Tip:
Highlight text to annotate it
X
the second part of this
fault alone problem is about the science
scientist or of quotes longer time you know he has been made
a lot of time rudy to the data collecting data
results are always going to be close to both sides will be frustrating at times
you'd ever want to take a look at it
another word you want to incorporate
science of twenty warnings
into the wards
that makes sense
phone a step back and take a look at what we've done i went back to
through the polygon warning air which begins october two thousand seven to
present
and these are the kind of numbers that we're looking at
q r c s
eighty jute tote total twenty super cells a hundred twenty eight we had some
tropical in there as well
of the interesting thing here is that
minus able
april two thousand lebanon's in q or c est wednesday are about all are was
super cell
and he has a lot of rhetoric about what how to handle qal cns tornadoes and
specifically and they're usually the weaker time
and i suppose someone that is true
but super cell weaker as well i mean honestly
taking a look at it
app fourth that fires
uh... we've only had five
uh... obviously we had had any for q r c s but
you know that's not a lot of disparity between mother i mean like is that most
tornadoes are going to be
zeros ones and maybe some clues in so
of us has something to keep in mind
about the unborn torn is we've had fifty of them
in uh... five years of landmarks
most of those have been killed c_s_
fourteen to twenty nine and there will be we've pretty much doubled on the
n_-word twenty as a fuel c_s_ is were super cell
how we want to have some on one tropical as well
uh... we've had one fatality
from uh...
etc you will see a story and that was back in nam nay
two thousand eight
so we have had that want to come from
our did it come from you'll see it
take take a closer look at some of these numbers here are the is the cal again q
r c s super cell in tropical
kantor is with the warning
i saw a little bit disparity there
fifty three eighty two of liqueur ceo something we need to do some improvement
there
well but the warnings themselves
every job way out
she died in a sense of what
the false alarm rate really look like her
uh... moto convection here
didn't have as much disparities that thought you know the way you the
rhetoric going around about you are serious you think it's probably double
defaults lol not really thirteen points
so still pissed ojai
on all counts all want to bring one thing that surprised me with the
tropical
false alarm rate way up to ninety one percent so
it's almost like you don't ever hear anything about tropical is almost like
you take those away
numbers just get better uh...
by ten points just
uh... by eliminating tropical but the inot sometimes i just kind of uh...
swap in the road
have appealed the
for each one
a little bit more disparity there between q assist the super cell so we
got our work cut out for us ah...
really
on all three counts to be totally honest with you appeal deeper tropical storm
autos fifty three percent
which i'm not a problem with uh... like totally times we had
uh... speculative personally comments what went on that sunny
uh... were were being pretty much beating the mandate on average
over also
even twelve minutes
twelve minutes for the luke you'll see a stuff is pretty doggone good so
uh... donato money issues will be terminal
looking at some kill csn that's really where we started
because that was getting the most attention of the times what to do but
you will see as a so
that's where the research list to begin with and we took a look at all these
events uh...
one-by-one
but some interesting thing to come up and again we looked at a lot of events
we've had uh... around a hundred nineteen i think it that's not how much
a repeat
some of these
and you could see especially kill c_s_ from going back to especially december
ten two thousand eight
twenty-two of c_s_ warnings
this is a bit number twenty one
within ninety percent false alarm rate
few months later has another twenty eight warnings twenty-eight warning for
the ninety six percent false alarm rate
and because they've been had two tornadoes
one of which didn't
even have a warning with it
uh... makes the event
may six
thirteen more
in the gym bag fifty six of all we have four on one twenty solo lord we got a
lot of work cut out for us
and a super cell as well
but uh... october twenty six twenty super cell warnings
hundred-percent false alarm rate
who want to read it we didn't have to be call and one of the twenty morning so
appealed newseum
say they'll enough to april twenty eleven april fifteenth of april twenty
seven he had to say that your march twenty six
with twenty-one warnings
uh... i think we had
one tornado
so that led to a ninety five percent but what we did get it
but still we had twenty warnings that went without
a tornado
and the day he was probably april eleventh this was a total q r c s event
we put out eleven warnings feels like a tuesday night
we put on eleven warnings
uh... what m ninety one percent also a radio the hundred-percent beauty we had
the one tornado
but all the other ones have alone so i think it's really
lead up to the culmination of uh... people being frustrated with the over
warning problem so we did have a workopolis
so i said to myself well ok can we look at any environmental parameter if you
are so yes that may lead to improve warnings
will it get want
events
we look at the top
can't and then we looked at the bottom tier
so ten killed c_s_ is the din produced
timid or heavy producer we're trying to get as much disparity
code between the two groups
say they were broken out for it and what this week or two that's
savers and warm season
are cold season is basically going to be uh...
