The FAA webcam image below shows the scene a few minutes after solar noon, with the sun just 5 degrees above the theoretical horizon.
Thursday, November 7, 2013
Ineffective Sunshine
The sun is shining brilliantly in Arctic Village today, but the temperature is holding steady at around -17 °F after an earlier low of -26 °F. With light winds, this suggests a near balance between absorbed incoming solar radiation and emitted outgoing longwave radiation.
The FAA webcam image below shows the scene a few minutes after solar noon, with the sun just 5 degrees above the theoretical horizon.
The FAA webcam image below shows the scene a few minutes after solar noon, with the sun just 5 degrees above the theoretical horizon.
Wednesday, November 6, 2013
Cooling Off
Interior Alaska finally caught up with the season yesterday and last night, as temperatures dropped rapidly in many locations to levels more characteristic of the time of year. There were some fairly impressive 24-hour and 48-hour temperature falls, including for example at Kaltag, which saw a change from 33 °F with rain at 8 pm AKST on Monday to -10 °F this morning. Compare the charts below, which show surface conditions (temperature in red digits) at 4 pm Monday (top) and 8 am today (bottom).
The lowest temperature reports that I came across include:
The lowest temperature reports that I came across include:
- -21 °F at the Norutak Lake RAWS, just south of the Brooks Range near the upper reaches of the Kobuk River
- -21 °F at the Noatak RAWS, in the Noatak National Preserve
- -18 °F at the Dahl Creek RAWS, near Shungnak
- -17 °F at Galbraith Lake, adjacent to the Dalton Highway on the north edge of the Brooks Range
- -16 °F at the Selawik RAWS (-13 at the ASOS)
October anomalies
[Updated to include Anchorage temperature at the forecast office, which is the official climate observing site - thanks Rick]
Below are maps showing the October temperature and precipitation anomalies at some primary observing stations across the state. The area of each circle indicates how unusual the month's conditions were, as measured by the anomaly in standard deviations for temperature and the percentile rank for precipitation (based on the 1981-2010 normals).
The month's average temperature was more than 1 SD above normal at all of the locations shown except Annette Island; the warmest place relative to normal was Cold Bay, with a remarkable monthly average of 3.5 SD above normal. The return period for an anomaly this large (either warm or cold) in any calendar month is about 185 years (i.e. if the climate were stationary and if the true distribution of monthly temperatures were Gaussian with mean and variance described by the 1981-2010 climatology, then a calendar month this unusual might be expected to occur once every 185 years on average at a single location).
Needless to say, the month was the warmest October on record at many locations, including (ranked from most significant to least):
Precipitation was well above normal in most locations and was the highest on record for October in McGrath. Barrow experienced its 8th consecutive month of above-normal precipitation. However, below-normal precipitation was observed at Fairbanks and Big Delta, where downsloping chinook winds kept things dry, and Annette Island, which was close to the apex of the big upper-level ridge. The map below shows the monthly departure from normal of the 500 mb heights:
Below are maps showing the October temperature and precipitation anomalies at some primary observing stations across the state. The area of each circle indicates how unusual the month's conditions were, as measured by the anomaly in standard deviations for temperature and the percentile rank for precipitation (based on the 1981-2010 normals).
The month's average temperature was more than 1 SD above normal at all of the locations shown except Annette Island; the warmest place relative to normal was Cold Bay, with a remarkable monthly average of 3.5 SD above normal. The return period for an anomaly this large (either warm or cold) in any calendar month is about 185 years (i.e. if the climate were stationary and if the true distribution of monthly temperatures were Gaussian with mean and variance described by the 1981-2010 climatology, then a calendar month this unusual might be expected to occur once every 185 years on average at a single location).
Needless to say, the month was the warmest October on record at many locations, including (ranked from most significant to least):
- Cold Bay, 46.0 F vs 44.0 F in 2002
- Gulkana, 37.8 F vs 35.4 F in 2002
- McGrath, 38.7 F vs 35.0 F in 2006
- Anchorage, 43.9 F (airport), 43.0 F (forecast office), vs 42.1 F in 1936
- Big Delta, 37.2 F vs 35.1 F in 1969
- Kotzebue, 34.5 F vs 32.6 F in 2006
- Bettles, 31.3 F vs 29.4 F in 2003
Precipitation was well above normal in most locations and was the highest on record for October in McGrath. Barrow experienced its 8th consecutive month of above-normal precipitation. However, below-normal precipitation was observed at Fairbanks and Big Delta, where downsloping chinook winds kept things dry, and Annette Island, which was close to the apex of the big upper-level ridge. The map below shows the monthly departure from normal of the 500 mb heights:
Tuesday, November 5, 2013
Single Digit Temperatures
With the imminent arrival of single digit temperatures (and below) in Interior Alaska, I put together a map of the average and latest dates of the first sub-10 degree readings around Fairbanks. Most stations are within a week or so of the latest first sub-10 degree occurrence. Only those stations with at least 15-years of winter temperatures with no more than 10 missing observations per winter were included. Most stations are a solid 2-3 weeks late for this measurement.
Note: Today is the 32nd consecutive day with an above normal daily temperature in Fairbanks. You have to go back to the winter of 200202003 to find a longer stretch (62 days). Before that, 1997 had two streatch longer than 32 days.
Note: Today is the 32nd consecutive day with an above normal daily temperature in Fairbanks. You have to go back to the winter of 200202003 to find a longer stretch (62 days). Before that, 1997 had two streatch longer than 32 days.
Aurora
The SNPP day night imagery (VIIRS sensor) caught a nice image of the aurora yesterday evening traversing the region between Eagle, Fairbanks, and Point Hope. The city lights for nearly all population centers are easily identifiable.
