Tuesday, December 31, 2013

A Thank You to KOMO 4

Tuesday, December 31, 2013
7:52 a.m.

I can't imagine a happier workplace.

Yes, that's how good KOMO has been to me. Every Wednesday and Friday this quarter, I looked forward to coming here. I knew I was going to have a great time with Steve Pool, and Shannon O'Donnell, and I did have a great time... greater than I could have ever imagined. But I didn't know I'd be developing relationships with Dan Lewis, Connie Thompson, and many of the other employees you see both on tv and behind the scenes. Yesterday, I got a lecture from one of the employees letting me know that I had stood behind anchor Molly Shen during the 4 p.m. show and the camera was catching me staring at her. One time, I realized two minutes before one of the main five or six o'clock weather segments started that I hadn't loaded the show up for Steve, and by the time he was on air, the show still wasn't completely loaded and he had to improvise. About halfway through his cut, the show fully loaded and he was able to run through it. He gave me a lecture, not a stern one though, and just laughed it off. You gotta screw up to learn how not to screw up, right?

I can have another quarter here, but I have to make way for the interns that are seniors. Remember, I'm only a junior. The minimum class standing for interns here is junior status, so I buzzed right in here as soon as I could. So I'm done for the rest of this year in order to let these seniors get a chance to do what I did. I'm definitely going to look into seeing if I can do other work here though that doesn't interfere with other interns (for example, on weekends) and I'm going to see if I can do anything with other stations. The tv meteorologists all know each other and Shannon used to work for KING, so maybe we can figure something out. It's always good to expose yourself to a variety of environments.

One thing's for sure though. I had the time of my life here. I'm still not sure if I want to be a tv meteorologist, but I'm more inclined to than before. Also, I shared a misconception that I think is held by much of the general population; that tv meteorologists are just pretty faces that don't know anything about the atmosphere. NOT TRUE. Steve doesn't have an atmospheric science degree, but he knows his stuff. Shannon has a degree, and she REALLY knows her stuff. These are real forecasters and real meteorologists. They are not reading off a teleprompter... everything is improvised. You have to know what you are talking about in order to do that.

I came in early today... 5:20 a.m. early... to work with Seth Wayne, the morning forecaster, and like Shannon vs. Steve, he has his own way about doing things, but just like both of them, he is incredibly nice and professional. It was a joy working with him this morning.

I can see why Steve has been KOMO's lead forecaster since 1984. This is a place where I feel comfortable. I don't want to leave, but at the same time, I'm gracious for the opportunity I was given, and I'm stoked, yes, STOKED, that others will get the same chance as me. It's just awesome. There's nothing like it.

If you want to go to KOMO or get a tour, contact me. I can hook you up.

Thank you so much everybody! Mom, dad, everybody at KOMO, all my friends for reading my blogs, it's been fantastic. I'll see you sometime soon!

Charlie =D

Saturday, December 21, 2013

The Truth About My Snow Forecast

Saturday, December 21, 2013
11:32 a.m.

It wasn't that good.

Simon Cowell is giving me the thumbs down: http://tinyviolets.wordpress.com/category/music/


I think my elitist response yesterday was an insecure attempt to retain my position as Seattle's best snow forecaster. For this snow event, that wasn't true. I was calling for what the models were showing... 2-6 inches of snow, with amounts mainly in the 2-4 inch range but a couple spots hitting 6 inches. Cliff Mass also called for 2-4 inches in the Seattle area. The National Weather Service, on the other hand, was calling for 1-3 inches over all of the lowlands, and they were dead on and they were much much closer. Nearly everywhere got between 1-3 inches. Fantastic work by the National Weather Service, because places north of Seattle looked like they could get 4+ inches based on the model initialized only 12 hours behind the actual event. Amazing.

What I was proud of was my radar interpolation of the incoming precipitation. I neglected to do some research to find out that the NWS had it right all along.

So here's the rating.

1. National Weather Service
2. Charlie Phillips
3. Cliff Mass

* I put myself ahead of Cliff because of my extra interpolation work. His initial forecast may have been ever so slightly more accurate than mine.

So screw the ego. I'm not always the best snow forecaster. 

