Cold Spring Day

Yesterday was a very chilly day for the time of year across interior Alaska, with breezy below-freezing conditions all day in most areas north of the Tanana and lower Yukon rivers.  Temperatures struggled to reach 20°F in the Yukon Flats despite plenty of sunshine, and Fairbanks only made it to 31°F.

The last time the daily high temperature was below freezing this late in the season in Fairbanks was during the record cold April of 2013 - you can read all about that in the blog archives here (Rick Thoman was the author then).  Prior to that it had been more than 10 years since a sub-freezing day in late April (2002).

The winds did not go calm last night, but temperatures dropped off sharply nonetheless, and on the North Slope this morning's temperatures were, well, Arctic.  Umiat made it to an impressive -27°F with still a light breeze this morning.  This is cold even by Umiat standards; in the combined history of data from Umiat sites, only 4 years were this cold so late in the season (but 1984 takes the cake, with -43°F on April 23).

Here are 24-hour low temperatures as of 5pm AKDT today:

To add insult to injury yesterday, a couple of inches of snow fell in parts of Fairbanks in the evening.  The airport's total of 1.8" was the most to fall this late in the spring since - you guessed it - 2013.  A 5-year return period is very consistent with the long term history for this kind of event, although there have been occasions with 6" or more at this late date or even into May.

The scene down on the Tanana River at Nenana was not encouraging for those with money on a very early breakup, and this morning looked less than spring-like at UAF:

The culprit for the renewed bout of winter was a sharp upper-level trough extending southwest from a cold low that dropped south to near Canada's northwest coast a couple of days earlier.

Sea Ice Update

One of the biggest Alaska-related climate stories of this winter has been the dramatic reduction in Bering Sea ice compared to recent decades (and probably much longer, as noted by this recent article on climate.gov).  The latest data show no late-season recovery, and in fact the ice extent deficit has "stabilized" at more than 500,000 km² below normal.  As of April 18, the Bering Sea ice extent of just 142,000 km² is equivalent to the 1981-2010 normal for the beginning of June.

In seeking an explanation for the absence of sea ice, I previously noted the role of reduced northerly flow.  The connection between Bering Sea ice and the local circulation pattern is also reflected in the following map, which shows the correlation between 500mb heights and sea ice extent for December-March.  When there is a strong trough over Alaska, Bering Sea ice expands, but when the winter is dominated by a ridge over Alaska, ice cover is reduced.  Not surprisingly, this winter's upper-level height pattern was very similar to the pattern that produces strong variations in Bering ice extent - see the second map below.

An interesting question that arises is "to what extent can this winter's ice loss be explained by the flow pattern?"  We can address this in a simple manner by estimating a linear relationship between sea ice extent and 500mb heights at the epicenter of (inverse) correlation near 65°N 160°W.  The chart below shows the reasonably good relationship (R=-0.67) and indicates that this winter's 500mb heights were the highest in the modern era of sea ice data.  However, the ice extent was clearly well below the level that would be expected solely from the strength of the ridge.

The chart below shows the same data in time series form.  This winter's 500mb heights were not much higher than in some recent years, but sea ice extent was far lower.

Performing the linear regression allows us to quantify the unusual shortfall of ice relative to what would might be expected from the flow pattern; see below.  Based on 500mb heights alone, the expected value of Bering Sea ice extent for 2017-2018 is 423,000 km², or 168,000 km² below the 1981-2010 mean.  The actual deficit this winter was 364,000 km², or more than twice what can be explained by the circulation pattern alone.

Perhaps the most interesting aspect of the analysis is that there has been an ice deficit relative to the regression model in each of the last four years, and it has grown steadily; here are the ice extent residuals from the last four years:

2014-15   -60,000 km²
2015-16   -107,000 km²
2016-17   -128,000 km²
2017-18   -195,000 km²

These results are consistent with the idea that the sea ice loss is not simply an expression of unusual weather patterns; rather, sea ice deficits have increased owing to persistent regional ocean warming and, presumably, Arctic-wide amplification of the background warming trend.

April Temperature Variance

Fairbanks saw its first 50°F temperature of the spring on Wednesday, and this was right on schedule according to the long-term history.  Based on data since 1930, the median date is April 11, but the first 50°F has typically arrived a few days earlier in recent decades.

Here's a chart of average daily high temperatures in each April since 1930 (blue markers) along with the date of the first 50+°F temperature (red markers).  Note that I've excluded 3 instances when 50°F occurred prior to March (1943, 1981, and 2009); these were exceptional winter chinook events, and for today's analysis I'm interested in what happens during spring.