Temperature and Snow Cover (Part 3)
My apologies to those who have had enough of this topic, but I think I've finally arrived at a more satisfactory answer to the question of whether and how much snow cover affects temperature at this time of year in Fairbanks. Hopefully I won't try anyone's patience. To recap briefly, my first attempt found a fairly strong connection between lack of snow cover and unusual warmth in late autumn, but later analysis prompted by comments from reader Eric helped to clarify that this is mostly a reflection of the larger-scale weather regime. I showed that a lack of snow cover is not correlated with warming at the surface relative to temperatures aloft, and for a while I thought that was the final answer.
However, I was troubled by the result, because it is almost an axiom among meteorologists that snow cover causes cooling relative to bare ground - how could this not be true in Fairbanks? I've often used this belief to successfully adjust temperature forecasts based on knowing what the snow cover is like, and I suspect NWS forecasters do the same in Alaska. Surely there is something I am missing in the data.
I began to consider the fact that warmer air masses are generally associated with a smaller drop-off in temperature with height (in summer) or greater inversion strength (in winter) in Fairbanks. Conversely, colder air masses tend to have a larger vertical temperature difference (in summer) or weaker inversion (in winter); see the chart below for proof of this. We also know that snow depth is tied to temperature variations aloft; so then why did I find no connection between snow depth and vertical temperature difference? Could it be that the two influences on vertical temperature difference are countering each other and masking the cooling effect of snow?
In light of this, it's clear now that any investigation of how snow cover affects surface temperature must first remove the effects of air mass variations, because the large-scale weather regime is by far the dominant influence on the surface temperature - and therefore it "contaminates" the analysis. To address this, I performed a simple linear regression of surface temperature (anomaly) with 850 mb temperature (anomaly) and then looked at the residual, i.e. the surface temperature variations that are not explained by changes in temperature aloft. This provides a much cleaner look at the problem. The chart below shows the median residual temperature anomaly in snow depth categories, and we see a result that finally resembles what we expect. When snow cover is missing in October, the surface temperature tends to become increasingly higher than expected based on the 850 mb temperature; and we see a similar effect in April when snow melts earlier than usual. Conversely, when snow is present outside the normal season, temperatures are slightly lower, both in early October and early May.
It's interesting to look at the past few weeks to see how temperatures have departed from the expected values based on 850 mb temperature alone. The chart below provides the answer: Fairbanks temperatures have been running up to 10 °F or even more above the temperatures that we would normally associate with the recent 850 mb temperatures. The smoothed difference is above +5 °F in recent weeks; this is consistent with the result shown above for the end of October, but is higher than the expected no-snow residual earlier in the month. Of course, there are many other aspects of the weather pattern (besides 850 mb temperature) that affect the surface temperature, such as cloud cover and wind speed and direction, so we don't expect the no-snow residual to explain everything. Note that the 850 mb regression uses data from 1948 through July 2013, and does not include the recent conditions.
Now that we have an estimate of the cooling caused by snow cover, or warming caused by lack thereof, we can imagine using this information to improve daily temperature forecasts. To illustrate this in a basic fashion, consider the two charts below: the first shows the October "predicted" temperatures based on 850 mb temperature alone (no snow adjustment and no other meteorological considerations), but the second chart includes the estimated warming effect of missing snow. Forecasts falling on the diagonal line are perfect. It's clear that the snow cover adjustment provides a modest improvement by helping alleviate the cool bias, although the adjusted forecasts are still too cool.
Finally, here are a couple of zoomed-in charts to show the distribution of the temperature residual associated with zero or trace snow cover in autumn and spring. There is not enough historical data to provide estimates after October 30 or before April 12 (I required at least 10 years since 1930), but it seems likely that the warming effect of no snow would be at least as great for nearby dates. And therefore until snow cover is properly established in Fairbanks this year, we expect that temperatures will (on average) stay much warmer - probably at least 5 °F warmer - than the airmass temperatures would normally dictate for this time of year.
However, I was troubled by the result, because it is almost an axiom among meteorologists that snow cover causes cooling relative to bare ground - how could this not be true in Fairbanks? I've often used this belief to successfully adjust temperature forecasts based on knowing what the snow cover is like, and I suspect NWS forecasters do the same in Alaska. Surely there is something I am missing in the data.
I began to consider the fact that warmer air masses are generally associated with a smaller drop-off in temperature with height (in summer) or greater inversion strength (in winter) in Fairbanks. Conversely, colder air masses tend to have a larger vertical temperature difference (in summer) or weaker inversion (in winter); see the chart below for proof of this. We also know that snow depth is tied to temperature variations aloft; so then why did I find no connection between snow depth and vertical temperature difference? Could it be that the two influences on vertical temperature difference are countering each other and masking the cooling effect of snow?
In light of this, it's clear now that any investigation of how snow cover affects surface temperature must first remove the effects of air mass variations, because the large-scale weather regime is by far the dominant influence on the surface temperature - and therefore it "contaminates" the analysis. To address this, I performed a simple linear regression of surface temperature (anomaly) with 850 mb temperature (anomaly) and then looked at the residual, i.e. the surface temperature variations that are not explained by changes in temperature aloft. This provides a much cleaner look at the problem. The chart below shows the median residual temperature anomaly in snow depth categories, and we see a result that finally resembles what we expect. When snow cover is missing in October, the surface temperature tends to become increasingly higher than expected based on the 850 mb temperature; and we see a similar effect in April when snow melts earlier than usual. Conversely, when snow is present outside the normal season, temperatures are slightly lower, both in early October and early May.
