Ice Fog and Surface Temperatures

Rick T, here. Just a quick follow-up on the question of ice fog and possible impacts on surface temperatures. Ice fog and its impact on surface temperature is a tough nut to crack to due many confounds: societal changes, e.g. people leaving vehicles running, urbanization changes, airport siting and activity, ASOS vs. human observed visibility, etc.

My sense is that, physically, the long wave radiational depth of the ice fog (analogous to optical depth) is what matters for surface temperatures, which I don't think has much correlation to reported (horizontal) visibility. I think the the temperature manifestation of this can be seen in the RAOBs with very low surface temperatures when lapse rates become neutral or negative (i.e temperatures are steady or following with height). 

Here are two very cold soundings, one from January 1969, the pre-pipeline era, and one from 1989, with a significantly expanded urban environment and many more automobiles operating.  The 1969 sounding looks like a myriad of other winter soundings in Fairbanks: the temperature at the airport was about 15C lower than 400 meters up. The 1989 sounding is much different and quite unusual, with slight cooling in the lowest 400 meters. No two situations are of course the same, but part of the reason for the dramatic difference is, I think, the depth of the ice fog. In spite of very low visiblities in both cases, I'd suggest that in the 1969 case the ice fog was very shallow, while in 1989 it was thick enough that the effective radiative surface was the ice fog top (~600), with weak mixing within the ice fog.

In any case, no hard science here, just my sense of some of the issues involved based on my years of trying to forecast in these conditions.

Arctic Update

Here's a brief update on temperatures at climate observing sites around the Arctic Ocean in the month of December.  According to monthly data from a set of 19 stations, December was nearly but not quite the warmest in the period of record I've used here (1971-present), with an average departure from normal of +3.3°C.  In this data set, December 2005 was slightly warmer than December 2016.

The relatively high temperatures in December continued the pattern of persistent warmth that was observed all year in 2016.  According to data from these particular stations, the average temperature was at least 1°C above the 1981-2010 normal for all 12 months of the year; it's the first time that has happened in this data set.  Perhaps more strikingly, the 19-station mean temperature was above normal for all but 5 days of the year.

If we look at the annual number of days with a mean daily temperature below normal (see below), there was a sudden drop-off in 2005, but for the past decade the number of below-normal days has been relatively stable; the 2005-2015 average was about 75 days or 20% of the year.  However, 2016 was quite different, with below-normal days becoming rare for the Arctic coastline as a whole, and even more so in the second half of the year; the last below-normal day was August 5, 2016 - more than 5 months ago.

The near-complete absence of below-normal temperatures in 2016 is entirely consistent with ECMWF model estimates of high Arctic (80-90°N) air temperatures, as calculated by the Danish Meteorological Institute; their 2016 chart (see below) shows no below-normal days at all outside the summer season.  The charts provided for earlier years on the DMI website show nothing comparable for sustained anomalous warmth in the reanalysis history back to 1958.

Historic Fairbanks Cold Snaps

Rick T. here with a short post for future reference. Here's an annotated plot of the most extreme cold snaps during the Weather Bureau/NWS era (since late 1929) in Fairbanks, from downtown, Weeks Field, International Airport, wherever the observations were being made. Here we're looking at all the cold snaps with a 7-day average temperature of 40 below or lower. Note: average temperature, not average low temperature. The dramatic dropoff in frequency and intensity of cold snaps post-1975 is partly PDO shift, partly trend and partly urban growth and the attendant ice fog, which serves as something of a "blanket" for the valley floor.

For those in Fairbanks in 1989: yes, the lowest weekly mean temperature reported from the Airport was -43.0F, and I agree, I think there was a problem with the temperature sensor on at least a few days. But based on the Ft. Wainwright observations, such as there were, it was not greatly lower than that.

Severe Cold Spell Possible

For those hardy residents of Alaska's interior who enjoy the occasional spell of intensely cold winter weather, and who have felt that the warmth of recent years has been over the top, the latest forecast indications may be welcome.  Others will feel less enthusiastic.  Recent forecasts have been coming into agreement with expectations for a period of very cold weather later this month, and there's a chance the cold may be extreme and prolonged.  Here are the 500mb height anomaly forecasts from 3 leading models for 7-10 days from now (click to enlarge): notice the extremely pronounced trough over Alaska and the good agreement between the models.

