Severe Plains

Totos Revenge Model Snow Forecasts

11:06 am on February 24th, 2013
Categories: Winter Weather

Here’s an stream of the latest model forecasts for the upcoming winter storm which will affect Kansas and Missouri on January 25-26, 2013.  The most recent is at the top.  For the latest please follow on Twitter (@SeverePlains) and this will be updated as quickly as possible once all model images are available.

Model Comparison 6pm Sunday, February 24, 2013

Model Comparison 6am Sunday, February 24, 2013

Model Comparison 6pm Saturday, February 23, 2013

Model Comparison 6am Saturday, February 23, 2013

Model Comparison 6pm Friday, February 21, 2013

No Comments »

Totos Revenge – Feb. 25-26 Winter Storm

8:19 am on February 24th, 2013
Categories: Winter Weather

Last Wednesday-Thursday a winter storm ravaged areas of Kansas and Missouri, bringing up to 16 inches of snow to some areas and brought the Kansas City Metro to a stop Thursday morning. Now, just 4 days later another winter storm is expected to strike and this one being as bad if not worse then the last. A few quick statements on the upcoming storm…

For Kansas City the snow will likely begin on Monday evening, meaning the Monday evening commute has the potential to be a very hazardous one.  Nearly everyone will likely be at work or in school so please be aware of the snow and when it is expected to start.  Take off a couple of hours early to avoid the rush and beat the snow home if needed.
Regardless, the heaviest snow at this point is expected Monday night with the potential for 8-14 inches and even some higher amounts based upon current model data.  This much snow while still have 8 inches on the ground will likely create big problems in attempting to clear snow, especially in residential areas.  If these totals come to fruition then it could easily be days before some streets are cleared.

This storm will not only bring copious snow, but wind speeds will likely be higher with ~20 mph sustained winds and the potential for gusts near 35 mph throughout the event.  While the last storm also had similar wind speeds they were only for a couple of hours.  This storm will bring gusty winds throughout the event which may lead to significant blowing and drifting.

Just how rare would it be for Kansas City to see two 8+ inch storms within a week?  Here’s some neat snow facts…

Kansas City’s 9.2″ of snow on February 21 was the 5th heaviest daily snow on February record.  It was also the heaviest snow since February 25, 1993.

While the event may span two days a majority of snow may fall on Tuesday, another calendar day record is not out of the question.

The February monthly records for snow is 20.7″ in 1960  – We are currently at 9.5″, meaning an 11.2″ event at KCI would tie 2013 for the snowiest February on record; 11.3″ or higher would break it obviously.

To the best of my knowledge there has never been two snow events within a week of each other that were both more than 7 inches.  The closet being January 9-20, 2011 when there was a 7.5″ event (Jan. 9-10) and a 7.2″ event (Jan. 19-20).

(hat tip to Eric Holthaus for the Totos Revenge name)

No Comments »

Week 9 BCS Rankings: Humans vs Computers

9:35 am on October 22nd, 2012
Categories: General Information

Last week I began a project that I hope to follow through with, comparing the human voters to the computer algorithms that help determine the BCS Standings.  You can see some background and last week’s data here.  Otherwise, lets dig right in with the teams from the Harris Poll and Coaches Poll which are determined to be over- or under-rated by the unbiased computers (positive numbers indicate human over-rated by negative numbers are vice-versa).

Harris Poll

Nevada  +21.00
Clemson  +13.25
Florida State  +11.75
TCU  +9.75
Duke  +8.50
Louisiana-Monroe  +8.25
USC  +7.75
Cinicinnati  +7.50

Iowa State  -14.00
Texas Tech  -7.50
West Virginia  -6.25

(Also note that Arizona is rated #23 by the computers, but is un-ranked in Harris making for a differential of at least -19.00)

Coaches Poll

Clemson  +13.25
Florida State  +11.75
Louisiana-Monroe  +11.25
Duke  +10.50
USC  +8.75
TCU  +7.75
Cincinnati  +7.50

Texas Tech  -7.50
Stanford  -7.00
West Virginia  -6.25

(Note that both Arizona and Iowa State are un-ranked by the Coaches Poll while ranked #23 and #27 respectively by computer average.  Leading to differentials of at least -13.00 for Iowa State and -17.00 for Arizona)

One additional thing to throw in this week is average differential for both polls.  Essentially how close are the human polls to that of the computers -