It's interesting to look at the past few weeks to see how temperatures have departed from the expected values based on 850 mb temperature alone. The chart below provides the answer: Fairbanks temperatures have been running up to 10 °F or even more above the temperatures that we would normally associate with the recent 850 mb temperatures. The smoothed difference is above +5 °F in recent weeks; this is consistent with the result shown above for the end of October, but is higher than the expected no-snow residual earlier in the month. Of course, there are many other aspects of the weather pattern (besides 850 mb temperature) that affect the surface temperature, such as cloud cover and wind speed and direction, so we don't expect the no-snow residual to explain everything. Note that the 850 mb regression uses data from 1948 through July 2013, and does not include the recent conditions.
Now that we have an estimate of the cooling caused by snow cover, or warming caused by lack thereof, we can imagine using this information to improve daily temperature forecasts. To illustrate this in a basic fashion, consider the two charts below: the first shows the October "predicted" temperatures based on 850 mb temperature alone (no snow adjustment and no other meteorological considerations), but the second chart includes the estimated warming effect of missing snow. Forecasts falling on the diagonal line are perfect. It's clear that the snow cover adjustment provides a modest improvement by helping alleviate the cool bias, although the adjusted forecasts are still too cool.
Finally, here are a couple of zoomed-in charts to show the distribution of the temperature residual associated with zero or trace snow cover in autumn and spring. There is not enough historical data to provide estimates after October 30 or before April 12 (I required at least 10 years since 1930), but it seems likely that the warming effect of no snow would be at least as great for nearby dates. And therefore until snow cover is properly established in Fairbanks this year, we expect that temperatures will (on average) stay much warmer - probably at least 5 °F warmer - than the airmass temperatures would normally dictate for this time of year.
Sunday, November 3, 2013
Another Very Warm October in Barrow
Another October has come and gone, and that means that once again Barrow had a very warm October: the average temperature was 24.7ºF, which is 7.5ºF above the 1981-2010 normal. This ties for the fifth warmest October of record and seven of the ten warmest have occurred in the past 15 years. Here's an updated plot of the October mean temperature at Barrow since the second order Weather Bureau station opened in 1920:
I've segregated the data into tercile values for the full 94 years: colder than 14.2F, warmer than 19.6 and then the middle tercile 14.2-19.6F. So this means that by definition, one third of the 94 Octobers are "warm", one third middle and one third "cold". The green line is a local regression fit (LOESS) that preserves timing of changes multi-year variation (as e.g. a centered running mean would). Obviously, the run of overheated Octobers since 2002 are unprecedented in the 94 years of unbroken instrumental observations.
In spite of the sea ice edge being closer to the Alaska coast at the end of the summer than it has been in some years, the ice was slow to return to the coast. Here was the view from the Barrow sea ice webcam late on the afternoon of October 30th. As you can see, no ice was visible looking north, and reports on the ground confirmed that no sea was visible the on the morning of November 1st.
I've segregated the data into tercile values for the full 94 years: colder than 14.2F, warmer than 19.6 and then the middle tercile 14.2-19.6F. So this means that by definition, one third of the 94 Octobers are "warm", one third middle and one third "cold". The green line is a local regression fit (LOESS) that preserves timing of changes multi-year variation (as e.g. a centered running mean would). Obviously, the run of overheated Octobers since 2002 are unprecedented in the 94 years of unbroken instrumental observations.
In spite of the sea ice edge being closer to the Alaska coast at the end of the summer than it has been in some years, the ice was slow to return to the coast. Here was the view from the Barrow sea ice webcam late on the afternoon of October 30th. As you can see, no ice was visible looking north, and reports on the ground confirmed that no sea was visible the on the morning of November 1st.
Saturday, November 2, 2013
October Summary
Here's the October summary from the National Weather Service in Fairbanks:
PUBLIC INFORMATION STATEMENT
NATIONAL WEATHER SERVICE FAIRBANKS AK
852 AM AKDT FRI NOV 1 2013
...OCTOBER 2013 WEATHER SUMMARY FOR FAIRBANKS ALASKA...
THE WEATHER FOR OCTOBER 2013 IN FAIRBANKS WAS DOMINATED BY WARM
AND DRY CHINOOK CONDITIONS AS A LARGE RIDGE FORMED OVER EASTERN
ALASKA. UPPER LEVEL WINDS AROUND THIS RIDGE DIRECTED MULTIPLE LOW
PRESSURE SYSTEMS WITH SUBTROPICAL ORIGINS NORTH OVER THE ALASKA
RANGE CREATING MULTIPLE EPISODES OF CHINOOK WARMING AND DRY
CONDITIONS ACROSS THE INTERIOR. THE 1ST OF OCTOBER WAS THE ONLY
DAY THAT THE AVERAGE TEMPERATURE WAS BELOW NORMAL. THE 4TH OF
OCTOBER AVERAGE TEMPERATURE WAS RIGHT AT NORMAL. EVERY OTHER DAY
IN THE MONTH WAS ABOVE NORMAL WITH THE THE AVERAGE TEMPERATURE ON
THE 28TH OF OCTOBER WARMING TO AN INCREDIBLE 31 DEGREES ABOVE
NORMAL. THE HIGH ON THAT DAY OF 51 DEGREES WHICH SET A NEW DAILY
HIGH TEMPERATURE RECORD FOR THAT DAY. THIS BROKE THE OLD RECORD
OF 48 DEGREES SET IN 1923. THE HIGHEST TEMPERATURE FOR THE
MONTH WAS 57 DEGREES WHICH OCCURRED ON BOTH THE 18TH AND THE 19TH
IN CONJUNCTION WITH A SIGNIFICANT CHINOOK WIND EVENT IN THE ALASKA
RANGE. THE LOWEST TEMPERATURE FOR THE MONTH WAS 16 DEGREES ABOVE
AND OCCURRED ON THE 26TH. OCTOBER 2013 RANKED AS THE 3RD WARMEST
OF 109 YEARS OF RECORD AND WAS THE 40TH DRIEST OF 103 YEARS OF
RECORD WITH ONLY 0.59 INCHES OF PRECIP FALLING AT THE AIRPORT. THE
PALTRY 0.70 INCHES OF SNOW THAT FELL IN OCTOBER WAS 10.1 INCHES
BELOW THE NORMAL SNOW FALL OF 10.8 INCHES. THIS RANKED AS THE
100TH SNOWIEST OCTOBER OF THE 104 YEARS OF RECORD...OR IT RANKED
AS THE 4TH LEAST SNOWY OCTOBER ON RECORD.
