Season of Freezing Temperatures

Every so often I do a quick climate analysis for a blogger in Houston, Texas, and one of the things that came up earlier in the winter was the frequency of freezing temperatures down there. In a few weeks, stations in Alaska will be approaching the date at which freezing temperatures are no longer expected. Therefore, I thought it would be interesting to look at 1) the number of days per season with sub-freezing temperatures and 2) the length of time between the average first freeze and the average last freeze. Specifically, I was interested in how Fairbanks compares to other places in Alaska and the Lower 48.

Number of Freezes

As the analysis unfolded, it became apparent that the Fairbanks International Airport station's values did not appear representative of the greater Fairbanks area. For example, the number of sub-freezing days per season at the airport's official station is 222.7. However, the Fairbanks Airport #2 station records 237.1 sub-freezing days. Therefore, I used an average of five stations to generate a single Fairbanks value of 226.6 sub-freezing days.

Figures 1 and 2 show the number of days with subfreezing temperatures for Alaska and for the entire U.S. respectively. Figure 3 shows the stations used in the Analysis. All stations with a value greater than Fairbanks' 226.6 (rounded to 227) are shown in blue dots. Note that the color scale is different on the two maps. I was a little surprised at the number of stations in Alaska that have more freezing days than Fairbanks – particularly places west of Fairbanks like Galena, Kaltag, and Tanana.

Figure 1. Map of Alaska showing the annual number of days with a temperature of 32°F or lower. Stations with a larger annual value than Fairbanks are highlighted in blue.

Figure 2. Map of the U.S. showing the annual number of days with a temperature of 32°F or lower. Stations with a larger annual value than Fairbanks are highlighted in blue.

Figure 3. Stations used in the analysis of sub-freezing temperatures. To qualify for the analysis, a station must have at least 10 complete years of data beginning in 1980-1981 (maximum of 33 years). Any year with more than 15 missing observations was excluded. There were 181 stations that met those qualifications in Alaska and 1025 in the rest of the U.S. The Lower 48 analysis did not include Cooperative or RAWS stations.

Length of Freezing Season

Another way of describing the frequency of freezing temperatures is to define a season where freezing temperatures are likely to occur. This is the opposite of a growing season. In this section I did a simple subtraction of the average date of the last freeze of the season minus the average date of the first freeze of the season. For Fairbanks, the average of the five stations that are collectively referred to a 'Fairbanks Valley' on the maps is 257 days (Sept 5 - May 20). Figures 4 and 5 show the number of days with in the freezing season for Alaska and for the entire U.S. respectively. All stations with a value larger than Fairbanks' are shown as blue dots (see Figure 3 for a map of all stations).

Interestingly, the number of stations in Alaska with a longer freezing season than Fairbanks is larger than the number of stations with a larger count of freezing temperatures. This implies that Fairbanks drops quickly into a freezing temperature regime in the Fall and bounces out of the freezing temperature regime quickly in the spring. This is also evident when comparing Figure 2 with Figure 5; that is, quite a few stations in the Lower 48 have a longer freezing season than Fairbanks (or most other Alaska stations for that matter). In the case of the Lower 48 stations, they are frequently at high elevations (less atmosphere to radiate longwave energy downward) and also have 8-10 hours of darkness in the summer (unlike Fairbanks) when heat can be radiated upward.

Figure 4. Map of Alaska showing the average number of days between the first freeze of the season and the last freeze of the season. Stations with a larger annual value than Fairbanks are highlighted in blue.

Figure 5. Map of the U.S. showing the annual number of days  between the first freeze of the season and the last freeze of the season. Stations with a larger annual value than Fairbanks are highlighted in blue.

Data Set

There are many interesting values that popped out. Far too many to mention. Therefore, for anyone who is interested, here is a LINK to the entire data set as an MS Excel file. The file is 2.8 megabytes.

