TechStuff Classic: Weather Tech, Part 2

How are weather predictions made? How does the tech used in meteorology work? And can we control the weather?

See omnystudio.com/listener for privacy information.

Episode Transcript

Speaker 1 (00:04):
Welcome to tech Stuff, a production from iHeartRadio. Hey there,
and welcome to tech Stuff. I'm your host, Jonathan Strickland.
I'm an executive producer with iHeartRadio and how the tech
are you. It's Friday, so it's time for another classic episode.
This is the second part of a two parter called

(00:26):
weather Tech Part two. So if you didn't listen to
last week's classic episode of weather Tech Part one, I
recommend you do that. That would have been last Friday.
This episode originally published on April twenty first, two sixteen. Enjoy.
We are now going to join the podcast already in progress.

(00:48):
We were able to gather a lot of information once
we had those basic tools available to us. But what
really pushed meteorology forward is when we could stop relying
upon the data that we can gather here on the
ground and supplement that with information from the atmosphere itself.

(01:08):
And that brings us to weather balloons. Weather balloons more
than just fodder for your roswell conspiracy right right, Swamp
gas and weather balloons and weather balloons do more than
just act as a subplot in an X Files episode. Right, Yes,
They're very important. Yeah, so they carry instrumentation that collects

(01:29):
data about atmospheric conditions and weather balloons have been around
for a long time, but more recently they typically carry
instruments called radiosond, which is a battery powered device that
can measure altitude, atmospheric pressure, temperature, humidity, wind speed. Sometimes
there's a GPS element to it, so it can so

(01:51):
people on the ground can track where the weather balloon is.
Normally the you tether these devices. You don't just release
a weather balloon and say sea, but sometimes you know
you need to have that GPS element there too. And
getting this information from the atmosphere is really important because
it can tell you about how conditions may soon change
on the ground. It's pretty interesting actually to ever if

(02:16):
you've ever had a chance to look at some of
the data pulled from these, because you see how different
conditions in the atmosphere are compared to what we experience here,
including some pretty intense winds at higher altitudes. So we've
got all this information being collected, and before we get

(02:38):
into space, because that'll be the next step outward, I
wanted to talk a little bit about what we do
with all that data. One of the things we do
is we create databases that have all this information. So
that let's say that we have a day with pretty
nice weather, we collect all the information about that. What
was the atmospheric pressure, what is the temperature, how much

(03:01):
humidity was in the air, what was the wind speed,
were there any higher low pressure systems nearby? Were the
what kind of front had just moved through? All this
sort of information, we feed it all into a database.
Collecting that over and over and over again allows us
to build a better virtual understanding of how weather works, right,

(03:22):
and we can supplement that with more information as we
learn more about the weather. Then we would end up
using that to help us make some predictions about how
weather might be in the future. And to do that,
really we use computer models. Typically we would build what
was called a numerical weather prediction model, the NWP. So

(03:49):
this is really a model that's made up of a
bunch of different calculations that take all of the different
variables into account and tell you, based upon all the
variables available to us, here's what it looks like the
weather is going to be like in X amount of time. Right, So,
whenever we're talking about forecasts, obviously we have to worry

(04:10):
about what are the current conditions and how far out
are we trying to predict the weather? And on TV
you might see five or seven or even these days
sometimes ten. Yeah. Like if you go to weather dot com,
they have a ten day forecast, which I always think
is hilarious. And the reason I think it's hilarious is
here's how those predictions work. You take all the information

(04:34):
available to you, you run it through your computer model,
which factors in these different variables and gives you sort
of a percentage of probability of what your weather is
going to be like in the next let's say hour.
What if you want to look two hours ahead, Well,
then what they do is they take the prediction that

(04:54):
they made for an hour from now and extrapolate from there, saying, well,
if in fact the weather is what we think it's
going to be like in an hour, this is what
should look like two hours from now. Well, if you
want to look at three hours from now, well let's
take what the results were for two hours from now

(05:14):
and extrapolate again, and that's what we think it's going
to be three hours from now. Extend that out to
ten days. Yeah, and it's going to become less and
less reliable, Yes, because you're basing your predictions upon the
results of a previous set of predictions, not upon a
previous set of actual conditions. Right, So when you're tracing