OK, well I'm almost always the best snow forecaster. I'll get 'em next time. ;)


Friday, December 20, 2013

Less Snow than Forecast

Friday, December 20, 2013
7:43 a.m.

Where's the glacier?

We got significantly less snow than forecast. I'm actually super proud of myself... I called this before several others, including but not limited to Cliff Mass and the Seattle NWS. That takes skill! (and a fair amount of luck)

As evidence, here was my Facebook status at 4:17.

I'm awake in my room after my alarm went off because I'm that much of a nerd. It's 4:17, and there is very light snow in the U-District. Heavier snow is on the way... you'll definitely notice it by 6. There is currently a light dusting on the cars.

I have to say, based on comparison of radar and 00z WRF-GFS model shots, this system looks to be taking a track further north than forecast, which would give Seattle less snow. We will see...

Here's what I saw. Look at the radar at 4 a.m., and look at the total predicted past 1-hour precip ending at 4 a.m.

Valid 04:00 am PST, Fri 20 Dec 2013 - 12hr Fcst: http://www.atmos.washington.edu/~ovens/wxloop.cgi?mm5d3_pcp1+2013122000///1

See how the precipitation on the radar is ever so slightly north than it was modeled on last night's run? Man I'm good. The National Weather Service was still sticking to their original forecast in their 3 a.m. forecast discussion. Granted, I had an extra hour 90 minutes, but still. I may have not nailed this event on the button well in advance (no one did), but when it came to short-range forecasting, I knocked it out of the park.

Snow will continue to fall up north. Guys north of Everett are gonna get the 2-6 inches Seattle was supposed to get. Things will transition to rain just after noon.

Time and time again, it has been shown that I am hands-down the best snow forecaster in Seattle. I'm humble about it too; I take my position with great grace.

Have fun in the inch or so there is!

Thursday, December 19, 2013

Oh Heck Yeah!

Thursday, December 19, 2013
8:32 p.m.

OK, so when I saw the latest model run, I absolutely jumped for joy. Why? Because the newest WRF-GFS model run upped the forecast snow amounts. At first, this model gave us a lot of snow, but then it sharply pulled back. Now, it’s trending whiter again. Let’s take a look at how this trending has turned out.
Valid 04:00 pm PST, Fri 20 Dec 2013 - 48hr Fcst: http://www.atmos.washington.edu/~ovens/wxloop.cgi?mm5d3_wa_snow3+2013121900///3

This was last night's run, and was as close to a Snowmageddon as the models ever came to showing. Half a foot just to the north of Seattle? Please sir, may I have another? 

But then came this morning's run. I cried.

Valid 04:00 pm PST, Fri 20 Dec 2013 - 36hr Fcst: http://www.atmos.washington.edu/~ovens/wxloop.cgi?mm5d3_wa_snow24+2013121912///3

Snow amounts decreased three-fold over north Seattle. Those poor folks in Tacoma weren't getting any. I got over myself fairly quickly... this was never expected to be the snow event of the century, and I honestly didn't even think it was gonna happen... but still. It's like walking out of a test feeling great about your performance, and then seeing that you failed it. OK, maybe not failed it, but didn't perform as well as you thought you would.

Valid 04:00 pm PST, Fri 20 Dec 2013 - 24hr Fcst: http://www.atmos.washington.edu/~ovens/wxloop.cgi?mm5d3_wa_snow24+2013122000///3

Then came this evening's run (and the numbers are still rolling in). It definitely wasn't as good as last night's run, but it was significantly better than this morning's run. There's that 6-inch bulls-eye over North Seattle. In any event, things look snowy tomorrow morning, especially from Northern Seattle northward. Kids - don't do your homework.

By the way, this snow event is brought to you by evaporative cooling. When the atmosphere is dry like it is right now, it acts like a "sponge" in that it absorbs some precipitation that falls from the sky. It takes energy to evaporate this precipitation and turn it into water vapor - heat energy - and this cools the atmosphere. If the atmosphere was moist, we'd be talking about a cold rain.

You can see that the moisture starts to come in around 4 a.m.