Cold West, Warm East

A strong east-west temperature gradient is evident across Alaska today, with a temperature difference of more than 40 °F between the central Interior and the Yukon-Kuskokwim Delta region.  In an interesting setup, cold air has invaded southwestern Alaska from the Bering Sea while eastern Alaska is basking in warmth under the influence of the seemingly semi-permanent ridge over western Canada.  Here are surface observations from the last hour (temperatures in red):

Upper-air maps from a model forecast for 4pm AKDT today are shown below, revealing the cold cyclone aloft over the southwest and the strong gradient of temperatures at 850 mb:

The following map shows pressure and vorticity (a measure of spin) in the middle levels of the atmosphere:

The warmth in the Interior today signals the seasonal trend towards relatively warmer conditions inland, compared to the west coast, as spring draws closer.  In fact, it just so happens that today is the day that Fairbanks becomes warmer than Nome in the climatological normals.  The chart below shows the annual cycle of daily normal temperatures at these two locations which have almost identical annual mean temperature.

Temperature Skewness Maps

In response to reader Eric's comment, Brian helpfully offered to produce some maps of seasonal temperature skewness across Alaska.  Brian used his expert GIS mapping skills, and I supplied him with the data.  Except for some of the Arctic coast stations (where I used hourly temperature data), all of the results are from the 1981-2010 GHCN daily data.

Maps are shown below for the four climatological "seasons" of December-February, March-May, June-August, and September-November.  These don't correspond particularly well to meteorological times of transition in Alaska, but they are conventional categories for dividing up the data.  Recall that positive skewness indicates a long upper tail, so that large warm anomalies are more common than large cold anomalies; and conversely, when there is negative skewness, large cold anomalies are more common.  Qualitatively, the warmer/red colors on the maps show regions and seasons that are more often "very warm" (relative to the mean), whereas colder/blue colors on the maps indicate locations that are more often "very cold" (relative to the mean).

Interesting features of the maps include the general lack of significant skewness in the central Interior, except in spring, the pronounced skewness in many places in spring, and the strong positive skewness in far southern Alaska in summer.  Each of these features describes characteristics of the climate that go beyond the traditional "seasonal norms" and even "seasonal variability/variance", but are nevertheless important and perceptible aspects of the local environment.

[Update March 16:]  The chart below shows the summer temperature distribution for Kodiak, where the skewness is strongly positive.  In the period 1981-2010, the daily mean temperature in June through August was never more than 10 °F below the mean, but daily anomalies of +15 °F or more occurred on a number of occasions.  It seems that the favored setup for the warm anomalies involves high pressure to the west, thus bringing warm air from the mainland to the north - and sometimes the air is very warm.  In winter the large-anomaly air from the mainland would be cold, and so the distribution is negatively skewed.

Skewed Winter Temperatures

Several days ago (February 28, to be precise) I noticed that temperatures at Cape Lisburne in the far northwest of Alaska were above freezing again; this was the 12th day since December 1 on which above-freezing temperatures were observed at that location.  A quick look at the historical data showed that while unusual, this number of warm days was not unprecedented for a single winter season.  I then got to thinking about the temperature distribution at Cape Lisburne, because temperatures above freezing represent an anomaly of 25-35 °F or more from the daily climatological mean in winter; but I have rarely if ever seen an equivalent anomaly there on the cold side, i.e. temperatures of -30°F to -40°F or lower.  This suggests a skewed distribution of winter temperature anomalies, whereby large positive anomalies are more likely than large negative anomalies.  The following chart confirms this fact, using the hourly temperature data from 1981-2010:

The chart shows a significantly positively skewed distribution, as anomalies of +15°F or higher occur on 7% of all days, whereas anomalies of -15°F or lower occur only 1% of the time.  The most common daily temperature anomalies are small negative anomalies: fully half of all days have a mean temperature of 0-10°F below "normal".

A more extensive look at the data from other locations indicates that positive skewness is observed in winter at all of the Arctic coast locations, but none show skewness as dramatic as Cape Lisburne.  Here is the skewness coefficient, i.e. the third moment of the winter temperature anomaly distribution, at several locations:

Cape Lisburne: +0.65

Point Lay: +0.59
Wainwright: +0.36

Point Hope: +0.34

Barrow: +0.31

Barter Island: +0.15

Deadhorse: +0.14

Note that the data are quite incomplete at some of these locations, notably for Barter Island, Wainwright (10-year gap in the 1990s), and Point Hope (no data before 1990).