Here are the CPC's latest 6-10 and 8-14 day forecasts, emphasizing the likelihood of unusual cold over southwestern portions of the interior in particular:

And here are a few words from Rick Thoman (note this is personal opinion, not official NWS guidance):

"Models having been pointing this way for some days, and now we have increasing agreement on the potential for a period of prolonged deep cold Jan 15-21 or so. The basic forecast pattern has similarities to portions of the 1989 and 1999 cold events. The devil is in the details (of course), especially east of 150W, where there is potential for clouds and snow to wrap around from the GoA, and so keep surface temps higher (just as we saw in the 1999 event).

North of the Alaska Range there is a plausible potential for multiple days with 850mb temps lower than -35C and surface temps lower than -60F, which in the past has resulted in some of the commuter airlines (and Everett's) shutting down service to rural communities."

For reference, here's a chart of the mean 500mb height anomaly during January 20-30, 1989, when extraordinarily cold conditions affected the interior.  It's hard to imagine, but the temperature in Bettles dropped below -60°F on 9 out of the 11 days in this period, including -69°F on January 26.  Fairbanks reached "only" -51°F but had many days that stayed at or below -40°.

Colder Again

The development of a strong high pressure system (1055mb+) across northern and interior Alaska sent temperatures into a dive in Fairbanks after snow ended on Thursday, with a drop from +22°F to -28°F in 24 hours.   Today the temperature struggled to get above -20°F at the airport.  It seems as if temperatures have been on quite a roller-coaster so far this winter, as seen in the chart below, but in fact the variability is not at all unusual.  Since November 1, the standard deviation of daily temperature anomalies has been 12.6°F, compared to the 1981-2010 normal of 13.3°F for the same period on the calendar.  Perhaps the volatility seems unusual to me because last winter was one of the least variable winters on record as far as temperature in Fairbanks is concerned.

It's interesting to look at how the winter-time temperature variance depends on the phases of ENSO, the PDO, and the NPM.  El Niño winters (like last winter) tend to bring reduced temperature variability to much of Alaska because of the persistent flow patterns associated with low pressure to the southwest of Alaska.  La Niña winters tend to be more variable as high-pressure blocking episodes develop and then dissipate in the vicinity of the Bering Sea.

The chart below shows a box-and-whisker representation of Fairbanks November-March temperature variability for different phases of ENSO.  For reference, last winter's standard deviation of daily temperature anomalies was only 10.8°F, the second lowest on record.

A parallel analysis for the PDO phase shows a similar relationship but with a more notable enhancement of variance on the "cold" side, i.e. negative PDO.

Finally, the NPM chart shows a surprisingly strong relationship, given that the NPM has little relationship with mean winter temperature in Fairbanks; negative NPM winters tend to have lower temperature variance.  Of course the fact that the NPM is uncorrelated with mean temperature does not at all mean that the circulation patterns are similar; the negative NPM flow pattern is more El Niño-like in the vicinity of Alaska, whereas the flow more resembles a La Niña pattern when the NPM is positive.

Given that this winter has so far been characterized by a modestly positive PDO and a strongly negative NPM, we would expect the temperature variance to be lower than normal in Fairbanks.  However, as we discussed with respect to the heavy December snowfall, this winter is not quite conforming to expectations.

Extremely Warm Day in Arctic Alaska

In the wake of the recent Bering Sea storm, a deep and very strong southerly flow has transported extremely warm air into northern Alaska, leading to exceptional warmth across the North Slope.  Remarkably, temperatures rose above freezing everywhere on the North Slope yesterday (or at least everywhere with a thermometer reporting in real-time), and today isn't much cooler.  The map below shows the maximum temperatures observed in the 24-hour period ending 5pm AKST yesterday.

Here is yesterday afternoon's 500mb analysis, showing the torrent of warm air flowing northward over the Bering Sea and far western Alaska.

Looking at the climate observing sites with long term histories, Utqiaġvik (Barrow) and Deadhorse both had a high temperature of 36°F yesterday, and Kotzebue hit 38°F for the midnight-to-midnight period.  This is the first time in the common period of record (December 1968-present) that all three stations have exceeded 35°F on the same day in winter; this has never before been observed between October 25 and April 13.  The chart below attempts to illustrate this in graphical form.

