Monday, October 24, 2016

Clear and Colder

With a little snow now on the ground and nights continuing to lengthen rapidly, temperatures have dropped sharply under clear skies in the past few nights.  Last night was the coldest yet, with temperatures below -15°F in quite a number of locations in the eastern interior.  Here are a few of the coldest measurements so far:

-24°F  Bolio RAWS at Fort Greely
-22°F  Chicken COOP
-20°F  Tok #2 COOP
-17°F  Circle Hot Springs COOP
-16°F  Salcha RAWS
-12°F  Northway AP

The coldest spot around Fairbanks seems to have been the sensor near Smith Lake on UAF's North Campus, which reached -13°F yesterday and -12°F today.  Fairbanks airport dropped below 0°F for the first time yesterday; this is just slightly earlier than normal.

The cold is not particularly unusual for the time of year, but the persistence of clear skies this month certainly is very unusual.  In just the past 2 weeks, Fairbanks has seen sunny and very warm weather (55°F on the 12th), sunny and windy weather, and now sunny and calm, cold weather.  The absence of cloud is remarkable, because this is normally one of the cloudiest times of year in Fairbanks (only early August is comparable).  Including today, 19 of 24 days so far in October have had less than 50% cloud cover on average, and this is easily a record for the entire month - the previous highest total was 10 of 31 days in 2009 (based on ASOS data since 1998).

The reason for the clear skies is plain to see in a reanalysis estimate of 500mb heights so far this month: an axis of high pressure has extended from the northeastern Pacific to northwestern Alaska and the Chukchi Sea.

Compared to normal, upper-level pressure has been well above normal to the northwest, and so the usual westerly flow across interior Alaska has been blocked.  Obviously this also explains the lack of precipitation; the light snowfall last week produced a mere 0.02" of liquid equivalent precipitation - the month's only precipitation.  The driest October on record in Fairbanks was 1954, with 0.08".

Here's the October 1-22 average relative humidity at 850mb, according to the NCEP/NCAR reanalysis: there's a large zone of sub-45% values across the interior.

The departure from normal shows where the dry conditions aloft have been most unusual:

Friday, October 21, 2016

Snow Cover - Will It Last

A fair portion of the interior now has at least a little snow on the ground after yesterday's modest snowfall.  Half an inch of accumulation was reported at Fairbanks airport, but the snow depth was rounded up to 1".  It's little more than a dusting of snow, so what are the odds that it will last and prove to be the start of the winter snow pack?

An analysis from a couple of years ago (see here) indicates that we're now late enough in October that there is less than a 20% chance that a one-inch snow cover will melt out to a trace or less.  This suggests that this year is likely to be one of the 10% or so in which the first measurable snowfall doesn't melt out.  On the other hand, the forecast for next week looks warm, with major storminess in the Bering Sea importing warm air over interior Alaska.  About half of all years reach 40°F or warmer after October 25 in Fairbanks, so if it doesn't snow any more - which looks likely in the short-term - then I'd say there's a fair chance of seeing bare ground again.

The chart below shows the number of days with zero or trace snow cover occurring after the first measurable snowfall, versus the date of the first snowfall on the horizontal axis.  Obviously the earlier the first snow occurs, the more subsequent snow-free days there tend to be before winter sets in.  But it's interesting to see that very late first snows do tend to melt out - which makes sense if the weather pattern is unusually warm in those years.

The outlier on the lower left is from the infamous early winter onset of 1992 - the only time the winter snow pack was established in September, and mid-September at that.

Thursday, October 20, 2016

MOS Forecast Bias

As Fairbanks awaits the much-delayed first accumulating snow of the season today (according to the NWS), I wanted to look back at the forecast discrepancy I noted a couple of weeks ago to see how things turned out.  The chart below shows the automated computer forecasts of daily high temperatures from the GFS MOS product (black line) and the actual outcome in red.  Note that I've taken the forecasts for "day 3", for example a forecast for Thursday's high temperature produced on Monday evening.

The results are amazing: the MOS forecast was about 18-20°F too cold for more than a week.  Given the usually strong performance of MOS, the error is very surprising.

I've done some digging on potential causes for the problem, and there seems to be a partial answer in the way the MOS equations are developed.  First, recall that the MOS (Model Output Statistics) technique uses multiple regression to estimate relationships between observed weather variables and predictors in the "raw" model forecasts (e.g. 850mb temperature).  To better capture differences in the relationships between summer and winter, the equations are developed separately for April through September and for October through March (reference here); so the regression switches on October 1 and April 1.

What does this mean for Fairbanks?  Well, the vast majority of days between October 1 and March 31 have snow on the ground, and so the "winter" model will reflect the physics of a snow-covered landscape; for example, when high pressure develops with low humidity and clear skies, the model will predict a substantial drop in temperatures.  However, in the minority of cases when snow is absent, then the model forecast will be too cold.  In other words, the model is effectively assuming snow cover to be present, although it does not use snow cover as a predictor.

This information partially explains the huge cold bias in early October this year, because the weather pattern was just the kind of set-up that is normally cold in winter - high pressure, clear skies, and low humidity.  However, if the snow-cover explanation were completely adequate, then we would expect that similar errors would have occurred in other years with no snow cover in October.  The charts below show the same analysis for selected years in the past.

First, in 2015 the MOS forecasts were too warm (as expected) during the period of unusual snow cover in late September and early October, but there was no obvious bias later in October after the early snow melted off.

October of 2013 was very unusual with its lack of snow, and MOS showed a modest cold bias then, but nothing like what we've seen recently.

Finally, I looked at two earlier years in which strong high pressure occurred in combination with zero snow on the ground during October.  In 2003, high pressure aloft peaked at the beginning of October, and there was a slight cold bias in the forecasts - but nothing too alarming.