Valid 04:00 am PST, Fri 20 Dec 2013 - 12hr Fcst: http://www.atmos.washington.edu/~ovens/wxloop.cgi?mm5d3_pcp3+2013122000///3

By 7 a.m., it has solidly engulfed our area.

Valid 07:00 am PST, Fri 20 Dec 2013 - 15hr Fcst: http://www.atmos.washington.edu/~ovens/wxloop.cgi?mm5d3_pcp3+2013122000///3

This model says the front will pass through Western Washington from south to north during mid-morning and probably cross Seattle around 10 a.m. Afterwards, it looks like some precipitation may keep falling in the form of rain as showers and a possible convergence zone set up. Showers will be sparse in nature up north, and Seattle and the central Puget Sound region will probably be shadowed by the Olympics. The westerly flow behind the front will be VERY conducive, however, to orographic snowfall in the Cascades, so hopefully the Summit and Snoqualmie can open before too long. All in all, I'm not expecting too much rain, which is a good thing, because rain on snow is just really, really depressing.

Southern Washington, however, will receive more rain. The same thing goes for Oregon. With the exception of the Longview-Vancouver area perhaps picking up an inch, however, the snowfall doesn't look to be all that spectacular. Excepting that region, Olympia southward to Salem look to pick up trace amounts.

I'll probably wake up early to see the snow and/or I'll stay up late to see what happens, either by choice or by instinct. Either way, I'll keep you posted.

At this point, my LAWES rating is 5.3d

5 = 80-100% chance of snow with snow observed in area (Victoria, BC earlier this evening)
.3 = 2-6 inches of snow
d = 3 or more models consistent

Have a good one folks!

Wednesday, December 18, 2013

Snow Update

Wednesday, December 18, 2013
11:26 a.m.

I was pretty happy when I saw the UW's WRF-GFS model this morning.


I wasn't quite stunned. If that was the case, this would be my reaction.


But hey, I'll take what I can get. This run was an improvement over yesterday's runs. I'm still a little skeptical for the same reason as I said in the previous post, as snow levels will be RIGHT on the edge, but the models have been pretty consistent, and as I inferred in the previous post, models can catch your eye, especially when the event is less than 48 hours out.

That being said, let's look at the setup.

Valid 04:00 am PST, Fri 20 Dec 2013 - 48hr Fcst: http://www.atmos.washington.edu/~ovens/wxloop.cgi?mm5d2_thick+//84/3

One thing we look at when making a forecast for snow is the thickness of the atmosphere. The colder the atmosphere, the denser it is, and the lesser the distance between pressure levels (the 1000-500 mb/hPa levels in this map). The warmer the atmosphere, the thicker it is. The highest thicknesses appear at the center of hurricanes, where the air is extremely warm throughout the atmosphere. Around Seattle, we generally need thicknesses to be around ~522 to start talking about snow. Here, they are 543. So how is this same model still forecasting snow?

Well, the upper atmosphere may be warming, but this moisture is forecast to arrive right in the early morning hours, when temperatures are the coldest at the surface.

Valid 04:00 am PST, Fri 20 Dec 2013 - 48hr Fcst: http://www.atmos.washington.edu/~ovens/wxloop.cgi?mm5d2_slp+//84/3

Another thing - Thursday is expected to be mainly cloudy, but the clouds look to decrease Thursday night. This keeps the temperatures down during the day and lets radiational cooling take over at night. It's like a 1-2 punch.

That's the thing though. This situation is SO dependent on timing. If things get shifted by a couple hours, then the lower atmosphere will likely be too warm, and we'll just get rain. Right now, the timing looks PERFECT - the precipitation starts to come in at 4 a.m. That means the models are showing the best-case scenario. But let's not think about that. :)

Valid 04:00 am PST, Fri 20 Dec 2013 - 48hr Fcst: http://www.atmos.washington.edu/~ovens/wxloop.cgi?mm5d2_pcp3+//84/3

Valid 04:00 pm PST, Fri 20 Dec 2013 - 60hr Fcst: http://www.atmos.washington.edu/~ovens/wxloop.cgi?mm5d3_ww_snow24+///3

Not too bad, eh? Three inches of snow over Seattle. If you want more, go north to the Everett area. If you want a lot more, go to the mountains. If you want less, go south. And if you don't want any, swim to China.