December Snowfall

Here's a quick update on yesterday's post: snowfall for December is now up to 32.9" in Fairbanks, which is the 4th highest on record (1930-present).  Nearly half of that fell in the last three days.  The airport recorded wind gusts as high as 52 mph early this morning, and sustained winds of 33 mph were measured; this sustained wind speed is the highest since February 2011.  It's also the highest for early-mid winter (pre-February) since 1970.

This morning's sounding showed an extraordinarily strong westerly flow aloft, with 850mb winds of 75 knots (86 mph).  This ties the all-time record for strongest 850mb wind speed in Fairbanks (1948-present); this has been an extreme and historic winter storm.

Returning to the topic of the relationship between snowfall and North Pacific sea surface temperature patterns, the following chart illustrates the unusual nature of this month's heavy snowfall.  The horizontal axis shows the difference between the PDO and NPM index values (PDO minus NPM), and December's snowfall is shown on the vertical axis.  When the PDO is positive and the NPM is negative, as they are now, it is unusual to have significantly above-normal snowfall.  Note that I calculated the linear trend line after transforming the snowfall data into an approximately normal distribution, because ordinary least-squares regression is not suitable for a skewed distribution like we have here (it weights the outliers too heavily).

Major Storm

A sprawling storm system stretching from far eastern Russia to interior and northern Alaska brought heavy snow to the Fairbanks area last night and is now producing strong winds (gusting to 50mph at the airport in the past couple of hours).  Here are preliminary snowfall totals from the NWS as of this morning:

Public Information Statement
National Weather Service Fairbanks AK
822 AM AKST Fri Dec 30 2016

...Snowfall Reports...

Location                     Amount    Time/Date       Lat/Lon              

...Middle Tanana Valley...
3 S Fox                      9.4 in    0500 AM 12/30   64.92N/147.62W       
20 NW Fox                    9.0 in    0500 AM 12/30   65.18N/148.07W       
1 NW Fairbanks               8.0 in    1100 PM 12/29   64.85N/147.69W       
2 NW College                 8.0 in    1000 PM 12/29   64.89N/147.88W       
2 N College                  8.0 in    0900 PM 12/29   64.90N/147.82W       
3 SSW Fox                    8.0 in    1000 AM 12/29   64.92N/147.65W       
4 NNW Fairbanks              7.6 in    1000 AM 12/29   64.89N/147.68W       
3 N Fairbanks                5.9 in    0830 PM 12/29   64.88N/147.65W       
Fox                          5.8 in    0845 PM 12/29   64.96N/147.63W       
2 W Two Rivers               5.5 in    1000 PM 12/29   64.87N/147.17W       
Two Rivers                   5.5 in    1000 PM 12/29   64.87N/147.09W       
North Pole                   4.0 in    0800 PM 12/29   64.75N/147.36W       
4 SSW College                12.9 in   0400 AM 12/30   64.82N/147.88W       
3 SW College                 11.8 in   0530 AM 12/30   64.84N/147.90W       
3 SW College                 11.8 in   0200 AM 12/30   64.84N/147.90W       
2 NNW College                11.7 in   0617 AM 12/30   64.89N/147.85W       
12 SW Ester                  11.0 in   0300 AM 12/30   64.74N/148.32W       
3 N Fairbanks                10.4 in   0530 AM 12/30   64.88N/147.65W       
1 E College                  10.0 in   0123 AM 12/30   64.87N/147.79W       

...Lower Koyukuk and Middle Yukon Valleys...
Kaltag                       9.0 in    0600 PM 12/29   64.34N/158.69W       

...Central Interior...
Nenana                       10.0 in   0300 AM 12/30   64.54N/149.09W

The surface and 500mb charts show the powerful southwesterly flow that is typical of heavy snow events in Fairbanks.  Here are the maps from 3pm AKST yesterday (click to enlarge):

and from 3am today:

The intensity of the flow into the interior is very pronounced and explains both the abundant moisture (snowfall) and windy conditions.  This afternoon's sounding from Fairbanks airport measured a wind speed of 47mph at 925 mb (~2000 feet above ground), which is the strongest in more than 6 years during the winter season.  The sounding (see below) also shows a remarkably strong lapse rate in the lower troposphere, meaning that the temperature profile is highly favorable for vertical mixing and downward transport of momentum.  It's very rare to see this kind of temperature profile in mid-winter: only one December day in Fairbanks history had a more unstable sounding from the surface to 700mb; that day was December 12, 1972, when 13.9" of snow fell over a 3-day period.