In 2009, a ridge of high pressure developed on about the 7th of October, and once again the MOS forecasts were too cold, but not drastically so.  Bear in mind that the GFS model and the MOS equations have changed over time, so the comparison to 2016 isn't quite fair, but if anything the newer forecasts should be better.

In conclusion, it's still not clear what caused the enormous errors in the MOS forecasts recently.  Part of the problem is that the winter MOS equations aren't suitable for predicting temperatures during snow-free conditions, but in previous years this hasn't caused a major issue.  The good news is the NWS forecasters ably detected the recent bias in the MOS output and adjusted their forecasts accordingly.  As a meteorologist myself, it's comforting to know there's still room to improve on what the computers provide.

Monday, October 17, 2016

Windy and Colder

Strong winds have been blowing in Fairbanks since yesterday morning, as high pressure to the north is creating an intense pressure gradient across the state.  The strong easterly flow has imported the season's first decently cold air mass, with 850mb temperatures dropping to -14°C above Fairbanks by yesterday afternoon.  Saturday morning's balloon sounding revealed the first entirely sub-freezing temperature profile of the season in Fairbanks, and this is much later than usual - the normal date for this event is September 28, and the record latest was in 1969 (October 20).  Fairbanks also recorded its first sub-freezing calendar day yesterday, with a high of 28° F - but no doubt it felt a good deal colder than that.

The graphics below show the sea-level pressure analysis (top) and 500mb height analysis (bottom) from 3pm AKST yesterday afternoon, courtesy of Environment Canada.  The development of an upper-level high pressure system over Chukotka is consistent with the high-latitude blocking patterns and negative Arctic Oscillation that have been observed so far this month.

Here's a surface weather display from about the same time.

Webcam images from Ester Dome yesterday showed an abundance of blowing dust in the Tanana River valley as a result of the strong winds.  The simple animation below shows the period when winds began to gust over 20 knots (23 mph) at Fairbanks airport; blowing dust was reported at the airport for several hours.

Sustained winds of 20 knots are quite unusual in Fairbanks at this time of year, as the following chart demonstrates.  In the nearly 20 years since the ASOS platform began reporting at Fairbanks airport, there have been only 2 other occasions with 20+ knot winds in the month of October, and September and November have seen zero and 1 such event respectively.  Winds like this are more commonly observed in late winter and spring.

In terms of wind direction, however, the current event (northeasterly winds) is much more typical in Fairbanks; the chart below shows that nearly all wind events of this magnitude are directed either from the northeast or from west to just south of west.  However, the easterly component is rarely observed in summer: since 1998, all of the June-September strong wind events have come from the western half of the compass (and one from the north).

Thursday, October 13, 2016

Freeze-Up Progress - Or Lack of It

The situation in Barrow deserves another quick mention today, because it's so extraordinary - the warmth is unprecedented for the time of year in the modern observational record.  There's no snow on the ground - not even a trace - and this is the latest date in the year when this has been observed (1920-present); and this isn't because it's been too dry, it's just been too warm.  Not a single day so far this autumn has seen a high temperature below freezing, and we're now 9 days past the previous record in this respect.  As the chart below shows, temperatures have not decreased at all in the last month - in fact the trend is slightly up.

Another perspective on the unprecedented warmth is to look at the accumulation of freezing degree days so far this season.  Barrow has managed a grand total of 4.0 FDDs so far this autumn, which is easily the lowest on record (2012 had 24.0 through October 12).  On the flip side, thawing degree days have accumulated at a record pace this month, with 24.5 TDDs for October 1-12.  The chart below shows the numbers for each year since 1930 (click to enlarge).

Here's the current webcam photo from Barrow - no sea ice and no snow.

Away from the warming influence of the Arctic (an odd concept, perhaps), freeze-up is beginning on interior rivers.  The webcam photos below show the situation on the Yukon River at Dawson YT and at Beaver.

For good measure, and for future reference, here are some shots from North Slope locations today.  Toolik Lake has frozen over in the past 48 hours.

Tuesday, October 11, 2016

Excessive Warmth in Barrow

It's that time of the year again when Barrow's temperatures reveal the impact of the loss of Arctic sea ice when compared to what used to be normal in the latter part of the 20th century.  Simply put, Octobers now are really warm relative to what they used to be, and there's little variance from year to year as the climate has become more or less maritime at this time of the year.  However, this month is standing out even relative to recent warm years, as Rick has noted in recent messages on Twitter:

To facilitate the comparison to how things used to be, I created climatological normals of daily high and low temperatures for 1961-1990 and 2002-2015; October of 2002 was the first in the modern run of very warm Octobers.  The chart below shows that the daily normals have risen for just about every day on the calendar, but the difference in October is most profound: from October 1 through October 29, the 2002-2015 normal low temperature is higher than the 1961-1990 normal high temperature.

Looking at the variance of daily temperatures, we find that the magnitude of daily variations has become much smaller from mid-September all the way through mid-winter, and this is especially true for daily maximum temperatures.  The peak percentage difference occurs on October 10, when the standard deviation of high temperatures has dropped by over 50% between the two periods shown here.  This is easy to understand physically: with sea ice now far to the north in October, it's just about impossible to get cold air into Barrow, so the lower part of the temperature distribution has been eliminated.

Here are some webcam photos from northern Alaska this afternoon, showing the relatively balmy scene.

Toolik Lake is still mostly open:

There's no snow on the ground at Kavik:

Teshekpuk Lake remains unfrozen, with bare tundra on the shore:

And there's just a little ice on one of the small lakes adjacent to Teshekpuk Lake:

Evidently the Colville River remains largely unfrozen at Umiat as well: