Elevation-Dependent Snow

Fairbanks airport reported light snow for more than 12 hours this morning, but with temperatures no lower than 34 °F there was little or no accumulation on the valley bottom.  However, on UAF West Ridge, at over 600 feet elevation, the temperature was close to 32 °F and there was accumulation for a while before diurnal warming led to melting.  Up on Keystone Ridge (1600'), the accumulation was 3.5" by mid-afternoon.  See below for photos: the first courtesy of the Alaska Climate Research Center webcam this morning, and the second from Keystone Ridge at about noon AKDT.

In the 18-year history of observations on Keystone Ridge, measurable snow has been on the ground on September 30 on 9 occasions; and this snow remained throughout the winter in 6 of 9 cases.  So we might estimate approximately a two-thirds chance that today's snowfall represents the establishment of the permanent winter snowpack in the hills around Fairbanks; although with a healthy 3.5 inches reported, and below-normal temperatures expected by Friday, it seems the odds may be higher than that.

Forecast Showing Cold Spell

The medium-range forecast for interior Alaska has taken on a cold look in the past few days and is now showing a cold upper-level trough being carved out over the state by the weekend.  If this verifies, it would bring daytime high temperatures near or below freezing in Fairbanks and of course a good chance of accumulating valley-level snow.  According to the long-term average, the establishment of the permanent winter snowcover is still more than two weeks away, but it can happen in early October - for example, the 2000-2001 winter snowpack arrived on October 5.

The maps below show the ECMWF and GFS model forecasts of 850 mb temperature anomaly (shaded) for Friday afternoon, along with the 500 mb height (dashed lines); the models agree in showing temperatures more than 6 °C below normal over parts of the interior.  Of course, this is still 5 days out, so the details will undoubtedly change.

Fairbanks Temperature Trends

Regular reader Mike asked if we could take a look at the long-term trends in 850 mb and surface-level temperatures over the interior, with a particular focus on whether surface temperatures are warming more rapidly than temperatures aloft.  For an initial look at this, I pulled out the data for Fairbanks airport since 1950, because a direct comparison between the two levels is possible at this location.  The charts below show the 64-year history of mean temperature anomalies for each of the conventional meteorological seasons.  The red and blue lines at the top of each chart show the 850 mb (blue) and surface (red) temperature anomaly relative to the 1981-2010 normal, and the green columns show the difference between the two (surface minus 850 mb anomaly).  If the surface has warmed more than the 850 mb level, then we would see a general upward trend in the green columns.

The long-term trends are summarized in the following chart, which demonstrates that winter has seen by far the most significant warming over the long haul, and autumn has seen almost none.  It's also clear that the surface has indeed warmed slightly more than the 850 mb level, but the difference is quite small.

There are a number of interesting aspects of the data that seem to be worth commenting on.  First, the great Pacific climate shift of 1976, i.e. the switch to a positive PDO phase, stands out very clearly in the winter temperature chart.  There has been no long-term winter warming since 1976, but then the PDO phase has been more often negative in recent years, so we might expect this.  In fact, if the PDO was all there was to it, then temperatures would have dropped more in recent years.

Second, the 1976 PDO shift also appears to be evident in the spring temperature anomalies, but much less so in the summer and autumn.

Third, it's interesting that autumn does not show more warming because of the loss of Arctic sea ice; we recently discussed the remarkable warming in October over much of Alaska, but Brian showed that Novembers have been cold lately; so the October and November changes have opposing contributions to the overall autumn anomaly in Fairbanks.

Fourth, there is a suggestion that the surface-850mb differences peaked in the 1970s and 1980s, and were more negative both early and late in the history - especially for spring and summer.  The Atlantic Multidecadal Oscillation (AMO) was negative in the 70s and 80s, and positive on either side, so I wonder if there is a causal relation.

Finally, I couldn't help but notice that the autumn temperature anomalies seem to show a rather regular zig-zag pattern from year to year at times, i.e. the inter-annual variations look less random than in the other seasons.  More analysis would be required to assess whether this is unlikely to be just a coincidence.