THE AVERAGE HIGH TEMPERATURE AT THE FAIRBANKS AIRPORT WAS 44 DEGREES
WHICH WAS 12 DEGREES ABOVE THE NORMAL HIGH TEMPERATURE OF 32 DEGREES.
THE AVERAGE LOW TEMPERATURE AT THE FAIRBANKS AIRPORT WAS 28 DEGREES
WHICH WAS 12 DEGREES ABOVE THE NORMAL LOW TEMPERATURE OF 16. THE AVERAGE
TEMPERATURE WAS 36 DEGREES WHICH WAS AN INCREDIBLE 12 DEGREES ABOVE
THE NORMAL AVERAGE TEMPERATURE OF 24 DEGREES.
OCTOBER 2013 WILL BE REMEMBERED FOR ITS LACK OF SNOW FALL AND LACK
OF SNOW DEPTH ON THE GROUND. ONLY A TENTH OF AN INCH OF SNOW HAD
FALLEN AT THE AIRPORT THROUGH THE MORNING OF THE 31ST AND ONLY
TRACE AMOUNTS HAD FALLEN AT THE AIRPORT ON SEVEN OTHER DAYS. WITH
A VERY WARM GROUND AND WARM CHINOOK WINDS WHAT LITTLE SNOW FELL
QUICKLY MELTED ALLOWING ZERO SNOW DEPTH ON THE GROUND THROUGH THE
30TH OF OCTOBER. SNOW DEPTH AT THE AIRPORT REMAINED ZERO UP UNTIL THE
EVENING OF THE 31ST WHEN 0.6 INCHES OF SNOWFALL IN THE AFTERNOON
AND EVENING. SNOW DEPTH ON THE GROUND IS ROUNDED TO THE NEAREST
INCH SO THE 0.6 INCHES OF SNOW ON THE GROUND ON THE EVENING OF
THE 31ST WAS RECORDED AS A SNOW DEPTH OF 1 INCH FOR THE 31ST.
THE LAST MINUTE SNOWFALL PREVENTED MATCHING A LONG STANDING SNOW
DEPTH RECORD FOR THE 31ST OF OCTOBER. THE LAST TIME SNOW DEPTH ON
THE 31ST OF OCTOBER WAS ZERO IN FAIRBANKS WAS IN 1925 AND IS THE
ONLY OCCURRENCE OF ZERO SNOW DEPTH IN THE CLIMATE RECORD. THERE
HAVE BEEN 5 OCCURRENCES OF A TRACE SNOW DEPTH IN THE CLIMATE
RECORD...THE LATEST OCCURRING IN 1962.
LOOKING FORWARD TO NOVEMBER...THE AVERAGE HIGH TEMPERATURE DECREASES
FROM 18 DEGREES ON THE 1ST TO 7 ABOVE ON THE 30TH. THE AVERAGE LOW
TEMPERATURE DECREASES FROM 3 ABOVE ON THE 1ST TO 11 BELOW ON THE
30TH. THE AVERAGE DAILY TEMPERATURE DECREASES FROM 11 ON THE 1ST TO
2 BELOW ON THE 30TH. OVER THE LAST 108 YEARS THE TEMPERATURES HAVE
RANGED FROM 54 DEGREES ON THE 25 IN 1936 TO 54 BELOW ON THE 29TH
IN 1909. NOVEMBER AVERAGES 0.67 INCHES OF PRECIP WITH ALL OF THAT
FALLING AS SNOW IN MOST YEARS...HOWEVER RAIN EVENTS DO HAPPEN AS
MANY WILL REMEMBER THE CRIPPLING ICE STORM IN NOVEMBER OF 2010.
NOVEMBER AVERAGES 13.2 INCHES OF SNOW. POSSIBLE SUN LIGHT CONTINUES
TO DECREASE AT ABOUT 6 MINUTES PER DAY...FALLING FROM 7 HOURS AND
48 MINUTES ON THE 1ST TO 4 HOURS AND 45 MINUTES ON THE 30TH.
THE OUTLOOK FOR FAIRBANKS FOR NOVEMBER FROM THE CLIMATE PREDICTION
CENTER CALLS FOR EQUAL CHANCES OF ABOVE...NEAR NORMAL OR BELOW
NORMAL TEMPERATURES. THE CLIMATE PREDICTION CENTER IS ALSO CALLING
FOR EQUAL CHANCES FOR ABOVE NORMAL...NEAR NORMAL OR BELOW NORMAL
PRECIPITATION.
OBSERVED LAST MONTH...10/2013
OBSERVED RANK
AVG MAX TEMP (F) 44.0 5TH WARMEST OF 109 YEARS
AVG MIN TEMP (F) 28.2 3RD WARMEST OF 109 YEARS
AVG TEMP (F) 36.1 3RD WARMEST OF 109 YEARS
TOTAL PRECIP (IN) 0.59 40TH DRIEST OF 103 YEARS
TOTAL SNOW (IN) 0.70 100TH SNOWIEST OF 104 YEARS
$$
CCC
Friday, November 1, 2013
Temperature and Snow Cover (Part 2)
Yesterday's post on snow cover and temperature prompted an interesting question from reader Eric about whether the data really showed what I was claiming, which was that deeper snow cover in Fairbanks at this time of year causes lower temperatures, and conversely a lack of snow cover causes warmer conditions. Eric suggested that the data pointed to different air masses and their temperatures as the leading cause of snow variations, rather than implying that snow has a pronounced influence on temperature. As I mentioned before, this is a chicken and egg problem: which comes first, snow or cold temperatures?