One might expect that similar skewness would exist for coastal stations south of Point Hope, but at least for Kotzebue and Nome this is not true: here are the winter skewness coefficients:

Kotzebue: -0.11

Nome: -0.37

The chart below for Nome shows that here the temperature distribution is quite negatively skewed, with large cold anomalies being much more common than large warm anomalies.  This is the kind of distribution I would intuitively expect in winter, with snow and sea ice at the surface tending to prevent temperatures from rising dramatically above the mean, but with no such constraints on cold anomalies.

A search through the GHCN database for Alaska reveals that negative skewness is indeed much more common than positive skewness, and in fact my calculations show only Barrow and Gulkana with positive skewness out of 71 stations with somewhat complete 1981-2010 daily data.  The winner on the negatively skewed side is Glacier Bay, with a skewness coefficient of -1.15; see the chart below.  Interestingly, 61% of all winter days in Glacier Bay are above the climatological normal (mean) temperature.

What might explain the unusual positive skewness that seems to be characteristic of Alaska's Arctic coast in winter?  I suspect it has something to do with downslope warming of already-warm southerly winds as they come off the high terrain of the Brooks Range; Cape Lisburne is flanked by high terrain immediately to the south and must experience downslope warming with some regularity.  The next step might be to examine Barrow upper-air temperatures and determine whether 850 mb temperatures behave the same way as the surface temperatures.

Umiat's Dubious February 1955 Records

*** Update on April 2, 2014: Rick has uncovered a note from the observer on the original microfiche that states the temperature observations are 17°F too low (cold). Between 5 and 5 p.m. on 2/18/55, the temperature was corrected by the observer and appear reasonable afterward. I have placed a before and after chart at the end of the blog post. ***

Last week we had a blog post that discussed wind chills and even wind chill records. At one point in the post there was a reference to a possible -106°F wind chill. After doing a little digging around, it appears that the -106°F wind chill is not credible. Specifically, I am talking about the Umiat station and the year was 1955. As many readers know, Umiat is probably the coldest station in Alaska with respect to winter temperatures. However, the data listed for Umiat in 1955 stretches credulity. In the GHCN database, there are two entries for Umiat, 1) Umiat (USW00026508), and 2) Umiat AFS (USW00026537). Unfortunately, the station history is quite a bit more complicated that just two station entries and other people are better equipped to discuss those differences.

The data for Umiat AFS only runs from 1953 to 1955 (it stops after Feb 22, 1955) – there is no overlapping data between the two stations. In Februray 1955, the data for Umiat AFS is exceptionally cold. As a comparison, I pulled all available February data for Barrow, Umiat, and Barter Island. All comparable years with data during that time period for those three locations was charted. There were 20 years in the comparison (1949-1955; 1976-1988). In addition, I plotted the 850 mb temperature from the Barrow upper air sounding during those time periods.

Fig. 1. Temperatures for 1949-1955 and 1976-1988 for Barrow, Barter Island, and Umiat. Barrow 850 mb temperature also plotted. The 1955 data is for February 1 to February 22 for all stations.

As you can see in Figure 1, the February 1955 data for Umiat is a significant outlier. In most years, the February temperatures are 5°to 8°F colder in Umiat than the other stations. However, in 1955 Umiat (AFS) was over 20°F colder. That discrepancy is far greater (by over 2x) than any other year. A good analog year for temperatures was 1984. In that year, the temperatures at Barrow and Barter Island were about 10°F warmer than Umiat – not 20°F+ like 1955. Again, that is a strong indicator that the 1955 temperatures in Umiat are not reliable.

The 1955 data for Umiat AFS is also colder than the acknowledged coldest month in Alaska's climate record (Ft. Yukon in December 1917). Figure 2 shows the list of every month in Alaska that has a monthly temperature lower than -40°F/C. Deering in December 1917 is crossed out due to the fact that its geographic location is not conducive to such an extreme monthly temperature and since the monthly LCD makes no mention of the station. In 1955, the second coldest month in Alaska was Barter Island with a monthly average temperature of -26.4°F.

Fig. 2. List of coldest months in Alaska.

Based on all available information, it appears that the measured temperatures for Umiat AFS in 1955 were approximately 10°F-15°F colder than the actual temperatures.

