My amateur approach to forecasting severe (tornadic) storms

[ChristofferB]

As you have already noticed I am trying to learn as much as possible before this season and my efforts are now focused on the forecasting of severe storm. Although I have read tons of forum posts, several SPC texts, plenty of excellent websites (such as WeatherPrediction.com), and the storm chasing handbook my greatest struggle was to find some sort of "basic recipe" to follow, or at least start with. I know that there is no such thing as a Basic Recipe and that there are exceptions to every rule but unless there were any broad guidelines, no-one would ever be able to forecast anything.

I needed to tie down the severe storm ingredients: moisture, lift, shear and instability into two important factors: maps and numbers. No matter how much I understood for example that instability was important and what this did in a storm I couldn't translate that into finding it on a map, and interpret it in terms of: is this good or bad?

So, the main questions I asked myself were:

- How can I find instability / moisture / lift / shear on a map?

and

- What values should I be looking for?

Through the resources mentioned above I had a decent understanding about each and every factor by itself. I knew, for example, I could find instability (CAPE) and moisture in a sounding - but how, for example, would I find the right station to check out in the first place!?

What I needed was first to find my ingredients translated into parameters and figured out the following:

- Instability: CAPE (and CIN as a limiting factor)
- Moisture: Dew points / LCL
- Shear: 0-6 km shear, as being the most important
- Lift: ?

I further realized CAPE and Wind Shear had a compound parameter in EHI.

I still couldn't find a parameter for "Lift" but my thought was that if the factors mentioned above create the gun powder keg, lift is the ignition - and it needs to be found in terms of a front or the likes of it where all other factors are met. I know there are some good information on WeatherPrediction regarding fronts that I will look further into.

My major breakthrough was finding the Powerpoint presentation describing the recipe for Tornadic Supercells, which explained and provided me with the numbers that could give me an idea of what to look for. WeatherPrediction.com was a great reference website but it is not the Forecasting Beginners Course it sort of claims to be, in my opinion. When finally realizing the parameters and corresponding values, THAT's when WeatherPrediction.com became my Go To-website to understand each parameter.

Given that PowerPoint-presentation and if my goal is to forecast tornadoes only, I imagine I can start a forecast now by trying to find areas that seem to fit:

0-1 km EHI > 1.0
0-6 km shear > 30 kts
LCL height < 1.500 m
0-3 km CAPE > 20
LFC height < 2.500 m

Further on, these following parameter values to use as a directive for minimum values:

CAPE > 1000
Dewpoints 55+ F

After having some parameters to look for I realized I didn't need to browse around for maps in 10's of different website (which I had assumed), the Mesoscale analysis maps (http://www.spc.noaa.gov/sfctest/new/viewsector.php?sector=19#) and the sounding analysis map (http://www.spc.noaa.gov/exper/soundings/) are really the only ones I need.

Now I finally have an approach that sends me to the maps looking for clues. Unless it is a Moderate/High-risk day I assume my found parameter values will hardly ever surpass my Tornadic Goal Values anywhere. But, now I can look at the Mesoscale Analysis map and see: "Oh, this place got some nice shear - I wonder if it has sufficient moisture! Hmmm, the LCL height is 2.200 m, that's not a good sign. I wonder if there are things that could help to bring it down!".

This was a huge step forward for me! Now, I had a forecasting approach - no matter how flawed - that I could start with. The biggest thing I am missing is a similar approach to finding Lift. I know from theWeatherPrediction-articles about factors that will cause lift, I just don't have an approach to find them on a map, and to determine if the Lift is "good enough" or "non-existing".

My list of disclaimers:

- I am very well aware of that this approach is narrow, flawed and possibly not even valid.

- I know there are plenty of exceptions to these rules.

- I know that an insufficient value in one parameter could be helped by another.

- I know that this approach, in terms of the values I have found, is for tornadic supercells - which is by far not the only thing we are looking for. I really enjoy LP supercells, for example.

- I know this doesn't really predict the timing of a storm. Winds will change the location of these parameter over time.

- It seems likely that compounded indices, like EHI, provides only estimates.

- I know that these values will hardly ever coincide at one certain point at a certain time.