The month of April has seen pronounced long-term warming in Fairbanks, with the linear trend line showing more than 5°F of warming over 87 years.  It's no surprise, then, to see that the first 50°F has become slightly earlier - although there's only a weak trend in the date of first 50°F.  The trend in April mean high temperature is highly significant (p<0.01), but the first 50°F trend is not significant at all (p~0.20).

At first glance it's surprising that the first 50°F trend is so much less robust, but this is partly because the variance of 50°F dates is so large.  Over the 87 years, the first 50°F date has advanced about 4-5 days, but the standard deviation is about 3 times as large (14 days).  In contrast, the interannual standard deviation of mean April high temperatures is almost exactly the same as the 5-6°F of warming that has occurred, so the long-term change has produced a very significant shift in the frequency distribution of mean temperature.

If we consider the normal rate of warming at this time of year in Fairbanks, we find that 5-6°F of warming is equivalent to about 8 days of climatological warming.  Naively, then, we might have expected the first 50°F to have advanced about 8 days; but even that would not correspond to a statistically significant trend in the date.  The fact is that the variance of the first 50°F date is so large that sampling variability could plausibly produce trends that are nearly zero or trends that are much greater than expected.  And so we might say that dates of "first warm day" or "first cold day", at least in Fairbanks, are not particularly useful as indicators of long-term change.

Another Warm Winter

For the fourth time in the past five years, the November-March mean temperature in Fairbanks was significantly warmer than the 1981-2010 average, i.e. the mean temperature was in the upper tercile of the 1981-2010 distribution.  Each calendar month was also warmer than normal, and in fact this warm "streak" goes back to last June; each of the past 10 calendar months has been at least 1°F warmer than normal.  This kind of persistent warmth is not unprecedented, but it's certainly unusual.

The chart below shows running annual mean temperatures (daily max/min/mean) in Fairbanks since the beginning of the NWS/Weather Bureau era.  The recent warmth has raised the annual mean to nearly 32°F, a threshold that was exceeded for part of 2016.

It's interesting to note that some very cold years in Fairbanks had annual daily maximum temperatures below freezing, but that hasn't happened for many years.  Also of note: the long-term trends in maximum and minimum temperatures are quite similar over the entire period, but minimum temperatures increased more dramatically in the climate shift of 1976-77.  Consequently, the linear trend in minimum temperatures since 1976 is actually negative (-0.15°F/decade), but daily maximum temperatures have continued to increase (+0.20°F/decade).  Of course we have to take these numbers with a grain of salt owing to location and instrumentation changes in the official climate record.

Warm April More Likely

With just a few days left to submit guesses for the Nenana Ice Classic, breakup prognosticators are pondering the extended-range weather forecast for the next few weeks.  The date of the Tanana River's breakup at Nenana is closely tied to local temperatures in April, so an early breakup will be likely if the unusual warmth of recent weeks and months persists.

Looking quickly at some of the forecast data, CPC's most recent outlook for April shows a slight warm tilt to the temperature probabilities over most of the interior. 

NOAA's CFSv2 model has recently signaled an increasingly pronounced warm anomaly centered over far eastern Russia and the Bering Sea.  The maps below show a progression of April forecasts from oldest at bottom to newest at top.  A few weeks ago it appeared possible that the pattern would favor a cold April in the Tanana valley, but this now seems unlikely.

As an aside, I'm in the middle of a very busy period of travel and work, so blog posts will continue to be thin for a while.

Late Winter Snow Showers

My visit to Fairbanks ended on Monday with a rare March instance of strong westerly breezes, adding further to the tally of "strong" westerly wind days in recent weeks (see here and here).  While it's typical for winds to pick up in late winter in Fairbanks, the seasonal breezes are usually from the northeast; in the past 20 years, there have only been 2 other days in March with similar winds (20-30 mph) from the western half of the compass.

In this week's event, the wind was associated with a strong cold front aloft; the 500mb chart below shows the trough axis near Fairbanks at 4pm AKDT on Monday.  Note the temperature of -39°C reported above Fairbanks; by Tuesday morning the 500mb temperature was down to -43.5°C, which was very nearly the coldest air of the entire winter at this level.

However, even as the cold air moved in aloft, temperatures at the surface were up near the freezing point on Monday afternoon.  The chart below shows the full sounding from 4pm on Monday; the (red) temperature trace shows a strikingly steep vertical temperature gradient, indicating an environment with potential for some convective overturning.