If the temperatures from Umiat AFS in February 1955 are no to be believed, then the -106°F wind chills from that month are also not to be believed. Figure 3 shows the time period with the minimum temperature and wind chill from based on the February 1955 data from Umiat AFS. Again, this data is now considered unreliable. If the temperatures are 10°F to 15°F too low, then the -106°F wind chill should be between -88°F and -94°F. These values are warmer than the statewide wind chill record. *** Update on 4/2/2014: The values are now considered to be 17°F too cold. Figure 4 shows the temperature and wind chill chart using the revised temperature values. ***

Fig. 3. Three-day temperature and wind chill chart for Umiat during February 1955 showing -106°F values. *** Modified on 4/2/2014 ***

Fig. 4. Three-day temperature and wind chill chart for Umiat during February 1955 accounting for the 17°F error identified by the original observer. The minimum wind chill is now -85°F. *** Added on 4/2/2014 ***

Out Colding Fairbanks

It is not often that anywhere in the Lower 48 has a colder winter than Fairbanks but this year was an exception. Looking at the primary stations across the U.S., 6 stations in northern Minnesota and North Dakota had a colder December through February than Fairbanks (see Fig. 1). This only includes so called 'primary' stations whose GCHN identifier begins with 'USW' (see Fig. 2).

The winter time period in Fairbanks saw an average temperature of -0.3°F. The normal winter value is -4.5°F. The 5 stations bested Fairbanks have winter normal temperatures in the +5°F +10°F range. While it may seems like stations in Minnesota and North Dakota would be colder than Fairbanks every couple of years, that is not the case. For the stations with the longest periods of record, Park Rapids, Pembia, Hibbing, and International Falls, you have to go back to the winter of 1976-1977 to find a season (with  complete records) that was colder than Fairbanks. That winter, Park Rapids was 0.8°F warmer than Fairbanks, Hibbing was 0.5°F colder, Pembia was likely 1.1°F colder (6 missing days), and International Falls was 3.8°F colder.

As many people know, International Falls calls themselves the "Icebox of the Nation. They have even trademarked the phrase (Reg. No. 3,375,139). Well, I can't help but note that most of Alaska is (considerably) colder than International Falls during an average year. Figure 3 shows those portions of Alaska that have a lower annual average temperature than International Falls.

 Fig. 1. Stations with a lower winter temperature than Fairbanks in 2013-2014.

Fig. 2. The 1043 stations evaluated in this study.

Fig. 3. Portions of Alaska that have a warmer/colder average annual temperature than International Falls, MN. This is based on NCDC 1981-2010 normal temperatures – not measured temperatures.

* Note: Figure 1 was updated on 3/8 after Hibbing reported data for the 8 days that had been missing.

Wind Chill Differences – Pre vs. Post 2001

As most readers of this blog know, the formula for calculating wind chill changed in 2001 (a good description of the new formula and the rationale behind the change cane be found here). The Old method was based on a study in 1945 that measured how long it took for water to freeze in various temperature and wind conditions in Antarctica. The study was not intended to determine equivalent temperatures but it was used for that purpose nonetheless and thus the Wind Chill Index (WCI) was born. Due to some glaring shortcomings with the application of the 1945 study to human perception, a series of wind tunnel experiments were conducted and a New Wind Chill Index was formulated and implemented by the NWS in 2001.

The New wind chill formula is actually an equation that was derived from a table with approximately 800 cells based on the wind tunnel experiments noted earlier. The air temperature in the original table ranged from +45°F to -40°F. The equation, however, can be used for any temperature and any wind speed. Clearly temperatures lower than -40°F occur frequently in Alaska. A comparison between the two formulas at 5°F is shown in Figure 1. In general, the New method's values are warmer than the Old values for most wind speeds. However, for low wind speeds, the New method produces slightly colder values since it assumes that the observer is walking at 1.5 m/s. The New method also takes into account the attenuation of wind from the standard measurement height to the space that people occupy.

Fig. 1. Old vs. New wind chill comparison.

Unfortunately many people in the general public do not know that this change was implemented and confuse Old and New values. For example, the football game in Cincinnati in January 1982 famously had a wind chill of -59°F at kickoff. However, few people know that using the New wind chill calculation, the kickoff value was 'only' -37°F. Figure 2 was produced by the Green Bay, WI, NWS office comparing the 1982 game with the 1967 Ice Bowl game. Conveniently, both Old and New wind chill values are given.