To help untangle the chain of causality, I repeated the analysis with 850 mb temperatures rather than surface temperatures. The idea here is that 850 mb temperature will not be influenced by the hypothesized low-level cooling from snow cover; and therefore if 850 mb temperature shows the same connection with snow depth, then the correlation is more likely caused by weather regime variations, i.e. a pattern with warm air drawn up from the south will favor less snow, but a cold pattern will favor more snow. The chart below shows the results, which remarkably (to me) are extremely similar to the original chart for surface temperature (included below for ease of comparison):
The result seems clear: year-to-year snow depth variations in Fairbanks at this time of year are just as closely connected to the 850 mb (free atmosphere) temperature as to the surface temperature. Warm air aloft is tied to a lack of snow, and cold air aloft is associated with deeper snow cover.
To dig in a little deeper, I examined the connection between snow depth and the difference between surface temperatures and 850 mb temperatures. This is useful because if deep snow cover causes cooling, then the surface temperatures ought to be relatively low compared to the 850 mb temperature; but if snow doesn't cause cooling, then the vertical temperature difference would not depend on snow depth. Remarkably, the latter is true - see the chart below. There seems to be no significant connection between the vertical temperature difference and snow depth; the downward trend is just the climatological tendency for strengthening inversion as the sun goes away. One could argue there is a small separation at the end of October between no-snow and deep snow, but it's a small and no doubt statistically insignificant difference.
In summary, the updated results show that deeper snow in late autumn and early winter is not significantly associated with surface cooling relative to temperatures aloft, and the correlation between snow depth and temperature extends to temperatures well above the surface. And so in terms of causation, it appears that snow depth responds to the large-scale weather pattern and is not itself a major influence on Fairbanks temperatures; this is just what Eric suggested.
I'd be glad to hear suggestions of other ways to look at this problem, or other physical mechanisms that could be at play to explain the data.
To help untangle the chain of causality, I repeated the analysis with 850 mb temperatures rather than surface temperatures. The idea here is that 850 mb temperature will not be influenced by the hypothesized low-level cooling from snow cover; and therefore if 850 mb temperature shows the same connection with snow depth, then the correlation is more likely caused by weather regime variations, i.e. a pattern with warm air drawn up from the south will favor less snow, but a cold pattern will favor more snow. The chart below shows the results, which remarkably (to me) are extremely similar to the original chart for surface temperature (included below for ease of comparison):
The result seems clear: year-to-year snow depth variations in Fairbanks at this time of year are just as closely connected to the 850 mb (free atmosphere) temperature as to the surface temperature. Warm air aloft is tied to a lack of snow, and cold air aloft is associated with deeper snow cover.
To dig in a little deeper, I examined the connection between snow depth and the difference between surface temperatures and 850 mb temperatures. This is useful because if deep snow cover causes cooling, then the surface temperatures ought to be relatively low compared to the 850 mb temperature; but if snow doesn't cause cooling, then the vertical temperature difference would not depend on snow depth. Remarkably, the latter is true - see the chart below. There seems to be no significant connection between the vertical temperature difference and snow depth; the downward trend is just the climatological tendency for strengthening inversion as the sun goes away. One could argue there is a small separation at the end of October between no-snow and deep snow, but it's a small and no doubt statistically insignificant difference.
In summary, the updated results show that deeper snow in late autumn and early winter is not significantly associated with surface cooling relative to temperatures aloft, and the correlation between snow depth and temperature extends to temperatures well above the surface. And so in terms of causation, it appears that snow depth responds to the large-scale weather pattern and is not itself a major influence on Fairbanks temperatures; this is just what Eric suggested.
I'd be glad to hear suggestions of other ways to look at this problem, or other physical mechanisms that could be at play to explain the data.
Thursday, October 31, 2013
Temperature Averaging
* Updated at bottom for clarity
Last week a discussion erupted regarding the calculation of average temperatures. Here is a little light reading on an example from down in Anchorage. The Anchorage forecast office is the official site where climate observations are made for the state's largest city. There is a thermometer on the site that publishes temperatures every 10 minutes. Some PAFC staff have indicated to me that it represents the official temperature but Rick has indicated that might not be the case. If we assume that it is, we can compare an average of the 144 daily readings with the official min plus max divided by two method for calculating average temps. For about 18 months I save the 10-minute readings for the PAFC site and put them into an Excel file. In June 2009 the web functionality changed any it only allowed the user to go back a few days before the data was purged.
The charts below show the daily difference between the 10-minute mean and the official daily mean. Over 90% of the time the two numbers were within 1 degree of each other. In the warm months, the 10-minute readings were typically larger than the official daily average. (*Update: the black line is a 10-day running mean). That is primarily due to a quick spike in temperatures at the official site late in the afternoon that pushed the official high upward for only a brief period. In the cold months, the pattern was much more chaotic and the magnitude was surprisingly smaller.
Last week a discussion erupted regarding the calculation of average temperatures. Here is a little light reading on an example from down in Anchorage. The Anchorage forecast office is the official site where climate observations are made for the state's largest city. There is a thermometer on the site that publishes temperatures every 10 minutes. Some PAFC staff have indicated to me that it represents the official temperature but Rick has indicated that might not be the case. If we assume that it is, we can compare an average of the 144 daily readings with the official min plus max divided by two method for calculating average temps. For about 18 months I save the 10-minute readings for the PAFC site and put them into an Excel file. In June 2009 the web functionality changed any it only allowed the user to go back a few days before the data was purged.
