Remarkable Run of Arctic Outbreaks along the Front Range

[Front Ranger]

Since 2008, the Colorado Front Range region, along with parts of WY to the north, have seen an incredible number of major Arctic cold waves. The period is remarkable not just for the number of events (7), but for the magnitude of cold they produced. This culminated in three top-tier cold waves in 2014, which is pretty much unprecedented.

 

I didn't really realize until I really looked at the numbers just how historic this period has been. I've ranked the cold waves below, with a few highlights illustrating how significant the cold temps were. There were many, many records set that I didn't list. Most of the records are from airports, though I have compared some to previous co-op records as well, depending on what data was available.

 

1. Nov 2014

 

This event blows all the others away. It was easily the greatest November Arctic blast for the region in the airport era, and perhaps only equaled by 1916 before that.

 

Casper:

 

Low of -27 on 11/12. Next earlierst low this cold was 12/5.

 

Cheyenne:

 

The 1/-14 day on 11/12 is tied with 1916 as the coldest day they've seen this early.

 

Laramie:

 

The low of -24 on 11/13 is their coldest temp on record this early. Even more impressively, the high of -1 on 11/12 is the coldest high they've seen in November.

 

Fort Collins:

 

The high of 5 on 11/12 is their coldest this early for any station.

 

Greeley:

 

The low of -13 on 11/13 is the coldest temp this early I can find for the area since 1911. The high of 9 on 11/12 is the coldest high temp for any Greeley station until of high of 8 on 11/16/55.

 

Boulder:

 

The high of 7 on 11/12 is by far their coldest high on record this early. The next earliest is 11/27/1919.

 

Denver:

 

The high of 6 on 11/12 is by far their coldest high this early. Next is 12/5/72. Same goes for the low of -14.

 

2. Oct 2009

 

This is the greatest early fall cold wave for the region in modern records. Most stations had never seen temperatures this cold, this early.

 

Casper:

 

Low of 11 on 10/10 is the coldest temp on record that early. The high of 22 on that same day is also their coldest max temp that early.

 

Cheyenne:

 

High of 17 on 10/10 is by far their coldest max this early. It's not until 10/29 that they've seen a colder high. The low of 10 the same day was the second earliest they've been that cold.

 

Laramie:

 

The high of 20 on 10/10 is their coldest high this early.

 

Fort Collins:

 

The 26/15 day on 10/10 produced both their coldest high and low this early.

 

Greeley:

 

Low of 14 on 10/10 is their coldest temp on record this early.

 

Boulder:

 

High of 29 on 10/11 is their coldest high this early.

 

Denver:

 

Low of 16 on 10/11 is their coldest temp on record this early.

 

3. Dec 2014

 

This event produced the most impressive December temps since 1990 for much of the region, and is probably the third or fourth greatest December Arctic outbreak for the area since 1950.

 

Cheyenne:

 

The record low of -22 on 12/30 was their coldest December low since 1990.

 

Laramie:

 

The -34 seen on 12/31 ties their coldest December temp on record (12/22/89). Dec 1990 produced -33.

 

Fort Collins:

 

The low of -17 on 12/31 was their coldest December temp since 1990.

 

Greeley:

 

Low of -29 at the Greeley airport on 12/31 is the coldest December temp I can find for the area since 1932.

 

Boulder:

 

High of 5 on 12/30 was their coldest December high since 1990.

 

Denver:

 

Low of -19 is their second coldest December temp on record, behind 1990.

 

4. Feb 2011

 

Overall, this was easily the greatest February Arctic event since 1989, which of course was one of the greatest on record. Probably the third or fourth most impressive Feb event since 1950.

 

Cheyenne: 

 

Record low of -24 on 2/2, which equaled how cold they got with the Feb 1989 blast. Only two colder events later in Feb, 1982 and 1936.

 

Laramie:

 

The high of -11 on 2/1 is their coldest max temp for February, beating out the -9 from 1989. The low of -39 on 2/2 is their coldest February temp on record.

 

Boulder:

 

Low of -17 on 2/2 was their coldest February temp since 1989.

 

Denver:

 

The high of -2 on 2/2 was their second coldest high this late, after 1989.

 

5. Feb 2014

 

Ranks right behind Feb 2011 for Feb Arctic outbreaks in the region.

 

Casper:

 

The low of -26 on 2/6 is their third coldest temp this late.

 

Cheyenne:

 

Had a record cold max of -4 on 2/5. Only one event had a colder high later, 1933.

 

Laramie:

 

The high of -5 on 2/5 is their second coldest high this late, after 1989.

 

Greeley:

 

Low of -27 on 2/5 is the coldest Feb temp in the area since 1989.

 

Denver:

 

Low of -19 on 2/5 was the coldest February temp since 1989, and third coldest Feb temp on record.

 

5. Dec 2008

 

In the airport era, probably the second most impressive 12/1-12/19 event, after Dec 1972.

 

Casper: 

 

High of -8 on Dec 14, tied with Dec 1972 for the coldest high on record before 12/20.

 

Greeley:

 

High of 5 on 12/15 was their coldest high this early since 1972.

 

Boulder:

 

The low of -15 on 12/15 was their earliest temp this cold since 1972.

 

Denver:

 

High of 1 on 12/16 is their coldest high this early.

 

6. Dec 2013

 

The second most impressive event for the first week of December, once again behind 1972.

 

Casper:

 

Low of -22 on 12/4. Only one year on record was this cold earlier, 1985.

 

Cheyenne:

 

The 4/-19 day on 12/5 was their third coldest day this early in the season.

 

Laramie:

 

The low of -31 on 12/5 is their second coldest this early, narrowly missing the -33 seen on the same date in 1972.

 

Denver:

 

The -15 on 12/5 tied 1972 for their coldest temp this early.

 

7. Dec 2009

 

Another early December event that produced quite a few daily records, but wasn't as impressive as any of the above events.

 

Casper:

 

Low of -29 on 12/8 is their coldest temp so early in the season, two degrees colder than they got in Dec 1972

[wx_statman]

Pretty impressive run of cold weather for you guys.

 

I do have a bit of an issue with the way you presented October 2009. The October 1969 event was only three days later in the calendar and produced much colder weather in Denver (3 degrees on 10/13/1969 vs. 17 degrees on 10/10/2009). So its not entirely accurate to call October 2009 the "greatest early fall cold wave for the region in modern records."

[wx_statman]

Also, what about the late 1950's-early 1960s?

 

1/4/1959: -28 in Pueblo, second coldest January reading on record

4/10/1959: -6 in Longmont, coldest so late in spring

2/28/1960: -11 in Denver, second coldest so late in winter behind 2/28/1962

3/3/1960: -8 in Denver, coldest March reading since 1886

1/10/1962: -24 in Denver, many all-time record lows in NM

2/28/1962: -18 in Denver, by far the coldest on record so late in winter

1/12/1963: -25 in Denver, many all-time record lows across the SW

[Front Ranger]

Pretty impressive run of cold weather for you guys.

 

I do have a bit of an issue with the way you presented October 2009. The October 1969 event was only three days later in the calendar and produced much colder weather in Denver (3 degrees on 10/13/1969 vs. 17 degrees on 10/10/2009). So its not entirely accurate to call October 2009 the "greatest early fall cold wave for the region in modern records."

 

Well, Oct 2009 produced a level of cold for the region as a whole at an earlier date than any other cold wave. Yes, it was only a few days before 1969, but it was still earlier. And you're talking about a period where the average temps are rapidly declining. Average temp from Oct 5 to Oct 15 drops about 3 degrees across the region.

 

Plus, I was talking about the region as a whole - not just Denver. The coldest high in Cheyenne was 4 degrees colder in 2009 than 1969. In general, the low temps were a little colder in 1969, but the high temps were close to the same or warmer.

 

It's certainly debatable, but like I said 2009 was certainly the earliest fall cold wave with that level of cold across the region.

[wx_statman]

Well, Oct 2009 produced a level of cold for the region as a whole at an earlier date than any other cold wave. Yes, it was only a few days before 1969, but it was still earlier. And you're talking about a period where the average temps are rapidly declining. Average temp from Oct 5 to Oct 15 drops about 3 degrees across the region.

 

Plus, I was talking about the region as a whole - not just Denver. The coldest high in Cheyenne was 4 degrees colder in 2009 than 1969. In general, the low temps were a little colder in 1969, but the high temps were close to the same or warmer.

 

It's certainly debatable, but like I said 2009 was certainly the earliest fall cold wave with that level of cold across the region.

 

I might look at things differently from you in the sense that I view the year in terms of "blocks" of time. To me October 10th and 13th fall within the same block of the calendar. The forcings are the same and the odds of a given weather event happening are the same whether you're talking 10/10 or 10/13. So I would grade the 1969 and 2009 events entirely on their merits. The only exceptions I would make are cases where the calendar rolls over, since monthly records are kind of a big deal. For example, if comparing the October 2002 event versus November 1935. Even then its only a technicality, since mother nature doesn't care that our calendar rolled over.

[Front Ranger]

Also, what about the late 1950's-early 1960s?

 

1/4/1959: -28 in Pueblo, second coldest January reading on record

4/10/1959: -6 in Longmont, coldest so late in spring

2/28/1960: -11 in Denver, second coldest so late in winter behind 2/28/1962

3/3/1960: -8 in Denver, coldest March reading since 1886

1/10/1962: -24 in Denver, many all-time record lows in NM

2/28/1962: -18 in Denver, by far the coldest on record so late in winter

1/12/1963: -25 in Denver, many all-time record lows across the SW

 

That period may very well be comparable for number and severity of cold waves. 

 

However, you are looking at a broader area (Pueblo and NM) with some of those cold waves. I was just looking from Denver to Casper. For example, with the Jan 1959 event, DEN bottomed out at -13 and the coldest day was -1/-8. That wasn't a top-tier cold wave on the same anomalous level for the Front Range as most of the events in recent years, when you look at when it occurred. DEN got down to -14 on 11/12/14. -13 in early January is considerably weaker.

 

That 4/10/59 number for Longmont seems suspect. DEN got down to 7 (still a very impressive cold wave for that late), but even Cheyenne didn't get as cold as -6.

[wx_statman]

That period may very well be comparable for number and severity of cold waves. 

 

However, you are looking at a broader area (Pueblo and NM) with some of those cold waves. I was just looking from Denver to Casper. For example, with the Jan 1959 event, DEN bottomed out at -13 and the coldest day was -1/-8. That wasn't a top-tier cold wave on the same anomalous level for the Front Range as most of the events in recent years, when you look at when it occurred. DEN got down to -14 on 11/12/14. -13 in early January is considerably weaker.

 

That 4/10/59 number for Longmont seems suspect. DEN got down to 7 (still a very impressive cold wave for that late), but even Cheyenne didn't get as cold as -6.

 

I just don't think the last 6 years have been unprecedented, which is what you said. Impressive, but not unprecedented.

[Front Ranger]

I just don't think the last 6 years have been unprecedented, which is what you said. Impressive, but not unprecedented.

 

Fair enough. The late 1950s to early 1960s were certainly comparable in some ways. If not unprecedented, darn close.

 

I would say for early season events (Oct-Dec), there is no doubt recent years reign supreme and were unprecedented. For the Jan-Apr period, late 1950s to early 1960s are definitely tops.

 

Though it's also worth noting, if you want to look at April, there's another event from recent years I left off the list: April 2013. On 4/10/13, DEN hit 6, which beats 1959 as the coldest temp that late in the spring. Cheyenne also had a high of 12 with this event, which is their latest max that cold.

 

It's been a crazy period.

[wx_statman]

Fair enough. The late 1950s to early 1960s were certainly comparable in some ways. If not unprecedented, darn close.

 

I would say for early season events (Oct-Dec), there is no doubt recent years reign supreme and were unprecedented. For the Jan-Apr period, late 1950s to early 1960s are definitely tops.

 

Though it's also worth noting, if you want to look at April, there's another event from recent years I left off the list: April 2013. On 4/10/13, DEN hit 6, which beats 1959 as the coldest temp that late in the spring. Cheyenne also had a high of 12 with this event, which is their latest max that cold.

 

It's been a crazy period.

 

No question. Especially the last 13 months.

[IbrChris]

Late Nov 1896 was pretty darn impressive at least for the PNW into the Northern Rockies.

[IbrChris]

Pretty impressive run of cold weather for you guys.

 

I do have a bit of an issue with the way you presented October 2009. The October 1969 event was only three days later in the calendar and produced much colder weather in Denver (3 degrees on 10/13/1969 vs. 17 degrees on 10/10/2009). So its not entirely accurate to call October 2009 the "greatest early fall cold wave for the region in modern records."

This is one reason you guys need to look at these events in terms of standard deviation rather than calendar days and event A to event B comparisons. Grab the 1981-2010 normals, grab daily data from the airport era and calculate the standard deviations for each calendar day for hi temp and lo temp. A -3 sigma event (-3 standard deviations) is much more impressive than a -2.5 sigma event.

 

Nov 1955 event in the PNW approached -4 sigma...which is incredible.

 

Standard deviations don't follow a neat curve either, they vary by day depending on the historical conditions in the POR on that calendar day. March 1st could have a standard deviation of 8.31 F for high temperature and March 2nd could be 6.76 F (just an example).