...and I know tornadoes can be produced in any other circumstance far outside of these values. Single cell thunderstorms, squall lines etc.


But, now I have something to work with and to bring up to you as something to discuss. What do you think? I would like to further discuss and ask:

* What would be the greatest flaws to this approach?

* In case I am just looking for (non-tornadic) supercell storms in general how should I look at these parameters instead? I assume the LCL height is not as important for example.

* What approach should I have in terms of looking for Lift? I know drylines and frontal boundaries cause lift but what map should I use and what should I be looking for?

 

[Taylor Wright]

I was in a similar position recently, trying to learn everything online, so I'll put in my two cents (correct me if I'm wrong about anything). One thing that helped me a lot was just taking multiple spotter courses from multiple sources. One might have a bit of information that the other didn't, and the repetition will help you memorize everything. This will give you a basic understanding of thunderstorm development, fronts, severe criteria, parameters, etc. My favorite course was the one I took from the Spotter Network website, which you might have already taken:

http://training.spotternetwork.org/

I also took an introductory level atmospheric science class at my university, which helped tremendously, so if you have something like this available to you you could try that.

Once you have a solid understanding of basic thunderstorm development, you can look more into the parameters you listed above. I'll give a brief summary of each one here (take with a grain of salt):

-EHI (Energy Helicity Index)-basically a combination of helicity and CAPE

-CAPE (Convective Available Potential Energy)- In a fluid, if something is warmer than it's surroundings, it rises due to buoyancy. So if you imagine a parcel of air rising through the atmosphere, if it stays warmer than it's surroundings, it continues to rise. CAPE is basically a measure of all of the potential energy possible IF the parcel is given a push upwards (past the LFC) until it is warmer than it's environment. Once it is past the LFC, in will continue to rise until it reaches the equilibrium level (where the parcel of air becomes as cool or cooler than the environment again). The greater the temperature difference between the parcel of air and its environment while rising, the greater the CAPE. The mathematics behind CAPE is somewhat complicated and you definitely don't need to learn it. Just learn how to visualize CAPE on a skew-T.

-LCL (Lifted Condensation Level)- the level at which a rising parcel of air becomes saturated (100% relative humidity). This is basically the cloud base height. This is so important because high LCLs indicate high cloud bases, which greatly reduces tornado potential, in general.

-LFC (Level of Free Convection)- the level at which a rising parcel of air can continue to rise without "help". Basically the point at which CAPE becomes a factor. You can have tons of CAPE, but if you don't have a source of lift (dryline, cold front, outflow boundary) to get you to the LFC, it won't matter. This is why low LFCs are important.

-CIN (Convective Inhibition)- You didn't say anything about this, but it's very important, and is another parameter calculated using a skew-t diagram. CIN (otherwise known as the "cap") acts like a lid on a pot of boiling water. CIN is caused by a temperature inversion, which means that temperature is increasing with decreasing pressure (height) in the atmosphere. This prevents parcels of air from rising because they will now be cooler than their environment. CIN is the reason you need a source of lift to get a parcel of air to the LFC. You want enough CIN so that storms don't pop off too early in the day before the atmosphere is ripe, killing your chase day. A little bit of CIN also helps keep supercells discrete. Too much CIN, however, can keep storms from initiating at all, resulting in a clear-sky bust.

Other skew-T based parameters you should learn: lifted index, lapse rate, equilibrium level

Once you have an understanding of the parameters above, you should learn how to find them on a skew-T diagram. MetEd offers a great online tutorial for skew-Ts:
https://www.meted.ucar.edu/search/search_results.php?hq=site%3Ameted.ucar.edu&cx=012446052473863902991%3Ar8nkgnwzzsc&cof=FORID%3A11&q=skew-T+mastery


I'm sure you've heard of TwisterData. Go there once you have an understanding of all of the parameters involved in forecasting severe weather to get your forecast data. It takes some time to get used to, but you can look at it during severe days that you aren't chasing and get an idea of why things happened, or maybe why they didn't. This will help you learn to forecast, but takes time. Forecasting is not something you can just read a book about and know how to do.

http://www.twisterdata.com/index.php?prog=forecast&model=GFS&grid=3&model_yyyy=2014&model_mm=03&model_dd=25&model_init_hh=18&fhour=00&parameter=WSPD&level=300&unit=MB&maximize=n&mode=singlemap&sounding=n&output=image&view=large&archive=false

As for you last three questions:

1) Your approach is just like everyone elses, look up parameters and see where the best place to go is. The only flaws are that obviously models aren't always correct, and sometimes it's hard to determine which parameters to go for. For instance, you might have good CAPE in Wichita, with okay shear, but better shear just north in Salina, with less CAPE. Where should you target? This is a problem I encounter EVERY TIME I chase. Making better decisions just comes with experience.

