Friday, November 14, 2014

U.S. Record Snowfall?


On February 7, 1963, an historic meteorological event happened – or maybe it did not. A remarkable 78” of snow was recorded by the Cooperative observer at the Mile 47 Camp station in southern Alaska. Not only is the 78” of snow in a single day the State of Alaska record, it also exceeds the current United States record. However, much debate exists as to the validity of the Mile 47 Camp record. The National Climate Data Center’s state climate extremes web page lists the Mile 47 Camp value as the Alaska record. However, the national climate extremes web page lists the 75.8” total at Silver Lake, Colorado, in April 1921 as the national record. In the following sections we will describe the climatological setting of the Mile 47 Camp station and the meteorological conditions present on the day of the record snow.

Where is Mile 47 Camp?

The Mile 47 Camp station was located along the Richardson Highway 47 road miles from Valdez, Alaska. Note the past tense description of the station. Mile 47 Camp was only in operation between 1959 and 1964. The station was located at an elevation of 1,250 feet above sea level in the Tiekel River and Stuart Creek valleys of the Chugach Mountains. Figure 1 shows the approximate location of the station, major highways, and all populated places in the area surrounding Prince William Sound. The station was located approximately 21 highway miles northeast of Thompson Pass, which is considered one of the snowiest locations on earth (approximately 535” annually). Thompson Pass is located at mile 26 of the Richardson Highway at an elevation of 2,650 feet above sea level.
Figure 1. Regional map of southern Alaska showing the Mile 47 Camp station location.

Interestingly, there is some confusion as to the exact placement of the station. The National Climate Data Center (NCDC) provides latitude/longitude coordinates for the station – but they are only an approximation as the coordinates place the station too close to the Tiekel River in a flooded area. Local knowledge indicates that the station was located along the west side of the Highway immediately north of the Stuart Creek bridge. Figures 2 and Figure 3 show the station location and a photograph of the site is shown in Figure 4.

The nomenclature for mileage descriptions is to refer to the upper milepost number for all areas between mileposts. For example, something located between mileposts 30 and 31 would be referred to as being at “mile 31.” This terminology suggests that the station was located between mileposts 46 and 47. However, the highway has been modified several times in the intervening years and thus the current mileposts do not necessarily reflect the milepost locations during the 1960s.

The precise station location notwithstanding, the physical settings are similar enough to render the distinction unimportant. The most important attributes of the site are the elevation and the exposure. As noted earlier, the station is located at an elevation of 1,250 feet above sea level. Also, the three-mile wide Tiekel River valley runs east-west for approximately fourteen miles eastward to the Copper River with an elevation drop of 900 vertical feet.

Figure 2. Station location for Mile 47 Camp. Background image from Google Earth (2004).


Figure 3. Station location for Mile 47 Camp. Background image from USGS Earth Explorer (1950).




Figure 4. Representative photograph of Mile 47 Camp location identified as “Actual Station Location” in Figures 2 and 3. This site is between mileposts 45 and 46 immediately north of the Stuart Creek bridge. Photograph by Brian Brettschneider. 

The Observation Form:

Very little information exists about the observations at Mile 47 Camp. The station was part of the Cooperative network of stations across Alaska and the rest of the U.S. To secure a contract for becoming an observer, training was required and quarterly inspections of equipment were undertaken. The observer, Ralph B. Lane, was the Foreman at the Alaska Department of Highway’s Ernestine Camp at mile 62 (15 miles north of Mile 47 Camp). We assume that Mr. Lane was appropriately trained as per standard procedures. 


Figure 5. Scanned Cooperative observer form for Mile 47 Camp in February 1963. Image from NCDC.

An inspection of the form in Figure 6 reveals several interesting details. First, and most obvious, are the 78” of snow on February 7th and the liquid equivalent precipitation of 6.02”. This volume of snow far exceeds anything else during the brief history of the Mile 47 Camp station. In addition, the temperature during the day was no greater than 8°F and the low was -4°F. The only comment recorded for this day was overcast.

A snow of this magnitude implies an airmass capable of holding a lot of moisture. The temperature observed that day was very low and by implication could not hold much moisture. This is a significant strike against a snow of this magnitude. Also, several other days that month saw 4 inches or more of snow but very little liquid equivalent. On the plus side, the daily snow depth corresponds with the observed snowfall for the entire rest of February and the remainder of the season (not shown).