(05:39):
it all the way back and your starting point is
right now, well, clearly, the further out we look, the
more unreliable the information is going to be, the more
likely some other variable that we have not anticipated will
play a larger role or a smaller role, and that
is going to affect the overall outcome of the of

(06:02):
what will actually happen. The forecast is the same, but
the actual thing we experience might be very different, which
is why if you're planning a picnic and you've got
ten days out from it and you're looking at the
weather and it says it's going to be absolutely perfect,
don't bet the house on it. Yes, that's not necessarily true.
Not to discredit numerical weather predictions, because a lot of

(06:25):
science and time goes into it. Yeah, but it's still
you know, It's the way that I see modern meteorology
is that over time we have continually built upon basically
what our ancestors did and just gotten it's gotten more scientific,
we've gotten better instruments, but it's still looking for patterns. Yeah,

(06:46):
exactly right. So you might look at the patterns of
when all of these conditions are in play. Out of
the last one hundred times that that happened, this is
how the weather turned out, and we're going to break
it down. So maybe eighty days out of those one
hundred days where the conditions were similar to today's, it

(07:08):
didn't rain at all. It was perfectly sunny, so eighty
out of one hundred it was lovely. The other twenty
days it rained and it was just steady rain, and
that's all there is to it. This is a super
oversimplified version of what could happen. This is what would
lead you to say there's a twenty percent chance of rain,

(07:28):
because he would say, all right, now, the last hundred
times the weather was exactly like it is today, twenty
of those times it rained, eighty of those times it
did not. Therefore, there is a twenty percent chance that
it will rain. That again is oversimplifying the way it works,
but generally speaking, that's kind of how they come to
those determinations, And in fact, there are ways of bolstering

(07:53):
the NWP by using something called model output statistics, which
is I kind of just talked about it a little bit.
I'll just go ahead and touch on it right now.
It's essentially doing what we were talking about, looking at
a specific region and the specific outcomes of days that
had similar conditions to the one you're looking at right now,

(08:13):
and then you're kind of making an educated guess based
upon a computer model and actual localized history. I got
that clear to something up for a lot of people.
I sure hope so because I've just kind of gone
with it. You know, I've just seen, oh, twenty percent
chance of rain, okay, and never really thought about what
goes into determining that. Well. And I know that there's

(08:33):
some people who had, you know, when they saw twenty
percent chance of rain, they thought it meant, oh, it's
going to rain over twenty percent of the forecast area,
which means that you know, that would be more like
scattered showers. That's really what scattered showers means. When you're
scattered showers, it means that parts of the forecast area
are expected to get rain, but it will not necessarily
rain over the entire forecast area. But if you hear

(08:57):
twenty percent chance of rain, it does not mean that
eighty percent of the forecast area is going to be
dry and the other twenty percent it's going to be wet.
Nor does it mean it will rain for twenty percent
of the day. In fact, that's part of the problem.
A prediction of precipitation, the good old pop, the pop
so pop. That requires a time element to it as well.

(09:20):
It doesn't mean anything without a time element. So if
you say there's a twenty percent chance to rain, you
also need to have an element of time attached to
that to make it meaningful. So twenty percent chance to
rain over the next six hours, then you know, all right,
So it's not saying that's going to be twenty percent
chance to rain or it's not gonna rain twenty percent
of the day, just that for the next six hours

(09:40):
there's a twenty percent chance it will be raining in
the forecast area. We will be back with more weather
technology after this quick break, so I hope that demystifies

(10:01):
some of it. Also, we can talk about radar. One
of my favorite things to talk about. Radar is awesome,
favorite character on mash me too. Yeah, well, I think
it's so adorable, right, Yeah, it's hard not to feel
for him, and the fact that he can anticipate everything
his commanding officer wants, and he can even say what

(10:23):
the commanding officer is saying before the commanding officer has
finished a sentence is obviously a key part of that operation.
But we're talking about actual radar, using radar to detect weather.
You've probably heard Doppler radar when looking at a weather report, like, well,
let's look at the Doppler radar and see where this