2) Generally days where you have good parameters, but high LCLs are the days you are going to have tornadoless supercells. Also, southerly as opposed to southeasterly surface winds reduce helicity and therefore can indicate it might be more of a linear event (squall line, MCS/MCV). Someone else might have a better answer for this question.

3) Lift is caused by drylines, cold fronts, convergence, and outflow boundaries. Luckily for storm chasers, these are usually associated with each other. A upper-level trough causes divergence (spreading out) in the upper-atmosphere, which is accompanied with air rushing upwards at the surface to replace the air moving above. Basically, upper-level divergence causes convergence at the surface. Convergence is indicated by a low-pressure system at the surface. Low pressure systems bring in cold-fronts behind them as they move eastward. They also suck moist air northward from the Gulf of Mexico, which is why you get enhanced moisture with a low, and is why the dryline follows the low like a cold front does. These fronts can be seen easily on surface observation maps. They act like a wedge, pushing the less dense, warm air upwards through the cap.

I hope this helped a little. Like I said, take everything with caution, as I have only been doing this for a few years. But I figured you might benefit from an explanation from a fellow amateur. If anything is incorrect, other members feel free to correct me.

-Taylor W.

 

[Stan Rose]

Christoffer, it looks like you are doing fine in your self-education; as Taylor said what you need to do is spend time in the field and match up your experience/observations
with your knowledge of the parameters and meteorology. One day in the field is worth a month staring at online guides.

One comment: the parameters you listed are probably more directed towards severe storms rather than tornadic. Probably most 'slight' risk days meet or exceed your values.
For tornadic cells, you are more interested in low level shear (hence, the importance of boundaries like fronts) and the quality of the shear to determine what mode of strom structure you might expect. Instead of focusing on 'shear of x knots' for example (though it certainly can be useful to look at maps of bulk shear) you might be better off looking at winds at various levels to get a feel of the vertical profile (a hodograph, for example). If you know basic vector math you can get a feel for the shear.
Learning to recognize boundaries on maps and images is obviously important, but also be aware that tornadic storms can form in the warm sector far away from any fronts--so in some cases you need to be aware of both the parameters and upper level features such as short waves or jet streaks.

Taylor, there's nothing blatantly incorrect (that will doom you as a storm chaser lol) in your response but there are some oversimplifications. If i have time i might be able to clarify later but as i said it aint fatal

 

[ChristofferB]

Thanks for your responses and clarifications. A few comments and responses:

Stan:
- Interesting that these numbers could work for non-tornadic cells as well. I used the lowest number in the range which may not cause a tornado.
- I would love to chase more but I live in Sweden so I'm quite limited to online learning at the moment. The second best I can do at the moment is to shadow chase / armchair chase. I'm preparing for a real chase in May with an organized tour but hoping to find means to do yet another though.
- Regarding your answer about tornadic supercells starting in a warm sector far from boundaries. What are typically the lifting reasons in that case?

Cross93:
- Thanks for the clarification of the different definitions. I actually understand them quite well but they are excellent for future reference! Maybe that's what you intended as well.
- Skew-T seems to really be the key to a lot and something I will look further into and really try to master - and visualize!
- Regarding your answer #2 about southerlies vs southeasterlies. You mention that southerlies are more likely to produce squall lines. Is that because they have drier air (from Mexico/Texas) rather than moist air (from the Gulf)? I would have guessed shear would be a more important factor in terms of whether a storm would become a supercell rather than a squall line?

