It seems like the first serious snow of the season usually catches us off guard. That’s human nature. And when I say us, this time it includes meteorologists. That’s the nature of the job. Forecasting the amount of precipitation is difficult enough — forecasting how much will be snow vs rain is that much harder. You have to get them both right to nail a snow forecast. The heavy northern panhandle snow Thursday and Friday last week (10/17-18) is a good example, and I’ll go into more detail below. But first, who got how much? Surprising.

What was the amount and extent of the snow?

Haines townsite was again* the bull’s-eye for this snow event. The official town climate station, known as Haines #2 and staffed by yours truly, measured 19.3 inches (49 cm) for the 2-day storm total, basically Thursday morning through Friday evening, with a little bit added overnight Friday night. Water equivalent was 1.36 inches (35 mm). Overall a very typical 14:1 ratio, but it started much wetter and ended much drier. The “Haines 0.5 NNE” CoCoRaHS station on Skyline Drive (371 ft/113 m elevation) reported 20.5″ (52 cm), a trained weather spotter report of 22.5″ (57 cm) came in from the northwest part of town, and the Ripinsky Ridge remote automated station at 2,500 ft/760 m) showed 24 inches (61 cm). Heading northwest up the valley, the precipitation dropped off sharply, and the border station had only 0.4″ (1 cm) of snow. Such is the localized nature of this kind of snow maker.

Our neighbor to the north, Skagway, only ~17 miles/27 km away had much smaller amounts as well. Exact figures are hard to come by there sometimes, but the amount of liquid equivalent at the airport was only about 25% of Haines’ and there was little on the ground after the weather cleared.

The cold north flow off the interior that set up the snow situation hit Skagway first, as would be expected. Check out this dramatic plot from the Marine Exchange of Alaska (MXAK) weather station there. Not hard to pinpoint when north wind picked up, which ushered in the snow:

At Gustavus, rain changed to snow about 11pm Thursday night about an hour after the wind shifted from southeast to northwest bringing temperatures down from an earlier high of 43F (6C) to just above freezing. (Snow often falls at above-freezing temperatures; very commonly at 33-35F (0-2C), many times in the upper 30s sometimes even around 40F (4C).) They picked up 7-8 inches (~20 cm) during the event. Across the strait Hoonah had only a couple inches (~5 cm). Out west, near the open ocean, rain was heavy but the cold air was too weak to create much snow: Elfin Cove pickup up 3.4 inches (9 cm) of very wet snow and Pelican just a trace. Look at this fun polar graph of the wind direction at Cape Spencer at the ocean entrance west of Glacier Bay, for 24 hours ending Friday at 10 am. The wind there was an onshore wind varying between NW and S early on (inner part of the circle) then swung to the SE then E as it built to 20-25 kts (10-12 m/s), an offshore “outflow” wind. Another calling card left by the cold air’s advance.

It took a while, but the cold air moving south out of Canada made it to Juneau Friday evening and the rain there gradually turned to snow at around 5pm at the airport, though out the road north of town up to around 6 inches (15 cm) of snow had fallen by Friday late morning. By the time the system dried out, the Mendenhall Valley was left with only a couple inches, and downtown less. But the rainfall totals were healthy in this southern part of the story, with reports varying from 1.36″ (35 mm) at Lena Pt, to 2.65″ (67 mm) in the upper Mendenhall Valley. Abut 30 miles SE of Juneau at the city’s Snettisham hydro project, the two day total was 8.00 inches (203 mm)! That’s liquid precipitation. Snowfall: zero. But think of what how much snow might have fallen at the higher elevations north and east. Chalk that up to terrain uplift more than frontal uplift (more on this below).

*That’s often the case but certainly not always. January of this year Juneau got clobbered with two big storms over 12 days, totaling 65 inches (165 cm) of snowfall, while Haines got only 16 inches (41 cm) during the same period.

Was this early/heavy snow typical? unusual? rare?

The local rule of thumb in Haines is that the rain will switch over to snow close to Halloween. How good is this rule?

It’s eerily good.

As you can see (keeping in mind the missing data from 1956-1972), the first measurable snow (not including traces) is almost always between Oct 15 and Nov 15 in Haines, and more often than not within a few days of Halloween. So 2024 was on the early side but still in the common range. No discernible trend in the date over the last ~100 years.

How big was this snow dump viewed against Haines’ fairly long history of snow measurements?

Bottom line: it was one of the biggest of the few big October snows on record, including just edging out a 99-year-old record.

Technically, we set a record for greatest 2-day snow in October, but there are lots of caveats. First, as noted above the records are gappy. About 1/4 of the last 99 years are missing. Second, our two day snowfall was so barely a record that I’d call it a tie. In October 14 and 15, 1926 the 2-day total snow was 19.0 inches (48 cm) and the water equivalent was 1.36 inches (35 mm). This year we had 19.3 inches (49 cm) with water equivalent of 1.36 inches/35 mm, yes exactly the same on the liquid. But in 1926 they seemed to be rounding the snowfall to the nearest inch most of the time, so who can say we really beat it. Plus, it’s pretty much impossible to accurately measure the snowfall to the tenth of an inch, especially with wind. Finally but perhaps most important, the measuring locations were different between the two event, we’ve certainly seen how much amounts can vary around town.