A total of nine days are missing from the monthly observation form. Interestingly, if more than ten days were missing, the monthly data would not have been published and the evaluation of the 78” snowfall observation would be moot.

Synoptic Observations:

Several other stations in the region reported observations that day. Most notably, Thompson Pass at mile 26 also published a daily summary. Figure 7 shows the February 1963 observations for Thompson Pass.


Figure 6. Scanned Cooperative observer form for Thompson Pass in February 1963. Image from NCDC.

Immediately we notice that Thompson Pass received only 2.2” of snow on February 7, 1963, and approximately 14” during the five-day period centered on February 7, 1963. Since Thompson Pass averages over 500” per year and Mile 47 Camp averages 150”-250”, we are tempted to immediately disregard the Mile 47 Camp based only on this comparison. However, if we widen our perspective to the entire Prince William Sound region, we see that much of the area saw significant precipitation – both rain and snow (see Figure 7). The regional observations reveal that the Thompson Pass data is not representative of the surrounding area.

Figure 7. Daily climate summary (max / min / precipitation / snow) for February 7, 1963, based on data from the GHCN v. 3 database. The Copper Valley School station did not report a daily snow amount but their snow depth reading indicates 8" of snow fell.

With the very notable exception of Thompson Pass, heavy precipitation was reported at every coastal station. The stations along the immediate coast were generally above freezing and the precipitation fell as rain. The stations west of Cordova all were below freezing and reported significant snow. A very sharp temperature gradient exists north of Valdez. Several stations reported significant snow (greater than 10 inches) with temperatures below 0°F. 

A strong temperature gradient in this region is not uncommon as the cold, continental air from the Copper River Basin (e.g., Gulkana) is drawn toward lower pressure in the Gulf of Alaska. The cold air is often dammed against the Chugach Mountain divide. Several stations in the region reported observations each hour on the day in question. In Figure 8 we can see the temperatures, weather, and wind conditions observed at 3 p.m. on February 7, 1963.

Figure 8. Observations at 3 p.m. (LST) on February 7, 1963, from the NOAA Integrated Surface Hourly (DS 3505) database.

Stations from Cordova eastward all reported rain with strong easterly winds. Stations west of Cordova and south of the Alaska Range reported snow and northerly winds. The temperature gradient and wind direction implies low pressure in the Gulf of Alaska with a warm or occluded front making landfall along the north Gulf Coast. Indeed, an inspection of the NOAA surface map library for the U.S. on February 7, 1963 (see Figure 9), shows a 976 millibar low pressure at 55°N and 143°W while air pressure values of 1012 millibars are shown in interior Alaska. This pattern is conducive for sending massive amounts of moisture to the areas east and north of the low pressure’s center.


Figure 9. Daily surface map on February 7, 1963, from the NOAA Map Library.

Looking at the upper level profile from Yakutat in Figure 10, temperatures were very warm for the time of year. This is not surprising based on the position of the low pressure identified in Figure 9 and the surface observations shown in Figure 8. The freezing level was greater than 1,500 meters above the surface and the lapse rate indicates general instability. The winds were very strong from the ESE and SE at every level. In fact, at 950 millibars (381 meters) the wind was already blowing at 54 knots and at 800 millibars (1781 meters) the wind reached 85 knots.


Figure 10. Upper air plot from 3 p.m. (LST) on February 7, 1963, at Yakutat, Alaska. Arrows show the direction of the wind.

Looking toward the other side of the low pressure center, we see a hint of the shallowness of the cold air mass that is being eroded from the southeast. Figure 11 shows the upper air temperature and wind from the Anchorage, Alaska, sounding at 3 p.m. LST on February 7, 1963. A sharp, and dramatic inversion is clearly present only a few hundred meters above the surface. In fact, the temperature is above freezing at 2,000'.

By early February, the 3 p.m. sun angle in Anchorage is high enough above to horizon to fully erode any valley inversion that might be present. Therefore, the inversion is indicative of an airmass change above the surface. The 20 knot northeast surface wind also lends credence to the inversion being airmass driven. The Fairbanks sounding (not shown) saw a similar, dramatic airmass inversion. Their temperature increased 21°C between the surface and 1,600 meters with a wind shift from northeast to southeast.


