Our region's extreme run

[PrinceFrederickWx]

2 hours ago, Fozz said:

I did a Python analysis on Baltimore's precip totals, just cause I felt like it, and it turns out that 2018 single-handedly wrecked the normal distribution. Precip totals are now skewed to the right. Sometimes one extreme outlier is all it takes.

Baltimore's final total of 71.82" is over 4 standard deviations above the long-term pre-2018 average.

Doesn’t NWS remove or smooth over extreme outliers when calculating averages though? I thought I remember reading they did something to Dec 09 and Feb 03 when calculating 1981-2010 snow averages, for example.

[Fozz]

4 hours ago, PrinceFrederickWx said:

Doesn’t NWS remove or smooth over extreme outliers when calculating averages though? I thought I remember reading they did something to Dec 09 and Feb 03 when calculating 1981-2010 snow averages, for example.

I'm not sure of their exact methods. I guess it would make sense to do that with their 30 year averages since the sample size isn't so big. I know they reduced the totals of Dec 2009 and Feb 2010 a few months after that winter, probably because of measurement issues or some other reason I can't clearly recall.

[RodneyS]

7 hours ago, Fozz said:

I'm not sure of their exact methods. I guess it would make sense to do that with their 30 year averages since the sample size isn't so big. I know they reduced the totals of Dec 2009 and Feb 2010 a few months after that winter, probably because of measurement issues or some other reason I can't clearly recall.

I can state with a high degree of confidence that the NWS does not smooth outliers in its 1981-2010 monthly and annual precipitation "normals" but rather uses the 1981-2010 averages, albeit with an unconventional rounding methodology.  I state this because I have computed the 1981-2010 averages for each month at DCA, and they either match the DCA normals exactly for each month or they are rounded up by 0.01 inches. For example, if you average the January 1981-2010 precipitation totals at DCA, you get 2.803 inches.  For some reason, NWS rounds this to 2.81, whereas using a conventional rounding methodology would result in 2.80.  The annual 1981-2010 precipitation normal at DCA is shown as 39.74 inches, whereas  I calculate 39.71 inches. However, there is no smoothing.  So, for example, the August 1981-2010 normal at DCA is only 2.93 inches, which is rounded up from the 1981-2010 2.924 inch average.  However, the historical August 1871-2018 DC precipitation average is 4.07 inches, and that is not taken into account. So, the 1981-2010 DCA August normal makes it appear that August is now a relatively dry month at DCA.  But is it really, or was the 1981-2010 August average there an aberration?  I would guess the latter, because beginning in 2011, August precipitation at DCA has averaged 3.77 inches -- closer to the long-term average than to the 1981-2010 average.  

In other words, you have to be careful with monthly precipitation normals because they may have little predictive value.   Regarding annual precipitation normals, 30-year annual precipitation averages have ranged between 38.62 inches during 1961-1990 to 43.01 inches during 1871-2010, with an 1871-2018 average of 40.92 inches.  So, again, using a 30-year annual precipitation "normal" can be misleading, as the most recent 30-years may be too short a period to conclude that anything other than a random variation is occurring. 

[Fozz]

22 minutes ago, RodneyS said:

I can state with a high degree of confidence that the NWS does not smooth outliers in its 1981-2010 monthly and annual precipitation "normals" but rather uses the 1981-2010 averages, albeit with an unconventional rounding methodology.  I state this because I have computed the 1981-2010 averages for each month at DCA, and they either match the DCA normals exactly for each month or they are rounded up by 0.01 inches. For example, if you average the January 1981-2010 precipitation totals at DCA, you get 2.803 inches.  For some reason, NWS rounds this to 2.81, whereas using a conventional rounding methodology would result in 2.80.  The annual 1981-2010 precipitation normal at DCA is shown as 39.74 inches, whereas  I calculate 39.71 inches. However, there is no smoothing.  So, for example, the August 1981-2010 normal at DCA is only 2.93 inches, which is rounded up from the 1981-2010 2.924 inch average.  However, the historical August 1871-2018 DC precipitation average is 4.07 inches, and that is not taken into account. So, the 1981-2010 DCA August normal makes it appear that August is now a relatively dry month at DCA.  But is it really, or was the 1981-2010 August average there an aberration?  I would guess the latter, because beginning in 2011, August precipitation at DCA has averaged 3.77 inches -- closer to the long-term average than to the 1981-2010 average.  