Figure 2. Football game comparisons. Source: http://www.crh.noaa.gov/news/display_cmsstory.php?wfo=grb&storyid=99351 

The Howard Pass event of mid-February possibly had a minimum wind chill of -97 (see Figure 3). The -97°F value is 1°F lower than a -96°F wind chill at Prudhoe Bay in January 1989 (Richard has done some excellent analysis and modeling of the Howard Pass event here and here). To confuse matters even more, the Prudhoe Bay wind chill value was originally -120°F since it occurred before the New wind chill formula was developed and implemented. In fact, there were several stations in January 1989 that had a wind chill of -120°F; including Barter Island and Cantwell. However, when using the New formula, the Prudhoe Bay value of -96°F bested the Barter Island and Cantwell readings.

Fig 3. Howard Pass RAWS data between Feb. 11 and Feb. 15, 2014. Air temperature, wind, and wind gust values are displayed. In addition, the New and Old windchill values are displayed. 

January 1989:

I did some digging around at the NOAA hourly observation site and decided to download all data for stations on the North Slope. None of the North Slope stations have ever recorded a wind chill lower than the -96°F in January 1989 at Prudhoe Bay. Of course the completeness of the data leaves a lot to be desired.

It was apparent that most North Slope stations recorded their lowest ever wind chill in January 1989. Therefore I also downloaded all data for stations across Alaska with hourly observations in January 1989.

Figure 4 shows the monthly minimum wind chill based on those observations using the Old formula in effect at the time and Figure 5 shows the monthly minimum wind chill based on the New wind chill formula. It is important to remember that the calculations were often, but not always, based on the same observation. These are minimum monthly values.

Fig.4. Minimum wind chills in January 1989 for all stations in Alaska that reported hourly observations using the Old formula. In many cases observations are missing so a few data point should be considered suspect.

Fig.5. Minimum wind chills in January 1989 for all stations in Alaska that reported hourly observations using the New formula. In many cases observations are missing so a few data point should be considered suspect.

As you can see, no portion of the state escaped the brutally cold wind chills. The color scale on both maps is identical. What stands out on on Figure 5 as compared to Figure 4 is, 1) the overall lower values, and 2) the evenness of the distribution. The -100s along the North Slope and in Bristol Bay were replaced with -80s and -90s. Also, the -70s in the Interior were replaced with -80s in many cases. As stated earlier, at low wind speeds, the New wind chill values are actually colder than the Old wind chill values.

Figure 6 is a map that shows the difference for the minimum wind chills using the two methods. Those areas in red saw their minimum wind chill values rise (warmer) with the New wind chill formula and those areas in blue saw their minimum wind chill values fall (cooler). Areas exposed to more wind generally saw their wind chill values rise with the implementation of the new formula and places that are less windy saw their wind chill values fall.

Fig. 6. Difference in the lowest monthly wind chill value in January 1989 using the Old and the New methods.

McGrath:

If you looked closely at Figure 5, you might have noticed the January 1989 minimum value for McGrath was -100°F. That is not a mistake. At 6 a.m. on January 27th 1989, the air temperature was a chilly -72°F with a wind speed of 7 miles per hour. According to the New wind chill formula, it felt like -100°F. Using the old formula, it felt like -94°F. Therefore, when the formula changed, McGrath went from middle of the pack to 1st place. Not just first place in January 1989 but possible first place all time. In addition to January 1989, I also downloaded hourly observations for any station that recorded an air temperature of -65°F or colder in any year (sub -100°F wind chills can occur with temperatures warmer than -65°F but the amount of data was far too large so I cut it off at -65°F). There was one other instance of a sub -100°F wind chill (-106°F to be precise) but its reliability is in doubt. Rick is doing some checking on it since that station does not have any other wind chills lower than -92°F. That will probably be the subject of another blog post.

Figures 7 and 8 show the hourly observations for McGrath and Prudhoe Bay respectively on January 27th and January 28th, 1989.

Fig 7. Hourly temperature and wind speed for McGrath, Alaska, on January 27 and January 28, 1989. Both the New and Old wind chill values are displayed. The -100°F value using the New formula is shown.