november through february
and what we looked at was altered
we want to stick with things that people know so if you look at as p_c_s page
mezzo analysis page that's what you got a fun looked at thirty nine total
parameters and
s_b_c_ does use a lot of love
data lives at the time they did
uh... to look at their gun total parameters and
we look at tempo ral stuff as well so we would six hours before the first
warnings through six hours after the first sweet ride to get some transworld
in a bit about nineteen thousand data points which you know that's that's
pretty sadness
soft pretty happy with the data gathering orgs
when we find well
let's go to to go to product everybody loves this is here to one palmer and
storm relative velocity
and what we found is interesting
special in the cold season and notice that in the cold season
as a share goes up
the higher helicity
you actually turn into non producers
so i want to look for in the cold season in the dead of winter
the more share you have the worst it's going to be for you activity a twenty
exactly opposite launches
so the more sure you have in the warm season the voters yet
andy's color boxes is is the twenty fifty seven eighths's percent on so it's
about the
you know they own advertisment you'd expect
ideas that are really interesting because
one
goal of this research is to break down this paradigm of what we've always
thought and that is
more *** cause more credits
so if you come and and u cn
elicited value zero one jump up three hundred four or five hundred even
getting close to six hundred picking it up
always going to be twenty day because we have a lot of share well
it turns out that may not be the case especially in the dead of winter so
not a bad finding their call
zero three same kind of stuff that they're obviously related
but you did this huge chill or
when you get these none producing events
zero three opposite
and the warm season
so jumps up in the one season looks like you need more cheap
how about the next big goto probably is
surface based kate
turned up cape is uh... especially in the warm season not a very good
predictor of anything lookout piggy
these error bars or
zero to a hundred percentiles are for
warm non producers and the warm season producers
solo artist
we've had a vents not producers
zero to over twenty five hundred so the more cape didn't mean anything
same thing with the warm producers as well
we've had a hundred to say two thousand jewels for killing her and they end up
being producers so in other words these to overlap significantly here
andrew tell you me
courses are thought to be a little bit better job so in other words
if you're down in this two to three hundred range in the cold season looks
like not much is happening you really need to jump it up to about
three eight two seven hundred so
just a little bit cable vietnam from the cold season to get the producers will so
that there's a little bit disparity think about that somewhere in the
twenties season situation
indicate which is directly
titus most unstable k directly tied to service pasted sort of the same thing
not a lot of the spirit in the warm season
cold season you don't get that take uptick
uh... from non producers to produce or so
uh... that was some things to learn their
l c l_ at the condensation level this was very interesting to me because
a lot of time l c l_ refers to
super cells never heard it in context if you will see s is but take a look
not a lot of
attraction in the cold season with warm season you get a huge day
so in other words
d_l_c_ able orators and this is an meters sold six hundred musical s looks
like he did a lot of
production it's above the six fifty are
when i can any producers
so that when a warm season
uh...
deborah something to take a close look at it as we
go through but unifil here a lot of things about what the conversation
levels associate with small as it is something the couple watching
turns out that we may have been missing all along
and then we got the sound and then i found very interesting high and highly
correlated line orientation this end up being the
the most quarrel aidid for amber anything we looked at all clear enough
and i found it interesting because a marathon to think about
what line orientation did too
school
so we have this while i'm here disoriented ad zero two one eighteen
another words is almost north-south
is standing straight up and down
savors his one that's leaned over even a little bit
i call this one thirty degrees to two t
so turned out the further
oriented
northeast the southwest
it is
the less likely it is to produce
so in other words
it standing straight up-and-down nor to set up
expect a lot of tornadoes
the further inflamed over
expect less
amount of activity tornado watch
photo that type of illness uh...