 

If you can get a simple .csv with the date and high/low you can easily separate them into calendar days and calculate standard deviation using Excel pivot tables.

[Front Ranger]

Late Nov 1896 was pretty darn impressive at least for the PNW into the Northern Rockies.

 

It doesn't look like that was a major one here. 1919 was probably the most impressive late November event for the Front Range.

[IbrChris]

It doesn't look like that was a major one here. 1919 was probably the most impressive late November event for the Front Range.

And then as we know Dec 1919 brought the hammer down hard on the PNW as well.

[BLI snowman]

Also, what about the late 1950's-early 1960s?

 

1/4/1959: -28 in Pueblo, second coldest January reading on record

4/10/1959: -6 in Longmont, coldest so late in spring

2/28/1960: -11 in Denver, second coldest so late in winter behind 2/28/1962

3/3/1960: -8 in Denver, coldest March reading since 1886

1/10/1962: -24 in Denver, many all-time record lows in NM

2/28/1962: -18 in Denver, by far the coldest on record so late in winter

1/12/1963: -25 in Denver, many all-time record lows across the SW

 

 

 

You could also include November 1959, December 1961, and then stretch it out to 1965 and include December 1964 (more of a northern Rockies event) and the March 1965 behemoth. 

 

I think the 1959-65 period is slightly more impressive than the last 6 years. 1961-62 in particular was just an amazing winter for cold in that region.

 

1989-1993 is fairly comparable as well, with the February 1989, December 1989, December 1990 (x2), October 1991, February 1993, and November 1993 events all being quite impressive in CO/WY.

 

Either way though, as with the PNW it's been a really great stretch for early and late season events in that area and a prolonged run of January mediocrity. Nice to see after the early 2000s stretch.

[Front Ranger]

This is one reason you guys need to look at these events in terms of standard deviation rather than calendar days and event A to event B comparisons. Grab the 1981-2010 normals, grab daily data from the airport era and calculate the standard deviations for each calendar day for hi temp and lo temp. A -3 sigma event (-3 standard deviations) is much more impressive than a -2.5 sigma event.

 

Nov 1955 event in the PNW approached -4 sigma...which is incredible.

 

Standard deviations don't follow a neat curve either, they vary by day depending on the historical conditions in the POR on that calendar day. March 1st could have a standard deviation of 8.31 F for high temperature and March 2nd could be 6.76 F (just an example).

 

If you can get a simple .csv with the date and high/low you can easily separate them into calendar days and calculate standard deviation using Excel pivot tables.

 

I see your point, but...that's kind of a lot of additional work, and I don't think the results would be all that different, at least when it comes to ranking events.

[Front Ranger]

You could also include December 1961 and then stretch it out to 1965 and include December 1964 (more of a northern Rockies event) and the March 1965 behemoth. 1961-62 in particular was just an amazing winter for cold in that area.

 

I think the 1959-65 period is slightly more impressive than the last 6 years.

 

1989-1993 is fairly comparable as well, with the February 1989, December 1989, December 1990 (x2), October 1991, February 1993, and November 1993 events all being quite impressive in CO/WY.

 

Either way though, as with the PNW it's been a really great stretch for early and late season events in that area and a prolonged run of January mediocrity. Nice to see after the early 2000s stretch.

 

Some of those events weren't as impressive for the Front Range, though. Dec 1989 was much more of a WY event (DEN only got down to -6 with that), the second Dec 1990 event was definitely not top-tier here, and the Feb 1993 event, while fairly impressive, was about on par with a second cold wave in Feb 2011 that I didn't even list.

 

2008-14 definitely had more top-tier, region-wide Arctic air than that period.

[BLI snowman]

Some of those events weren't as impressive for the Front Range, though. Dec 1989 was much more of a WY event (DEN only got down to -6 with that), the second Dec 1990 event was definitely not top-tier here, and the Feb 1993 event, while fairly impressive, was about on par with a second cold wave in Feb 2011 that I didn't even list.

 

2008-14 definitely had more top-tier, region-wide Arctic air than that period.

 

Not sure where you're looking (Wunderground?) but Stapleton had a -18 with the December 1989 event. Fort Collins had -22. That event was legit in CO. December 1990 had -1/-18 in Loveland with the 2nd cold shot, seems comparable to what was just seen. 

 

The last few years might be a little more impressive, but it's very comparable IMO. Mostly because the 2/89, 12/90, and 10/91 events were flabbergastingly impressive and then there were several other pretty nice events. 

