can change between now and summer, but these temperatures are more typical for July than February. (Weather Bell)

Seeing this in February is truly remarkable, as these readings, specifically in the Main Development Region are more normal for July. It’s February 21st, so a lot can change between now and June or July or August. But, suffice to say, this is as strong a signal for an active hurricane season as you would likely ever see. So, the SST checkbox is checked for “Active” for another month.

What about El Niño? Well, there are ample signs that it is beginning to wane as expected. The animation below shows SST anomalies from the Equatorial Pacific Ocean from the ocean surface (top) to 450 meters below the surface (bottom). Notice how a lot of the deep warm water, particularly on the right side of the image (the Eastern Pacific) gets eroded away in the last few frames. That’s a sign of colder water making a return to the ocean’s subsurface in the Pacific. It won’t take much to begin to wind down the El Niño from this point.

El Niño is being slowly devoured from below by colder water in the Pacific. (NOAA)

Again, it’s not uncommon for an El Niño as strong as this one to quickly weaken in springtime. And it appears this year will fit that trend. But where it goes from here is equally important in terms of how it may impact hurricane season. The latest long range forecast from a slew of government agencies shows about 2/3 of ensemble members taking us to La Niña by mid-summer and about 1/3 holding us back.

The latest C3S multi-system forecast for ENSO shows about a 65/35 split among various model ensemble members taking us into La Niña this summer. (ECMWF/Copernicus)

Ensemble forecasting is the best approach most times, especially at this timeframe because instead of one deterministic outcome, we get to see generally what the consensus is and where any reasonable subset of outliers exists. I look at these charts frequently, and a couple items stand out right now. First, the majority take us into La Niña by July, if not June. Second, some of these model members really take us into La Niña, perhaps moderately so by mid-summer. However, there is also a subset of members that pull a Lee Corso, trying to drift us back into El Niño after dragging us back to near neutral. The takeaway here is that the odds seem to favor La Niña, and indeed, NOAA did recently issue a La Niña Watch. But there is still some inherent uncertainty, and oh yeah, we are in what is termed the “spring predictability barrier,” which is a fascinating study in how models can be truly imperfect. I do think they’re getting better at this and perhaps this barrier isn’t as strong as it once was. But the stench of busted forecasts past does still linger in close enough proximity for me to proceed with caution.

So, the verdict? An exceptionally warm Atlantic Ocean, and odds favoring La Niña development this summer continue to suggest an active hurricane season ahead. But where those storms go, how strong they are, and when they’ll be at their worst is a forecast that is close to impossible to make at this juncture. As always, if you live in a hurricane-prone region, the best advice in any hurricane season is to be prepared as if that season were the one.

Snowstorms like the one on Tuesday offer some of the most difficult challenges for consistently accurate weather forecasting

One sentence summary

Snow forecasting is hard, and we’ll use yesterday’s storm in the Eastern U.S. to show off some of the inherent challenges that exist.

Yesterday’s storm exposed the limitations of making consistently good snow forecasts

As a native of New Jersey and having spent the first 5 winters of my career in the lake effect snow belt, I know how tricky snow forecasting can be. There are endless amounts of complications with snow: Elevation, air temperature and snow ratio, dry slots, ground temperatures, sun angle, mixing, vertical velocity and relationship to the snow growth zone in the atmosphere, and on and on. It remains one of the more elusive forecast problems. My own opinion on this is that our ability to forecast snow has hit a bit of a plateau. I will admit that I don’t know if statistics actually bear that out. Hurricanes continue to show improvement, temperatures show improvement, but snow? Precipitation in general? I don’t know how much better it is today than 5 or 10 years ago. Feel free to blast me in the comments if that’s not accurate.

But, look, even if forecasts have improved considerably in the last 5 to 10 years, the inherent challenges with snow forecasting will always remain. Here’s an example.

An impressive snowfall gradient in northeast Pennsylvania from yesterday’s frustrating snowstorm. (NOAA)

Yesterday morning, in northeast Pennsylvania, east of Scranton there was a pretty interesting scenario. In Shohola, on the Delaware River border with New York, 15 inches of snow was reported at 10 AM. About an hour later, roughly 10 miles west of Shohola, PA a total of 3.5 inches was reported just northeast of Lake Wallenpaupack. Doing some quick math, that comes out to near a foot of difference over 10 miles, or an average of over 1 inch of change in snow accumulation per mile.

