Autumn Climate Data

Temperatures have been dropping off sharply in the past 10 days in the interior, and the change has felt more acute because of the unseasonable warmth before that.  Fairbanks didn't see anything more than a mild freeze until three weeks ago, when the temperature first dropped into the low 20s Fahrenheit (about 3 weeks later than usual).

The last week or so has felt very different.  Starting on November 4, high temperatures have been below 20°F; the first sub-zero low temperature occurred on the 9th; and then yesterday the high was -1°F.  Winter has most definitely arrived.

Last month I missed my usual monthly climate summary owing to the focus on ex-Typhoon Halong; so let's look back at September and October.  Both months brought unusually low pressure to northern Alaska and stronger than normal westerly flow across the state.  There was a particularly strong trough over the western Bering Sea in October, and this bears part of the blame for pulling Halong up from the south.  Both months saw above-normal high pressure to the south of the Aleutians.

The southwesterly orientation of the flow in October explains the unusual warmth in Fairbanks, and the anomaly was widespread; but September was somewhat cool in western Alaska, as the air had a more northwesterly origin.  Both months were very unusually warm in the southeastern interior and along the Gulf coast, according to ERA5 data:

Recall that early September saw a very unusual heatwave in the southeastern interior, and there was another rare warm extreme in early October: Northway reached 62°F on the 8th, with a low temperature of 40°F (easily the latest on record for such a high daily mean temperature).  This was followed by unrelenting warm anomalies for several weeks:

Another theme of the two months was the unusual excess of cloud cover across large parts of Alaska.  October was especially cloudy for the interior, according to ERA5 data.

Looking at cloud cover data from Fairbanks airport, the situation has been dreadful since the middle of August.  Until this week only one day (September 17) reported mostly clear skies for more than 20% of the day, and the fraction of hourly reports with mostly clear skies is easily the lowest on record for the September-October window.  This is so anomalous that I think it's worth a separate post after looking at data from the CRN solar radiation instruments.

Not surprisingly, precipitation was very abundant across much of the state in October, and northern Alaska was substantially wetter than normal for three months in a row (August through October).  Bettles saw 10.6 inches of precipitation in the three month period, the second highest on record for that period.

The active westerly pattern produced above-normal wind speeds in October (in part due to Halong of course), but September was relatively calm for the central and eastern interior.

Finally, it's worth noting that the enhanced wind speeds across the northern North Pacific produced a dramatic drop-off in sea surface temperatures from just north of Japan to the Gulf of Alaska.  Compare the two figures below (August versus October):

The North Pacific Mode (NPM) SST index, which I revisited recently here, has dropped off precipitously:

Given that La Niña will prevail during at least the first part of this winter, the disappearance of the positive NPM probably means a greater chance of cold in Alaska; but much will depend on the high-latitude flow, which is about to become very perturbed in connection with a rare upper-atmosphere disturbance over Arctic Canada.  More on that later.

Ocean Temperatures and Halong

In an earlier comparison of ex-typhoon Halong versus other historic autumn Bering Sea storms, I noted an intriguing similarity: Halong, Merbok, and the major storms of 2011 and 1974 all occurred during prolonged La Niña episodes in the tropical Pacific.  It's worth digging into this a bit more to see what might be going on.

Looking back at historical data since 1950, there have been 11 years - including this year - when La Niña occurred in late autumn for the second or third consecutive year.  (It's not uncommon, by the way, for major La Niña episodes to extend over multiple years, in contrast to El Niño, which tends to be "one and done".)  Here's the average sea surface temperature (SST) anomaly pattern in September through November of those years (relative to the long-term trend):

For comparison, here's the September-October SST anomaly this year:

There's a lot more warmth in the northeastern Pacific this year compared to the historical pattern, but there is a similar contrast between a cool tropical Pacific and warmth extending eastward from Japan.  The warmth from Japan eastward is very characteristic of a negative PDO phase, and indeed the PDO has been strongly negative in recent months.

Zooming in on the North Pacific, it's interesting to see that Halong and Merbok both passed over very warm ocean water (relative to normal) to the south of the Aleutians before moving up into the Bering Sea.  Based on the analysis above, that region of warm water is typical of persistent La Niña and negative PDO regimes.

The storm of 2011 also originated over a region of warm water, although the North Pacific as a whole was much cooler back then.