[Front Ranger]

Not sure where you're looking (Wunderground?) but Stapleton had a -18 with the December 1989 event. Fort Collins had -22. That event was legit in CO. December 1990 had -1/-18 in Loveland with the 2nd cold shot, seems comparable to what was just seen. 

 

The last few years might be a little more impressive, but it's very comparable IMO. Mostly because the 2/89, 12/90, and 10/91 events were flabbergastingly impressive and then there were several other pretty nice events. 

 

Ok, you're right about Dec 1989. However, if you look across the board for the second Dec 1990 event, it really wasn't that impressive. Didn't even get to -10 at DEN. That Loveland reading seems like an outlier, similar to wxstatman's Longmont number for 4/10/59 earlier.

 

The 12/90 and 2/89 events were undoubtedly two of the most impressive Arctic blasts for the region on record, but 11/14 was just as flabbergasting for when it occurred. Overall, if you add in the second Feb 2011 event and the April 2013 event, 2008-14 easily had more significant events. That makes 9.

[BLI snowman]

Ok, you're right about Dec 1989. However, if you look across the board for the second Dec 1990 event, it really wasn't that impressive. Didn't even get to -10 at DEN. That Loveland reading seems like an outlier, similar to wxstatman's Longmont number for 4/10/59 earlier.

 

The 12/90 and 2/89 events were undoubtedly two of the most impressive Arctic blasts for the region on record, but 11/14 was just as flabbergasting for when it occurred. Overall, if you add in the second Feb 2011 event and the April 2013 event, 2008-14 easily had more significant events. That makes 9.

 

Not an outlier. Cheyenne was -6/-16 with the late December 1990 event and Casper was -7/-25. That event was more latitude oriented, but it was just as impressive as what just occurred. 

 

If you're going to add those events then you'd probably need to add March 1989, since it was impressive in WY, -15 in Casper on 3/4.

 

8 significant WY/CO events in 6 years versus 9 in 7 years. Seems pretty comparable, IMO.

[wx_statman]

You could also include November 1959, December 1961, and then stretch it out to 1965 and include December 1964 (more of a northern Rockies event) and the March 1965 behemoth. 

 

I think the 1959-65 period is slightly more impressive than the last 6 years. 1961-62 in particular was just an amazing winter for cold in that region.

 

1989-1993 is fairly comparable as well, with the February 1989, December 1989, December 1990 (x2), October 1991, February 1993, and November 1993 events all being quite impressive in CO/WY.

 

Either way though, as with the PNW it's been a really great stretch for early and late season events in that area and a prolonged run of January mediocrity. Nice to see after the early 2000s stretch.

 

I would also give the nod to that era over 2008-14. We haven't seen any events on the level of January 1962 or January 1963 recently in CO and WY, although NM did it in February 2011.

 

For example Laramie Airport hit -50 on 1/12/1963 and -47 on 1/10/1962, which happen to be their two coldest readings since 1948 (beginning of POR at airport). Denver hit -25 and -24 in those two events while they haven't even hit -20 in recent years.

[wx_statman]

Not an outlier. Cheyenne was -6/-16 with the late December 1990 event and Casper was -7/-25. That event was more latitude oriented, but it was just as impressive as what just occurred. 

 

If you're going to add those events then you'd probably need to add March 1989, since it was impressive in WY, -15 in Casper on 3/4.

 

8 significant WY/CO events in 6 years versus 9 in 7 years. Seems pretty comparable, IMO.

 

Very comparable. This thread is a good example of recency bias on the part of the original poster.

[bainbridgekid]

This is one reason you guys need to look at these events in terms of standard deviation rather than calendar days and event A to event B comparisons. Grab the 1981-2010 normals, grab daily data from the airport era and calculate the standard deviations for each calendar day for hi temp and lo temp. A -3 sigma event (-3 standard deviations) is much more impressive than a -2.5 sigma event.

 

Nov 1955 event in the PNW approached -4 sigma...which is incredible.

 

Standard deviations don't follow a neat curve either, they vary by day depending on the historical conditions in the POR on that calendar day. March 1st could have a standard deviation of 8.31 F for high temperature and March 2nd could be 6.76 F (just an example).

 

If you can get a simple .csv with the date and high/low you can easily separate them into calendar days and calculate standard deviation using Excel pivot tables.

Great post. Definitely more work, but real statistical analysis will always be more meaningful and objective in discussions like this. Comparing standard deviations is an excellent way to compare the relative rareness of different types of events throughout the year.