Now, the moment, you've all been waiting for... my LAWES rating.


Meaning a snow event within 3 days with up to 4 inches of snow with one consistent model showing it. Other models besides the WRF-GFS are also showing snow, but I decided not to include them because they are not showing as much.

This event, save perhaps a snowy Puget Sound Convergence Zone, is the hardest type of snow event to predict here in Western Washington.



Tuesday, December 17, 2013

Snow on Friday and Another LAWES Scale Revision

Tuesday, December 12, 2013
11:13 a.m.

I'm back! Sorry it's been so long. The end of the quarter was quite stressful. Heck, I had a seizure on the night of my final. That sure threw a wrench in my studying/sleeping plans. But, I survived, and that's what really matters. Food. Water. Shelter. Survival.

And when it comes to weather, let's face it; the only thing that REALLY matters to us around here is snow. We pretend to be fascinated by windstorms and floods, and I would say that we actually are interested in them, but nowhere near the extent as snow. And when it comes to snow, it's only REALLY worth it if we get an arctic blast that makes the roads icy for days on end (though this was much more applicable when I was in school). I remember in 2008 when all of Seattle was covered in snow, and I skied down Madrona and Alder Hills. Alder Hill was actually pretty steep... I was going really fast... and since this was at night and much of the road is covered by trees, there were times when I couldn't see anything, which was awesome and nerve-racking. I had also been watching a whole bunch of Warren Miller skiing videos lately, and I got the bright idea to ski down a steep hill of blackberry bushes across the street from my house. I survived with only a few minor scratches. Good times, good times.

Our last arctic outbreak was definitely our chance to get some snow, and considering how cold it got, the snow would have stuck around for a while. It's a shame we didn't get any. Thankfully, all hope is NOT lost, as some snow looks to be in our possible future for Friday!

Valid 04:00 pm PST, Fri 20 Dec 2013 - 84hr Fcst: http://www.atmos.washington.edu/~ovens/wxloop.cgi?mm5d3_ww_snow24+///3

Now, as you can see, this isn't going to be the storm of the century, or even the decade. Earlier models had the highest amounts over the North Interior, but this morning's latest UW WRF-GFS model has the highest amounts over the Kitsap Peninsula. This period of snow is not embedded in an arctic outbreak, so I regret to say it won't be sticking around, but it should make for some G-Rated fun anyway, if you know what I mean. When a blizzard is around, don't take the kids, grandparents, or even Navy SEALS outside. Leave that task to me and my camera crew.

Jim Cantore covering a winter storm in Chicago. He gets surprised by thundersnow - watch the video here! http://www.metatube.com/en/videos/48757/Jim-Cantore-Thunder-Snow/

By the way, if I ever become a TV meteorologist, I am SO doing the whole Jim Cantore gig. Getting paid to go to where the severe weather is? Sounds good to me. Unfortunately, The Weather Channel is stationed in Atlanta.

But in the meantime, we won't have any record thundersnows, but we should be on the lookout for perhaps a slushy half-inch or so. I made a scale way back in high school called the LAWES (Likelihood of an Arctic Weather Event With Snow) Scale that I didn't get to use at ALL last year, so here's a recap of the scale and my rating for the upcoming event. I usually don't pull out this scale for these minuscule snow events like the one we may see where there is nothing arctic about them, but I've been dying to use it, so let's go ahead and put this bad boy to work. I was taking a look at it a while back and decided it could use some revision. So revise I shall.

Here's what the old scale looked like.