It's tempting to speculate that the extra warmth and moisture available from the unusually warm ocean to the south of the Aleutians provided more fuel for these storms than would be derived from cooler oceans.  Rick Thoman mentioned to me that modeling research is already under way to quantify the role of the unusually warm SSTs for Halong, with preliminary results showing a significant impact.

Another aspect of the persistent La Niña/negative PDO SST pattern is that the north-south ocean temperature gradient is greater than normal near the Aleutians, and this gradient tends to enhance the North Pacific jet stream, providing more upper-atmosphere support for strong extratropical cyclones.  Here's the average 500mb height anomaly in the same "persistent La Niña" years; notice the strong trough from the East Siberian Sea to Alaska.

Again, this is not too dissimilar to what we've seen this autumn so far:

It's worth noting that autumns in which La Niña is just developing do not have the same magnitude of warmth to the east of Japan, and they lack an upper-level trough near the Bering Sea that would support (and reflect) strong storm activity.  The PDO tends to be less negative.

Part of the reason for the difference is that La Niña favors high pressure ridging over the central North Pacific that gradually produces warmer ocean temperatures over time, and so the warm anomaly east of Japan becomes more amplified by the second or third consecutive La Niña winter.  In contrast, an initial La Niña is often developing after El Niño, and so the mid-latitude North Pacific SSTs tend to be cooler owing to the lingering El Niño influence (lower pressure with more wind and cooler conditions to the south of the Aleutians).

An interesting corollary to this discussion is that the long-term trend seems to be favoring more frequent and persistent La Niña and negative PDO conditions in the Pacific; and this therefore seems to raise the risk of more frequent severe Bering Sea storms in autumn.  Last year's UAF ACCAP report "Alaska's Changing Environment" indeed documented a recent increase in Bering Sea storms, but also indicated that no clear long-term trend has yet emerged.  It will be interesting to see how this assessment evolves in the coming years.

https://uaf-accap.org/alaskas-changing-environment/

Turning Wintry

Snow has been on the ground in Fairbanks since late last week, and at this point it is unlikely to melt out, despite a warm forecast for Halloween.  (Temperatures are expected to get above freezing in the next couple of days, which is unusual for the time of year, but not extreme - it was 48°F on Halloween 4 years ago.)

Other than the great chinook and "brown Christmas" events of 1934, the latest date in the year with at least two inches of snow on the ground that subsequently melted out in Fairbanks was October 26 in 1940.  So if two inches of snow is observed at the end of October, it's almost certainly safe to remain there until spring.

Here's the history of snowpack onset date in Fairbanks; this year's presumed date of October 24 is not unusual.  1934 and 1992 were the two big outliers.

Temperatures have been dropping off statewide since I posted about "lack of cold" on the 18th, driven by rather strong troughing from the Bering Sea to the Gulf of Alaska:

Interestingly, Utqiaġvik yesterday dropped to -10°F, and the daily average temperature was below 0°F for the first time in October since 2018, and before that since 2002.  Until the mid-1990s, a daily mean temperature below 0°F in October wasn't at all unusual, occurring in more years than not, but this is only the fifth such day since 1996.

Looking back at autumn in Utqiaġvik, it's interesting to observe how little unusual warmth there has been in comparison to the past decade.  It looks like the October and September-October averages will be the coolest since 2015, but of course that's still well above the normal of the 20th century.

Halong vs Merbok

Now that ERA5 data is in from last week, we can compare ex-Typhoon Halong with other historical storms on a level playing field, i.e. with the same model reanalysis.  The most obvious comparison to begin is with ex-Typhoon Merbok from September 2022.  First, here are the ERA5 tracks and minimum MSLP of the two storms, at 3-hour intervals:

Merbok tracked much farther west but was much bigger, with more total energy, and it moved much more slowly as it reached the northern Bering Sea, prolonging the period of high wind and flooding for western Alaska.  Here's a comparison of ERA5 wind gusts at a grid point near Hooper Bay, on the far western end of the Y-K Delta, where the peak wind was almost the same in both storms (according to the model).  Notice how quickly the winds dropped off in Halong, versus the prolonged period of high winds during Merbok.

A map comparison of peak wind gust shows that Halong was worse in terms of wind across the Y-K Delta and up to Norton Sound and the western interior; but Merbok was worse for most of the southern Seward Peninsula, including of course Nome.

The small size but ferocious wind of Halong made its impacts much more akin to those of a landfalling hurricane (or typhoon), although both storms were thoroughly extratropical by the time they impacted western mainland Alaska.

Halong was so intense that ERA5 indicates all-time record high wind gusts along a fairly lengthy swath extending northeastward from Nunivak Island.