 

I do take issue with your point about each day having a different standard deviation though. There is no reasonable reason for one day to have a statistically significantly different standard deviation than the next day. That would have to just be coincidence caused by a relatively small sample size. I think it would be more meaningful to use the average standard deviation from several surrounding days than for one specific date for these types of calculations.

[IbrChris]

I see your point, but...that's kind of a lot of additional work, and I don't think the results would be all that different, at least when it comes to ranking events.

You might want to see my thread on Portland that I posted yesterday evening. The results were...quite different than I expected. For instance is a high of 28 on Nov 15th more anomalous than a high of 18 on Dec 15th? You can blurt out raw anomaly values such as "Nov 15th value is a -23 deg anomaly and the Dec 15th value is a -25 deg anomaly, so Dec 15th" but that doesn't take into account the variability in high temperatures during those differing times of year, which standard deviations do. It's much more robust than the back-and-forth that is typical on these forums, and yes it requires some basic statistical knowledge and a little knowledge of spreadsheet applications. Both should be in the repertoire of anyone who considers themselves a climate statistics junkie...just sayin.

 

No pain, no gain.

[IbrChris]

Great post. Definitely more work, but real statistical analysis will always be more meaningful and objective in discussions like this. Comparing standard deviations is an excellent way to compare the relative rareness of different types of events throughout the year.

 

I do take issue with your point about each day having a different standard deviation though. There is no reasonable reason for one day to have a statistically significantly different standard deviation than the next day. That would have to just be coincidence caused by a relatively small sample size. I think it would be more meaningful to use the average standard deviation from several surrounding days than for one specific date for these types of calculations.

This is due to the raw standard deviations for a relatively small sample size...about 74 records per parameter per calendar day in the case of PDX. Even a sample size double this would not produce smoothly curved standard deviation values over the course of the year.

 

I do agree with your averaging idea though...what would be a good interval of days? A month? A month with 30 days would have over 2200 data points represented by the mean standard deviation.

 

To get an idea of the daily standard deviation values for high temps they generally range from 5.5 to 10, highest in spring and fall.

For low temps the daily standard deviation values range from 3 to 8.5 with much more variability in winter.

Also using monthly values makes this a WHOLE lot easier as you can find them for each month and each parameter over at Utah Climate Center.

[IbrChris]

Ya, no the monthly averaging method gives me unrealistic sigmas of in some cases nearly -6...basically impossible. This is because some months of the year have much different variability regimes in early month as opposed to mid or late month. I was seeing much more reasonable values with a daily standard deviation as opposed to a mean monthly one.

The other thing is since arctic events occur so infrequently in the period of record and otherwise temps during the year are fairly tight around the climatological means given it's a maritime climate, the arctic events can't be adequately modeled as part of a normal distribution...they skew it and thus the sigma values aren't as useful as a measure of rarity...ie normal distribution return periods are incorrect (since it's not a normal distribution).

Any thoughts on how to remedy this Bainbridge Kid? Averaging a month doesn't produce realistic sigma values as I noted and averaging smaller sections will still be relatively small sample size. I know meteorologists use standard deviation as a measure of a threshold being achieved...such as a "3 sigma anomaly in 500 mb height"...although 500 mb height has less variance than surface temp so a smaller sample size would behave better.

[Front Ranger]

Very comparable. This thread is a good example of recency bias on the part of the original poster.

 

Not going to argue with you about it. I've been very fair in my assessments and acknowledged fair points from others - I agreed that the late 1950s to early 1960s was comparable. I could just as easily say others have "good ol' days bias".

 

Let's just stick to the stats, statman. Accusations of bias are unfounded and accomplish nothing. Why derail good discussion with stuff like that?

[Front Ranger]

I would also give the nod to that era over 2008-14. We haven't seen any events on the level of January 1962 or January 1963 recently in CO and WY, although NM did it in February 2011.

 

For example Laramie Airport hit -50 on 1/12/1963 and -47 on 1/10/1962, which happen to be their two coldest readings since 1948 (beginning of POR at airport). Denver hit -25 and -24 in those two events while they haven't even hit -20 in recent years.

 

You're measuring things by absolutes. Obviously, cold events in January are going to have more absolute cold.

 

But as far as top 5 regional events for when they occurred, 2008-14 scores at least as high. Nov 2014 was as anomalous as it gets, and at least 4 other events were top 5 for the month in which they occurred.

[Front Ranger]

Not an outlier. Cheyenne was -6/-16 with the late December 1990 event and Casper was -7/-25. That event was more latitude oriented, but it was just as impressive as what just occurred. 