Fall Temperature Drop Off

As Richard noted yesterday, temperatures are dropping in the Interior. This should not come as a surprise to anyone who lives in most of Alaska. As the sub-solar point drops south of the equator (Autumnal Equinox) there is less and less solar energy to prop up the temperatures. October is traditionally known as the time of greatest temperature decline but that is not the case everywhere.

Let us begin with the amount of temperature change. I looked at the 1981-2010 NCDC normals for all stations in Alaska and identified the maximum decline for the daily normal temperature during any 31-day period. The results of that analysis are shown in Figure 1. Approximately 25% of Alaska observes a 31-day temperate drop in the fall/winter of 23°F or greater and half of the state sees a drop of at least 20°F. The statewide "winner" is actually the Woodsmoke Cooperative station near Fairbanks with a seasonal drop of 27.9°F. Central is next with 27.2°F followed by Chicken with 26.6°F. The smallest 31-day change in the Fairbanks area is Keystone Ridge with a value of 19.3°F. At the other end of the spectrum is Adak in the far western Aleutian Islands. Their seasonal temperature drop (31-days) is only 6.0°F. Dutch Harbor, Shemya, St. George, and St. Paul are all at, or under, 7.0°F.

Figure 1. Largest 31-day decline in normal temperatures throughout the year based on NCDC normals.

Anecdotal, we think of October 1st as the beginning of this seasonal drop off but that is not necessarily the case. Figure 2 shows the first day of the greatest 31-day temperature decline for all of Alaska based on stations with NCDC published normal temperatures for 1981-2010. As it turns out, the October 1st line runs right through the Fairbanks area. However, nearly half of Alaska does not get going on their fall temperature swoon until the first week of October. In southwestern Alaska, it begins in the middle to later portions of September.

Figure 2. Starting date of the largest 31-day decline in normal temperatures throughout the year depicted in Figure 1.

Cooling Off

It won't have escaped anyone's notice that temperatures are dropping fast across interior and northern Alaska (especially after the mixed rain and snow in Fairbanks today).  Bettles recorded their first freezing degree days of the winter yesterday (daily mean temperature below freezing), which is just about on schedule for the time of year.  However, the low temperature was 14 °F, and it's still a little early for that kind of chill.  Up at Barrow, yesterday was the first day that the temperature failed to rise above freezing; this is later than normal, as we would expect with the reduced ice cover.

The chart below shows the recent daily high and low temperatures from Fairbanks airport, Keystone Ridge, and the secondary airport site (FAUA2).  It's interesting to note that Keystone Ridge has seen the temperature drop below freezing on six nights so far this autumn, but the primary airport site has only had two frosts so far.  Looking at the joint history of observations since 1996, I expected to see that this is unusual, thinking that overnight inversions would typically produce more frost in the valley than on the hill in September.  However, the overall frequency of frost is just about equal in September at the two locations; some years favor more early frost events in the valley, and others, like this year, favor more frost on the hill.  It's not until October and the advent of snow cover that daily low temperatures become significantly lower in the valley (see climatological chart here).

The chart also shows the much cooler temperatures reported from the airport #2 location, where the temperature has dropped below freezing 14 times already this autumn.  We discussed this thermometer's characteristics at some length here, although without arriving at a clear resolution as to whether the #2 temperatures are truly legitimate.  However, Brian showed that the 2000-2013 temperature differences between the two airport sites were largest in the month of September, and so the recent differences are consistent with this fact.

Fairbanks First Snow Stats

The forecast for Fairbanks on Tuesday includes a mention of snow for higher elevations. Here is the text-based forecast issued at 4:00 p.m. AKDT for Zone 222:

TUESDAY...MOSTLY CLOUDY. RAIN LIKELY...MIXED WITH SNOW ABOVE

1500 FEET. UP TO 1 INCH OF SNOW ACCUMULATION. HIGHS IN THE MID

40S. LIGHT WINDS.

TUESDAY NIGHT...MOSTLY CLOUDY WITH RAIN AND SNOW LIKELY IN THE

EVENING...THEN DECREASING CLOUDS AFTER MIDNIGHT. NO SNOW

ACCUMULATION. LOWS IN THE 20S. WEST WINDS TO 10 MPH.

