Friday, November 25, 2022

Thanksgiving Climate

Happy Thanksgiving to US readers.  It was a warm one in Fairbanks: about 15°F above normal for the date, and that makes it 4 out of the last 5 years that the holiday has been uncharacteristically mild.  But last year was different, with temperatures more typical of mid-January.

Here's a chart of the daily temperature range on Thanksgiving Day since 1930 in Fairbanks.  There are huge variations from year to year, of course, but somewhat less so in recent decades, as nearly all the really cold Thanksgivings (daily mean temperature below -20°F) were pre-1970.  (1994 was a notable exception, however.)

As for snowfall, it's interesting to note that even though only 19 of 93 Thanksgivings since 1930 have produced an inch or more of snow (20%), this is beating the odds for the time of year as a whole: only 13% of all November-December days see an inch of snow.  As for a more substantial snowfall of 3", this has been accomplished on Thanksgiving only twice: in 1975 and 1996.

At least Thanksgiving has always been "white": there has always been snow on the ground in the recorded climate history.  The same cannot be said of Christmas (1934).

Here are parallel charts for Anchorage and Utqiaġvik.  Interestingly the only Thanksgiving with a temperature above freezing in Utqiaġvik was way back in the volatile 1930s (1931 in this case), although 2017 and 2019 came pretty close (25°F and 27°F respectively).

Monday, November 21, 2022

Historical Context for Warmth, Freeze-Up Delays

Last week's warmth on the North Slope was exceptionally unusual, one of the most extreme temperature anomalies on record for that part of the world in winter.  Amazingly, the Ivotuk CRN (1900' elevation) reported a daily mean temperature of 43°F on Thursday.

To facilitate a consistent historical comparison, I calculated the area-average daily mean temperature for the North Slope climate division using ERA5 reanalysis data back to 1959, and according to this data there's never been a daily mean temperature above freezing for the division as a whole in winter (November through March).  We don't yet have ERA5 data from last week for comparison, but the operational ECMWF model (similar to ERA5) shows a daily mean temperature just below freezing: see the dashed extension to the ERA5 2022 line in the chart below.

The warmest winter day for the North Slope in the 1959-present ERA5 data was January 29, 1963, which had an enormous ridge over western Alaska - only slightly different from last week's anomaly - see below.



While the North Slope extreme has subsided for now, the state as a whole is very much warmer than normal, as evidenced by UAF's statewide temperature index: the chart below shows the last 14 months or so.

The contrast to last year is becoming more significant by the day, as last year the second half of November was very cold indeed.  The brief cold that we saw back at the end of October and the beginning of November this year now looks very mild and fleeting in comparison.

Neither the Tanana at Nenana nor the Yukon at Dawson are properly frozen over yet:

It's a bad situation along the Kuskokwim too, as documented by Bethel Search and Rescue a few days ago in their fifth annual freeze-up survey:

"Summary: For the past several years BSAR has been scheduling our first aerial freeze up survey of the Kuskokwim River on No­vember 18th. Flying the same day each year allows us to look back at previous years for comparison. In 2021, with a record cold November there were well established snow machine trails to nearby villages and even a limited amount of truck traffic on the River by this date. For November 18, 2022 conditions are very different. The lack of cold weather and early snowfall has slowed the freeze up process on the Kuskokwim River and it’s tributaries. THERE IS CURRENTLY TOO MUCH OPEN WATER ON THE KUSKOKWIM RIVER FOR SAFE TRAVEL NO MAIN RIVER TRAVEL RECOMMENDED AT THIS TIME"

Lots of pictures here:

Thursday, November 17, 2022

Like Yesterday But More So

This morning's sounding from Yakutat reported a 500mb height of 5820m - higher than has ever been observed in October through March (data back to 1946).  Even in April there's only been one occasion with a stronger ridge, and none in May.  Here's a graphic from my simple monitoring page:

And it's a bona fide thaw for much of the North Slope: temperatures are well above freezing across large areas to the north of the Brooks Range.  The CRN site at Toolik reported 44°F today, and Ivotuk made it up to an incredible 47°F.  The Ivotuk CRN has only been operating since 2014, but this is the warmest observed there between October 5 and May 9.

Wednesday, November 16, 2022

Extreme Ridge

Records have fallen as upper-air heights have risen over Alaska today, reflecting a really extreme ridge of high pressure aloft (i.e. elevated "heights" of constant pressure surfaces like 500mb).  Here's a view of the mid-atmosphere situation at 3pm today:

The ridge is centered near Yakutat, and accordingly this afternoon's balloon sounding observed Yakutat's highest 500mb height on record for November, by some margin.  It was very close to the record for winter (November through March) as well; that record is just a smidge higher, from February 1989.