The above plot shows snow depth history for the Haines #2 climate station (only operating since 2000). This year is in red, the average in black and the other years in the fainter lines. This highlights how unusual the the recent event is since 2000. How about earlier? Here are top 15 events for both the cooperative observer station (Haines #2) since 2000 and the “airport” which is really an amalgam of various locations around Haines, mostly not at the airport, from 1911 to ~1997 with lots of gaps. These are records restricted to September and October only:

rankings from XMACIS

OK, how does this compare on a state-wide basis? Here are the high performers again only allowing events up to October 31:

Note how the top of the list is almost all SE AK or northwest along the coast to Prince William Sound. More surprising is some of the big snows in the interior and at Utqiagvik (Barrow). This list is not necessarily complete or guaranteed accurate, though I don’t see in obvious bogus data on this screen.

What was behind all the snow?

The most common cause of heavy snow in the northern panhandle is overrunning. It’s basically a front: warm air being forced aloft when confronted by cold (denser) air. The movement of this interface (front) determines what we call it: a cold front is when cold air is advancing into where the warm air was, and a warm front is when the warmer air takes over cold territory. Often its a standoff, a stationary front.

This stationary condition is common in this part of the world where large mountain barriers can affect the balance of power of the air masses. The graphic above shows a stationary front as it is found in the interior. Now add the coastal mountains to buttress battle line. In winter, cold interior air does not easily rise to cross mountain ranges, so gets funneled through gaps and passes. This cold air now is in the typical territory of the warm, wet marine air. This battle tends to happen close to that mountainous coast. The cold air usually can’t push on south over hundred’s of miles of ocean, and conversely the warm air has a very hard time dislodging the cold air with it’s back to the mountains and more cold air rushing down to reinforce. The rising motion of the warm air hitting the cold air (and the mountains) wrings precipitation out of the air. If the temperature is right for snow, that’s what we get.

Why so much, and why over such a limited area?

Think about the front described above. Both sides are getting reinforcements. The warm moist air being supplied by ocean winds blowing from the south or east while the cold interior air is getting sucked off the interior through the gaps and passes, blowing from the north or west. Here’s what it looked like Thursday morning after the cold air had recently made it past Eldred Rock (just above center of map).

The opposing directions set up the overrunning situation. Here’s a plot that illustrates this, valid Friday morning at 10 am over Haines. This is simulating what a traditional balloon sounding would look like, but based on model data. (The nearest actual balloon soundings are Yakutat, Whitehorse, and Annette Island, all too far from the action to see the Haines situation).

The red line is the temperature, the green the dew point. The fact that they are close together tells us the air is very moist. The bottom of the temperature trace is labeled with the surface temperature: 27F or about -3C. The slope of the line tells us that temperature decreases with height for a distance, then curves back to the right, showing a layer of warmer air above cooler. This is where the warm, moist ocean air is pushing in from the south. Over the ocean it was at the surface, but when it meets the cold air coming down from the north it can’t hold the low ground. The cold air is considerably denser, plus it is backed up by the mountains, so the warm air slides up over the cold, and over the mountains. Rising, moist air makes precipitation. This happens all the time with the mountains alone, as a look at average precipitation maps would confirm.

Now look at the wind barbs on the right vertical axis of the plot. The winds near the surface are from the north at 10 knots (5 m/s), then somewhere around 4-5,000 ft (~1,400 m) above ground level they go light and then switch to south at 30 knots (15 m/s). There are your two opposing air masses each trying to get to the other side of the line. Note how all this is happening in the lower part of the atmosphere, below 8-10,000 ft. (~3,000 m) This is why the mountains are so important, as they are right in the way at these levels.

But with the this front, the area of strong uplift and cold enough temperatures moved south from the main mountain backbone and parked over Haines long enough to dump almost two feet of October snow, then drifted south to bless other areas to lesser degrees, petering out around Juneau. If the balance had been a little different, the front might have held back, and Skagway would have gotten more snow, or the Haines border station, if the alignment of the pressure gradient caused the cold winds to favor different channels. Or it could have pushed south quicker, leaving a lesser amount in Haines but stalling over the Icy Strait area, or Juneau, for instance, changing more of their rain to snow at ground level. Or if the cold air were less cold, or the warm air warmer, much or all of the snow could have melted aloft and fallen as rain. I think you can see how  complicated and sensitive to small variations the situation can be, and thus how hard to hit the forecast every time.

Thanks for reading, and please, I love to hear feedback, questions, etc. If you are curious about how to find some of the data and graphics I use in these posts, check out my curated Alaska Weather Links where by choosing categories, or with a keyword search (I’ve provided good keywords in the attributions under the graphics in this post) you can find each of the sources I’ve used today, plus much more.