In other words, you have to be careful with monthly precipitation normals because they may have little predictive value.   Regarding annual precipitation normals, 30-year annual precipitation averages have ranged between 38.62 inches during 1961-1990 to 43.01 inches during 1871-2010, with an 1871-2018 average of 40.92 inches.  So, again, using a 30-year annual precipitation "normal" can be misleading, as the most recent 30-years may be too short a period to conclude that anything other than a random variation is occurring. 

I agree, which is why I ran the numbers using the entire period from 1871 to 2017. I'm not all that comfortable with a sample size of only 30 if I can do better.

[cae]

On ‎1‎/‎1‎/‎2019 at 6:20 AM, Fozz said:

I did a Python analysis on Baltimore's precip totals, just cause I felt like it, and it turns out that 2018 single-handedly wrecked the normal distribution. Precip totals are now skewed to the right. Sometimes one extreme outlier is all it takes.

Baltimore's final total of 71.82" is over 4 standard deviations above the long-term pre-2018 average.

From what I can tell, nobody knows the appropriate probability distribution for precipitation data.  It appears to vary by region.  The normal distribution isn't widely used because it allows for negative precipitation values and underestimates the probability of large totals.  The gamma distribution is widely used, but it too can underestimate the probability of extreme events. 

[Fozz]

17 minutes ago, cae said:

From what I can tell, nobody knows the appropriate probability distribution for precipitation data.  It appears to vary by region.  The normal distribution isn't widely used because it allows for negative precipitation values and underestimates the probability of large totals.  The gamma distribution is widely used, but it too can underestimate the probability of extreme events. 

I did a Shapiro-Wilk test on the precip data through 2017, and it confirmed a normal distribution (or close enough to it). But with 2018 included, that was no longer the case.

[RodneyS]

This morning both DCA and IAD recorded their daily high temperatures.  However, DCA's high was 56 degrees vs only 40 at IAD.  As best as I can tell, that 16-degree high temperature difference at the two airports was the second greatest ever, behind only December 9, 1975, when DCA recorded a high of 53 and IAD recorded a high of 36.

[MN Transplant]

13 minutes ago, RodneyS said:

This morning both DCA and IAD recorded their daily high temperatures.  However, DCA's high was 56 degrees vs only 40 at IAD.  As best as I can tell, that 16-degree high temperature difference at the two airports was the second greatest ever, behind only December 9, 1975, when DCA recorded a high of 53 and IAD recorded a high of 36.

I’m surprised we haven’t had more days like this.  Nice catch.

[RodneyS]

On 1/14/2018 at 10:17 AM, RodneyS said:

According to my records, yesterday's 41-degree temperature spread at DCA (62 maximum, 21 minimum) was the greatest since December 22, 1998 (67/24).

And the January 20, 2019 38-degree temperature spread at DCA (56 maximum, 18 minimum) was the greatest since the January 14, 2018 41-degree spread.

[RodneyS]

On 1/1/2019 at 8:08 AM, RodneyS said:

The final 2018 figures are 66.74 inches at IAD and 66.28 inches at DCA.  The IAD total broke the 2003 record by 1.07 inches and the DCA total broke the 1889 record by 4.95 inches. 

DCA just recorded its first 1-inch or more precipitation day of 2019, after a record-breaking 24 such days last year.  More remarkably, the amount of precipitation there since July 1, 2018 is now more than 44 inches.  The previous record for any July-March period was only 43.81 inches, set during July 1934-March 1935.  So, we have exceeded that total with more than two months still to go. 