There are several places around Fairbanks that are greater than 1,500' above sea level. The only one with at least 15 years of snowfall data is Keystone Ridge. If they receive the forecasted 1" of snow, it would be 2 days earlier than the median date of September 25th. All other stations have a median first snow date between September 30th and October 4th. As the map in Figure 1 indicates, the first snow has been as late as the end of October or even early November!

Figure 1. Dates of first snowfalls for stations around Fairbanks with at least 15 years of complete snowfall data since 1950.

Figure 2. Photo from the Alaska NWS Facebook page showing light snow on the UAF Campus. Image Source:    https://www.facebook.com/NWSAlaska/photos/a.146147645438637.42652.137686999618035/739650926088303/?type=1&permPage=1

Arctic Ice Minimum

Arctic sea ice extent appears to have reached its seasonal minimum in the past day or two.  The National Snow and Ice Data Center has not yet "called" the minimum but indicated a few days ago that the minimum extent would be slightly lower than last year.  However, the Danish Meteorological Institute graphic suggests that the ice extent is very slightly higher than last year (see below).  Regardless, this year's ice extent has followed a similar path to 2013 throughout the melt season.

It's worth noting that the CFSv2 computer model forecast of significantly increased ice extent this year was flat-out wrong.  We discussed the forecast here and here.  Given the magnitude of the anomaly in the forecast, this is quite a bust.

Here are some other figures to document this year's ice cover at seasonal minimum, courtesy of the University of Illinois and NSIDC.

September Precipitation Days

After the tremendous deluge on September 1st and 2nd, no measurable precipitation has been observed at the Fairbanks International Airport. Is this unusual? Since 1915, the average amount of precipitation between September 3rd and September 17th is 0.49". This is only the 4th year (out of 99) with no precipitation during those dates. The other years were 1963, 1974, and 2006. Also, there are normally 4.3 days with measurable precipitation during these 15 days; but of course none this year.

For the entire month of September, the long-term average number of days with measurable precipitation at the Fairbanks International Airport is 9.8. When days with Trace values are included, the number jumps up to 16.0 days. The 9.8 days with measurable precipitation is actually on the low end of other readings around Fairbanks. Figures 1 shows the normal (average) number of days with measurable precipitation in the Fairbanks area for all stations with at least 10 complete years of precipitation data since 1950. Figure 2 shows all days with at least a Trace of precipitation.

Figure 1. Long-term average number of days with measurable precipitation in the greater Fairbanks area for all stations with at least 10 complete years of summer precipitation data since 1950.

Figure 2. Long-term average number of days with any precipitation (Trace or measurable) in the greater Fairbanks area for all stations with at least 10 complete years of summer precipitation data since 1950.

Readers of this blog probably know that September is past the peak of the wet season for Fairbanks. July is typically the month that observes the most rainfall and the most days with measurable precipitation. Figure 3 show which portions of Alaska are wettest during which month. Most areas south of the Alaska Range are wettest in September through December.

Figure 3. The wettest month of the year across Alaska based on NCDC normals.

At the statewide level, we see that a significant gradient exists with regard to the number of days with measurable precipitation during the month of September from northeast to southwest (see Figure 4). Approximately 1/3 of days see measurable precipitation northeast of a line from Glenallen to Wainwright. Southwest of that line, significantly more days with measurable precipitation are observed.

Figure 4. Long-term average number of days with measurable precipitation in the state of Alaska area for all stations with at least 10 complete years of summer precipitation data since 1950.

What is really interesting is to see the number of days with any precipitation (Trace values included) at the statewide level (see Figure 5). Note the very large number of Trace+ days along the North Slope and the west coast. The portion of the state that sees half or fewer days with a Trace or more of precipitation is now much smaller – with a minimum around Tok.

Figure 5. Long-term average number of days with any precipitation (Trace or measurable) in the state of Alaska for all stations with at least 10 complete years of summer precipitation data since 1950.