line graph here
so in other words at one eighty right here is when the storm is finished
radovan and look between
once said to be in one ninety one is almost north-south is all of our red
producers
one is tilted over
as we get the more like two tended to fifty
hitting closer to east-west orientation it really drops off big time
now i will be the first to amend this is only came events open so i'll have a lot
of answers but
why that there is no makes it all
okay it is a definite
uh... changeover once you get to a certain angle
who plotted on a roll bar graphs
and another thing that came up
and that is it only work in the warm season i can't explain you why
in the cold season they didn't seem to work
home study that a little bit further but for now
warm producers have this one
eighty to one ninety
or intention sort of work
are basically zero to one eighty
or ten degrees to one ninety you're standing straight up in that
and the first leaned over
you start to get into this non produces a look at a huge gap in between these
two so i think there's some
some products and scientific credibility listeners bet we're going to be
something that we look at the future
hasn't been tested out live
role world stuck it was all looking at the store orbit
quite interesting to see this line or intake and stuff
did make a lot of sense to me so
i went back and i started thinking about it
by the way this is ninety three percent court that's not that's incredible
that's a very high correlation
i started thinking about it but you can't be line orientation someone else
has to be going on it turns out as the zero danny climber sheer vector
and i don't know anybody
in the research paper to anywhere else to look at
zero two eight climbers share vector
for
kill c_s_
spun her what they were thinking about zero one zero three called the coast of
the lower boundary layer
i've never thought look
deep in the layer to see there's something out that in these areas
turns out that
this back if it's blowing into the back of the line at ninety degrees
and i don't care how allies oriented
it has already east-west sanders
vector is blowing into the back of a from them straight north another words
as long as it's angle in between the line
and the vector is that ninety degrees you're in good shape so
in other words
what you get to about really the cutoff was about forty five and higher
did you get up in the nineties here angles around the site
so once we get to a certain place it turned red army's
become producers
at the line of oriented more toward like twenty degrees or more parallel
to pick your old csr the squall line
the closer it is two zero
becomes a non producer
tomorrow
zero e-commerce your orientation going nine degrees behind the line
is much more favorable careers during the war in season remembers only once
the one st
some other words the more i had a right angle it becomes or ninety degrees
a better chance it is for producer
but i have a lot i thought they ship rather than the fact that probably
winds blowing into the back of the storms
allow laufer updraft separation solo order still looks like on on radar with
a squall line
but the winter actually allowing oppressed as house separate what you get
a separate updraft
other storms become much more beds from a rare perspective it's very hard to
pick this out
but i think that's exactly what is happening
uh... what the storms
so no desire to see these little kings like this
all over the place
because of the way the
sheer vectors lined up compared to the squall line itself
very interesting
no reply that when i was well susan again
is not busier a comparable sheer it's the orientation
it is
too
align itself
so that the higher the morgue dangled comes like ninety degrees
uh... you definitely at least producers
and when it's more parallel
to the line itself
uh... you get these warm non producers
coats and it just didn't seem to work out all the study this on a lot further
incidently it'll take us anywhere
uh... as we get more events under her belt
are so that's what we can look at what's that you'll see as though i think we're
getting some
cake but i think that was enough i thought we needed to look at everything
was throw everything into a pot
tropical super cell and you'll see it's and uh... one of the polar bears a
hundred and nineteen when we went back to two thousand five
why do we stop it well
uh... basically whereas pc
middle middle announces that with the database co-stars about two thousand
late two thousand five but
sweater nineteen events this cover two hundred fifty nine twenty dash
anatole warnings were eight hundred and forty one of the same time
so this is kind of all the got out of it first of all the uh...
disparity between
different point of types you hear about this and nick i wonder maybe what it is
for our here's
observes an f_-one certainly dominate uh...
quite a few attitudes entries for uh... just a seven year period
uh... but certainly uh...
the disparity is on that
smaller in as you would expect bfc roseanne ones
jennifer's perimeter we compared all these events too
were surface based care
how they compare
okay this is the number of tornadoes in percentage
over on the left hand side
and it turns out like a is
pretty much worthless when it comes to
uh... tornado occurrence
believe it or not
eighty percent
avar tornado's come with
service based kate fallows and or less
eighty percent
and when you go from thousands two thousand you only jump up about ten
percent see you go from seven am point nine to ninety one point in time
despite handing another thousand
joules per kilogram
so what it is really saying is saying it
you're defending and looking at
tape
surface pasted as your
distinctive parameter for tornadoes you can forget
is not going to happen
we have a lot of tornadoes
and five hundred range
we have a lot of thousand
and we don't have we actually have less as you get higher kissinger
sobha overlay the number of f zero an f_-one corners and again
most of those fall from about twelve hundred
or less
we think well okay
stronger tornadoes not true
same thing with after having read
we had a lot of f f two or greater tornadoes with a thousand last
and i think there's a significant bob after twenty five hundred
and this is all going to april twenty seventh that one day
and allow for the significant bulb out here at the end
at the way that day in this bomb pretty much goes away it's all of our point is
basically live
from about twelve hundred jaws per kilogram earlist
so in other words if you get an updated if it's not zero
he's trying to get around all three hundred five hundred especially the
seven hundred range
you can forget favors the