The charts below show the daily difference between the 10-minute mean and the official daily mean. Over 90% of the time the two numbers were within 1 degree of each other. In the warm months, the 10-minute readings were typically larger than the official daily average. (*Update: the black line is a 10-day running mean). That is primarily due to a quick spike in temperatures at the official site late in the afternoon that pushed the official high upward for only a brief period. In the cold months, the pattern was much more chaotic and the magnitude was surprisingly smaller.
* Update: Here is an example of a single day's temperature profile. On 2/4/2008, the official high was +2 and the official low was -12. The strict arithmetic average of those numbers yields an average -5°F. What I am interested in knowing is exactly how representative this average temperature is as a proxy for the temperature at a much finer temporal resolution. Since there are 144 measurements throughout the day, we can compute an average in much greater detail. On this date, the average of the 144 measurements was -6.17°F. In my opinion, that is actually really good when compared to the simple high+low/2 method. 65% of the 10-minute observations were below the average but the short, intense peak nearly balanced the prolonged cold temperatures nearly perfectly. Over the course of the year, this value is pretty representative. In summer it is greater and in winter it is smaller. Hope that helps.
Temperature and Snow Cover
Temperatures in Fairbanks continue to run far above normal, as illustrated by the fact that this morning's 5 a.m. temperature of 19 °F is equivalent to the normal high temperature for this date. An increasingly large contributor to this anomaly is the lack of snow cover, which profoundly influences the balance of infrared radiation and temperature at the earth's surface. I thought it would be interesting to see how strongly snow depth is correlated with temperature in the autumn and early winter in Fairbanks - see the chart below. Note that I've calculated each category's median temperature only for dates with 10 or more years available; and the anomalies are all with respect to the 1981-2010 normals.
When snow cover is absent or merely a trace in Fairbanks (blue line), the temperature tends to be increasingly above normal as October proceeds; this reflects the increasing departure from normal that the lack of snow represents. Of course, this is a chicken and egg problem: do the warm temperatures cause lack of snow, or vice versa? It's both, but as the season advances, a lack of snow becomes an increasingly unusual influence that prevents valley-level temperatures from dropping as they normally do. Also, by late October even well above normal temperatures are not normally a hindrance to snow accumulation (this year being an extreme outlier with such abnormal warmth).
It's also interesting to see that prior to about 18 October, a thin snow cover (1-3") is associated with below-normal temperatures, but after that date it is more often observed with unusual warmth. This is simply because thin snow cover is greater than is usually observed at the earlier dates, but becomes less than usual later. The same is true for 4-6 inch snow depth; but snow more than 6 inches deep is associated with colder than normal temperatures throughout the autumn and early winter, and especially in October.
When snow cover is absent or merely a trace in Fairbanks (blue line), the temperature tends to be increasingly above normal as October proceeds; this reflects the increasing departure from normal that the lack of snow represents. Of course, this is a chicken and egg problem: do the warm temperatures cause lack of snow, or vice versa? It's both, but as the season advances, a lack of snow becomes an increasingly unusual influence that prevents valley-level temperatures from dropping as they normally do. Also, by late October even well above normal temperatures are not normally a hindrance to snow accumulation (this year being an extreme outlier with such abnormal warmth).
It's also interesting to see that prior to about 18 October, a thin snow cover (1-3") is associated with below-normal temperatures, but after that date it is more often observed with unusual warmth. This is simply because thin snow cover is greater than is usually observed at the earlier dates, but becomes less than usual later. The same is true for 4-6 inch snow depth; but snow more than 6 inches deep is associated with colder than normal temperatures throughout the autumn and early winter, and especially in October.
Tuesday, October 29, 2013
Feels Like Equivalency
Through October 28th, the daily average temperature in Fairbanks has been warmer than normal for 26 of the 28 days. One way to describe the actual temperature is to come up with an analog of which calendar day it most closely approximates. For example, the daily average temperature of 44.5 degrees yesterday would be expected to occur on September 17th. Therefore, a "feels like" equivalency of September 17th can be assigned to yesterday's climate. September 17th is 41 days earlier than October 28th. The following table shows the date, the actual temperature, and the feels like analog date. On average the analog date has been 15 days earlier that the actual date. So if it feels like the climate has been running about two weeks behind lately, you are correct.
Record Warmth Again
The warm, rainy conditions across interior Alaska have broken a number of longstanding records; here are a few, and I'm sure Rick and Brian can add more.
The chart below shows the updated daily temperature anomalies for Fairbanks since August 1. Remarkably, October so far is less than 1 °F cooler than the second half of September; and the October average high temperature is actually higher than in the second half of September.
- The 850 mb temperature above Fairbanks at 4 pm AKDT yesterday (9.4 °C) was the highest of record for so late in the season. The previous record for this late or later was 8.6 °C on December 12, 1985.
- The high temperature of 62 °F at Big Delta yesterday was by far the warmest recorded so late in the season; the previous record was 55 °F on (again) December 11, 1985. Yesterday's high would have been the warmest so late in the season if it had occurred two weeks ago, because a temperature above 60 °F has never been recorded in Big Delta after October 13 (records back to 1941).
- In Fairbanks, the high of 51 °F and low of 38 °F yesterday broke the Weather Bureau/NWS era daily records by 7 °F each. Temperatures higher than this have been observed later in the year, but not since the 1930s.
- Rick pointed out in an e-mail that this is the latest in the autumn that Fairbanks has ever observed measurable rain with no snow at all on the ground. Rain has occurred later with snow already on the ground, but the rain with completely bare ground is unprecedented for the time of year. The latest this has occurred before (since 1930) is October 24, in 1981. This morning's soggy scene captured by the Alaska Climate Research Center webcam on UAF West Ridge (below) has never been seen before at this time of year.