LAWES Level 0: You are on a planet that has the ingredients necessary for snow.
LAWES Level 1: Models are hinting a chance of snow beyond a week out, 0-20% chance of snow
LAWES Level 2: Models are showing a chance of snow within a week, 20-40% chance of snow
LAWES Level 3: Models are showing a scenario that would produce snow within 5 days, 40-60% chance of snow
LAWES Level 4: Models are in consensus about a snow event 1-3 days out, 60-80% chance of snow
LAWES Level 5: Places around the area are already getting snow AND snow is forecast for Seattle, 80-99% chance of snow
LAWES Level 6: Snow is occurring, 100% of of snow (derp)

Ending in:

.0 - the event is too far out to pin down snow totals
.1 - up to 2 inches of snow
.2 - up to 4 inches of snow
.3 - 2-6 inches of snow
.4 - 4-8 inches of snow
.5 - 6-12 inches of snow
.6 - board the next plane to Panama


The second part of the scale is fine, but it's the first part that needs an overhaul. There isn't a direct correlation between how far in the future the snow event is forecast to occur and the probability of it occurring in the first place. Heck, sometimes, we look outside, see the white stuff, and sit back, bewildered, and try to remember if we even expected the white stuff in the first place. Therefore, I propose we drop the probabilities from the first number and add an alphabetical suffix to the second number while keeping the second number the same. Here are those suffixes.

a - very little consistent model consensus
b - one model consistent
c - two models consistent
d - three or more models consistent

Bottom line: there may be other models I'm not counting that are showing snow events that would hypothetically "increase" the alphabetical suffixes. 

You noticed that I didn't put any probabilities in. I'm going to put the probabilities on a case-by-case basis, because all snow events, just like all people, and just like all snowflakes for that matter, are all special.

This applies to operationals and ensembles, and either one can work. For example, if the GFS operational model is inconsistent but the ensemble mean is consistent, then we count the GFS as consistent. If the operational was consistent but the ensembles were inconsistent, we'd still count the GFS as consistent.

So that's the new and improved LAWES scale!

LAWES Level 0: You are on a planet that has the ingredients necessary for snow.
LAWES Level 1: Models are hinting a chance of snow beyond a week out
LAWES Level 2: Models are showing a chance of snow within a week
LAWES Level 3: Models are showing a scenario that would produce snow within 5 days
LAWES Level 4: Models are in consensus about a snow event 1-3 days out
LAWES Level 5: Places around the area are already getting snow AND snow is forecast for Seattle
LAWES Level 6: Snow is occurring

Ending in:

.0 - the event is too far out to pin down snow totals
.1 - up to 2 inches of snow
.2 - up to 4 inches of snow
.3 - 2-6 inches of snow
.4 - 4-8 inches of snow
.5 - 6-12 inches of snow
.6 - board the next plane to Panama

With suffix of

a - very little consistent model consensus
b - one model consistent
c - two models consistent
d - three or more models consistent

*I don't have access to all models

Ladies and gentlemen, I give our chance of snow on Friday a LAWES 4.1b. However, I'm honestly not overly keen on it actually happening. Temperatures are going to be just above freezing in most places, and how cold will probably be decided by the extent of radiational cooling earlier that evening. I'm giving a 20% chance here for all you UW kids, but for those of you on Queen Anne or Capitol Hills, I'm giving you a 40% chance. Those to the north have a better shot, and those on the Kitsap Peninsula are above the 50% bar in my book. These percentages will change as the event comes closer.

One thing is for sure though. Those New England folks, who have been pummeled by feet of snow in recent days, think we are absolute wimps. So please keep this whole snow thing on the DL, and for once, don't share this blog with your friends. When it comes to freaking out about snow... being a Seattlite is downright embarrassing.

Happy holidays!

Monday, December 9, 2013

ATMOS 301 (Final) - Radiation and Climate

Monday, December 9, 2013
12:44 a.m.

Let's start out with some definitions:

Radiant Flux = F,   units are in Watts (Joules/sec)
Irradiance = E, units are F per unit area (W/m^2)

Energy of a photon: h*v,     where h = Planck's Constant and v = Frequency
c*/v=lambda       lambda = wavelength, c* = 3*10^8 m/s

A Blackbody is a hypothetical substance that absorbs all incident radiation and emits the max possible at all wavelengths. The relationship below is known as Planck's Law.

 Through this law, we can discover that the wavelength is inversely proportional to the temperature. This is known as Wien's Law.

When we integrate Planck's Law, we get the Stefan-Boltzman Law. I just put the whole slide down for this one because I thought it explained a lot.