you don't need to worry about instability you're you're good to go
uh... kind of funny thing too is that
able twenty-seven we talk about that a lot
twenty-five hundred k
april fifteenth we had a major outbreak eight q two guests
looked over a lot
thirty-one toward natives
thirty warren
tornadoes on april fifteenth
what was the cake on april fifteenth
seven hundred
so that's probably where this bomb comes from
seven hundred kate on april fifteenth
with thirty-one tornadoes
so you actually have no chance by looking at cape as long as you have
enough
three hundred four to five hundred
or more you're going to go i'll worry about it anymore
moving along zero one common or shared knowledge about elicited is this year
vector what is the magnitude here
a lot better disparity along with it
so it's looking right once you get over thirty
knots you're good to go
question is what happens
below thirty knots well only sixteen percent of our clients have been thirty
or less
if you have twenty
it's two point three percent of artworks
so looking is looking like you to get into your thirty didn thirty five range
you're good to go there as well
fail when it when i was with that
much as expected
a significant percentage really starts at about thirty
or ten percent
and these are all the week tornados and it really goes up from that
same thing with the uh... number of
a greater d_f_ two twenty items
so it's looking like independently anyway at thirty knots is going to be
uh... i'm nice little cut off
anything less than that
you're getting the less than ten percent
so i got interested and i thought well
this may be where we can start to approve some
and the lower numbers
so change a graph a little bit is the same exact line of desert toward
co-owner share in the hospital bag app put the number of tornadoes we've
actually had not the percentage at each dot
sort forty-two of art when it is a calm with thirty lewis
you get down here to fifteen
not so we haven't had any historical
now overlay taiwan overweight though number of warnings
from the same chart
so distorted started telling me l a big part of the story
so with thirty knots we've had four teacher tornadoes but two hundred and
sixty-eight warnings
commentator fifteen we've had zero twenty is the leadership twenty-five
warnings
twenty-six
with issued
fifty-nine warnings
so what i'm saying to myself that is okay if you walk in and were put
together a game plan of how we want to handle this event and this year is only
twenty knots zero one
i consider myself historically arab two hundred fifty nine tornadoes we've only
had six in our history
which went nuts
so and we had nearly sixty warns
so we can back off just a little bit
same goes for twenty five
and even thirty
so we're going to say is in between these two risers and some significant
improvement that probably should be going on here
if we have that this historical perspective under about
you walk in here
anderson again this is something that you can look at calden species page
pc ever thirty not to share historically we've only had it
forty two tornadoes but a two hundred sixty-eight warnings
ups and maybe we can just ease up a trigger finger a little bit
with the issue and so many mornings
if we know this number is on the lower end
which a lot of times it is on the lauren
now want to get up near forty fifty there's a whole different ball game
things change dramatically
thirty-year less i think we can start to look at some place a little bit more
carefully before
uh... issuing a lot
how l c l_
and again we look at about thirty-nine paramus level i picked up the three
years
l c l_
very interesting and well huge disparity
percentage of toys
really takes a major job at a hundred meters
or less
smells he'll get to a hundred meters alaska
big-time jill
soared nine hundred
you're talking about ten percent of all of our tornadoes
have occurred at nine hundred or above our or higher
healthier
it's almost like you did get into
the quadruple digits five thousand last lecture eleven hundred you can almost
forget it
you walk in a day in the l c l_ is fourteen hundred
there is no way you're going to poison according to today that we looked at his
store
take a look at it or is associated with the weaker ones are all around that
range
especially a thousand or lower
the stronger tornadoes sampanna deu
big spike there about seven hundred so you again you really look at about nine
hundred or so
or less really gets going about eight
a hundred two seven fifty
just below that you're going to go
so did the same thing with the warning so i'd like to put the tornadoes down
here
we have to read total tornadoes one is eleven hundred or higher l c o
only fourteen twenty thousand
how many warnings the same numbers
missus i think we can really do some significant improvements okay
twelve hundred us here we've had three twenty nos
forty-six warning same thing eleven three and forty eight
fallibility
fourteen total tornadoes will found the higher a hundred and thirty-nine
warnings
nine hundred sent a huge jump here and i think we can
to bring this that's ok you walk in the so fasten historically we don't have
that many corners with the files less than ten percent
so we need to bring that one thirteen of them i think that's where we can start
makes and smith again improvements
so i went back and it was ran all the numbers from two thousand seven polygon
morning air
all about whether server that rich p_r_d_ seventy six percent are way up
thirty percent lead time seventeen happen remember at the beginning of the
show and said that
the hypothesis is
that if you try to bring down the falls along that you also that or repeal the
lead time
matter which party implements some of these techniques are described to you
which has been this year
we're starting to see someone else
appealed either up
ninety two percent have rather been present a false alarm decrease with
decreasing behalf
and within hertel economy
seventeen with that is actually though not just to take saw them pretty
impressed
so far with these preliminary results in again
but we've got a lot of work doing a lot of work to study we had even begun to
look at the radar yet
so they're certainly going to be something probably to the radar data
a looking at is well
but just what the techniques of describe a special person virumail stuff huard is
set to see some part of results not able to continue in that direction
and these numbers of two thousand twelve these are kind of numbers i can look
what i can live with
a false alarm rate is forty i think we could do still do a little bit better
beyond that
but it's certainly a significant improvement from what we've had a
all right so with that i will take your questions