(10:44):
precipitation is moving in. Doppler radar for weather is different
from Doppler radar used by say, police officers, who are
trying to detect if you are speeding. The Doppler radar
that meteorologists use actually shoots out radio waves in very
short bursts called pulses, and then the radar listens for

(11:07):
any echoing pulses coming back to the antenna, and the
short pulses indicate not just the presence of something out there,
but whether it's moving and which direction is it moving in.
Is it moving toward the radar station or away from it.
If a Doppler radar receiver detect waves of a higher frequency,

(11:27):
the precipitation particles are moving toward the radar exactly, and
lower frequencies they're moving away. Yes, Because what's happening is
it's similar to a Doppler shift. And anyone who's ever
heard a vehicle with a siren go past yeah, is
familiar with this. It's a higher pitch as the vehicles
coming toward you, and a lower pitch as it's moving away.

(11:49):
What's actually happening is as a vehicle is moving toward you,
the sound waves it's emitting are being compressed. Now that
compression creates a higher frequency, which means we detect higher pitch.
As the vehicle passes, those frequencies are elongated, which means
a lower pitch. Same thing is true with the radar
accept Instead of it being a pitch, it's a radio frequency.

(12:11):
If so, a higher frequency will tell you, yeah, something's
coming towards you, and a lower frequency will tell you
something's moving away from you. And also the time between
when the pulse goes out and when you detect it
tells you the distance from the radar detection system and
the precipitation. So you could even say there's a storm

(12:33):
system that's five miles to the west, it's moving easterly
at this speed because you've detected it through a series
of pulses. If you have enough radar detection stations, you
can even describe the shape of the weather system and
talk about how some areas are more intense than others.

(12:53):
You can get all of that information from this approach,
and it's amazing how this thing works. First of all,
it's super high power. These radar stations are they're they're
generating or they're transmitting I should say, at four hundred
and fifty thousand watts. So your typical microwave oven is

(13:14):
a thousand watts. Wow, so you need a four hundred
and fifty of those to equal one of these radar systems.
So four and fifty thousand watts, and the pulse lasts
so short as to be unimaginable. It is point zero
zero zero zero zero one five seven seconds long, or

(13:35):
one point five seven times ten to the minus six seconds.
So if you hear the weatherman on TV bragging about
Doppler radar, there's a reason. It's very impressive. Yeah. I mean,
you're sitting at a burst of radio signals at such
a fraction of a second that it is again impossible
to even imagine. Meanwhile, then it listens for a longer period,

(13:59):
and by longer I mean relatively longer. It's still a
fraction of a second. It's point zero zero zero nine
nine eight four three seconds. So it shoots out a pulse,
listens for a little while, so it can detect when
the pulse comes back and what frequency it's at, so
it knows whether or not a body is moving toward
it or away from it, and then it does it again.

(14:21):
But that means, with that amount of time and the
comparatively large amount of time of listening, for every hour
of operation, the radio or the radar antenna is only
shooting out signals for seven seconds out of an entire hour. Wow,
that means for the fifty nine minutes fifty three seconds,

(14:42):
it is not sending out a signal. It is listening.
So for almost a full hour it's listening and only
for seven seconds is it actively shooting out a signal.
Can you mentioned it a conversation for someone who talks
seven seconds out of an hour, it would I any
conversation with me would last like a decade and before
you could get a word in edgewise. Yeah, it's it's

(15:05):
pretty amazing, and you usually would have one of these
stations shooting out these radio bursts at different angles of elevation.
These are called elevation slices, and when you go through
the entire range, you get what was called volume coverage
pattern or VCP, and that's what tells you what the

(15:25):
activity is, not just at ground level, but up in
the atmosphere as well. Toper radar can also detect tornadoes. Yeah. Yeah,
if if the particles switch from moving toward and then
away over a small distance, there's a good chance it
could be a tornado. Yeah, we know a lot about
those here in the southeast too. We get a lot

(15:47):
of tornadoes, not as many as places in the you know,
like in the Midwestern states, things like you know, Oklahoma
and stuff. I mean, you guys get tornadoes even more
frequently than we do, but we get them pretty serious.
Actually this year hasn't been too bad. No, but there
was one in November, yeah, which is weird because typically
we get them in the spring. Yes, usually between March