Figure 11. Upper air plot from 3 p.m. (LST) on February 7, 1963, at Anchorage, Alaska. Arrows show the direction of the wind.

Ernestine Camp:

The Alaska Department of Transportation & Public Facilities (ADOT&PF) has several maintenance camps along the Richardson Highway from Valdez to Glennallen. Each camp is responsible for a defined stretch of highway. The area of responsibility for each station has moved over time but currently the Ernestine Camp, located at mile 62, is responsible for the area between mile 42 and mile 82. This includes the location where the Mile 47 Camp station was located. The highway south of mile 42 falls under the responsibility of the Thompson Pass maintenance camp.  

A Cooperative station was located at Ernestine Camp from 1965 to 1977. Unfortunately the frequency of observations was not very consistent. However, the staff at the Camp has kept meticulous snow and road condition records during the October through April time period from 1981 through the present. The author of this blog post visited the station in October 2014 and discussed snow climatology with the Ernestine Camp Foreman. In addition, the weather records since 1981 were made available for inspection. The following section contains some of the informal observations from the Camp Foreman:
  • The area around Mile 47 Camp is generally warmer than the rest of the stretch of highway Ernestine Camp is responsible for (note: the Mile 47 Camp is the lowest elevation on the Richardson Highway for over 100 miles)
  • The amount of snow usually, but not always, increases as you head to the southern portion of their area of responsibility, including Mile 47 Camp.
  • “Crazy things” can happen in the Stuart Creek (Mile 47) part of the highway. Winds can howl through the Tiekel valley coming from the east and southeast off the Copper River. The result is often surprising snow totals in this region.
  • Usually Thompson Pass gets the most snow but sometimes the Ernestine area of responsibility will get 2’-3’ of snow and Thompson Pass barely receives any snow.
  • If he had to pick a spot where a highly out of the ordinary snow would occur, he would choose either the area around Mile 47 Camp or the area at the head of Ernestine Creek (near Ernestine Camp).
Analog Events:

If similar events to the one in question are identified, we will feel much more comfortable about the validity of the February 7, 1963, snowfall. Specifically, we want to answer three questions with the analog analysis. 
  1. Have very large snowfalls been observed in the Ernestine Camp area of responsibility? 
  2. How common are heavy snowfall events concurrent with very cold temperatures (0°F or less)? 
  3. Have large snows been recorded in the Ernestine Camp area of responsibility but not at Thompson Pass?
As stated earlier, the staff at the Ernestine Camp have collected daily snowfall data continuously since 1981. These data are unpublished and reside in the Ernestine Camp office. The Camp Foreman generously allowed the records to be inspected for this analysis. 

Analog Question #1:

Have very large snowfalls occurred north of Thompson Pass? The answer to this question is Yes. Several very large snowfalls were recorded at the Ernestine camp during the thirty-three years for which daily snow records exist. Two snowfalls produced in excess of 40”. On March 16, 1995, the Ernestine Camp reported 52” of snow. Unfortunately no temperature observations were made that day but based on the temperatures on the days before and after the storm, the temperature may have been below zero for much of the snow event. On March 15, 1995, the high temperature was +1°F and the low was -19°F. On March 17, 1995, the low was -10°F and the high was +34°F. The handwritten form from March 1995 is shown in Figure 12. The 52” snow is clearly visible. The back of the form (not shown) has a monthly plot of snow that confirms the 52” notation on the front of the form. A comparison with Thompson Pass indicates this was not an isolated event. Figure 13 shows the unpublished March 1995 daily observation form for Thompson Pass courtesy of the Alaska Department of Transportation & Public Facilities staff at the Thompson Pass Camp. The form shows 51” fell on March 15, 1995, and an additional 46” fell on March 16, 1995 (in the column with the heading "N/S"). For this event, we see very large snowfall amounts reported on all sections of the highway – including Thompson Pass and Ernestine Camp.


Figure 12. March 1995 weather and road conditions form for Ernestine Camp.


Figure 13. March 1995 weather and road conditions form for Thompson Pass. The observations were continued from the previous page (not shown) so the date was not noted for each line. A printed transcription of this form (also not shown) confirms the dates and the snowfall amounts.