In response to the increasingly unstable environment, heavy snow showers emerged on Monday afternoon and moved through the Fairbanks area.  See below for a sequence of radar images at hourly intervals from 2pm to 7pm AKDT.

Interestingly a similar round of showers occurred on Tuesday afternoon as surface temperatures made their way up to the freezing point again, while very cold temperatures persisted aloft.  Tuesday afternoon's temperature difference (42.6°C) between the surface and 500mb was one of the greatest on record for March in Fairbanks and was actually the largest on record for so early in the year.  Here's a Tuesday afternoon radar image:

Reader Bill suggested that it's not uncommon to see heavy snow squalls like this in late winter in Fairbanks, but without going through a long-term archive of radar data, it's difficult to say what the frequency actually is or whether it might be changing over time.

Interior Snow Pack

I'm up in Fairbanks this week, getting a taste of late winter conditions - although warmer and snowier than normal for the time of year.  Today I took a trip up to the Chena River State Recreation Area to get a first-hand look at the snow pack in the hills, and I can state unambiguously that there is a lot of snow on the ground.  The photo below was taken at about 800' elevation.

I wrote about the above-normal Chena Basin snowpack back in January, and the chart below updates that analysis through today.  Remarkably, the average of 5 sites that are reporting snow water equivalent is close to double the 1981-2010 median for this date, and it's close to a record.  The only year with more snow on the ground at this date was 1993 - no other year comes close in the SNOTEL data for these 5 sites (1981-present).  The winter of 1992-1993 brought 139" of snow to Fairbanks (3rd greatest on record), and March 1993 was the last time the measured snow depth reached 3 feet at the airport.  For comparison, Fairbanks has had nearly 80" of snow so far this winter, with a peak snow depth of 32".

It's worth noting too that the snow water equivalent (SWE) is at or very close to record levels for the time of year at 3 of these 5 SNOTEL sites.  The SWE of 9.7" at Mt. Ryan (2800') is not far off the all-time record of 10.5" (April 1993), and the remarkable 14.8" that's currently being reported from Munson Ridge (3100') was exceeded only in 1991 and 1993.

The map below provides a broader look at snow pack in regions where data are available.  Significant shortfalls are evident in south-central and southeast Alaska, and this is related to La Niña, which reduces the number of storms in the Gulf of Alaska.  On the other hand, snowfall has been above normal in the interior and at least as far north as the Brooks Range; the Atigun Pass SNOTEL is currently reporting a snow depth of 57", which is a new record for March and ties the all-time record for the site (1983-present).

Excessive Warmth in Barrow

Much has been written in various outlets about the extremely warm winter that the Arctic has been having this year, and so I haven't felt the need to draw attention to it here, but the persistence of the warmth has been so extraordinary of late that I would be remiss not to make some mention of it.  Here's my usual chart of daily mean temperatures compared to the standard 30-year normal for Utqiaġvik (formerly Barrow); the warmth has been extremely persistent and pronounced, with only a handful of days below normal since autumn.

We've become accustomed to seeing extreme warm anomalies during autumn in Barrow in recent years, owing to the loss of sea ice at that time of year, but the kind of winter-long warmth we are seeing now is really remarkable.  The north coast of Alaska was finally iced in by mid-December 2017 (much later than normal, of course), but that didn't bring an end to the anomalous warmth; it seems that the entire Arctic basin has been unable to cool off this winter, even though there's not that much more open water than normal in percentage terms.

In comparing temperatures this winter to the climate of previous decades, it's remarkable to observe that the respective histograms of Barrow temperatures are non-overlapping to a significant extent - see the chart below.  In comparison to the extreme warmth of this winter, even the notable differences between the 1951-1980 and 1981-2010 climates seem quite minimal; and even the then-record warmth of 2016-2017 seems rather less significant compared to this year.

Another perspective on the anomaly is gained by looking at the annual fraction of December-March days with daily mean temperature at or above 0°F - see below.  This winter's outcome (so far) is an astonishingly large deviation from historical climate normals.

Sea Ice and Southerly Winds

Last month's look at ice loss in the Bering Sea provoked a comment from reader Eric about the impact of southerly winds on ice extent, so I thought it would be interesting to make a quick comparison of this winter to previous years in this respect.

The chart below shows some November-February average data from the north-central part of the Bering Sea.  Based on reanalysis data from a grid point near St. Lawrence Island, the black line shows mean temperature at 925 mb (about 2000' above sea level), and the blue line shows the southerly component of the wind at the same height - so negative wind values indicate northerly flow (the usual situation).  The red line shows the southerly wind component observed at the surface at Gambell and Savoonga, based on hourly observations since 1996.