[RodneyS]

On 1/7/2018 at 10:29 AM, RodneyS said:

Dulles fell below zero this morning for the first time since February 24, 2015.  This morning's reading of -1 broke the record for January 7th of +1 that was set in 2014.

This morning, Dulles fell below zero for the first time since January 7, 2018.  However, this morning's reading of -2 fell short of the January 31st record of -7 set in 1965.

[WxUSAF]

BWI set a record low max today of 21.

[RodneyS]

Today's high temperature of 74 at DCA was 50 degrees warmer than the high of 24 recorded there just four days ago.  As best as I can tell, that 4-day warm-up of 50 degrees in the DC maximum was second only to the 52-degree 4-day warm-up in maximum temperatures that occurred from February 16, 1930 (24) to February 20, 1930 (76).

[PrinceFrederickWx]

Not sure if @RodneyS had noticed this, but for DCA, the 10.3" recorded from 1/12-14 this year would make #7 on the list of top 3-day snowfalls, per their criteria (i.e. snowfall must have occurred on all three days). This would be on the second chart, not the first one that lists top snowfalls overall.

https://www.weather.gov/lwx/winter_storm-pr#TopDaySnowfall

[RodneyS]

5 hours ago, PrinceFrederickWx said:

Not sure if @RodneyS had noticed this, but for DCA, the 10.3" recorded from 1/12-14 this year would make #7 on the list of top 3-day snowfalls, per their criteria (i.e. snowfall must have occurred on all three days). This would be on the second chart, not the first one that lists top snowfalls overall.

https://www.weather.gov/lwx/winter_storm-pr#TopDaySnowfall

I had not noticed that, but had noticed that overall that snowstorm tied for 23rd on the all-time DC list (with February 22-23, 1987).  Also, the 10.6 inches received at Dulles in that storm was 16th all-time there, displacing the 10.3 inches that fell at Dulles during the infamous (poorly-forecast) 10.3-inch snowstorm of January 25, 2000. 

[RodneyS]

April 2019 averaged 62.4 degrees at DCA, second all-time in DC to the 63.8 recorded in April 2017.  However, precipitation was only 2.24 inches -- the first below normal month at DCA since October 2018.

[Subtropics]

On 5/1/2019 at 12:28 PM, RodneyS said:

April 2019 averaged 62.4 degrees at DCA, second all-time in DC to the 63.8 recorded in April 2017.  However, precipitation was only 2.24 inches -- the first below normal month at DCA since October 2018.

Two historically warm years out of 3, with last year being historically chilly if I remember correctly? Doesn't seem like DC has "average" months all that often in the spring.

[RodneyS]

16 hours ago, Subtropics said:

Two historically warm years out of 3, with last year being historically chilly if I remember correctly? Doesn't seem like DC has "average" months all that often in the spring.

You're right that last April was relatively cool at DCA -- 54.9 degree average, which was 1.9 degrees below the 1981-2010 normal, and 7.5 degrees cooler than this year. 

[RodneyS]

On 1/24/2019 at 2:16 PM, RodneyS said:

DCA just recorded its first 1-inch or more precipitation day of 2019, after a record-breaking 24 such days last year.  More remarkably, the amount of precipitation there since July 1, 2018 is now more than 44 inches.  The previous record for any July-March period was only 43.81 inches, set during July 1934-March 1935.  So, we have exceeded that total with more than two months still to go. 

Another DC precipitation record has just fallen.  The previous DC record precipitation from the second half of one year to the first half of the following year was 62.16 inches, set during July 1885 to June 1886.  With 0.62 inches already recorded at DCA tonight, the July 2018-June 2019 total there is 62.53 inches and counting. 