Alright, now that we have those equations established, we should get onto establishing the equation for the Earth's radiative equilibrium. We've got (or at least we had) the same amount of heat coming in and coming out, and the Earth's temperature has been staying more or less the same over the past 10,000 years. Of course, the last 50 years have NOT continued this trend, as the Earth is not in steady state and more radiation is being absorbed than emitted and we are heating up as a result.

But if we let F_i = F_o,


The emissivity is the ratio of the actual to the max possible radiation that can be emitted at wavelength lambda.

Kirchoff's Law says that emissivity exactly equals absorptivity.

Here's a diagram of the eart and how this stuff works.

Here are the equations to figure out the temperature of the Earth at atmosphere. a_sol was given to be .1 and a_ir was given to be .8. E was calculated using the equation earlier.

Here's another example.

As you might be able to infer, I'm kinda getting tired, but I think I understand most of this stuff. I'm going to review some more notes, go to bed, wake up and review some hw problems, and then do the test. How's that sound?


ATMOS 301 (Final) - Numerical Weather Prediction

Sunday, December 8, 2013
11:39 p.m.

Numerical weather prediction has gotten pretty darn good over the past few decades, but it wasn't always that way. But taking this class has made me realize how incredible numerical weather prediction really is. I mean, c'mon now... we take a whole bunch of numbers, put them through these crazy equations that I am only beginning to understand, and we magically get these pretty accurate forecasts of what's going to happen. It's like looking into the future. It's weird. It's also 11:45 p.m., and I've got a long way to go.

First, we've got three "prognostic equations" - one for T, one for u, and one for v, that we use for weather modeling. Here they are.

We will let the atmosphere be adiabatic, frictionless, and dry, so H=0, a=0, and T_v = T. The flow will be geostrophic (for simplicity purposes).

In this case, the state of a sample of air is described by six dependent variables: T, rho, Z, u, v, and omega. Each dependent variable has a value at each point (x, y, p, t). These points are in a grid in the atmosphere called the "model domain."

Because we have six unknowns, we need three more equations. We already have these.

These are called the diagnostic equations because they don't contain time derivatives and don't have any prognostic value. Therefore, they allow us to find rho, Z, and omega from the predicted variables T, u, and v.


Why do weather predictions suck?

Well first off, they don't, thank you very much. But they've got one big thing working against them, and that is that the atmosphere is a chaotic system; it is one whose future states are sensitive to its initial state. That's kind of the nature of the beast itself. 

There are four other main points that Professor Houze brought up in his lectures that talk about some of the reasons models aren't as accurate as they could be. Here they are:

1.) Input data are incomplete
    -   this is especially true over oceans, deserts, polar ice caps, Siberia, the Congo, etc. Certainly not over Oklahoma City.

2.) Small scale processes are not observed (you can't get in between every grid point)

3.) Observational errors - some people are dumb, but nobody's perfect

4.) Errors with friction and heating in the mathematical representations of the equations.


Sunday, December 8, 2013

ATMOS 301 (Final) - The Dynamics of Atmospheric Wind Flows

Sunday, December 8, 2013
7:56 p.m.

Were all familiar with Isaac Newton. You know, that "apple" guy. Or was it a peach? Either way, I'm sure most of you are also familiar with his second law, which says that F=ma, or that Force = mass*acceleration. There are three types of horizontal forces in the atmosphere that sum up to make the forces that we observe with the weather everyday, so let's start out by talking about these individual components.

First, let us define our coordinate system so I don't have to keep running through it again. If you know your calculus, you should understand this. If you don't know calculus, just pretend that the u axis is the x axis and the v axis is the y axis.

Here is our equation for the sum of the total horizontal forces in the atmosphere:

When do some math and derivations that I don't think we'll have to do on the exam, we get the following equation as the end-all-be-all vector equation of horizontal motion.


f = Coriolis Parameter = 2*omega*sin(phi)

phi = latitude
omega = angular velocity of earth's rotation

C_x= fv
C-y= -fu

The air can only get started moving by means of the pressure gradient force, and it moves from high pressure to low pressure. Once it is moving, the Coriolis and frictional forces can act on it.