Fig. 8. Hourly temperature and wind speed for Prudhoe Bay, Alaska, on January 27 and January 28, 1989. Both the New and Old wind chill values are displayed. The -96°F value using the New formula is shown.

Is the -100°F at McGrath a populated station record?

That is an interesting question. For starters, it is worth noting that the original research that went into developing the new wind chill formula did not use air temperatures lower than -40°F and that the fitting of a formula to the observational data is the only reason it can be extended backward. Nevertheless, one of the pioneers of the New wind chill research, Randall Osczevski, discusses the New formula (which was actually developed in the 1990s) in this paper and uses values as low as -100°F with air temperatures as low as -60°F. He also discusses how the New formula is more realistic at extreme low temperatures. Also, since NOAA calculates wind chill values for any temperature and any wind speed, we can assume an implicit endorsement of the formula at extreme low temperatures.

The Alaska NWS offices do not issue Wind Chill Advisories or Wind Chill Warnings unless the sustained winds are, or are forecasted to be, 15 mph or greater for at least three hours. Maps of all Alaska advisory criteria can be found HERE. So, can a wind chill be a record if it wouldn't even qualify for a Wind Chill Advisory? Should there be a minimum wind speed criteria?

Several NWS offices have lists of statewide wind chill records. When the Minnesota Climatological Working Group discussed extreme wind chills, several examples they gave used wind speeds of 6 or 7 mph. The Montana Climate Atlas uses 10 mph as a threshold for their monthly probability maps. However, it is unclear whether the record lowest wind chills (see Fig. 9) also used the 10 mph threshold. Since it was unstated, my guess is that they did not.

Fig. 9. A table from the Montana Climate Atlas showing statewide lowest wind chills.

The NWS office in Lacross has a climatology of wind chills in the Northern U.S (Lower 48). They use a 10 mph filter do develop their probability maps. In their report, they state, "[t]his speed was chosen as it is the minimum threshold currently used throughout most of the NWS for the issuance of Wind Chill Advisories or Warnings." Since their report is looking at climatological probabilities, it makes sense that they have a wind speed cutoff. Their report does not contain a list of extreme wind chills.

The NCDC's State Climate Extremes Committee does not keep track of wind chill records and as far as I can tell, no standards exist. Wind chill is a human construct and is therefore not directly measurable with a calibrated instrument. My guess is that wind chill records will always remain unofficial.

To my knowledge, no one has questioned the measurements in McGrath. It is a first-order station and the NWS had an observer in McGrath at the time. Therefore, we can assume that the measurements and calculations are correct. In my opinion, any list of extreme wind chills should have all valid calculations included, regardless of the wind speed. If the wind chill made the temperature feel like -100°F, then that should be on the list ahead of the -96°F at Prudhoe Bay and the -97°F at Howard Pass (if verified).

Higher Resolution Howard Pass Modeling

One of the highlights of last month's weather in Alaska was the reported extreme wind chill from the Howard Pass RAWS in the western Brooks Range; the event was discussed at some length on this blog.  I presented a cursory look at two relatively high-resolution model forecasts of the event, which showed very low wind chill but did not reflect wind speeds as high as reported from the RAWS anemometer (up to 91 mph sustained).  We speculated that if the extreme wind speeds really occurred, then higher resolution modeling might be required to capture the local flow processes.  Prompted by this notion, I recently ran the WRF model over a limited domain at higher resolution

To obtain a high resolution simulation without requiring prohibitive computing resources, I used a traditional nesting technique to embed progressively higher resolution domains within larger coarse-resolution domains.  The figure below shows the three smallest simulation domains at 9 km (white), 3 km (blue), and 1 km (red) grid spacing; an outer domain at 27 km grid spacing was also used, with initial and boundary conditions from the GFS model.  The location of Howard Pass is indicated with a white dot.  I performed the simulation on an Amazon cloud virtual machine, which is available at a relatively low hourly cost.

The model topography over the inner domain, with 1 km (horizontal) grid spacing, is shown below, along with the locations of 4 RAWS installations.  The model was initialized at 18 UTC (9am AKST) on February 14, and run for 30 hours.