Harris Poll Average Differential:  4.75
Coaches Poll Average Differential:  4.47

1 Comment »

BCS Rankings: Humans vs. Computers

5:24 pm on October 15th, 2012
Categories: General Information

There has been plenty of complaints about the BCS computer rankings and how they can’t always choose the top 2 team correctly.  At least that is how it appears in the eyes of a majority of college football fans.  However, who can really say that an unbiased computer calculation, or six of them for that matter, are really wrong?  Humans, whether they want to admit it or not, have a bias when it comes to rankings.  Either you were previously affiliated with a college or conference, like the underdog, or just don’t have enough knowledge of every team to make an unbiased ranking.  The first BCS rankings of the season were released last night, which is calculated by an average between the Harris Poll (human), Coaches Poll (human) and the BCS Computer average.  The computer average takes the 6 computer rankings, drops the highest and lowest, then takes the average of the 4 middle rankings.

I’ve done this in the past, but this go around I thought I’d share it with the rest of you.  How do these initial BCS rankings appear when it is the two human polls compared to the computer average.  I simply took the rank of the Harris and Coaches polls and compared to the Computer Average to see which teams are over-rated or under-rated in the eyes of the unbiased computers. Take a look, positive numbers mean they are over-rated by the humans while negative numbers indicate they are under-rated by the humans:

Harris Poll Extremes:

Florida State  20.250
Virginia Tech  19.750
Clemson  18.500
Oklahoma State  16.750
Louisiana-Monroe  16.000
North Carolina State  11.500

Iowa State  -13.250
Texas Tech  -12.500
Western Kentucky  -8.250
Tulsa  -6.250
Texas A&M  -5.750

Coaches Poll Extremes:

Louisiana-Monroe  21.000
Clemson  18.500
Florida State  18.250
Oklahoma State  17.750

Iowa State  -17.250
Texas Tech  -11.500
Tulsa  -8.250
Stanford  -7.750
Texas A&M  -5.750

What do you think?  Is there a better way to get rid of bias?  Should the computers have more pull than just a third of the rankings?

2 Comments »

Kansas City Snowfall Probabilities

6:24 am on September 24th, 2012
Categories: General Information, Winter Weather

Meteorological summer ended on August 31 and on Saturday we had the Autumnal Equinox, marking the official start of Fall for Kansas City and the United States for that matter.  The colder air is already showing itself with low temperatures well into the 30s across parts of Kansas and Missouri Sunday morning.  Colder air will continue to gradually work south and the usual rain (or lack thereof this year) will turn to snow!  The extended range models agree with this idea, pushing strong arctic cold fronts into the Plains in early October.  In the meantime though, I’ve gathered up a few snowfall facts for Kansas City as well as some daily probability graphs for snowfall in KC too for your enjoyment.

Earliest Snowfall  -  October 13, 2001 (Trace)

Earliest Measurable Snowfall  -  October 15, 1898 (3.3 inches)

Latest First Snowfall  -  Winter of 1933-34 on January 21, 1934 (0.6 inches)

Latest First Measurable Snowfall  -  Winter of 1933-34 on January 21, 1934 (0.6 inches)

Latest Snowfall  -  May 3, 1907 (1.7 inches)

Latest Measurable Snowfall  -  May 3, 1907 (1.7 inches)

Largest Single Day Snowfall  -  March 23, 1912 (20.5 inches)

Now onto some graphs, each graph contains the daily probability based upon data from 1889-2011 (123 years) as well as a 7 day moving average for a ‘smoothed’ line of these probabilities.

A few noteworthy details given by the graphics, January 25th and 26th are the days most likely to have snow with 25 of 123 years recording at least a trace of snowfall.  Those two dates obviously fall around the peak of when snow is most likely in Kansas City for the winter season, late January into early February.  There are three days which share the distinction of most likely for accumulating snowfall, January 26th, February 4th and February 21st all have a 19% chance of snow (23 of 123 days).  Looking for those bigger snowfall amounts? February 28th has had the most snowfalls of an inch or greater since 1889; December 24th and February 23rd both have had the most snowfalls of 3 inches or greater.

Daily Probability for Snow

Daily Probability for Measurable Snow

Daily Probability for 1″ or More Snow

Daily Probability for 3″ or More Snow

There is a lot more you can discern from the graphs if you look at them closely.  Have a question about the data or a comment, leave it here or send it to me on Twitter @SeverePlains.

