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.

Periodic Thermal Waves

*Updated to add 500mb anomaly reanalysis maps at the bottom *

A few days ago I thought it would be interesting to see what would happen when calculating what day of the week was the warmest in the U.S. It was intended to be a silly exercise to demonstrate that even if one day was the highest it would be statistically trivial. I looked at 'primary' stations only since my poor computer could only handle two years at once before crashing. So I queried 2012 and 2013.  Well, much to my surprise, there were very distinct patterns. Not only were the patterns geographical, but the were temporal and propagated from west to east. Figure 1 shows the warmest days of the week for all primary stations in the U.S. Across areas with minimal longitudinal variation, the daily percentages are staggeringly variable. In Alaska for example (see Figure 2), an amazing 41% of stations were warmest on Thursday and only 4% are warmest on Tuesday or Wednesday. To me, that is a signal indicating a possible climate connection.

Fig 1. All primary stations in the U.S. mapped according to which day of the week during 2012 and 2013 was the warmest (1=Sunday, 2 = Monday, and so on.) Only stations with 95% complete data were used.

Fig 2. All primary and RAWS stations in Alaska mapped according to which day of the week during 2012 and 2013 was the warmest (1=Sunday, 2 = Monday, and so on.). Only stations with 90% complete data were used.

I then set out to check time periods other than seven days. Of course a seven day period corresponds very nicely with a calendar week. I checked time periods between 2 and 30 days. For the 30-day period think of it as grouping Jan 1, Feb 1, March 1, etc. and calling those "Group 1". Then take Jan 2, Feb 2 March 2, etc. and call those "Group 2". 

After a little trial and error, it was pretty obvious that 4-days was a very prominent pattern for the Lower 48 and modestly prominent for Alaska. Figure 3 shows which of the repeating 4 days is the warmest for the Lower 48 and Figure 4 shows the same but for Alaska.

Fig 3. All primary stations in the U.S. mapped according to which day, of a repeating 4-day pattern, during 2012 and 2013 was the warmest. Only stations with 95% complete data were used.

Fig 4. All primary and RAWS stations in Alaska mapped according to which day, of a repeating 4-day pattern, during 2012 and 2013 was the warmest. Only stations with 90% complete data were used.

Looking at the ESRL Reanalysis data (see Figures 5, 6, 7, and 8 below), a very clearly defined 4-day thermal wave is evident. I compared each of the four repeating days with the other days in the 4-day set to track the atmospheric patterns for 2012 and 2013. I am quite amazed that given the number of days used in the analysis (728) that any pattern whatsoever holds up. If the 4-day pattern was 3.5 or 4.5 days it would break down in the reanalysis data after a while. When I ran the same analysis at the 7-day interval (not shown) a noticeable, but less prominent thermal wave was also evident. I can share those images with anyone who may be interested.

So what recurs at 4 and 7 days. One obvious answer is planetary (Rossby) waves. They tend to recur at 7-day intervals according to the literature but their period is pretty variable and is highly dependent on the number of waves. The fact that the 7-day pattern in Figure 1 is more prominent between 40°N and 50°N lends credence to the planetary wave origin for that length time period. But what about the 4-day pattern? It has a much stronger signal. Is it related to planetary waves? What drives the thermal push depicted in Figures 5-8? Also, might the 7-day pattern actually be a 1/2 strength version of the 4-day pattern since 7 is almost, but not exactly, a multiple of 4?

Surely this is not a new discovery. My brief search of the literature didn't lead to any obvious answers beyond the Rossby wave solution. Any ideas would be much appreciated.

Fig 5. Day 1 of the 4-day thermal wave as depicted by the ESRL daily composite reanalysis. A wave axis is superimposed as a red line. This is a compilation of all Day 1s minus the average of Days 2, 3, and 4. 728 days were used in the analysis.

Fig 6. Day 2 of the 4-day thermal wave as depicted by the ESRL daily composite reanalysis. A wave axis is superimposed as a red line. This is a compilation of all Day 2s minus the average of Days 1, 3, and 4. 728 days were used in the analysis.