The chart below shows the updated daily temperature anomalies for Fairbanks since August 1. Remarkably, October so far is less than 1 °F cooler than the second half of September; and the October average high temperature is actually higher than in the second half of September.
Monday, October 28, 2013
Deep Warmth
A powerful southerly flow of air aloft is once again transporting extremely warm air for the time of year up to Alaska. The morning balloon soundings from Anchorage and Fairbanks show a deep layer of above-freezing air, extending up to nearly 8000 feet above sea level at Fairbanks and an extraordinary 9500 feet at Anchorage (see below). The 700 mb temperature of -0.1 °C at Anchorage is close to a record for the time of year.
The warmth won't last long in the interior this time, as the surge of warm air heads quickly east later today and the freezing level drops rapidly later today and tonight. It will be interesting to see if valley locations cool enough for accumulating snow tomorrow; if not, then October may end as the least snowy on record in Fairbanks (0.1 inches so far).
The warmth won't last long in the interior this time, as the surge of warm air heads quickly east later today and the freezing level drops rapidly later today and tonight. It will be interesting to see if valley locations cool enough for accumulating snow tomorrow; if not, then October may end as the least snowy on record in Fairbanks (0.1 inches so far).
Sunday, October 27, 2013
Most and Least Difficult Temperature Records to Break
Several of tomorrow's temperature records in Fairbanks are relatively easy to break. The low minimum record of -17 is only 1.97 standard deviations below the normal low of 6.3 degrees. Only two calendar days have record low temperature that are closer to their respective daily normal temperatures than October 28th. Therefore, the low minimum record for tomorrow is statistically, the third easiest low minimum record to break. Conversely, the record low minimum temperature for May 9th is pretty safe. It is 6.33 standard deviations below the 1981-2010 daily normal of 35.1 degrees.
The low maximum record for tomorrow is 4 degrees. That is only 1.61 standard deviations below the daily normal maximum of 21.4 degrees. No other low maximum record is closer to its respective daily normal temperature for any day of the year. Therefore, the October 28th low maximum record is the least difficult daily low maximum record to break during the course of the year. For reference, the coldest October 27 maximum is -1 and the coldest October 29 maximum is -6. Those records are more difficult to break.
The chart (graphic) below shows the most difficult daily records to break (left side) and least difficult daily records to break (right side) over the course of the year. There is a noticeable, and interesting, seasonality to the chart.
Four types of records are shown.
1) High maximum (a.k.a. record high) --> shown in pink
2) Low maximum (a.k.a record coldest high temperature) --> shown in yellow
3) Low minimum (a.k.a. record low) --> shown in blue
4) High minimum (a.k.a. record warmest low temperature) --> shown in green
Note 1: the standard deviations from normal are all based on the 1981-2010 climate normal period. The relative ease or difficulty to break a record should be compared to the current climate regime.
Note 2: The May 9 (1964) record was 5.32 SD below the mean for the 1931-1960 climate normal period.
The low maximum record for tomorrow is 4 degrees. That is only 1.61 standard deviations below the daily normal maximum of 21.4 degrees. No other low maximum record is closer to its respective daily normal temperature for any day of the year. Therefore, the October 28th low maximum record is the least difficult daily low maximum record to break during the course of the year. For reference, the coldest October 27 maximum is -1 and the coldest October 29 maximum is -6. Those records are more difficult to break.
The chart (graphic) below shows the most difficult daily records to break (left side) and least difficult daily records to break (right side) over the course of the year. There is a noticeable, and interesting, seasonality to the chart.
Four types of records are shown.
1) High maximum (a.k.a. record high) --> shown in pink
2) Low maximum (a.k.a record coldest high temperature) --> shown in yellow
3) Low minimum (a.k.a. record low) --> shown in blue
4) High minimum (a.k.a. record warmest low temperature) --> shown in green
Note 1: the standard deviations from normal are all based on the 1981-2010 climate normal period. The relative ease or difficulty to break a record should be compared to the current climate regime.
Note 2: The May 9 (1964) record was 5.32 SD below the mean for the 1931-1960 climate normal period.
Saturday, October 26, 2013
Minimum Diurnal Range
Fairbanks has just passed the date on the calendar when the climatological average diurnal range (the difference between the daily high and low temperatures) is smallest. At the airport observing site, the minimum diurnal range is just under 15 °F and is smallest around October 21, according to the NCDC 1981-2010 climate normals.
The chart below shows the annual cycle in diurnal range for Fairbanks, Anchorage, and Barrow. Of these three locations, the average diurnal range is highest at Fairbanks throughout the year, because cloudiness and relative humidity are usually much lower in the Alaskan interior than on the coasts; the relatively clear skies and low humidity in Fairbanks allow the temperature to show a greater response to the daily cycle in solar insolation (except in mid-winter). Also, in Fairbanks in winter, changes in cloudiness within a 24-hour period can bring large changes in temperature, whereas in Anchorage and Barrow the generally more persistent cloudiness makes large daily changes of this kind less common.
But how do we explain the October minimum in diurnal range in Fairbanks? I can think of three ingredients for a small average diurnal range:
Finally, snow depth is typically not great in October in Fairbanks - and of course in some years may be non-existent. If the ground remained snow-free for the rest of the year, then it's likely that the diurnal range would continue to decrease; but as snow depth builds in early winter in Fairbanks, then nighttime cooling increases and the diurnal range picks up again. At least, that's my interpretation of what's going on.
The rapid increase in average diurnal range in late winter in Fairbanks is undoubtedly due to the return of the sun, combined with a typically deep snow cover and generally clear skies; all three factors favor large temperature swings between day and night.