[PrinceFrederickWx]

BWI is currently tied for 10th hottest July (and 10th hottest month). That’s probably going to be an outright win for 10th (or possibly even a tie for 9th) after today.

Four of BWI’s top ten hottest July’s (and top ten hottest months) have occurred in this decade.

[PrinceFrederickWx]

September finished tied for 5th hottest and also 2nd driest at BWI. 

This is the decade of hot Septembers at BWI- 2018 is now tied for 11th hottest and 2016 is now ranked 16th hottest. It will be interesting to see how the 1991-2020 norms affect September, as there as only been one this decade with below average temps- 2013, which was still only -0.1.

@RodneyS or @MN Transplant how did DCA/IAD do?

[Winter Wizard]

6 hours ago, PrinceFrederickWx said:

September finished tied for 5th hottest and also 2nd driest at BWI. 

This is the decade of hot Septembers at BWI- 2018 is now tied for 11th hottest and 2016 is now ranked 16th hottest. It will be interesting to see how the 1991-2020 norms affect September, as there as only been one this decade with below average temps- 2013, which was still only -0.1.

@RodneyS or @MN Transplant how did DCA/IAD do?

DCA: +5.5 and 0.25", good for 2nd warmest and 3rd driest
IAD: +4.6 and 0.41", good for 3rd warmest and 2nd driest

[RodneyS]

1 hour ago, Winter Wizard said:

DCA: +5.5 and 0.25", good for 2nd warmest and 3rd driest
IAD: +4.6 and 0.41", good for 3rd warmest and 2nd driest

Minor correction for DCA:  According to the Capital Weather Gang, September 2019 was the 3rd warmest and 4th driest September in DC history; see https://www.washingtonpost.com/weather/2019/10/01/washington-just-posted-its-rd-warmest-th-driest-september-record

[yoda]

From today's AFD re September rainfall records:

Quote

Here are the rankings for driest Septembers on record.

Washington, DC area (Ronald Reagan Washington National Airport, DCA)
1. 0.11 inches (2005)
2. 0.14 inches (1884)
3. 0.20 inches (1967)
4. 0.25 inches (2019)
5. 0.32 inches (1977)
1981-2010 normal monthly rainfall for September: 3.72 inches

Note: Weather records for the Washington, DC area have been kept at
what is now Ronald Reagan Washington National Airport (DCA) since
1941. The official weather record consists of that data back to
1945, and additional precipitation records observed in downtown
Washington, DC extend the period of record back to 1871.

Baltimore MD area (Baltimore-Washington International Thurgood
Marshall Airport, BWI)
1. 0.09 inches (1884)
2. 0.16 inches (2019)
3. 0.21 inches (1967)
4. 0.32 inches (1906)
5. 0.35 inches (2007)
1981-2010 normal monthly rainfall for September: 4.03 inches

Note: Weather records for the Baltimore, MD area have been kept at
what is now Baltimore-Washington International Thurgood Marshall
Airport (BWI) since 1950. Additional precipitation records observed
in downtown Baltimore extend the period of record back to 1871.

Sterling/Dulles VA area (Washington Dulles International Airport,
IAD)
1. 0.15 inches (2005)
2. 0.41 inches (2019)
3. 0.62 inches (1967)
4. 0.78 inches (1978)
5. 0.93 inches (2014)
1981-2010 normal monthly rainfall for September: 3.92 inches

Note: Weather records for the Sterling/Dulles VA area have been kept
at what is now Washington Dulles International Airport (IAD) since
1960.

[RodneyS]

October 2, 2019 is in the record books for the hottest October maximum in DC, Baltimore, and Dulles.  The respective maximums today were 98, 98, and 96. The previous record for DC and Baltimore was set on October 5, 1941, with respective maximums of 96 and 97.  The previous record for Dulles was set October 9, 2007, at 94. 

[PrinceFrederickWx]

Tied for 11th hottest October at BWI.