Fig 7. Day 3 of the 4-day thermal wave as depicted by the ESRL daily composite reanalysis. A wave axis is superimposed as a red line. This is a compilation of all Day 3s minus the average of Days 1, 2, and 4. 728 days were used in the analysis.

Fig 8. Day 4 of the 4-day thermal wave as depicted by the ESRL daily composite reanalysis. A wave axis is superimposed as a red line. This is a compilation of all Day 4s minus the average of Days 1, 2, and 3. 728 days were used in the analysis.

* Update section

After reading a section from a paper from 1976 (Blackmon, M. L. 1976. A climatological spectral study of the 500 mb geopotential height of the Northern Hemisphere. J. Atmos. Sci. 33, 1607 -1623.) that describes the propagation period of Rossby waves as 20° longitude per day, I decided to make reanalysis plots of the 500 mb height anomalies. The 20° value corresponds to 1/4 of the width of the Lower 48 states. Therefore, the 4-day thermal wave may be entirely explained by that.The wave train of anomalies stands out very nicely and clearly propagate from west to east. 

Fig 9. 500mb height anomaly on Day 1 of the 4-day thermal wave as depicted by the ESRL daily composite reanalysis. This is a compilation of all Day 1s minus the average of Days 2, 3, and 4. 728 days were used in the analysis.

Fig 10. 500mb height anomaly on Day 2 of the 4-day thermal wave as depicted by the ESRL daily composite reanalysis. This is a compilation of all Day 2s minus the average of Days 1, 3, and 4. 728 days were used in the analysis.

Fig 11. 500mb height anomaly on Day 3 of the 4-day thermal wave as depicted by the ESRL daily composite reanalysis. This is a compilation of all Day 3s minus the average of Days 1, 2, and 4. 728 days were used in the analysis.

Fig 12. 500mb height anomaly on Day 4 of the 4-day thermal wave as depicted by the ESRL daily composite reanalysis. This is a compilation of all Day 4s minus the average of Days 1, 2, and 3. 728 days were used in the analysis.

Near-Record Inversion

The temperature inversion above Fairbanks reached near-record strength yesterday morning, with a temperature difference of more than 30 °C between the surface and the warmest level aloft.  This has been observed only three times before in February, on three consecutive days in 1989.  The February record for strongest inversion (from the surface to the warmest level aloft) is 32.9 °C.

While we're on the topic of upper-air temperatures, I created a graphic to visualize the evolution of temperatures aloft through the winter thus far - see below.  This is a time-height cross-section of the balloon sounding temperatures at Fairbanks; the freezing line is denoted with a black contour.  The remarkable warmth in October and January is readily apparent, as is the relative scarcity of deep cold air.  Of course this graphic might be more useful if it showed temperature anomaly rather than actual temperature, but it will take a bit more work to develop the daily climatology at each level.

Strong Morning Inversion

A building ridge aloft is a good set-up for strong valley based inversions in the winter in Interior Alaska, as temperatures this Saturday morning illustrate. Here's the temperature trace from the lowest 1000 meters from the early morning sounding at the Fairbanks Airport.

As usual, most of the vertical temperature gradient is in the lowest 100 meter (-29C to -15C). The pressure gradient is increasing across the area this morning, with 10 to 20 mph east winds blowing at elevation, and the wind is starting to reach the valley floor in places. Surface temperatures from Fairbanks-land from 7 to 8am AST Saturday show a similar spread:

Goldstream Creek: -31F
Fort Wainwright RAWS: -30F
Woodsmoke CWOP: -30F (near North Pole)
Eielson AFB: -28F

UAF West Ridge -14F
Fairbanks Airport -15F (winds NE 5 mph)
Fort Wainwright Airport: -19F

Wickersham Dome: +10F (2200' MSL)
Caribou Peak RAWS: +10F (2500' MSL)
Keystone Ridge: +12F (1600' MSL)
Fox Ridge (CWOP)" +14F (2100' MSL)

A 40 degree spread is pretty impressive even by Fairbanks standards. Being two months past winter solstice, clear skies, and some wind I'd expect temperatures will rise dramatically at valley elevations today.