December 2016 is now in 5th place for total snowfall (1930-present) and will probably reach 3rd place by tomorrow.

Surprising Snowfall

In a change of scene that is no doubt refreshing for some, Fairbanks has seen a significant amount of snow in the past several weeks.  The total so far this month is 16.9", compared to a normal accumulation of just under 10" for the full month of December.  With more snow in the forecast, it looks like this month will beat the 18.6" of December 2012 and become the snowiest December since 1992 (28.5").

Here's a chart of the season-to-date snow accumulation (green line) compared to normal (brown line), recent years, and the historical extremes.  I created the chart on the xmACIS2 interface, which provides a wide variety of handy analysis options.  Fairbanks is still in the lower tercile for season-to-date snowfall owing to the very dry start to winter.

A couple of weeks ago I mentioned that the PDO was still moderately positive, which was interesting in view of the notable cold spell that managed to envelop most of the state earlier this month.  The positive PDO phase also tends to be associated with below-normal snowfall in interior Alaska, as we saw in the past couple of winters; and so the current snowy spell does not reflect the PDO influence.  Furthermore, the NPM (North Pacific Mode) is strongly negative at present, and this too is normally associated with dry weather in winter.

The chart below shows the rank correlation coefficient by month between snowfall and the PDO and NPM index values.  For the November-March seasonal average, the NPM actually has a slightly stronger correlation with Fairbanks snowfall than the PDO.  Given that the current NPM index is among the most negative on record, and the PDO is also positive, it's quite surprising to see a snowy weather pattern emerge in Fairbanks.

Here's a recent SST analysis, showing North Pacific temperature anomaly patterns that are characteristic of the positive PDO and especially the negative NPM:

See these earlier posts for more perspective on the PDO and NPM and their impacts:

http://ak-wx.blogspot.com/2015/11/summary-of-sst-impacts.html

http://ak-wx.blogspot.com/2016/11/north-pacific-temperature-update.html

River Ice, Satellite Data, and Solstice Sun

I've been busy lately, but here are a few miscellaneous items that may or may not interest a few readers.  First - the webcam from Dawson, Yukon, is still showing a rather considerable amount of open water on the Yukon River more than two months after ice began forming.  This is despite daily mean temperatures in the -30s Fahrenheit for a week earlier this month.  The incomplete freeze-up could reflect warmer than normal temperatures back in November, although it's not necessarily that unusual to see open leads on northern rivers even in the depths of winter.  Those more familiar with the Yukon River at Dawson would be able to say how unusual this particular situation is.  [Update: this article gives some perspective: http://www.cbc.ca/news/canada/north/dawson-city-freeze-up-late-1.3871343]

The Tanana River at Nenana finally froze over completely in the past week or so, according to webcam views.  The first image below, from December 15, shows the last vestiges of the central channel that was last to freeze over.  Yesterday's image (second below) showed a more satisfactory scene.

On another note, I've recently figured out how to decode the land surface temperature estimates from the Suomi NPP satellite.  This satellite orbits the earth about 14 times each day and measures (among other things) emitted infrared radiation, leading to an estimate of surface temperatures when clouds are absent.  The data are available at a resolution of 750m, which allows for some really nice spatial detail.

I thought it would be interesting to examine the minimum surface temperature estimates from the recent cold spell, so I extracted the lowest value produced by the various scans available each day from December 6-10; see below for maps.  Note that there are some obvious artifacts in the data (e.g. on the 10th), and there is missing data in some areas owing to cloud and fog (including probably ice fog in the coldest locations); but overall the data look quite good.  The very coldest temperatures on these days (for this map domain) were found over the Old Crow Flats of northern Yukon.

 

For comparison, the observed daily minimum temperatures at the 3 locations marked on the maps were as follows (Fairbanks, Chalkyitsik RAWS, and Chicken respectively).  A detailed long-term investigation of the quality of the satellite estimates over Alaska would be most interesting.