This afternoon's Fairbanks sounding also saw the highest November 500mb height of record, and again it was only just behind the all-time winter record, set in late December 1983.

The most obvious effect of the monster ridge has been to deliver extreme warmth to the coastal periphery of Alaska, from the southwest to the North Slope.  Strong winds lifted temperatures to around the freezing mark across the entire North Slope, and above freezing in areas with additional downslope warming.

Three of the four CRN sites on the North Slope rose above freezing:

Deadhorse: 34°F

Toolik: 37°F

Ivotuk: 39°F

It looks like the entire length of the Haul Road from Galbraith Lake northward rose above freezing, and there are even a couple of 40°F readings at USARRAY instruments in coastal ANWR.

Of course it's very warm at higher elevations of the interior as well - Denali's Eielson Visitor Center is currently sitting at 44°F - but the strong inversion has kept valley-level temperatures lower, as is typical in the absence of a chinook setup.  Here's this afternoon's sounding from McGrath: below 0°F at the surface, but +45°F about 1800 feet above the surface.

Wednesday, November 9, 2022

October Climate Data

First, an update on freeze-up at a couple of points of interest (at least to me).  The Tanana River at Nenana is partially but not fully frozen over, and the Yukon River over at Dawson appears to be frozen just downstream of town, but surface water is still moving in front of the webcam.  Beautiful light in both scenes:

And now looking back at October's climate anomalies: NOAA/NCEI data indicates that the month's temperatures were (a) far above normal in Southeast Alaska, (b) modestly above normal along the rest of the Gulf Coast, and (c) near the normal of the last 30 years for the rest of the state.

For the South Panhandle division, October tied the 1986 record for warmest October in the entire history since 1925.  This outcome is closely linked to the extremely warm sea surface temperatures across the northern North Pacific:

As for precipitation, October was another wet month compared to normal, the fourth in a row for the state as a whole.  The statewide July-October total precipitation was easily the highest on record, and the year-to-date statewide precipitation has also moved into first place in the history (1925-present).

Interestingly, this is the first year since 1981 to fall in the top 10 for January-October precipitation statewide.

The ERA5 gridded reanalysis data agrees quite well this month, with the exception of precipitation for the North Slope (see below).  Of course the error bars on North Slope precipitation estimates are very large because of the regrettable paucity of ground-truth observations.  For what it's worth, Utqiaġvik reported 1.01" of liquid-equivalent precipitation, the fifth highest in the past 30 years, and the SNOTEL sites along the Haul Road recorded near-normal to above-normal precipitation.

The Gulf Coast and southeast interior were much windier than usual, according to ERA5, but apparently western Alaska was relatively calm.  This is a bit of a surprise, given the lower-than-normal pressure over the Chukchi Sea.

As for sunshine, it was in short supply again for southern and eastern Alaska.

I'll update this post tomorrow with October's temperature reports from around the wider Arctic region.

Here we go: October's departure from the 1991-2020 normal temperature at 32 Arctic sites:

Here's the anomaly in terms of standard deviations: it was a significantly unusual month in central Arctic Russia.

The year to-date is running 2.8°C above the 1991-2020 normal at Russia's Tazovsky station, and that's good for the 3rd warmest year on record so far, after 2020 and 2016.  On the cool side, Kotzebue is running 1.1°C below normal, which is similar to last year through October.

Monday, November 7, 2022

The 2010 Similarity

Yesterday I mentioned 2010 as a similar year for climate in the past, noting that the very high MSLP in Fairbanks yesterday nearly matched the November record high pressure that was set in 2010.  That record was set on November 17, and within a week Fairbanks was shut down by a tremendous ice storm caused by record-breaking rainfall.  Here's one in a sequence of blog posts by Rick Thoman during that 2010 event:

Alarmingly, Rick noted today that the forecast for the middle of next week looks very indicative of rain.  It's still some way out, and the forecast could change a lot, but as we've seen many times in recent years, any amount of rain would be very bad news.

Let's hope the 2010 similarity doesn't play out any further.

Sunday, November 6, 2022

Cold High Pressure

A strong ridge moving across the Bering Sea and western Alaska since Friday has produced very high pressure at the surface for interior and eastern Alaska, and so, like clockwork, we have cold in the valleys, with strong temperature inversions.