8-14 Day Forecast Visualization for Alaska

A few weeks ago I decided to download the GIS shapefiles than NCEP makes available on their website for the 6-10 Day and 8-14 Day extended forecasts. I am particularly interested in these because the push the boundary between a weather forecast and a climate forecast. By downloading the shapefiles and then generating temperature anomaly maps from reanalysis data, we can visualize how effective the forecast actually was (note: this is not an actual verification). The YouTube video below is a compilation of 30 consecutive 8-14 Day forecasts beginning with the December 31 to January 6th time period and ending with the the February 11th to February 17th time period (no forecasts are issued on the weekends). The GIS maps (left) are placed next to the reanalysis data (right) for comparison. There are a few days where the GIS data may not have been processed correctly (repeating polygons on consecutive days). Nevertheless, it makes for an interesting visualization. With the exception of a few days in mid-January, they 8-14 Day forecasts were actually pretty good.

The Howard Pass Wind Chil and Meteorlogical Records

The discussion here on Deep Cold and elsewhere concerning the wind chill values reported from the Howard Pass RAWS platform over the weekend have been useful in helping to clarify the way the meteor/climatological community thinks about "records" and what constitutes a record.

From an operational viewpoint, "on the fly" data quality evaluation is a routine, every day, indeed every hour part of the job. Doing it correctly though requires a detailed understanding of not only the general meteorological situation but details through the meso and mircoscales, the dominant physical processes, the impact of orography on the variables in question and the details of instrumentation measuring and reporting the physical parameters of interest.

Alaska, with it's complex terrain and high latitude location is an exceptionally difficult arena for meteorological quality control. In Interior Alaska, it's boringly common in mid-winter for temperatures in two locations five miles apart to be 30ºF different, and for such differences to persist for days. How would that fly in Minnesota? Now in this case we happen to have a good physical understanding of how that happens. But temperatures are simple compared to winds.

Details of terrain and instrumentation are even more important for winds than temperatures. Consider the variability of winds but relative uniformity of temperatures in the heavily instrumented areas along the Delta River south of Fort Greely or in the Healy/Denali Park Entrance area. No one who has ever driven through Glitter Gulch in winter in  during a Chinook doubts the 70 mph winds reported at Antler Creek from the RWIS and at the same time the 15 mph or less winds at the AWOS at the Denali Park train station.

So how does all this play into the Howard Pass report? We know that outlets of constricted terrain can be remarkably windy in the appropriate meteorological conditions. We know the details of the instrumentation and know it is research grade equipment. We know it was installed and is maintained by experts in high latitude meteorological instrumentation (thanks Ken and Pam!). We know that the highest resolution numerical models routinely run over Alaska do capture something of the extreme winds and temperatures observed (thanks to Richard for this useful post illustrating this). I made an educated but incomplete scour of observations some years ago for this kind of information and Brian has now systematically checked hourly observations for most of the North Slope sites that might have ever have wind chills in the 90s below.

So, as I see it, to the best of our ability to determine, Howard Pass RAWS did observe the lowest wind chill reported in real-time in Alaska, and I think we can say so with confidence. That does not mean it is certain. And certainly we can argue whether the wind speeds should be adjusted due to the low height. But this is detail: the base observation has verisimilitude, and with no other observations anywhere close (KTZA2 is 67 miles away, and IMYA2, in a completely different physiographic setting, is 42 miles away), the best hope we have for verification would be higher, say 1km, horizontal resolution modeling. Even this will simply make the case more or less solid. There can be no proof.

In the past ten years we've seen a dramatic increase in the amount of quality surface data available in Alaska. but in winter. I expect we'll see more of these kind of "oddball records".

January Post-Mortem

I've been meaning to post a brief analysis of January climate conditions over Alaska, especially in view of what proved to be a very misleading analog forecast for the month.  Here's what I wrote in December:

http://ak-wx.blogspot.com/2013/12/analog-seasonal-outlook.html

"While the uncertainty is great, and we can't have high confidence in this kind of forecast, the analogs suggest a cold January may be in store for Fairbanks."