Analog Question #2:

Have heavy snowfalls occurred with very cold temperatures? The answer to this question is also Yes. A surprising number of snowfalls in the 6” to 20” range have occurred when the high temperature was below 0°F. This is quite uncommon for nearly any place in the world. An example of a heavy, cold snow at Ernestine Camp is shown in Figure 14. The temperature on December 14, 1981, did not exceed -10°F. That same day, 30” of new snow was reported. Similarly, many, many large snows are reported at Thompson Pass with temperatures below zero. While this is unusual in most places, it is not at all uncommon along the Richardson Highway. We suspect that cold, continental air frequently resides in the lowest several thousand feet while very warm and moist air rides over the top of the cold airmass and the precipitation falls through this cold layer. If the cold air is continuously replenished, airmass modification due to vertical thermal advection need not occur. 


Figure 14. Portion of the Ernestine Camp observation form from December 1981. 

Analog Question #3:

Finally, have heavy snows occurred at Ernestine Camp but not at Thompson Pass? The anecdotal evidence stated earlier indicates that it happens occasionally. Looking through the Ernestine Camp and Thompson Pass records, we found a close analogy on January 7, 2001. On that date, the Ernestine Camp form shows 40” to 60” falling (see Figure 15). We infer from our conversation with the Camp Foreman that the 60” total represents the southern portion of the highway that Ernestine Camp is responsible for (the Mile 47 Camp region). Looking at the Thompson Pass form (see Figure 16), only 10” of snow fell on January 7, 2001 (16” more fell the following day). While the snow at Thompson Pass was significant, it paled in comparison to the amounts in the Ernestine Camp portion of the highway. We therefore feel confident that large snow events can occur in the Mile 47 Camp region independent of what is observed at Thompson Pass.


Figure 15. January 2001 weather and road conditions form for Ernestine Camp.



Figure 16. Portion of January 2001 weather and road conditions form for Thompson Pass.

The Observer:

Very little is known about the observer, Mr. Ralph B. Lane. We know that he was the Ernestine Camp Foreman and that he lived in Copper Center. We checked with NCDC but they did not have any information about him; even if they did have information they are prevented from disclosing it. We were provided a document by the current Alaska DOT avalanche forecaster at the Thompson Pass Camp that concerned the observations by Mr. Lane. In the letter, dated May 23, 1963, the Weather Bureau State Climatologist is concerned about a rain gauge that he found tipped over at the site as well as the observation time difference between temperatures and precipitation. His primary concern appeared to be the number of missing days and even discussed moving the station because of this issue. The letter is shown in Figures 17 and 18 below.

A reference is made to the quality of a May 9th observation in the letter but no mention is made of any other data quality problems. Presumably the State Climatologist had seen the 78" snowfall measurement by this time. This letter would have been the perfect opportunity to list a series of data quality problems if the State Climatologist was especially concerned about his observations.


Figure 17. Page 1 of a letter from the Weather Bureau State Climatologist, C. E. Watson, to Mr. Ralph B. Lane dated May 23, 1963.


Figure 18. Page 2 of a letter from the Weather Bureau State Climatologist, C. E. Watson, to Mr. Ralph B. Lane dated May 23, 1963.

Site Visit with Mr. Lane:

While researching this record, I spoke with a professional meteorologist, Richard Lomire, who spent an entire day with the Mile 47 Camp observer, Ralph B. Lane, and the Thompson Pass observer from the early 1950s, John R. Hagberg, during the mid 1990s. The purpose of the 1990s visit was to discuss the snow climatology of the region. The following bullet points describe some of the highlights of  the February 1963 snow event that were relayed to Mr. Lomire nearly 20 years ago.
  • Mr. Lane remembered the event as if it were yesterday.
  • The snowflakes were as large as silver dollars.
  • The wind was perfectly calm.
  • He had never seen snow that deep before.
  • Mr Lane was caught in, and survived, an avalanche on that day while in his snowplow several miles south of Mile 47 Camp (note: there is an avalanche path at mile 42).
  • There was a measuring board near the camp that he would stop by on his plowing route and take snow measurements.
  • It was typical for snow totals to be largest at that location.
  • Mr. Lane wrote a description of the event and gave it to Mr. Lowrie. Unfortunately, it is in a box, in a storage building, in another state at this time.
  • Mr. Lane retired to Baja California, Mexico, due to health reasons and has since passed away.