 

If you're going to add those events then you'd probably need to add March 1989, since it was impressive in WY, -15 in Casper on 3/4.

 

8 significant WY/CO events in 6 years versus 9 in 7 years. Seems pretty comparable, IMO.

 

Not top-tier numbers for late December. Cheyenne has been colder with 11 other December events, most coming earlier in the month, obviously. Casper's numbers are more impressive, but there have still been 6 colder December events. 

[Front Ranger]

You might want to see my thread on Portland that I posted yesterday evening. The results were...quite different than I expected. For instance is a high of 28 on Nov 15th more anomalous than a high of 18 on Dec 15th? You can blurt out raw anomaly values such as "Nov 15th value is a -23 deg anomaly and the Dec 15th value is a -25 deg anomaly, so Dec 15th" but that doesn't take into account the variability in high temperatures during those differing times of year, which standard deviations do. It's much more robust than the back-and-forth that is typical on these forums, and yes it requires some basic statistical knowledge and a little knowledge of spreadsheet applications. Both should be in the repertoire of anyone who considers themselves a climate statistics junkie...just sayin.

 

No pain, no gain.

 

Well, you'll notice in my comments for these events I didn't go by anomaly - I gave context to the numbers based on how often other events of a comparable nature had occurred around that time.

[BLI snowman]

Not top-tier numbers for late December. Cheyenne has been colder with 11 other December events, most coming earlier in the month, obviously. Casper's numbers are more impressive, but there have still been 6 colder December events. 

 

Cheyenne was -2/-22 with the recent December event. Casper was -3/-16. Pretty comparable. Not understanding how something like December 2008 is top tier and that 1990 event isn't. The late December 1990 airmass was impressive, particularly in WY. 

[wx_statman]

Not going to argue with you about it. I've been very fair in my assessments and acknowledged fair points from others - I agreed that the late 1950s to early 1960s was comparable. I could just as easily say others have "good ol' days bias".

 

Let's just stick to the stats, statman. Accusations of bias are unfounded and accomplish nothing. Why derail good discussion with stuff like that?

 

It is what it is dude. The original post was interesting and I enjoyed your stats. Thank you for that. But you also made it sound like the last 6 years were some sort of unprecedented barrage of cold waves for the Front Range, which isn't true. Similar stretches occurred twice in just the last 50-60 years.

 

Which begs the question, did you know any of that or did you have to be informed? Would you still have used words like "incredible" and "unprecedented" to describe the recent stretch had you been fully informed of the climate history of your own backyard before you made this thread? Those are the questions I'm left with.

 

Recency bias is the tendency to default to recent experience at the expense of analyzing the past. Its more convenient. Happens all the time. It was easier for you to throw out superlatives about the 2008-14 stretch than to fully explore similar stretches in the past in order to provide a clear perspective of exactly how historic (or not) the last 6 years have been. It took other people to do that.

 

Hence the accusations!

[wx_statman]

You're measuring things by absolutes. Obviously, cold events in January are going to have more absolute cold.

 

But as far as top 5 regional events for when they occurred, 2008-14 scores at least as high. Nov 2014 was as anomalous as it gets, and at least 4 other events were top 5 for the month in which they occurred.

 

Then that begs the question - exactly what is more impressive? An event with absolute minimums in January or an event in November that has to be graded on a curve? Its pretty subjective.

[Front Ranger]

Then that begs the question - exactly what is more impressive? An event with absolute minimums in January or an event in November that has to be graded on a curve? Its pretty subjective.

 

I go by what's most anomalous, as it sounds like Chris does as well.

 

When you're talking about a bunch of events in a short time frame, having a lot of very anomalous events is what is statistically significant, and really what this post was all about.

[Front Ranger]

It is what it is dude. The original post was interesting and I enjoyed your stats. Thank you for that. But you also made it sound like the last 6 years were some sort of unprecedented barrage of cold waves for the Front Range, which isn't true. Similar stretches occurred twice in just the last 50-60 years.

 

Which begs the question, did you know any of that or did you have to be informed? Would you still have used words like "incredible" and "unprecedented" to describe the recent stretch had you been fully informed of the climate history of your own backyard before you made this thread? Those are the questions I'm left with.

 

Recency bias is the tendency to default to recent experience at the expense of analyzing the past. Its more convenient. Happens all the time. It was easier for you to throw out superlatives about the 2008-14 stretch than to fully explore similar stretches in the past in order to provide a clear perspective of exactly how historic (or not) the last 6 years have been. It took other people to do that.