Now, does this prove anything? No. We don’t know whether the measurement in Shohola was perfectly reliable. We don’t know what elevation it was taken from. I suspect the 3.5 inch report was from about 1,200 feet in elevation, and I don’t suspect the Shohola report was from much higher an elevation. But, you never know. Other reports nearby seem to corroborate it enough. But even if it’s 12 inches instead of 15, the point is still clear here: Snowfall gradients like these are common in Eastern U.S. snowstorms. They are the rule, not the exception.

Radar image from northeast Pennsylvania around 6:30 AM ET on Tuesday. (RadarScope)

If you look at radar from yesterday morning in Pennsylvania, one of the key hurdles in snowfall forecasts shows up. We call this mesoscale banding. Weather modeling has come a long way in the last 10 to 15 years, but one of the trouble spots can occasionally be these mesoscale bands. You’ll get widespread, large scale snowstorms, like the one yesterday in the East. But embedded within the overall precipitation will be these narrow, sometimes stationary bands of snow. For those of you on the Gulf Coast, think of what happens when you get training lines of thunderstorms that cause flash flooding. In storms like the one yesterday, by a different process, you get a similar outcome but with snow. On the map above, you can see the roughly 10 mile separation between Shohola and Kimbles (where the 3.5 inch report came near). Over that distance at 6:30 AM ET, the radar shows heavy snow near or just southeast of Shohola, diminishing to light to moderate snow 10 miles west of there. The difference between the heavier snowband in yellow and the generic light to moderate snow can easily be 1 or 2 inches an hour. So, if that band parks there for 3 hours, you could easily see 5 to 7 inches of snow under it with 1 to 3 inches just west. Extrapolate that over a couple more hours or increase the snow rates in the heavier mesoscale band and, well, you get the idea.

Let’s look at another example. This one comes from Connecticut yesterday.

Radar loop from Tuesday morning in Connecticut. Note the heavy, nearly stationary band of snow just northwest of Hartford. (RadarScope)

A report of 15.5 inches of snow occurred near Farmington, CT, which was placed precariously under that heavy band of snow you can see northwest of Hartford and north of I-84 if you look at the radar loop above. Notice how the area northwest of that snowband into Litchfield County, CT sees almost none of that. Well, that area to the northwest ended up with plenty of snow, about 4 to 8 inches worth in many spots.

Snow totals in northern and western Connecticut (a few of these are incomplete) shows 4 to 8 inches in northwest Connecticut with a very well defined corridor of 10-15 inches just north of I-84 and northwest of Hartford. (NOAA)

But you come just southeast of there, in the area from near Danbury through Waterbury through Bristol and Farmington out toward Tolland, and you see 10 to 15 inches of snow. Over short distances, you could have gone from 7 or 8 inches to 14 or 15 inches. Granted, both cause impacts, but nearly doubling up snow at those amounts can have severe ramifications on travel or forecast perception.

A rough estimate of snow totals yesterday based on available reports. (Tomer Burg)

So why am I devoting a whole post to this topic? Well, for one, the East has seen a bit of a snow drought the last couple years, and it’s nice to revisit the idea that this stuff happens. But also, I saw a number of posts about threats, real or exaggerated, received by meteorologists because the forecast of this event busted in spots. Just a couple days ago, it looked like a broad, major New England storm. Some accused meteorologists of lying or just not being honest.

Here’s the truth: Snow forecasting is inherently difficult. You see the images above. Imagine now if that mesoscale band is forecast to be over Long Island and Central and Southern New Jersey in modeling ahead of the storm. Instead, when the storm happens, the band shifts northwest by about 25-35 miles and ends up over Philadelphia and New York City instead. Not to discount the southern half of Jersey (my homeland) or Long Island, but you’re talking about a shift where millions of people see considerably more snow than perhaps they expected. Or, even worse for perception, it ends up out to sea by 25 miles and now a few million fish get heavy snow instead of a few million people? Now, imagine if the models just simply give you little to no indication that the mesoscale snowband is going to setup at all? In these events, everything is connected. Trying to pin down storm track, trying to predict if and where banding will occur, trying to factor in temperatures in the air and on the ground (like yesterday where many places saw a lot of falling snow but not a lot of accumulating snow), or forecasting things like sleet or rain mixing in: It all matters. And the reality is, we’re decent at it all, but we’re not consistently great at it, and neither are models.