It's interesting to note that this winter so far has produced the record highest mean temperature and southerly wind component at 925mb, according to the reanalysis; the average 925mb wind still had a northerly component, but barely.  The surface winds from the two observing sites on the island are highly correlated with the 925mb winds, and the surface data reveal a positive mean southerly component for the first time on record.  It's no wonder that ice extent has suffered with such an absence of cold northerly flow.

Here's an update of the chart from the last post, showing that the ice extent staged a modest recovery after its collapse in February, but in recent days it has turned down again.

Obviously the ice situation is now well outside the historical range from the satellite era, and this is partly because there was such little ice cover prior to the recent episode of ice loss.  The magnitude of the recent loss is not in itself unprecedented, as 6 previous years since 1979 saw at least as much ice loss over a 15-day period - although presumably it's "easier" to lose a lot of ice when it's being removed from warmer locations in the southern Bering Sea.

The maps below show sea-level pressure anomalies from the aforementioned 15-day periods in which ice loss was greater than the recent episode, and the last map shows this year's anomaly.  The common feature is obviously an enhanced southerly flow between high pressure over the Gulf of Alaska and strong low pressure over the western Bering Sea or far eastern Russia.

The unusual flow pattern appears to be partly related to the modest La Niña anomaly that continues to be evident in the tropical Pacific; the map below shows the average February-March sea-level pressure anomaly for 10 previous years when weak-moderate (but not strong) La Niña conditions were observed in late winter.  Strong La Niña episodes tend not to be as favorable for ice loss, because the PDO is more often negative in those winters and the Bering Sea tends to be more dominated by high pressure (and thus it's colder in Alaska).

Winter Winds

After suggesting last Saturday that Fairbanks' February snow total probably wouldn't match last year's number, another round of very stormy conditions moved in on Monday, delivering enough snow to beat last February by just 0.1".  The airport's snow depth of 32" is now exactly the same as last year's number on February 28th.

As in my last post, I want to focus on the very unusual winds that have occurred in conjunction with the recent repetitive storms.  Remarkably, the peak 2-minute wind speed at Fairbanks airport reached 20 mph or higher on 9 out of 15 days in the latter part of February, and this is easily a record for any time in the winter, including March (based on data since 1996).  In fact, this many "windy" days has not previously occurred in anything less than a 65-day period in the last ~20 winters in Fairbanks; it's just extremely unusual to have such a focused period of repeated "high winds".

The rarity of recent events raises the question of whether there's a long-term trend towards more wind in winter in Fairbanks.  Based on ASOS data from the past 20 years, the answer seems to be clearly yes.  The two charts below show the average wind speed (10m above ground) for November-December and January-February, along with the maximum values for both 2-minute and 5-second averages.

The sample size is obviously very small for the seasonal maximum wind gust values, so I dug a bit deeper by looking at frequencies of exceeding the 98th percentile of the climatological distribution.  I estimated the 98th percentiles separately for November-December and January-February and then combined the results for the following chart:

The results indicate that both 2013-2014 and this winter had 8 days with peak wind speed above the 98th percentile on both the 2-minute and 5-second time scales; the long-term expected value would be 2-3 days.  The suggestion of an increasing trend is rather striking, although 20 years is a short period, and if we take out the two big outliers the picture would be less clear.

Dividing the results into early and mid-late winter (see below) emphasizes the outliers even more.  The 98th percentile estimates are included on the legends; they're higher for January-February as a result of the normal annual cycle of wind speed in Fairbanks.

Looking farther back in Fairbanks' climate history is not simple, because the pre-ASOS wind measurements are not directly comparable with the modern data.  We can, however, gain a longer-term view from the winds reported by balloon soundings; the charts below show results for mean and maximum 850mb wind speed in November-December (top) and January-February (bottom).

The November-December data show some indication of an uptick in 850mb wind speeds in the past 20 years, but in a longer-term context there's no obvious trend, and recent conditions do not seem to be outside the range of historical norms.

The situation for January-February is a bit more interesting, as the 850mb wind speed seems quite stable, including in the past 2 decades, and there's little evidence of a recent increase as suggested by the surface data.  Could it be that surface winds have increased while winds at around 4000' above ground level have not?  This might be consistent with a change in preferred flow direction, or perhaps it's a reflection of weaker surface-based inversions that allow winds aloft to mix down more often.  It would be worth looking into this more closely, and it would also be interesting to see if the same thing has happened at other locations, including in the Arctic where recent warming has been so pronounced.