(16:09):
and June. That's kind of like our let's play it easy. Yeah,
but you don't wash that on anybody. So no, I
have been through a close call with a tornado while
wearing Renaissance Festival gear. That sounds surreal. That was my
final day when I did my first run at the
festival in two thousand and one. Wow. Yeah, we had

(16:33):
a really massive thunderstorm and at one point someone said
that there was a tornado a tornado watch, but not
a tornado warning watch, being that the conditions for a
tornado forming are present, warning being that a tornado has
actually been spotted in the region, in case you were wondering.
So now let's talk about satellites and meteorology. So the

(16:55):
computers are really good for building out those models and
giving us predictions. The double radars really good at tracking precipitation.
What do weather satellites do well, they're keeping an eye
on global weather patterns. But there are two different types
of weather satellites and they do this in different ways.
So one is the geostationary weather satellite. Now, geostationary weather

(17:20):
satellites maintain their relative position over a specific point on
the Earth. They are at a very high orbit over
the equator and they're always looking at the same thing
because their orbit is at the same speed as Earth's rotation,
not really the same speed, but relative speed. Because it's
able to stay in that same point over that part

(17:43):
of the Earth, and so they have to be on
an equatorial orbit and they have to be at a
particular altitude for this to work. It's great because it
means they can keep an eye on a specific region.
It's lousy because one, they're really far away, so the
instrumentation you have to put on the satellites has to
be incredibly sophisticated in order to get good readings from
that altitude. Plus they have a limited view, right, They're

(18:06):
always looking at one part of the Earth. They can't
see anything else outside of that view. Always geostationary yep.
So the other type you have are satellites that are
in a polar orbit around the Earth. Polar orbits are interesting.
So if you think of the Earth on its axis,
the polar orbit is going parallel to the axis of
the Earth. It's going perpendicular to the equator. So you

(18:29):
would think of it as going from north to south
and then south to north because once it crosses the
south pole, you can only go north at that point.
That's the only direction left to you, and it goes
in that circle, which means these satellites get a full
view of the entire Earth, because the Earth is rotating
while it's going in this orbit north south orbit. But

(18:52):
it also means that you only get a look at
the same part of the Earth twice in a twenty
four hour period, since once every twelve hours, you could
always put another satellite up there, and that way you
could get you put it on the opposite side of
the Earth where it's in the same orbit, and then
you get a look every six hours, just one from

(19:13):
one satellite and then six hours later one from another satellite.
But you also get to see everything on the planet,
So there's your trade off is that you get a
more comprehensive view, but you don't get a consistent view
of any one part of the Earth with these kind
of weather satellites. So a lot of weather surfaces depend

(19:33):
upon both, yeah, so it's good to have both, yeah.
And typically they carry devices called radiometers, which usually have
a small telescope or some sort of antenna, a scanning
device of some sort, and one or more detectors that
can pick up visible, infrared or microwave radiation, and they
use that to take measurements of the Earth and send

(19:53):
that down to the planet's surface, so that weather stations
around the world can take that data and crunch it
and figure out what the heck is going out on
out there when the frogs are raining from the sky
apart from amphibious assault, and all of those measurements are
actually done through little electrical voltages which then get digitized,

(20:14):
so transformed into digital information before transmitted down to Earth,
because you know, zapping electricity through space down to the planet,
it's not the most efficient way of getting information across
You certainly don't want to have a power chord stretch
all the way there. That would just be such a pain,
very inefficient. Yeah. Also with aircraft patterns. Yeah, and if
you don't have geostationary orbit, it gets wrapped up around

(20:35):
the planet pretty quickly. Yeah. We will conclude our two
part episode series about weather attack after this quick break.
So how do meteorologists do things like predict temperature, like

(20:58):
predict highs and lows and that kind of stuff? For
this I went to a website that was written by
a meteorologist named Jeff Haby, and boy howdy, did it
suddenly dawn on me how much more complicated this was
than I had even anticipated. But according to Haby, he
looks at everything from thermal advection, wind speed, cloud cover,

(21:22):
dew point, and the number of daylight hours expected for
that region in order to come up with the prediction
for the high temperature of the day in the low
temperature of the day. So what the heck does all
that mean? So thermal advection, what is that? That refers
to the transportation of heat by a moving fluid. So
typically the stuff that affects the advection include the strength