My major question regarding lift / fronts etc:

a. Which would be your preferred map, or maps, to look for this on?
b. What exactly would you be looking for on that map? Like if you see a trough on a map, where would you be looking for boundaries and fronts in relation to that. I know troughs have a cold front and a warm front (and sometimes also a dryline) related, and drawn out on the map, to the trough. Is it those boundaries we are aiming for?
c. How do you determine whether a boundary is strong or not? I.e. if it has "good" lift, or not.

 

[Mike Johnston]

Remember that the strength of lift should be viewed relative to the other ingredients for severe storms. If the atmosphere is very juicy, sometimes all it takes is a nudge to provide a spark. To locate main synoptic features such as fronts and dry lines, look at the temp/wind/dewpoint map under the "basic surface" tab on the SPC mesoanalysis page. Look for tight temperature and/or dewpoint gradients and areas of abrupt changes in wind direction.

Beyond the main synoptic map, you will find that more subtle boundaries and sources of lift can be found out in the warm sector such as:

1. Pre-frontal trough: where there is an elongated area of low pressure out ahead of the main front. Look for waves or kinks in the isobars.

2. Mesolow: a more localized area of low pressure. You may have to search more detailed sources such as actual current barometer readings from local stations to identify these.

3. Outflow boundary: a mini-cold front generated from prior thunderstorm outflow, many times from the previous day! Sometimes, you can see clues on visible satellite pictures. Even on radar, these may show up as fine lines when daytime heating starts up the next day. These lines will often be moving in a direction independent of the mean flow or expected storm motion. What is the radar actually seeing? Often, flying insects caught up in the local change of air masses!

4. Thermals: Simply where heated air rises from near the ground. The same effect that allows a hawk to soar seemingly without effort on a warm, sunny day. Often, thermals can emerge in areas of breaks in the clouds, where radiation works its magic, which you can observe via satellite. Otherwise, don't look for them on any weather map.

 

[Rob H]

Lift is by far the hardest ingredient to assess. It's probably easiest to infer based on the type of lift. Have you been to haby's hints yet? There is a page with just four links to lift, shear, instability, moisture. On the lift page there are examples of sources of lift with further links to more detailed topics.

Generally speaking drylines and outflow boundaries are subtle enough to give you isolated storms. Warm fronts are good for one or two isolated storms for a bit, cold fronts are usually too strong to promote isolated cells. All sorts of other things can factor in that we still don't understand. Google for Parkersburg and gravity waves, or 5/3/99 and cirrus holes/convective rolls.

Most of the lift analysis is going where the other 3 ingredients are present, and hoping that the lift is enough without being too much. I would honestly ignore most of the charts/maps for analyzing mid-atmosphere lift unless you're getting hardcore into the science. The simulated reflectivity on models like the HRRR do a much better job of figuring out PVA/divergence/etc in regards to capping and other factors than any human can.

edit: here's the page, http://www.theweatherprediction.com/severe/ingredients/

 

[RRogers]

You should definitely check out www.meted.ucar.edu. The site requires registration to view the modules but is free. There are tons of applicable modules that cover severe weather but check out the one on skew-t mastery (http://www.meted.ucar.edu/mesoprim/skewt/) as a starting point. Good luck!

 

[ChristofferB]

Thanks once again, it seems like Lift will be something I need to read up more of. I have read plenty of Haby Hints, some MetEd-courses (they are awesome!) and I am still trying to get it all together.

Mike: Thanks for the tips, I will try to google examples of those. This is typically my main challenge in learning: how to translate a theoretical concept into the equivalent map/visual look.

RobH: That was some fantastic videos and photos that showed up for gravity rolls, cirrus holes. I recognize seeing stuff like that on "Cool link of the day"-type of websites. So, is it a common approach to lift by first trying to find the shear, instability and moisture and then check for lift - rather than the opposite way around? It seems easier and it seems also that lift might be the most difficult and most dynamic variable to track down.

RRogers: Skew-T seems like the best tool there is out there (as long as you know where to look) and I think I have it figured out quite well now (thanks to e.g. MetEd).

Since we have two Slight Risk-days coming up, I think it is a perfect time to get to practise!