A Case For and Against the Observation:

A proposed rationale supporting the observation:

If the observation is to be believed, here is a possible scenario: a very strong, moist flow of the intense storm (see Figure 10) funneled moisture over and though the fourteen-mile long Tiekel River valley where it was lifted 1,000’ for an extended period of time while cold air was advected southward at the surface. The Yakutat sounding from 12 hours before the one depicted in Figure 10 (not shown) indicates similarly warm temperatures and strong winds at all levels originating from the same direction. Large flakes, as relayed by Mr. Lowrie, promote prodigious accumulation rates when the dendrites are not broken. The calm winds at the surface strengthen the likelihood of this possibility. Finally, the fact that Mr. Lane was caught in an avalanche during this event is indicative of a very heavy snow event. Presumably he was aware of the avalanche path and being caught in the avalanche might attest to the exceptional quality of the event.

Overall, the monthly snow total for February 1963 and the snow depths are not unrealistic based on the monthly forms provided by the current Ernestine Camp Foreman. 

Problems with the observation:

There are a number of questionable aspects to the observation. First, the extremity of the event itself is a significant red flag. One would expect that a comment or some other notation would be made on the form about the event, as was done on the Silver Lake, CO, form from April 1921, but no comment was written.

An inspection of the form shows that several snow events in excess of 4" yielded unrealistically low liquid equivalent snow amounts. In fact, three snows that measured 4" all reported 0.02" of liquid equivalent. An 8" snow reported 0.04" liquid equivalent and a 2" snow was shown to be 0.01" liquid equivalent. In each case, the snow ratio was 200:1. This is clearly not possible. Perhaps he guessed after the fact or he did not melt the snow properly. Also, if the snow was measured every time he passed the Camp and the snow board was cleared each time, the minimum separation of snow observations may not be present –  which would have the effect of inflating the total.

Several of the markings on the form appear in a different shade of ink. The snow depth on the 1st through the 6th, the 16th, 17th, 24th, and 25th are all in a lighter ink. These observations were likely filled-in all at the same time at the end of the month. While this may be problematic, it may also represent a transferring of notes from an informal form to the official form.

Conclusion:

The goal of this project was to determine if the remarkable 78” snow total from Mile 47 Camp was possible. Based on our analysis, and the conversation relayed by Mr. Lowrie, the unequivocal answer is Yes, it is well within the realm of possibility. This is not the same as declaring the observation to be correct. But, despite the magnitude of the event, there is enough circumstantial evidence to assume the observation is plausible.

Is the Mile 47 Camp value a U.S. record? Well, it exceeds any other “official” value ever measured. Somewhat confusingly, NCDC considers the Mile 47 Camp observation the Alaska record but not the national record – even though it is larger than the acknowledged national record. In fact, the GHCN database flags the value with an “X” indicating that it failed a bounds test. Given the short period of record for the station and the extremeness of the event, this is not surprising.

We searched newspaper records for Valdez, Copper Center, Glennallen, Anchorage, and Fairbanks for a reference to the storm but none was found. Not a single narrative of the storm appears to exist anywhere. However, if the extreme snowfall amount was localized to a remote section of highway, the lack of notoriety is not unexpected.

In theory, it is possible to model this event using the Weather Research and Forecasting (WFR) model that is available to the public. A nested grid of increasingly detailed topography may allow for a simulation of the event. However, the input for the model would be limited to the excellent, but coarse, ESRL reanalysis data and this is unlikely to achieve the desired results. An ensemble of model runs with slightly different initial conditions may improve the model output. This is an area for future research.

The Alaska Region National Weather Service (NWS) Office has reviewed the Mile 47 Camp record during the last 12 months and did not take any action to remove the value from the record books. After reviewing the data and summarizing it in this narrative, we thank the NWS for conducting due diligence and not disqualifying the record.

We respectfully request that the National Climate Extremes Committee evaluate the validity of the observation.

Postscript: This blog post is an informal version of a manuscript in preparation for publication. The journal and publication date is unknown at this time.