We can actually write this equation in two different ways. The way just mentioned is used for application to a constant geopotential height in the atmosphere with the pressure changing, while the second equation below with the -g_0 term is used to apply a constant pressure surface to the atmosphere when the height is changing.

Geostrophic Balance:
Geostrophic balance refers to when the red (either one) and fk x v terms are equal, meaning there is no friction in the atmosphere. In this case, the velocity of the wind is directly parallel to the isobars.

Wind in the Boundary Layer:
The lowest part of the atmosphere (the boundary layer) has lots of friction, so the Equation of Motion goes changes to suit it. The vector r must balance the PGF.

Below is a series of diagrams to explain how this process works. They are taken from Professor Houze's presentation on dynamics and should be able to explain how it works better than I could ever dream of doing so with. First, we start out with some wind going to the right, and I'll let the diagrams explain the rest from there.

The Thermal Wind:
This is were things got interesting and difficult for me, so I'm going to spend a bit of time on it. First off... let's recall the first law of thermodynamics, which says that the change in internal energy is equal to the heat added to the system minus the work done by the system.

We can let a variable H=dq/dt, however, and we can get the following relationship.

In this form, the 1st law of thermodynamics takes the form of an equation for predicting the rate of change of temperature of a parcel of air. We want to predict the rate of rate of temperature at a point, so we do a lot of transformation of equation stuff (what else is new) and end up with the following equation:

Dry Static Stability:

The dry static stability equation gives us the stability of the atmosphere. The atmosphere is usually either stable or conditionally unstable, so

and omega is greater than 0, which means there is local warming. If omega is less than 0, than there is local cooling and adiabatic expansion.

Horizontal Advection:

There are two types of horizontal advection: cold advection and warm advection. In cold advection, wind blows from lower to higher temperature, and in warm advection, the opposite happens. Warm advection takes place ahead of a cold front whereas cold advection takes place behind it.

If we let delta T = change in temperature and delta s = change in distance, warm advection occurs when delta T/delta S is less than 0 (wind is blowing from an area of higher temperature to an area of colder temperature, and the temperature is therefore getting colder with distance) and cold advection occurs when it is greater than 0.

A good deal more confusing than thunderstorms, and not nearly as fun to write about. Well, let's move on to our next topic: numerical weather prediction!


ATMOS 301 (Final) - Deep Convection and Thunderstorms

Saturday, December 7, 2013
8:38 p.m.

Well, I'm back! I've been through heck and back over the past few weeks with schoolwork, but I now have some time to write on this blog, and write I shall.

OK, well that wasn't really true. I actually have an atmospheric science 301 FINAL on Monday, and I need to study for it. I studied for my midterm by doing this same sort of blogging method, and I actually did quite well on the midterm. And it wasn't just because of my previous exposure to weather... I'm learning many things in this class that I've never even come across before.

Alright, so we got several topics to cover. This blog will cover thunderstorms, but then we'll move onto the dynamics of horizontal atmospheric flow, numerical weather prediction, and radiation/the greenhouse effect/why we are glad we aren't in a perpetual nuclear winter (yet). Here's thunderstorms.


I'm sure we're all familiar with cumulus clouds. Every thunderstorm starts as a cumulus cloud. There are two types of cumulus clouds that we tend to talk about: cumulus humilis and cumulus congestus. Humilis are known as "fair-weather cumulus," because they are pretty short and fluffy due to a stable atmosphere overhead preventing them from growing them much further. 

Cumulus congestus, on the other hand, are the cumulus that have the potential to grow into cumulonimbus clouds that have the potential to produce lightning, tornadoes, and even sharknadoes. I haven't seen that movie yet, but I need to see it immediately. The atmosphere above their LCL (lifting condensation level) is much more unstable, as as a result, the air parcel visibly denoted by the cumulus congestus cloud is a lot more buoyant. 

Take a look at the two skew-T plots below. The one on the left is for humilis, and the one on the right is for congestus. Notice how much more unstable the congestus skew-T is.