(21:44):
of wind, So how hard is the wind blowing in
that region? The temperature gradient between the warmer and colder areas.
So if one area is warmer than the other, is
it warmer by like a couple of degrees or is
it more significant than that is ten degrees fahrenheit, that
would be a much larger gradient, right, And the angle

(22:06):
between the wind direction and the temperature gradient, if that
angle is more narrow, you're going to see a greater
thermal advection, meaning you'll see more temperature changes moving into
an area from a different region. So that's just advection.
That's a lot of play. The other one that you
other term you might be a little confused by. I mean,

(22:28):
wind speed makes sense, cloudcover makes sense daylight hours. All
of that makes sense, but what about dew point? That
refers to the temperature at which air must be cooled
at constant barometric pressure for water vapor to condense. So
that temperature, again is dependent upon things like the actual
air pressure, right, and so the dew point changes based

(22:49):
upon those other factors as well. So all those have
to be taken into account before a meteorologist can forecast
what the temperature is going to be the next day.
This is why we're so happy to have those complicated
computers now, because if you were to keep track of
this yourself, you probably go bonkers. And then we have
like the idea of the probability of precipitation, which we

(23:11):
kind of talked about already, but generally speaking, there's some
weather services that will only predict rainfall if it's expected
to be over a certain amount, like point two five millimeters.
If it's going to be less than point two five millimeters,
it doesn't even register as rainfall in predictions. You would
say there's a zero percent chance or whatever, if that's

(23:35):
what you think is going to be the accumulation. Some
other ones are like, no any rain at all counts
if it's one drop of rain it rained in that region,
so it really depends upon the service. But that we
already talked about the percentages and what those means, so
hopefully that clears things up. And again that kind of

(23:56):
goes into that concept of model output statistics, where you
correct for your predictions based upon past conditions for a
particular region. All of this comes together to create the
weather report that you see. So I think the real
takeaway here is it's incredible the amount of schooling and

(24:23):
expertise a meteorologist has to have in order to do
his or her job properly. Yes, right, like, because you
see how complicated this is, and you start to have
an appreciation of all right, they said it was a
sixty percent chance of rain. I brought my umbrella and
never rained. Now you realize, well, when you're talking about
the system, this complicated and this unpredictable, something that can

(24:45):
change dramatically just because something you did not anticipate happened,
you start to appreciate more the challenge that they have
to do their jobs properly. So give your meteorologist a
hug and say thank you, because this stuff is hard. Yeah,
and my you know if weather is the state of

(25:08):
the atmosphere from day to day, and the atmosphere is
super complex, you know, it's it's my favorite analogy from
our article on our website about meteorologies, that the atmosphere
is like a soup with too many cooks. Yeah. Yeah,
there's a lot at play and so many different variables

(25:28):
are working behind the scenes to give you a simple
a simple weather forecast that is easily digestible. Yeah. This
is also why when you hear about supercomputers running weather simulations,
that's why you need a supercomputer because of this too
many cooks. I mean, it takes a lot to make

(25:50):
us stew. I hope, I hope at least some of
you are singing along now, But yeah, it's it really
does explain why you need that massive amount of computing
power just to do something that you would think would
be fairly simple. You know, you're thinking like, oh, it's
like six or seven factors, and then you realize, oh, wait, no,
there are these other things that also have an effect,

(26:11):
and in some cases a measurable and meaningful effect, not
just a potential effect. And it really does drive home
that it's amazing we can have relatively accurate weather predictions
in the first place. And also it makes me kind
of sad that I no longer I used to be

(26:34):
on television with a local weather guy. He did a
show at five thirty in the morning, and I would
show up on television and do a gadget segment with him,
and it was great. He was very nice, and his
ability to break down complicated concepts of weather in a
way that was helpful to people so that they could

(26:55):
plan their day was really amazing, especially when you start
really thinking about all the things that come into play
to make that, you know possible. So our hats are
off to you meteorologists out there, keep doing the good work.
I look forward to learning more about you know, when
we when we figure out even more details about the