So it’s important to recognize that meteorologists try very hard to do good work during these events. Sometimes things change. Sometimes it’s at the last minute too. And that can lead to a significant forecast bust in some places. Or less a bust and more of an over or underestimate. It will happen again, and because of the geography of the Northeastern Megalopolis, there will be times where millions of people get outcomes that are somewhat unexpected. Just remember: No one is misleading you or lying to you. The atmosphere likes to challenge us, and some days we are left with egg on our face.

As California looks to a stretch of quieter weather, some signs that more rain is on the horizon

One-sentence summary

After an impressive rainstorm in California, quieter weather will settle in for a few days before we watch the potential for another rain event in the region beginning around next weekend.

California cleaning up

After a wild sequence of rain in California, the time for clean up has begun.

Click to enlarge a map of the last week’s worth of rain totals in the LA Basin and surrounding areas. (NOAA)

The amount of rain that fell in some places is staggering. California is always a place of extremes, and with these extremes seemingly happening more often and at greater amplitudes, it’s making for a lot of whiplash.

In quite a change, Sierra snowpack now ranges from 68 percent of normal in the south to 78 percent of normal in the north. Those values were 36 and 60 percent respectively a week ago. While these storms come with damage and a lot of headaches, on net this is probably good news for California. The Sierra needed snowpack replenishment desperately. A year ago, after another sequence of ferocious storms, the amount of water in the snowpack ranged from 170 to 240 percent of normal.

Sierra snowpack has made major gains over the last week, but it’s still below average for this time of year and severely lagging last year’s high-end pace. (California DWR)

For the rainy season so far, San Diego has seen 7.88″ so far, about two and a half inches above normal. Downtown Los Angeles near the USC campus has recorded 15.43″, which is over seven inches above normal. Almost 90 other years back through the 1870s did not even see that much rain during the entirety of the rainy season in Los Angeles. Downtown San Francisco has seen over 16 inches this rainy season, which is about 3 inches more than usual to this point. As is typical for El Nino, the anomalies are strongest in SoCal.

What’s next? A break, but then potentially more rain

Things really do calm down for California over the next few days with one more round of rain, then some showers, then a dry stretch.

Additional rainfall expected over the next couple days in southern California. (Pivotal Weather)

Rainfall over the next week or so should be manageable, though some residual issues due to mudslides and high water will continue. But new problems should be minimal.

Beyond early next week, things start to change again, and it’s possible we have another round of significant weather aiming for California beginning around next weekend. Why? If we look up at the jet stream level on today’s run of the European model, we can see what’s happening. Initially, California is in a calm bubble of sorts. This is what will help restore fair weather to the region.

The jet stream relaxes into next week, but another surge of Pacific moisture may be on the horizon, aiming for Mexico or California. (Tropical Tidbits)

If you look out over the Pacific Ocean, you’ll see the jet stream winds again stretching and extending toward the U.S. West Coast around late next week or weekend. This shows up on most modeling and has for a few days now. Thus, it seems reasonable to think that another round of storminess may approach California around that time. In fact, the Climate Prediction Center has again highlighted California and the West for another round of heavy rain and snow risk beginning around February 17th.

The Western U.S. will again be in line for the risk of another round of heavy precipitation beginning around next weekend. (NOAA)

Modeling backs this up, with the GFS Ensemble from earlier today showing anomalies of about 1 to 2 inches on average across California from February 16th through the 23rd. Ensemble means tend to “wash out” extremes by design. In other words, because you’re looking at an average of a 30 member ensemble, you will lose some of those extremes in favor of a dominant signal. What we know? It will likely get wet again in California. What we don’t know? How wet.

GFS Ensemble mean forecast precipitation anomaly for February 16-23, showing above normal precipitation in California, Arizona, and Texas. (Tropical Tidbits)

All that said, this is looking like another potentially robust multi-day, multi-inch rain event next weekend into the following week in California. A good idea to stay tuned.