Date	PAFA	CIKA2	CQNA2
Dec 6	-29°F	-51°F	-42°F
Dec 7	-29°F	-52°F	-27°F
Dec 8	-25°F	-51°F	-18°F
Dec 9	-28°F	-49°F	-42°F
Dec 10	-26°F	-32°F	-52°F

Finally, the FAA webcam at Ambler yesterday provided an opportunity to see the brief sunrise that occurs on the winter solstice even above the Arctic Circle (owing to atmospheric refraction).  Ambler is about 35 miles north of the Arctic Circle but still experiences over an hour of full daylight even on the shortest day of the year.  One has to go about 20 miles farther north to have no sunrise on December 21.

Winter Solstice in Fairbanks

Hi, Rick T. here, and happy winter solstice to you and yours. Outside, winter solstice is often thought of as the first day or winter, but in Alaska, this day is a big deal because of the solar aspect. After all, we've had "winter" weather for months, and no sourdough would confuse the dim of late December with blinding sunshine of late-February. Symbolically, winter solstice is the darkest day of the year, though in fact several days around solstice are effectively the same. Here in Fairbanks, the sun will be above the horizon for 3 hours and 43 minutes, the brightest stars are visible on a clear solstice morning until almost 10am, and reappear by 345pm. Early in the 20th century, winter solstice fell on December 22 more often than not, becoming more common on December 21 as the century progressed (as reckoned as the day of occurrence in Fairbanks based on the time zone in use at the time). This is due in part to the slow precession of the earth's axial tilt and in part to the multiple time zone changes.

So, just for fun, here's a plot of daily temperatures on winter solstice day in Fairbanks since 1904. Some solstice factoids: highest temperature was 42F in 1939; lowest temperature -49F in 1961. Six winter solstices have highs above freezing, and ten have had lows of 40 below or lower. [Note: I've put up a similar version of the plot below on twitter, #akwx]

Cold Spell Ends

After nearly two weeks of significantly colder than normal temperatures in Fairbanks, clouds finally rolled in this morning and produced warming at the surface.  The temperature rose from -21°F at midnight to +9°F at 3pm as infrared radiation from the clouds above warmed the air at the surface and thereby weakened the surface-based inversion.

The chart below (click to enlarge) shows the hourly temperature data from the airport in the past 19 days, along with the 850mb and 925mb temperatures as well as the sea-level pressure (MSLP).  The strong inversion began on the 3rd when the pressure rose up through 1030mb, and the inversion weakened today as the pressure dropped back below 1030mb.  Of course the pressure itself is not directly responsible for the temperature changes, but it's correlated with the cloud cover: high pressure conditions tend to be clear and calm, allowing for large net radiation losses from the surface.

With the temperature staying below 0°F for 17 consecutive days and below -13°F (-25°C) for 11 days, this cold spell marks quite a change from recent warm winters; the last comparable cold period was in early December 2012.

It's interesting to note that the PDO phase is still moderately positive; its influence has certainly not been felt lately over interior Alaska.

Here's a chart of recent daily high temperatures, which have been more unusually cold than daily low temperatures.  Prior to today, 27 consecutive days had a high temperature below normal; that's the most since the very cold spring of 2013.  It will be interesting to see if the exceptional and persistent warmth that began so dramatically in May 2013 has ended for now, or if warmth will soon return.

Arctic Update

Last month I introduced a set of 19 climate observing sites around the periphery of the Arctic Ocean that allows us to get a reasonable picture of basin-wide temperature variations in the context of historical observations stretching back several decades.  This is rather useful because reanalysis estimates of temperature are subject to model error and bias, including sometimes errors in the long-term trends.  Of course measurements at surface observing sites have their own quality concerns including sometimes significant urbanization influences (e.g. at Utqiaġvik/Barrow).

The chart below shows the November mean temperature for each year since 1971, and just as for October the warmth this year stands out as unprecedented in the modern era.  The November mean anomaly (departure from normal) of +6.7°C was nearly 3°C higher than the previous warmest November (2012, +3.9°C).  This November's anomaly was also greater than October's anomaly (+5.6°C) and higher than that of January 2016 (+6.5°C); this makes November 2016 the warmest calendar month relative to normal in the data set.