Cumulus humilis
Cumulus congestus

 Oh yeah, at some point I should probably show you what they look like. I don't know where these were taken, but they probably weren't taken in the 21st century. It's good to have some old-school photographs now and then in this digital age of ones and zeros. Can you guess which cloud is which?

The top one is humilis, and the bottom is congestus.

Now that we've got our cumulus background, lets looks at the environmental controls for the formation of deep convection and thunderstorms.

Temperature and Moisture Stratification:

There are three conditions necessary for deep convection:
   1.)   A deep, conditionally unstable environmental lapse rate
   2.)   Deep boundary-layer moisture
   3.)   Low-level convergence (or lifting) sufficient to "release" the instability

Convection feeds on the potential energy inherent in the temperature in the moisture stratification, and this potential energy is known affectionately by the meteorological community as CAPE (Convective Available Potential Energy) and is in J/kg.

It is defined as the integral from the LFC (level of free convection) to the EL (equilibrium level) of F/(rho)'dz. F is the upward buoyancy force per unit volume on the rising air parcel due to the temperature difference between the parcel and its environment, rho' is the density of the parcel, and dz indicates that we are taking the integral with respect to height. We do some weird derivation stuff, and we find out that the integral is essentially the area on the skew-T plot from the LFC to EL and pounded by the temperature sounding on the left and the moist adiabat on the right. CAPE values of 0-1000 are considered marginal for convection, but when values get above 4000, look out.

*When it comes to wind, storms generally move at the average direction and speed of all the horizontal wind vectors in their component of the atmosphere. But it's not just horizontal wind profiles that drive storms... vertical ones have a huge one too, and I'll discuss those shortly.

Structure and Evolution of Convective Storms:

There are three types of storms: single-cell, multi-cell, and supercell.

1.) Single-cell storms

These types of storms are the most common type of storm, and they are also the weakest. If multi-cell storms are like "Daniel's Broiler" and supercells are "Canlis," these guys are just "Papa John's." These storms are short-lived because they "self-destruct" very quickly in that their downdraft quickly cuts off their updraft and the supply of warm air necessary to keep the storm alive.

2.) Multi-cell storms

The gust front is the boundary between the warm, moist air from within the boundary layer and the denser, evaporatively-cooled downdraft from the storm itself and can be thought of as the “leading edge” of the storm. This advancing gust front is an example of what would be the leading edge of a “gravity current,” which is where a mass of high-density fluid flowing along a horizontal bottom displaces a fluid of lower density. This displacement causes lift, allowing new cells to continuously form.

This allows for the whole multi-cellular system to exist for quite some time. The picture below shows this "conveyor belt" effect pretty well.

Multi-cell storms can be pretty damaging. "Derechos" are a type of multi-cell system that can devastate areas for hundreds of miles. Just because they don't produce torndaoes doesn't mean they can't cause massive destruction. The June 2012 North American Derecho killed 28 people and caused 2.9 billion dollars in damage as it tracked from the Midwest to the Atlantic Coast.

3.) Supercell storms:

Supercell storms are defined by having a rotating updraft. This rotating updraft forms when wind shear in the lower atmosphere gets a vortex of air parallel to the ground rotating in the atmosphere. This vortex is then lifted into an upright column by an updraft in the unstable atmosphere, and this rotating column of air becomes the rotating updraft.

When the updraft lifts this vortex up, we now have two vortices in the storm. The high pressure prefers to be on the upshear side of the low pressure, so the storm actually ends up splitting into a "left-mover" and a "right-mover" so that this preference is conserved. 

The right mover is generally the "favored" mover in the US. I give myself a headache trying to visualize and explain it, so I'll just give you the diagrams and let you do the dirty work yourselves. The vertical pressure gradient helps to reorganize the storm so that the updraft is on the right and the downdraft is on the left once the vortex of spinning air has been lifted upright by the initial updraft. I'll just leave it at that.

Supercells are pretty complex in their anatomy and have a lot of little (and not-so-little) quirks that set them apart from other thunderstorms, so take a look at the drawing below, copy it, etc. and you should be fine. You'll know you'll be in the green when you can draw it using an Etch-A-Sketch

That wasn't too bad... mostly qualitative stuff. Now it's time to move on to something a little more intimidating...