El Nino and Positive PDO in Winter

Back at the end of July, reader Tracy asked if I could look at the relationship between Alaska winter climate patterns and El Niño or the PDO phase.  I've been very remiss in not getting to this earlier, but now with the change of seasons the topic is more immediately relevant.  Earlier in the year I posted some maps illustrating the El Niño connection to summer and spring conditions, so we can do the same thing for winter.  For the purposes of this post, I've defined winter as November through February: the season that is both cold and dark.

First, the maps below show the distribution of temperature and precipitation anomalies in the top 10 El Niño winters from the period 1951-2013.  The colored columns represent the fraction of years falling in the three climatological terciles, i.e. below-normal, near-normal, and above-normal.  We see that strong El Niño winters most often bring unusual warmth to southeastern and south-central Alaska, but temperatures are most often close to normal overall in the interior and west.  Below-normal precipitation is a common occurrence from Barrow and Bettles down to Cold Bay in these winters.

What if El Niño conditions are combined with a positive PDO phase?  This is the situation that we are currently looking at, and it seems fairly likely to persist into winter, as a weak El Niño episode finally seems to be taking hold in the tropical Pacific, and the PDO phase is moderately positive.  As the maps below indicate, a warm winter becomes more likely in most of Alaska - except for the west and at Barrow - when the PDO phase "lines up" with El Niño.  Also, the precipitation signal is more mixed and shows no clear large-scale signal when the PDO is decidedly positive along with El Niño.

If the PDO were not positive, then the odds would favor near-normal temperatures in many locations (see below) - and the Niño-related warmth in the southeast would be gone.  Recall that we saw the same thing with the earlier maps for spring and summer: El Niño years look vastly different in Alaska depending on the PDO phase.

To confirm the strong connection with the PDO by itself, the maps for the top 10 positive PDO winters show an overwhelming signal for warmth in the interior, south-central, and southeast.

Finally, below are the mean anomaly maps for winter sea-level pressure and 850 mb temperature in the top El Niño winters and top positive PDO winters.  These maps suggest, interestingly, that the patterns are really quite similar between El Niño and the positive PDO, but the surface warmth is much more dominant in the case of the positive PDO.  Cloud cover differences may be an important part of this, but that will have to be a topic for another day.

MSLP, El Niño:

MSLP, Positive PDO:

850 mb Temperature, El Niño:

850 mb Temperature, Positive PDO:

Summer in September

It doesn't get any better than this in mid-September in Fairbanks: mid-70s F, abundant sunshine, and light winds.  Fairbanks International Airport is reporting 76 °F, which is higher than the normal high temperature at the climatological peak in early July.  It's also just two degrees shy of the warmest ever observed so late in the season.

The scene at the Golden Heart Plaza is simply glorious:

Chinook and Record Warmth

Southerly flow ahead of the Bering Sea storm has brought an intense chinook and very warm conditions to interior Alaska in the past two days, and if you thought records were being broken, you were correct.

At Delta Junction, yesterday's midnight-to-midnight low temperature was in the 60s °F, probably 63 or 64 °F (the official value is not reported yet), and this is a record for September - the previous record highest low temperature for the month was 60 °F in the great chinook of late September 1995.  The latest in the year that a low temperature as high as 63 °F has been reported before was August 28, 1989 (with mostly complete data back to 1943).

The lack of nighttime cooling in the Delta Junction area was of course associated with high winds, which were sustained near 30 knots and gusting over 40 knots for many hours yesterday.  The National Weather Service reported many high wind observations, including a gust of 76 mph from a station 11 miles SSW of Fort Greely.  Blowing dust was reported yesterday and Friday on the METAR observations from PABI; this is confirmed by some FAA webcam photos that I saved yesterday afternoon:

At Fairbanks the high temperature of 74 °F yesterday tied the record for the date, but the warmth was not as extreme as in the 1995 chinook, which took temperatures to 78 °F in Fairbanks on September 21.

Yesterday's low temperature in Kotzebue was 52 °F, which ties the record from 1933 for warmest low temperature ever observed so late in the season.  The high temperature of 65 °F on Friday set a new record for the date, but in terms of absolute anomaly this event again falls short of the 1995 event, which brought 66 °F to Kotzebue on September 21.