It seems that some of the same basic physics comes into play in Barrow and Anchorage, as the diurnal range minimizes near the onset of winter and then increases in mid-winter (slightly in Anchorage) and again in spring. However, I'm not sure what to make of the early summer variations in Barrow; any suggestions would be welcome.
To round out the analysis, below are maps showing the minimum diurnal range and the date of the minimum. The annual minimum in diurnal range tends to progress southeastward across the state through the autumn and winter, with some Pacific coastal locations seeing their least variable daily temperatures after the turn of the year.
The chart below shows the annual cycle in diurnal range for Fairbanks, Anchorage, and Barrow. Of these three locations, the average diurnal range is highest at Fairbanks throughout the year, because cloudiness and relative humidity are usually much lower in the Alaskan interior than on the coasts; the relatively clear skies and low humidity in Fairbanks allow the temperature to show a greater response to the daily cycle in solar insolation (except in mid-winter). Also, in Fairbanks in winter, changes in cloudiness within a 24-hour period can bring large changes in temperature, whereas in Anchorage and Barrow the generally more persistent cloudiness makes large daily changes of this kind less common.
But how do we explain the October minimum in diurnal range in Fairbanks? I can think of three ingredients for a small average diurnal range:
- a low sun angle and thus small solar insolation, minimizing daytime warming
- high cloudiness, minimizing both daytime warming and nighttime cooling
- lack of snow cover, minimizing nighttime cooling
Finally, snow depth is typically not great in October in Fairbanks - and of course in some years may be non-existent. If the ground remained snow-free for the rest of the year, then it's likely that the diurnal range would continue to decrease; but as snow depth builds in early winter in Fairbanks, then nighttime cooling increases and the diurnal range picks up again. At least, that's my interpretation of what's going on.
The rapid increase in average diurnal range in late winter in Fairbanks is undoubtedly due to the return of the sun, combined with a typically deep snow cover and generally clear skies; all three factors favor large temperature swings between day and night.
It seems that some of the same basic physics comes into play in Barrow and Anchorage, as the diurnal range minimizes near the onset of winter and then increases in mid-winter (slightly in Anchorage) and again in spring. However, I'm not sure what to make of the early summer variations in Barrow; any suggestions would be welcome.
To round out the analysis, below are maps showing the minimum diurnal range and the date of the minimum. The annual minimum in diurnal range tends to progress southeastward across the state through the autumn and winter, with some Pacific coastal locations seeing their least variable daily temperatures after the turn of the year.
Thursday, October 24, 2013
Start of Freeze-Up
In a rapid response to the cooler - though still much warmer than normal - conditions, ice has begun to form on inland waterways across northern Alaska. The webcam photo below, courtesy of Pro Music in downtown Fairbanks, shows some thin ice on the Chena River, and FAA webcams reveal a similar situation on the Yukon River at Beaver and the Koyuk River at Koyuk (also shown below). It is interesting that in the case of Koyuk, the lowest the temperature has dropped is 27 °F, but with a stiff breeze today, that is all it took to begin the process of freeze-up.
Also of interest is the view from Kivalina, on the Chukchi Sea coast northwest of Kotzebue. Despite temperatures only dropping below freezing yesterday evening, and a morning low temperature of only 24 °F, the lagoon behind Kivalina shows a substantial amount of ice today. There's just no getting around the inevitable for this time of year.
Wednesday, October 23, 2013
First Sub-Freezing High Temperature
It appears that Fairbanks has recorded its first day this season with a high temperature at or below freezing. This is a full two weeks later than normal for the Airport station and only three days away from the station record (Note: the Fairbanks Airport station has been recording observations since 1948). The stations on this map all have at least 15 years worth of winter seasons. Seasons with more than 10 missing observations were not included. Today's date is at, or later, than the record latest first sub-freezing day for a few locations. It is possible that some stations broke this record. The analysis on this map does not include and data from this winter.
Gradual Cooling
Surface temperatures have cooled off quite a bit across the eastern interior in the past couple of days, with overnight minimum temperatures in the single digits and teens in many places, as befits the season. The Chicken COOP observer reported 0 °F for the first time this morning, which is about 10 days later than normal; the climatological normal low temperature drops to zero in only three days from now at Chicken. A number of higher elevation RAWS locations also reported 5 °F or lower, and Arctic Village in the northeast reached 0 °F.
Despite the surface cooling, above-freezing air was still observed in this morning's 4 a.m. sounding at Fairbanks. The chart below shows that the 925 mb temperature (around 700 m above ground) has been above freezing at each 12-hourly measurement since early on October 11. With gradual cooling continuing aloft today, and surface temperatures just below freezing, the column will soon drop below freezing if it hasn't already. However, this won't last long, with warmer air returning soon.
Tuesday, October 22, 2013
October Warmth
- Fort McPherson, in the far northwest of Canada's Northwest Territories, experienced its warmest October week of record for the 7 days ending yesterday (Oct 15-21), with an average temperature of 44.1 °F. The old record, in a fairly complete period of record back to 1893, was 43.9 °F for Oct 1-7, 2003. It is quite astonishing that this record could be broken in the second half of the month, considering the rapid pace of seasonal cooling at this time of year.
- In Anchorage, the week ending October 20 was the second warmest October week of record (weekly mean temperature of 49.0 °F, compared to 49.6 °F for Oct 1-7, 2003).
- McGrath saw its third warmest (non-overlapping) October week of record. More remarkably, the Oct 13-19 average low temperature of 40.9 °F was the highest on record for any week either wholly or partially in October; the previous record was 40.7 °F for Sep 25 - Oct 1, 2002. The chart below shows the recent history of daily minimum temperatures in McGrath.
Subscribe to:
Posts (Atom)