(27:16):
complexities of the atmosphere and perhaps are able to make
even more accurate models, maybe we will one day reach
the back to the future too level where minute by
minute it tells you what the weather is going to be.
Of course, then I think they were actually suggesting that
we would have weather control, which is a whole other

(27:38):
thing that's opening and cannle worms. Yeah, I talked to
Dylan that I said I had thought about doing a
little discussion about weather control. But obviously we've gone pretty
long already. So what I will say about weather control
is weather systems represent a huge amount of energy, and
in order for us to affect or manufacture weather events

(28:03):
on a large scale, we would have to be able
to generate that amount of energy and pour it into
the atmosphere in a way that actually does what we
wanted to do. And we're so far away from any
of those things that it's absolutely unrealistic to think of
weather control, even if you are a Cobra commander. Yeah,
you know, going back to the beginning of this episode

(28:24):
with that listener request for this, and yeah, their father
said that in the sixties that they felt like the
weather report was just a joke. And you fast forward
to now and how you know, maybe some days you
grab an umbrella and you don't need it, but that
it's you know, you can track major weather patterns and
incoming storms. That we have a pretty good ability to

(28:47):
track hurricanes and like flash floods and things like that.
I can't imagine where I'll be in forty or fifty years. Yeah,
the fact that we can get at least heads up
on stuff before it becomes critical to us, is really important.
I mean, Dylan, you probably remember it wasn't that long

(29:09):
ago here in Atlanta when we had the snow apocalypse. Yes,
and because it was one of those things where the
initial weather report suggested that the weather was going to
miss the city and it didn't, that's an indication that, yeah,
our predictions are not one hundred percent accurate, they're not infallible.
And it also taught us a valuable lesson, which is

(29:31):
that even when you feel like there's a pretty good
chance that you're going to miss out on that bad weather,
it doesn't hurt to prepare because the alternative is to
spend eight hours on two eighty five, even if it
is an inch of snow, because it's Atlanta. Yeah, and
because we don't have a system in place to deal

(29:51):
with an inch of snow, and we got a lot
of hills. None of us have snow tires. Why would you?
And you let out the private in the public sector
at the same time, that was particularly bad. I remember
I actually stayed here, not here, but in our old
office location. I stayed there pretty much through the full
day because it was like, I'm gonna take Martha, I'm

(30:12):
gonna take the train. It's not gonna be a big deal.
It took me three hours to get home, usually would
take me forty five minutes, and I was getting I
got online, got ready to complain, and then I started
reading messages for my friends who were stuck in their
cars and had been for six hours. I thought, yeah, okay,
we're going to back away from the community slowly. Kids
stuck in school buses overnight. Yeah. Yeah, pretty rough stuff.

(30:34):
So yeah, we're We're not perfect, but it is getting better,
and it is pretty impressive to see the amount of
information you can get. I love, I mean I love.
I find watching Doppler radar readouts to be fascinating, Like
I could have that open on my desk all day
if I if I didn't have other stuff I need
to do. So Drees, thank you so much for sending

(30:56):
that request in. It's a lot of fun to kind
of read up on it and to go over Dylan,
thank you for joining me in the studio. Thanks for
having me greatly appreciate it. Well, that was it for WeatherTech,
at least from back in twenty sixteen. Again, I'll probably
need to do an update to this whenever I do
these classic episode intros naltros. I'm always reminded, Oh yeah,

(31:18):
I should really revisit this. This is one of those.
But if there are topics like new ones, or maybe
there's a topic you think I should revisit that I
haven't really talked about in a long time, you should
let me know. And there are a couple of different
ways of doing that. You can go to Twitter and
you can tweet to me. The show handle is tech
Stuff HSW or if you prefer to actually talk to me,

(31:42):
you can download the iHeartRadio app. It's free to downloads
free to use. You can navigate over to tech Stuff
by putting that into the little search field and you
will see on the tech Stuff podcast page there's a
little microphone icon. If you click on that icon, you
can leave a voice message up to thirty seconds in
length and let me know what you would like me
to talk about in the future, and I will talk

(32:03):
to you again really soon. Y text Stuff is an
iHeartRadio production. For more podcasts from iHeartRadio, visit the iHeartRadio app,
Apple Podcasts, or wherever you listen to your favorite shows.