Upper Air This Summer: Part I - Temperatures

Summer 2014 will long be remembered in the Interior for the amount of rain that fell. Most stations around Fairbanks had their wettest summer on record. Bettles had their second wettest summer and their greatest number of days with measurable precipitation. In addition to the moisture, the temperatures were generally a degree or two below normal. Those are all surface based statistics. How did the upper levels look?

Figure 1 provides a point-specific summary of conditions for the Fairbanks upper level balloon soundings. For all categories, the 2014 temperatures were slightly below the normal value. Somewhat surprisingly, the precipitable water (PW) value was also slightly below normal. This phenomenon was discussed earlier in the summer. A summary can be found here.

Figure 1. Upper level summary at Fairbanks during June-August time period.

If we move beyond Fairbanks, a more complete picture of upper level conditions can be assessed. In regards to temperature, the upper levels were slightly below the 1981-2010 normals. Figures 2 and 3 show the normal 850 mb temperatures and the 2014 850 mb temperatures and Figure 4 shows the 2014 value as compared to normal (units are Celsius). Only Nome, Annette Island and St. Paul had appreciable warmer than normal 850 mb temperatures in 2014. As Figure 5 indicates, with the exception of Cold Bay, the 850 mb temperature anomalies mirrored the surface anomalies. Surprisingly, Cold Bay had near normal 850 mb temperatures even though they experienced their warmest summer on record. Kotzebue's surface departures were also somewhat disconnected from their 850 mb temperatures. In the case of Cold Bay the significantly warmer than normal sea surface temperatures kept temperatures in the lowest levels significantly above normal.

Figure 2. Normal (1981-2010) 850 mb temperature for the June-August time period.

 Figure 3. 850 mb temperature in 2014 for the June-August time period.

Figure 4. 850 mb temperature difference from 1981-2010 normal for the June-August time period.

Figure 5. Surface temperature departure from normal for the June-August time period.

At the 500 mb level, we see a similar pattern; i.e., most of Alaska was cooler at 500 mb except for Nome, Annette Island and St. Paul (see Figures 6 and 7). Again, Cold Bay and Kotzebue were extremely warm this summer but at 500 mb they were below normal. This would seem to indicate that their atmosphere was more unstable (steeper lapse rates) but both stations were below normal for the summer.

Figure 6. Normal (1981-2010) 500 mb temperature for the June-August time period.

Figure 7. 500 mb temperature in 2014 for the June-August time period.

Part II of this post will look at the precipitable water values statewide and how well it correlated to observed precipitation.

Tropical Cyclone Energy

After a chilly start to September, the temperature trend has gone into reverse and brought unseasonably late warmth to interior and northern Alaska.  The temperature exceeded 70°F in Fairbanks today, which happens this late in the season in only about 15% of years.  A strong low pressure system in the Bering Sea is largely responsible for the deep southerly flow that has brought warmth northward; heavy rain and strong winds were experienced in southwestern Alaska today.  The visible satellite image below shows the enormous comma-cloud appearance of the storm this afternoon as it was getting wound up.

The rapid intensification of the storm is evident in two sea-level pressure analyses from Environment Canada - see below.  The first is from 4 pm AKDT yesterday and the second is from 10 am today; in the 18 hours between these maps, two distinct low pressure centers merged and a rapid deepening was observed.

The more southerly of the two low pressure centers at 4 pm yesterday was the remnants of Tropical Storm Fengshen, which passed to the south of Japan on Monday as a 60-knot storm.  I'm no expert on the extratropical transition of tropical cyclones, but it's likely that the deep tropical moisture associated with the southern storm helped feed the intensification of the combined system.  Ironic as it may seem, the climate of Alaska is closely tied to events in the tropics.

Update: here's a wonderful infrared satellite image of the storm on Saturday morning, as linked by Gary in a comment.  The image is from the water vapor band, so light areas show regions of cloud and/or high humidity in the middle to upper atmosphere, while darker areas represent low humidity aloft (not close to the surface).