Early Season October / November Snows. Be Careful For What You Wish For?

[Rainshadow]

There are lies, gosh darn lies, and statistics, so read this into it what you want.  First, something is happening with the xmacis site, I can’t go back prior to 1940 on the redirected northeast climate center site for PHL.  So, this may be a good or bad thing.  This post is based on 79 winters at PHL International since 1940.

Snowfall is not a normal distribution; it is skewed to the right. Based on the current normal of 22.4” for PHL, since 1940-41, 48 seasons have had below normal snowfall, 31 above.  A 61%/39% split.  If one uses the current 1990-2019 running average for the past three decades, it is getting even more feast/famine/skewed: 19 below, 11 above the running 30-year average of 23.8”.  In spite of the differences the snowfall median (half above/half below) in both samples has remained between 18” and 19.5”.

Throwing normals/averages around gets confusing, so I am going to stick with the current 22.4” normal.

First October.  If it snows (even a trace) in October, generally speaking it looks like it doesn’t matter.  Six of the ensuing seasons were below average, five were above average, pretty close to climo distribution. 

However, delving into the October weeds further, snow occurred in five Octobers in which the ensuing winters had a moderate or strong el nino and four of them had above current normal snowfall.  1972-73 the notable exception.   The remaining it snowed in October seasons were five below, two above the seasonal average.

Onto November and using a measurable threshold. Off the top, it doesn’t look that different from snowfall climo, 10 ensuing seasons below, 8 ensuing seasons above.  If one were to subtract the November snowfall from the seasonal totals (or use the current 30 year running average), it becomes 13 below / 5 above.   The four snowiest Novembers had below average snowfall the ensuing cold season months, three were outright below average even including their November snowfall.  Among the five exceptions, one can say three are in play as a possible enso outcome for this winter: 1961-2, 1977-8 & 1978-9.

Three of the snowiest seasons in which it did measurably snow in November had that snow occur late in the month (27th-29th).  There were some late in the month not so good measurables in there also, so there is no calendar day threshold to sound the all clear.  Here is the table:

 

[PreserveJon]

As always, great statistical analysis Tony.  Thank you.

[susqushawn]

Good stuff thanks!  Quells my anecdotal concerns

[colonel_kurtz]

nice Tony

get this message when trying to access NOWdata

Due to fiscal year contract processing restrictions, NOWData will be unavailable for a short period. In the meantime, please use the National Centers for Environmental Information's (NCEI) Climate Data Online (CDO) resource. If you are not automatically redirected there within 15 seconds or have pop-ups blocked, please go to https://www.ncdc.noaa.gov/cdo-web.

Here's the entire snowfall period of record:

 

 

[Rainshadow]

11 hours ago, colonel_kurtz said:

nice Tony

get this message when trying to access NOWdata

Due to fiscal year contract processing restrictions, NOWData will be unavailable for a short period. In the meantime, please use the National Centers for Environmental Information's (NCEI) Climate Data Online (CDO) resource. If you are not automatically redirected there within 15 seconds or have pop-ups blocked, please go to https://www.ncdc.noaa.gov/cdo-web.

Here's the entire snowfall period of record:

 

 

Thanks Carl.   Prior to 1940 it was much more common for November snows to not matter going forward.  Eyeballing looks like 60 below/ 40 above which is close to snowfall climo.  Even so 5 of the 6 snowiest Novembers still had a below average snowfall season. 

[susqushawn]

On 10/18/2019 at 11:56 AM, Rainshadow said:

There are lies, gosh darn lies, and statistics, so read this into it what you want.  First, something is happening with the xmacis site, I can’t go back prior to 1940 on the redirected northeast climate center site for PHL.  So, this may be a good or bad thing.  This post is based on 79 winters at PHL International since 1940.

Snowfall is not a normal distribution; it is skewed to the right. Based on the current normal of 22.4” for PHL, since 1940-41, 48 seasons have had below normal snowfall, 31 above.  A 61%/39% split.  If one uses the current 1990-2019 running average for the past three decades, it is getting even more feast/famine/skewed: 19 below, 11 above the running 30-year average of 23.8”.  In spite of the differences the snowfall median (half above/half below) in both samples has remained between 18” and 19.5”.

Throwing normals/averages around gets confusing, so I am going to stick with the current 22.4” normal.

First October.  If it snows (even a trace) in October, generally speaking it looks like it doesn’t matter.  Six of the ensuing seasons were below average, five were above average, pretty close to climo distribution. 

However, delving into the October weeds further, snow occurred in five Octobers in which the ensuing winters had a moderate or strong el nino and four of them had above current normal snowfall.  1972-73 the notable exception.   The remaining it snowed in October seasons were five below, two above the seasonal average.

Onto November and using a measurable threshold. Off the top, it doesn’t look that different from snowfall climo, 10 ensuing seasons below, 8 ensuing seasons above.  If one were to subtract the November snowfall from the seasonal totals (or use the current 30 year running average), it becomes 13 below / 5 above.   The four snowiest Novembers had below average snowfall the ensuing cold season months, three were outright below average even including their November snowfall.  Among the five exceptions, one can say three are in play as a possible enso outcome for this winter: 1961-2, 1977-8 & 1978-9.

Three of the snowiest seasons in which it did measurably snow in November had that snow occur late in the month (27th-29th).  There were some late in the month not so good measurables in there also, so there is no calendar day threshold to sound the all clear.  Here is the table:

 

That July 2009 trace was a doozy!  Thanks again for putting this together and dont mind my sarcasm