No Comments »

Struggling Shower

7:52 pm on September 3rd, 2012
Categories: Photography

Going to try and include more photography to keep the updates more frequent. Here is the sunset tonight as a shower struggles to the west of Gardner, Kansas.

No Comments »

Kansas City: Summer of 2012

7:11 pm on September 1st, 2012
Categories: General Information, Weather Recap

What many will remember as the hottest and driest summer of their lives came to an end yesterday and for quite a few the numbers support that claim.  Kansas City finished out the meteorological summer (June 1 – August 31) with the 6th direst and 16th warmest on the record books.  It could have been much worse, but some August rain and cool spells provided a quick drop to those rankings which were both in the top 10.  In fact, the last day of August was able to tally up an additional 0.89 inches of rain which took Summer 2012 from 5th to 6th.  I’ve posted a scatterplot of precipitation vs temperature in the past and figured it is a good way to show just how warm and dry it was:

Kansas City Summer 2012 Temperatures & Precipitation Scatterplot

In the scatterplot above, 2012 is red to help it stand out from the others.  We definitely weren’t the extreme in temperatures that many probably thought it felt like on those hot afternoons and a look into the numbers shows some interesting information.  The average high temperature this summer in Kansas City was a warm 92.68 degrees, the average low temperature this summer was 66.96 degrees.  While those numbers alone may not mean much, their rankings show the story.  The average high of 92.68 degrees ranks 5th warmest on record behind 1936, 1934, 1954 and 1901.  The average low of 66.96 degrees ranks only 75th warmest on record!  The dry conditions that we had this summer allowed overnight lows to drop significantly when combined with calm winds and clear skies.

Take a look at this table below which shows the largest spreads between average high and low temperatures during the summer months:

Only 1936 had a larger differential between the high and low temperatures than 2012, also of note is that 1936 was the warmest summer in terms of average highs and despite the differential was 3rd warmest in terms of average lows!  Another little stat for you, this summer featured 61 days with a high of 90 degrees or warmer which ties for fifth most all-time behind 1936 (73), 1934 (70), 1901 (65), 1988 (62) and tied with 1954.

In the end,  if you remember the Summer of 2012 in Kansas City as hot and dry, the numbers certainly back up your statement.

No Comments »

Smoky Skies

2:27 pm on August 18th, 2012
Categories: General Information, Photography

Wildfires continue across the Inter-mountain West, which continue to spew smoke into the atmosphere as they burn largely out of control. With the current weather pattern, this smoke that has been lofted up into the atmosphere is being pushed down into the Western High Plains and nearby areas. Hard to believe, but areas even as far away as Kansas City are seeing the affects of this smoke. The smoke on the horizon is providing very dull, but colorful sunsets as shown below.

Sunset near Kansas City – August 17, 2012

Sunset near Kansas City – August 16, 2012

No Comments »

Kansas City: How Hot & How Dry

7:48 am on July 25th, 2012
Categories: General Information, Weather Recap

There has been plenty of discussion and comparison to how hot and dry it has been not only in Kansas City, but across much of the nation this summer and for the entire year.  Through June, 2012 has been the hottest year on record (118 years) for the contiguous United States with 28 states also recording their hottest year to date on record.  Those include Kansas, Missouri, Oklahoma, Nebraska, Iowa and South Dakota among others.  To make the heat worse, it has been dry, according to the NCDC it is the 18th driest year on record through June for the contiguous United States.  Missouri and Nebraska were both ranked 11th driest, Kansas ranked 27th driest.  Colorado, the scene of many wildfires, has had it’s second warmest and second driest year on record.  The Central Plains and Midwest aren’t alone though, the East Coast is in a similar boat with Delaware chiming in at their warmest and driest first six months of the year on record!

Let’s dig into some specifics though, this for Kansas City where records used date back through 1889, how does this year compare to the dust bowl years and many more that have seen drought and heat.  Here are some quick stats on Kansas City’s year to date:

And now where each of those numbers ranks when comparing it to the previous 123 years of data for Kansas City.  It is important to note a couple of facts here, the location of observed data has obviously changed over the years along with its’ surroundings.  From 1888 to 1938 the observations were taken from locations that were in the downtown area, then from 1938 to 1972 they were taken at the Downtown Airport located along the Missouri River.  Since 1972 the data has been taken from the Kansas City International Airport which is located around 14 miles northwest of the Downtown Airport and is largely surrounded by more rural landscape rather than city buildings and streets.