The charts below show the hourly evolution of temperature and wind speed at the location of the Howard Pass RAWS, for the 1km simulation (red line) as well as the 5km WRF forecasts that were examined earlier.  The black lines indicate the reported conditions from the RAWS.  Clearly, the high resolution simulation showed wind speeds only slightly higher than the lower resolution runs in the latter stages, and still far below the RAWS reports.  The 1km WRF forecast temperatures were also considerably higher than the reported temperatures.

Looking at the other three RAWS sites within the inner domain, the 1km wind forecasts generally showed less discrepancy from the observations, although the forecast wind speeds were also much too low at the Noatak RAWS for most of the time.  It is interesting, however, to see that the 1km model produced a jump in wind speeds to observed levels for a few hours at the Noatak RAWS; the 5km WRF runs were unable to reproduce the higher wind speeds.

The modeled spatial distribution of wind speeds in the vicinity of Howard Pass is shown in the maps below at intervals of six hours; the RAWS location is indicated with a white dot.  Note that the wind vectors are not shown for all the grid points, as the grid is much finer than the spacing of the arrows suggests.  For scale, the plot area covers roughly 50x50 km.

It's immediately apparent that the model produced considerable variability in wind speeds within 10 or 20 km of Howard Pass, as we would expect in an environment of complex terrain.  According to these results, the highest wind speeds were located just downwind of the highest terrain but did not extend to the Howard Pass RAWS location; and the highest speeds were still well short of the RAWS observations.

For comparison, the wind speed in the lower resolution simulations at 30 hours is shown below for a region of approximately 200x200 km.  It seems that the area-average wind speed near Howard Pass was only slightly higher in the 1km simulation, but the higher resolution allowed more spatial variability to develop.

In view of these new results, is it less likely now that the reported extremes from Howard Pass represented reality?  Well, perhaps; but I would suggest that the 1km simulation may still be inadequate to capture what actually happened.  The major reason for this is that a 1km simulation is still too coarse to explicitly simulate the "boundary layer", which is the turbulent layer of air next to the ground (airflow higher aloft is usually laminar, not turbulent).  Numerical models like WRF use so-called parameterization schemes to calculate and represent the effects of sub-grid-scale transfers of heat and momentum in the boundary layer.  It probably goes without saying that the details of the boundary layer scheme will make an enormous difference for the model's predictions of wind speed near the ground - and I think it's possible that the WRF boundary layer schemes are ill-suited to capturing the kind of flow that was occurring near Howard Pass.  [Note that I re-ran the simulation through 12 hours with an alternative boundary layer scheme and obtained very similar results.]

To illustrate the flow environment over Howard Pass, the chart below shows the vertical profile of temperature and wind speed at 12 hours into the 1km simulation.  A strong inversion was present as very cold surface-level air flowed up and over the pass from the North Slope, and wind speeds were highest just above the surface.

Unfortunately a mistake in the model setup meant that I didn't obtain the fine-resolution vertical profile data for later times in the model run, but we know that the winds above the surface strengthened dramatically in the next 18 hours.  The map below shows the wind speed at 825 mb, which was about 750 m above the level of Howard Pass RAWS and near the level of maximum wind speed.  Note that the 825mb pressure surface intersects the ground in the lower right, hence the lack of data.  The predicted 825mb wind speed was 80-90 mph or higher over a wide area above and west of Howard Pass; this is supported by similar plots for 850mb from the 5km WRF runs (see maps below).

 

According to the model, then, very high wind speeds occurred not very far above the surface on the lee side of the high terrain, but the model does not show the high momentum reaching the surface.  Presumably this is because the strong low-level temperature inversion caused the boundary layer scheme to produce relatively little vertical mixing of momentum; but it's an open question as to whether this is realistic or not.

To summarize, 1 km modeling of the Howard Pass event fails to reproduce the extreme conditions reported by the RAWS.  However, the discrepancy between the model and the observations still doesn't necessarily invalidate the RAWS data; in the words of a famous British comedy film, "it's only a model".  The obvious way to test whether the boundary layer scheme is artificially damping the surface wind speeds would be to re-run the model with still higher resolution (close to 100m) so that the boundary layer scheme can be dispensed with; this type of simulation is called Large Eddy Simulation.  Unfortunately, however, this would probably require some dedicated research funding to obtain the necessary computing resources.