Here are storm tracks and ERA5 peak wind gusts for a couple of other historic autumn storms in the Bering Sea (but in these cases having no relation to pre-existing typhoons, as far as I know).  First, the 2011 Bering Sea "Superstorm", a really massive storm:

And the infamous storm of 1974, which was catastrophic for Nome:

It is very interesting that all four of these storms occurred during La Niña conditions, and not only that, during second-year or third-year La Niña events (i.e. the second or third consecutive La Niña winter after an El Niño).  More on this "coincidence" another time.

Lack of Cold

Things are about to change, but the absence of cold in much of Alaska has been a pronounced anomaly for weeks - recall that Fairbanks tied its record for latest first freeze.  The coldest temperature so far at the Fairbanks airport is only 26°F, which is a record for highest minimum temperature up to this date in the autumn.

The loss of early cold in recent decades is striking.  Prior to about 2000, it wasn't uncommon at all to see temperatures drop into the single digits Fahrenheit by this date, but that's only happened once in the past decade.

Anchorage is yet to see its first freeze at the airport, only the second time that has happened since 1952.  The year 2000 also seems to mark the point at which more severe cold stopped occuring by mid-October in Anchorage.

For completeness, here's the chart for Juneau.  The last three years failed to see a freeze by this date, but not this year.

The statewide temperature index shows the sustained anomalous warmth of the past 10 days or so.

This is of course linked to the mid-level circulation pattern, involving a strong ridge centered to the south of Alaska and a strong trough over Kamchatka and the western Aleutians.  Warm southwesterly winds have prevailed between the two features, and this pronounced southerly flow is what transported ex-Typhoon Halong up to western Alaska.  The temperature contrast between the trough and the ridge also played into the re-intensification of Halong as it interacted with the mid-latitude jet stream.

Halong Analog

Prompted by a question from Rick Thoman, I looked back at historical data since 1950 to search for past storms with comparable path and strength to ex-Typhoon Halong.  The task was made much easier using the ERA5-based extratropical cyclone track summaries from the University of Manitoba and NSIDC.  I searched for sub-970mb storms that passed within a box between the Y-K Delta and the Seward Peninsula (see below).

The closest analog in terms of path and ferocity of the wind - as estimated by ERA5 - was a February storm in 1951.  Here's the estimated track (a very good match for Halong) and minimum MSLP (about 10mb lower than Halong at the point of closest approach to the Delta):

And here's the model-estimated maximum 10m wind gust:

This 1951 event produced Bethel's highest sustained wind speed measurement on record, 54 knots or 62 mph.  Wind gusts weren't reported back then, but 80+ mph (as indicated by ERA5) seems entirely possible.  Halong's winds peaked at 48 mph (sustained) and 76 mph (gust) in Bethel, although the measurement techniques aren't directly comparable.

The 1951 storm warranted the following comment in the monthly climatological summary for Bethel (click to enlarge).

The big difference with the 1951 storm is that it occurred in February, so it almost certainly wasn't a redeveloping West Pacific typhoon, and the impacts would have been different for coastal areas where ice was present (but I'm not aware of ice extent information from that time).

In just a couple of days we'll have the ERA5 data for Halong, allowing for a direct comparison of the model data for these and other historical events.

Storm Catastrophe

"Absolute devastation... worst-case scenario" - words chosen no doubt carefully by the Coast Guard's Arctic commanding officer in today's briefing on the impacts of ex-Typhoon Halong.

https://www.adn.com/alaska-news/rural-alaska/2025/10/13/dozens-of-people-rescued-at-least-3-still-missing-after-storm-devastates-western-alaska/

The hardest-hit locations turned out to be on the south side of the Y-K delta, between the Kuskokwim estuary and Nunivak Island.  This is where the storm roared ashore in the early hours of yesterday morning, the extreme winds directed perpendicular to the shallow, low-lying coast.  Here's the GFS analysis from 3am.

Here are reported maximum wind gusts in mph, click to enlarge:

The 6-hourly GFS data shows the storm's evolution.

A meteorological and climatological perspective on the causes of the storm will be worthwhile in due course.  For now, here's a link for those with desire and ability to help with relief and recovery costs:

https://www.adn.com/alaska-news/rural-alaska/2025/10/13/how-to-help-residents-and-communities-after-catastrophic-western-alaska-storm/

https://alaskacf.fcsuite.com/erp/donate/create/fund?funit_id=14833