Here's the MSLP chart from 3am AKST this morning:

In Fairbanks the airport reported MSLP of 1050.2mb at 6am, which according to Rick Thoman is "just shy of the November record of 1051.4mb set in 2010".  Interestingly, 2010 had a number of similarities to this year in terms of global climate, including a strong La Niña with a negative Indian Ocean Dipole in autumn, and terrible monsoon flooding in Pakistan in late summer.  It may be no coincidence that we're seeing a similar weather event for Alaska at the moment.

Rick commented on the 2010 MSLP record at the time:

Here's an animation of the 500mb ridge progression from Friday morning to Tuesday morning (predicted):

Not surprisingly, temperatures dropped off impressively last night, with -20s seen widely in the valleys of the eastern interior.  As usual the coldest spots around Fairbanks-land were Goldstream Creek (-20°F), UAF Smith Lake (-23°F), and North Pole (-24°F).  Farther afield, the CRN site southeast of Northway reached -29°F and Chicken saw -31°F.

The morning sounding from Fairbanks shows the hefty inversion, and indeed the temperature hasn't dropped below 0°F at the higher elevations around Fairbanks (Cleary Summit, for instance).


Surprisingly, the Yukon is not yet frozen upstream at Dawson:

Below is a chart of temperatures over the last two weeks at Smith Lake on UAF's North Campus, illustrating the sharp drop-off yesterday.  Note too the difference between the blue and green lines: blue shows the temperature at 8m above ground, and green the temperature at the standard 2m above ground.  A difference of about 5°F is typical on clear, calm, cold nights, demonstrating how shallow the strong surface-based inversion is in these particularly cold locations.


Wednesday, November 2, 2022

Follow-Up on Temperature Extremes

Temperatures have dropped off very sharply in the last 10 days for most of Alaska, and the first well-below-zero readings have appeared in the usual cold spots.  So far the Chalkyitsik RAWS is "in the lead" with -30°F yesterday morning.

This is only slightly unusual for the date: in the last 20 years, it has been typical for -20s to appear somewhere in the state by early November.  Just two years ago on this date it was -40°F in Chicken.

For the state as a whole, though, the cold is unusual; it's the most significant below-normal spell for the statewide temperature index since early August.  Of course, the absolute magnitude of the departure from normal is now much greater than it was in summer, because temperature variance is so much higher in the cold season.

It's worth reviewing data from the last couple of years to see how abundant "extremes" of both warm and cold have been.  For background, in March 2021 I used ERA5 reanalysis data to look at the imbalance in warm versus cold extremes for (most of) Alaska's land area, and it was interesting to note that the occurrence of cold extremes had dropped off in recent years even after detrending the data.  In other words, even relative to the modern warmer baseline, there had been an absence of significant cold for several years, in both summer and winter.

Note that here I'm using a threshold of 2 standard deviations above or below normal (or trend) to define an "extreme".  For reference, Sunday's notably chilly high temperature of only +7°F in Fairbanks was 1.35 SD below the 1991-2020 normal; a 2 SD cold anomaly would have meant a high temperature near 0°F, or a low temperature near -18F.

Here are updated charts for the land area coverage of these warm and cold extremes in June-August, relative to a trend-line normal.  The land area box includes all of Alaska west of 141°W, i.e. excluding Southeast Alaska (where ERA5 does not do well with surface temperatures because of topography).

 This summer brought a notable uptick in cold extremes relative to trend, and somewhat surprisingly (because southern and southwestern Alaska were very warm in June) there wasn't an over-abundance of +2SD daily warm extremes.

Below are the equivalent charts using a fixed 1981-2020 baseline; as expected, this summer's cool extremes did not rise to the level of many years prior to 2010.  I think it's also interesting to note just how severe the 2022 wildfire season was despite an excess of extreme warmth; I surmise that this is because it was driven by extreme drought, persistent but less-extreme warmth, and numerous lightning ignitions.

As for winter, the ERA5 data shows a relative lack of warm extremes in the past two winters, even relative to a fixed climatology.  This certainly reflects the influence of a persistent La Niña with a negative PDO phase; these large-scale climate phenomena prevented a recurrence of the excessive warmth of 2013-14 through 2018-19.

Last winter also produced more cold extremes than in the prior 8 winters - mostly in November last year, which was the coldest November statewide since 2011.  But just like this summer, overall the winter of 2021-2022 wasn't particularly unusual compared to the colder winters of earlier decades.

It will be interesting to see where we go from here.  With La Niña still hanging around, there's a good chance we'll see at least some episodes of quite unusual cold in the coming months, and a return to pervasive warmth seems a bit unlikely in the near future.  However, the CPC's latest seasonal forecast for December-February only hints at increased odds for significant cold in Southeast, and the warm tilt has increased since last month for western Alaska.