However, January proved to be very warm over all of Alaska, with a prolonged chinook event in the second half of the month.  Daily mean temperatures were below normal on only four days in Fairbanks.  All-time records for January high temperatures were set in some locations, the state record was tied, and Nome and Barrow set very significant records as discussed on this blog.  Precipitation was well above normal at most locations, with notable exceptions being Fairbanks where the chinook effect prevailed, and Annette Island, which was close to the apex of the anomalous ridge; the third chart below shows the average 500 mb height anomaly for the month.

So what went wrong with the long-range forecast?  Well, in retrospect it was unwise to focus on just a single predictor (2013 conditions in Barrow) rather than also examining potential predictability from other sources.  A major influence on Alaska climate for some months now has been the unusually warm ocean temperatures in the northeast Pacific.  The maps below show (top) the sea surface temperature anomaly for January and (bottom) the mean temperature anomaly between the surface and 100m depth.  These are remarkably large anomalies over a large area and represent a major perturbation in the climate system.  The unusual ocean warmth has developed in tandem with a persistent circulation pattern that keeps bringing warm air to the same areas over the northeast Pacific, thereby reinforcing the ocean anomaly; and of course it's a positive feedback, as the unusual ocean warmth also reinforces the strength of the atmospheric ridge.  The persistent ocean anomaly seems to be at least part of the reason why a similar weather pattern keeps developing near and over Alaska: compare the third chart below, showing the October 500 mb height anomaly, to the January map above.

We can go all the way back to May 2013 and find a very similar ocean temperature pattern, as shown in the next two charts; and of course the summer brought record-breaking warmth to interior Alaska.  The northeast Pacific sea surface anomaly did go away for a time in the early fall, as colder air swept down and erased the surface warmth (third chart below), but a subsurface chart from September shows that there was still anomalous warmth lurking beneath the surface (fourth chart below) and the overall pattern remained unbroken.

In summary, the forecast failure for January was a good lesson that we should not rely too heavily on one analog predictor - especially one without a strong physical basis.  This is not to say that the Barrow analog was without value; but it seems the ocean conditions south of Alaska were a more potent factor and prevented the January climate from following the pattern of the Barrow analog years.

Chena River Ice Thickness

**  Update section added at the bottom of the post on 2/18 **

In two of the last three years I have traveled to Fairbanks in February or March. Whenever I go I always stay at Pike's Waterfront Lodge. One year the Chena River was open and the other year the thickness was only 5". However, upstream at the Steese Highway the thickness was quite healthy. This is apparently a common occurrence. But why?

The simple answer is that a power plant between the two sites puts warm, clear water into the River thereby reducing the ice downstream at Pike's Landing. However, the ice thickness is pretty close early in the season at the two sites. Also, Rick noted that another powerplant is upstream of the Steese Highway that generates power for Fort Wainwright. So, there is probably a simple answer but I don't know what it is. Thoughts?

Fig 1. Ice thickness measurements (dots) 2001 to present for Chena River at Pikes Landing and at Fairbanks (Steese Highway).

Fig 2. Smoothed ice thickness 2001 to present for Chena River at Pikes Landing and at Fairbanks (Steese Highway) using monthly averages.

Fig 3. Google Earth screenshot of the two measurement locations. The GVEA Aurora power plant is also shown.

Fig 4. Picture I took of a van stuck on the Chena River Ice Bridge in March 2012.


* Update section  *

Google Earth has a 'historical photography' utility that I just decided to check out. Fortunately there is some photography from March 9, 2009, that we can look at. The following four screenshots are ordered from east (Steese Highway) to west (Pike's Landing). The effect of the Aurora Plant is quite apparent. What is still unresolved is why the early season ice at Pike's is 'normal' but it deteriorates quickly in February. Presumably the power plant is in full production mode in December but the ice hasn't started to melt out yet. Perhaps the thermal 'front' takes a long time to make it down that distance from when the pkant operations kick into high gear earlier in the winter.

Fig 5. Google Earth historical image from March 9, 2009 (1 of 4).

Fig 6. Google Earth historical image from March 9, 2009 (2 of 4).

Fig 7. Google Earth historical image from March 9, 2009 (3 of 4).

Fig 8. Google Earth historical image from March 9, 2009 (4 of 4).