The daily temperature analysis (see below) shows that peaks of basin-wide warmth occurred at the beginning of the month and around the 20th of the month.  By the end of November the anomaly was fading as more normal conditions returned to the Arctic overall, but nevertheless the mean temperature has remained above normal every day.

A couple of the stations (on Russia's central Arctic coast) have become more than 10°C colder than normal in the past couple of weeks, but other sites have remained extremely warm relative to normal; for example, Ambarchik on the coast of the East Siberian Sea reported a daily high temperature well above freezing  (+40°F) last Thursday (December 8).  For Ambarchik last Thursday, the daily mean temperature's departure from normal was an extraordinary +27.1°C (+49°F), which ties the most extreme daily warm anomaly for any of the 19 stations since 1971; the record was previously held by Deadhorse (Alaska) on January 15, 1997 (49°F warmer than normal).

The NSIDC's latest report provides additional context on November's Arctic weather patterns and the state of sea ice.  November's Arctic sea ice extent was the lowest on record for the month, and the monthly ice extent remains more than 3 standard deviations below normal (although not quite as anomalous as during October).

Although not directly related, the ice conditions in the Antarctic are so unusual that they also deserve a mention; the November mean ice extent in the south was by far the most anomalous on record (for any month) at 5.7 standard deviations below the 1981-2010 normal.  This year's conditions are an extreme outlier compared to the climate of the last few decades, as seen in the NSIDC's graphic below.  On a global basis, sea ice extent in November was an astonishing 7.2 standard deviations below the 1981-2010 normal.

50 Below Reached

The Chalkyitsik RAWS reported -51°F this morning, which is the lowest observed temperature of the winter so far in Alaska (except for a dubious reading from the Bettles SNOTEL, which seems to have a cold bias).  The Chalkyitsik RAWS is located about half way between Fort Yukon and the village of Chalkyitsik, and is more than 100 feet lower in elevation than the latter.  Climate observations from Chalkyitsik itself were recorded only from 1962 to 1972, and the data are quite incomplete.

Temperatures in Fort Yukon were nearly steady at around -45°F today.  Below is a simple animation of the south-pointing FAA webcam view during the 3 hours of sunshine.  It's interesting to see the apparent fluctuations in the concentration of condensate or particulates in the shallow surface layer.  I'm not quite sure if the haze is composed of smoke or ice crystals, although the airport reported ice crystals in the air for most of the day, and the reported visibility varied from 0.5 to 2.5 miles.  With a strong temperature inversion undoubtedly present in the lowest levels, the haze was trapped very close to the surface - even below the tops of the trees at times, judging from the visual appearance.  A lot of interesting physics goes on in the surface layer of a winter day north of the Arctic Circle.

Cold Spell Persists

It's been quite a while since interior Alaska has seen persistently cold conditions during winter, so the current lengthening cold spell is a notable change of scene.  Including today, Fairbanks has been colder than normal for 17 of the last 18 days.  Here's an update to my chart showing the daily statewide minimum temperature; this is based on 24-hour minimum temperature reports ending at various times of day, so the daily measurement windows are not the same everywhere, but the trend is clear: the colder spots in Alaska have produced -40s for more than a week now.

The coldest area last week was the north-central interior, with the Norutak Lake RAWS reaching -40° or colder for 8 consecutive days ending yesterday.  The coldest conditions have now shifted to the northeast, with the Chalkyitsik RAWS sitting at -48°F this evening and apparently on track to break -50°F tonight.  Last winter no station in Alaska quite reached -50°F, although the Kanuti Lake RAWS got very close.

Environment Canada's surface chart from yesterday afternoon showed strong high pressure over northeastern Alaska, and the Fairbanks sounding measured a cold column with 850mb temperatures well below -20°C.  However, strong warming has been occurring aloft in the past 24 hours as the high pressure produces subsidence; note the difference between the two soundings below (3pm yesterday and 3pm today).

The 1.5-meter temperature measurements from UAF's Poker Flat Research Range show the warm-up nicely; this morning the temperature jumped from -28°F to -7°F in only half an hour.  The Poker Flat site is located near Chatanika on the Steese Highway north of Fairbanks at an elevation of about 700 feet above sea level.