[RodneyS]

On 2/28/2018 at 6:55 PM, RodneyS said:

February 2018 at DCA was not as warm as last year's record-setter of 47.7 degrees, but its 45.3 degree average was good enough for 3rd place all-time in DC, behind only last year and February 1976 (46.9).  The warm finish to the meteorological winter brought the 3-month average to 40.0 degrees, with a true 90-day average of 39.9.  Each average was 3.9 degrees below last year's respective averages, but still warm enough for the 29th all-time warmest meteorological winter in DC.

February 2020 averaged 43.8 degrees at DCA, with the meteorological winter there at 42.8 degrees, each 7th on the all-time DC list, although February 1984 also averaged 43.8.

[RodneyS]

On 3/20/2012 at 12:42 PM, RodneyS said:

 

Eight years ago today, I posted on this thread -- http://www.americanw...h-on-record-dca: "The final numbers for the 2011/12 astronomical and composite winters are 45.9 and 45.6 degrees, respectively. Thus, the record for the warmest astronomical winter was broken by 2.2 degrees and the record for the warmest composite winter was broken by 1.6 degrees. When the astronomical winter warm record was last broken (in 1989-90), it was only by 0.2 degrees, and when the composite winter warm record was last broken (in 2001-02), it was only by 0.9 degrees."  

This season, the astronomical winter (December 21-March 20, on average) averaged 45.5 at DCA , and so it fell only 0.4 degrees short of 2011/12.  And the composite winter (December 1-March 20, on average) averaged 44.6 at DCA, and so it fell 1.0 degree short of 2011/12.  Both averages are good enough for second warmest in DC history.

 

 

[RodneyS]

On 3/20/2020 at 9:52 PM, RodneyS said:

Eight years ago today, I posted on this thread -- http://www.americanw...h-on-record-dca: "The final numbers for the 2011/12 astronomical and composite winters are 45.9 and 45.6 degrees, respectively. Thus, the record for the warmest astronomical winter was broken by 2.2 degrees and the record for the warmest composite winter was broken by 1.6 degrees. When the astronomical winter warm record was last broken (in 1989-90), it was only by 0.2 degrees, and when the composite winter warm record was last broken (in 2001-02), it was only by 0.9 degrees."  

This season, the astronomical winter (December 21-March 20, on average) averaged 45.5 at DCA , and so it fell only 0.4 degrees short of 2011/12.  And the composite winter (December 1-March 20, on average) averaged 44.6 at DCA, and so it fell 1.0 degree short of 2011/12.  Both averages are good enough for second warmest in DC history.

March 2020 averaged 53.2 degrees at DCA, good enough for 5th on the all-time warmest March DC list, with March 2012 warmest at 56.8.  Further, 1Q 2020 averaged 46.5 at DCA, which was second on the all-time warmest 1Q DC list, with 2012 again warmest, at 47.3.  Moreover, for the 12 months ending March 31st, April 2019-March 2020 set the all-time record at 61.7 degrees, edging out April 2011-March 2012, which averaged 61.6.

 

 

 

[RodneyS]

I just posted this on the Capital Weather Gang webpage, in response to Ian Livingston's article at https://www.washingtonpost.com/weather/2020/04/28/pm-update-some-more-showers-tonight-much-warmer-wednesday

Today marked the 15th consecutive day at DCA without a 70+ degree reading.  The only other April 14-28 period where that happened in DC weather history was 1875.  However, as Ian notes, the current run appears over, as tomorrow is forecast to be at least 70.  In 1875, the run ended one day later, on April 30th, with a reading of 72.  Still, there is a big difference between these two Aprils.  April 1875 averaged only 48.0 degrees, the second coldest April in DC history, whereas this April has averaged 54.8 so far, only slightly below the 1871-2019 historical April average of 55.3.  By the way, 1874 was the coldest April in DC history, at 47.1, and the only April day that reached 70+ that year was the 15th, at 72 degrees.