 

Hence the accusations!

 

I already agreed with you that there was arguably one other comparable period on record. Using "unprecedented" might have been overstating. However, I would still give the nod to 2008-14 based on the number of top 5 events for when they occurred: Nov 2014, Oct 2009, Dec 2014, Feb 2011, Feb 2014, and April 2013. I don't believe any other comparable period had that many top 5 events.

 

For someone who calls himself "statman", you sure are spending a lot of time analyzing my superlatives! 

 

As far as the rest, I'm not going to argue with you about who is more familiar with this climate or how much time I've spent looking at local data. I'll just say that your assumptions that I have recency bias because it's "easier" are just plain wrong. And I had to point out to you how unlikely that Longmont number you quoted was. Because of my knowledge of local climate, I know Longmont seeing -6 when DEN only sees 7 is highly suspect. I stuck to certain stations when doing my analysis because I know which ones tend to be the most reliable. I didn't just throw random ones in there that would support a certain conclusion.

 

Again, let's focus on the stats.

[bainbridgekid]

Ya, no the monthly averaging method gives me unrealistic sigmas of in some cases nearly -6...basically impossible. This is because some months of the year have much different variability regimes in early month as opposed to mid or late month. I was seeing much more reasonable values with a daily standard deviation as opposed to a mean monthly one.

 

The other thing is since arctic events occur so infrequently in the period of record and otherwise temps during the year are fairly tight around the climatological means given it's a maritime climate, the arctic events can't be adequately modeled as part of a normal distribution...they skew it and thus the sigma values aren't as useful as a measure of rarity...ie normal distribution return periods are incorrect (since it's not a normal distribution).

 

Any thoughts on how to remedy this Bainbridge Kid? Averaging a month doesn't produce realistic sigma values as I noted and averaging smaller sections will still be relatively small sample size. I know meteorologists use standard deviation as a measure of a threshold being achieved...such as a "3 sigma anomaly in 500 mb height"...although 500 mb height has less variance than surface temp so a smaller sample size would behave better.

I wouldn't use a month. Like you said, true standard deviation values really do change over the course of an entire month.

 

I would say using the 7 surrounding days (3 before and 3 after) would probably give you the most meaningful values. That would give you a sample size around 500 for PDX (7 times 74) which should create a fairly smooth standard deviation distribution and wouldn't include days that were significantly different from the date being analyzed.

 

Increasing the sample size in studies like this is extremely important. Makes the results so much more meaningful and requires much less variable results to remain statistically significant. With a relatively small sample size under 100, you require pretty large effects to remain statistically significant. This makes you miss the many variables that have smaller effect sizes.

[BLI snowman]

I already agreed with you that there was arguably one other comparable period on record. Using "unprecedented" might have been overstating. However, I would still give the nod to 2008-14 based on the number of top 5 events for when they occurred: Nov 2014, Oct 2009, Dec 2014, Feb 2011, Feb 2014, and April 2013. I don't believe any other comparable period had that many top 5 events.

 

 

Again, not sure I would agree. 1959-65 had quite a few top tier events for when they occurred. April 1959, late February 1960, January 1962, late February 1962, January 1963, March 1965. All of those stand way out in your period of record.

[SilverFallsAndrew]

Interesting topic. I was in Denver on Wednesday morning. It was -11 when my plane touched down. 

[bainbridgekid]

Then that begs the question - exactly what is more impressive? An event with absolute minimums in January or an event in November that has to be graded on a curve? Its pretty subjective.

I would argue that grading November and January cold events on a curve is the only fair way to do it.

 

If 0 degrees in November is significantly more rare than -10 in January, then the November event is more impressive.

[Front Ranger]

Again, not sure I would agree. 1959-65 had quite a few top tier events for when they occurred. April 1959, late February 1960, January 1962, late February 1962, January 1963, March 1965. All of those stand way out in your period of record.

 

I agree with you on all of those. Nov 2014 was probably more anomalous than any other single event, but Feb 1962 was close.

 

But I also just realized I overlooked one more remarkable event for 2008-14: the very beginning of May 2013.

 

Cheyenne: low of 9 on 5/2 is the coldest temp on record this late in the spring

 

Denver: low of 19, also by far the latest for a temp in the teens

[Front Ranger]

I would argue that grading November and January cold events on a curve is the only fair way to do it.

 

If 0 degrees in November is significantly more rare than -10 in January, then the November event is more impressive.

 

Yeah, it seems like a no-brainer to me.