As you can tell by the rankings, no matter which time period you take it is in the top ten regardless.  Data taken from April 1 through July 24 is the most impressive, ranking third warmest and second driest since 1889.  To give you a neat graphical representation of the data, scatter plots for each time period are below.  I’ve highlighted 2012 in red font for easier finding, obviously the more normal years are cluttered together pretty well.  The plots have total precipitation as the y-axis with average temperature in the x-axis, so the warmer and drier it is the closer to the bottom-right corner the year will be.  The opposite is of course true for the cooler and wetter years, which will then be located closer to the upper-left.  Take a look below (click to see larger image):

No Comments »

Period of Records

9:36 am on July 7th, 2012
Categories: General Forecast, General Information, Weather Recap

It has been obvious to everyone across the Central and Eastern United States for the past couple of weeks that it has been hot, on occasions oppressively hot.  After a series of storms across the Midwest to the East Coast that left hundreds of thousands to millions without power for days the heat was likely unbearable.  Dozens of deaths have been contributed to this heat wave that has obviously been described as record-breaking.  Perhaps some have even titled it as the ‘worst-ever’.  News stories across the nation have been using the numbers courtesy of NCDC and NOAA to describe just how hot it has been, with over 23,000 daily high temperature records broken this year through July 5.  However, there is an important caveat to remember and that is a variable known as period of record.  The period of record indicates how long that station has been in operation, and NCDC requires a station to have a period of record (POR) of at least 30 years to be included in their records database.  But, do these stations with only 30-50 years of data skew the records?

In my opinion, yes.  Sure there have been record-setting temperatures in the recent decades for stations that date back to 1900, but on the same token some of the hottest summers’ occurred during the 1950s and 1930s.  For an example, take a look at Kansas City who period of record dates back to 1889, here are the top 10 warmest summers:

1) 1934
2) 1936
3) 1954
4) 1901
5) 1953
6) 1913
7) 1918
8) 1952
9) 1937
10) 1980

Only one of those (1980) have occurred within the past 58 years!  This will be an important factor when I dig into some numbers below. But first, another detail that can skew these record numbers.  The NCDC records database is based upon the exact station, for example Kansas City International Airport is the current official observation station for Kansas City.  While records from the local NWS climate report date back to 1889, the NCDC records database only goes back 39 years for Kansas City as that is the time in which the airport become an observation station.  This is the case in many areas where the official observing station has switched locations and taken on an new identifier, the period of record in the NCDC database does not incorporate an entire record such as one you would find in a local climate report.  The NCDC records database indicated that in June 2012 that Kansas City International Airport had 5 daily record high temperatures. How many of those were actual records with the actual data that dates back to 1889?  Zero!

Let’s dig into some of the data and numbers that you may have seen thrown around lately.  In June 2012 there were 3,282 daily high temperature records either broken or tied across the United States.  Those 3,282 records account for 1.9% of the records that were possible, in other words there were still 168,160 daily high temperature records that remained unbroken.  While that may be significant, I don’t have the research done to conclude whether that 1.9% is indeed statistically significant or not.  The main focus is the period of record, so of these 3,282 records how many of them had a period of record that reaches back through the 1930s?  How  many of them don’t even touch the 1950s?

Only 697 of the 3,282 came from a station with a period of record dating back at least 80 years; that is only 21.2% of the daily high temperature records that include the dust bowl era.  Of those 3,282 there were a whopping 1,331 of them that were from a station with 50 years or less of records under their belt; that accounts for 40.5% of the daily high temperatures which don’t include the 1930s or even 1950s!  One can’t say how many of those 1,331 records would have still been records if data was available through the 1930s, but one can reasonably presume that a majority probably would not make the cut.  The same discrepancy can be found as you head into the first five day of July with daily high temperature records, with 1,570 either broken or tied of a possible 29,022.  While that does equate to a solid 5.4% of daily high temperature records broken out of the possible only 324 (20.6%) have a period of record of at least 80 years.  And once again, there are more stations that have a period of record of 50 years or less, 569 or 36.2% than those with the 80+ years of record.

In the end, I’m not saying the recent heat wave wasn’t really that hot or that it wasn’t record breaking. I’m simply informing you that you need to take into consideration other variable before spitting out raw numbers and a little research is always a good thing!

No Comments »