 

[ChristofferB]

I was reading the Forecast discussion at http://www.stormtrack.org/forum/showthread.php?30352-2014-03-28-FCST-TX-LA-AR-MS and was really excited I understood pretty much everything!

One thing I wonder though, from this discussion, is that they were discussing 1.000 - 1.500 CAPE as being quite bad. I thought that was decent/good CAPE? So, what kind of values do we want on CAPE to be a bit excited? (I know it has to be in relation to shear etc).

 

[Mike Johnston]

I was reading the Forecast discussion at http://www.stormtrack.org/forum/showthread.php?30352-2014-03-28-FCST-TX-LA-AR-MS and was really excited I understood pretty much everything!

One thing I wonder though, from this discussion, is that they were discussing 1.000 - 1.500 CAPE as being quite bad. I thought that was decent/good CAPE? So, what kind of values do we want on CAPE to be a bit excited? (I know it has to be in relation to shear etc).

Well, just as a very general guideline:

0 - 1,000 j/kg = Marginally unstable
1,000 - 2,500 j/kg = Moderately unstable
2,500 - 3,500 j/kg = Very unstable
Over 3,500 j/kg = Extremely unstable

Of course, these are just general guidelines. And there are other factors to consider, such as the CAPE profile (eg. is it tall and skinny or fat?). Also, remember the key word in CAPE is "potential." If the surface parcels don't rise to their LFC and there are no storms, the values turn out to be not so important.

 

[Dan Robinson]

The necessary CAPE needed for tornadoes is different in every situation. Yesterday, I'd call 1,500 J/kg very good for that particular setup in MO. The better the directional shear (especially in the low levels), the less CAPE is needed for a tornado threat. It is common for cold-core events to produce tornadoes with 500 J/kg. In general, the closer you get to a low pressure center, the less CAPE you need. The farther away from a low you get, the more you need.

It's important to distinguish between suface-based CAPE (SBCAPE) and other types (MLCAPE, MUCAPE). Instability rooted closer to the ground is more able to stretch vorticity in boundaries and "spin up" tornadoes. 500 J/kg of SBCAPE on an occluded front very close to the center of a low (a cold-core setup) is more chaseable than 2,000 J/kg of MUCAPE where the bulk of the instability is elevated.

It's common for chasers to get fooled by CAPE bullseyes in mesoanalysis and model charts. Intuition tells you to go into the CAPE bullseye - when more often than not, the better tornado threat is at a boundary on the fringes of the CAPE axis.

 

[JamesCaruso]

Cross93:

- Regarding your answer #2 about southerlies vs southeasterlies. You mention that southerlies are more likely to produce squall lines. Is that because they have drier air (from Mexico/Texas) rather than moist air (from the Gulf)? I would have guessed shear would be a more important factor in terms of whether a storm would become a supercell rather than a squall line?


Didn't see anyone answer the above question for you, so figured I would give it a go.

Southerly winds can be quite moist and are not necessarily drier. Shear is indeed a factor and that in turn is related to surface wind direction (and speed). You want clockwise turning with height. A southerly wind at the surface, veering to southwesterly at mid-levels, produces less shear than a southeasterly surface wind veering to southwesterly up above.

Another factor in convective mode is the amount of forcing/lifting. Cold fronts tend to produce squall lines, while drylines tend to generate more isolated storms.

 

[ChristofferB]

Thanks for the clarifications regarding the CAPE. I guessed there would be tons of exceptions to my "model" but I guess it is just a matter of practice and experience. I have not yet been able to make a "full circle" in terms of looking at a forecast, trying to determine which locations might have the right settings and then also be able to watch the storms develop to see how right I was.

I have only used the mesoscale analysis map at SPC to look at Shear, Moisture etc. but which map (url?) do you look at to forecast parameters like helicity, dewpoints, cape etc? The Mesoscale analysis only provide a Now-analysis (+6h), right?

 

[Mike Johnston]

I have only used the mesoscale analysis map at SPC to look at Shear, Moisture etc. but which map (url?) do you look at to forecast parameters like helicity, dewpoints, cape etc? The Mesoscale analysis only provide a Now-analysis (+6h), right?

There are many different forecast maps on many different web sites. For example, here is a very basic one:

http://weather.rap.ucar.edu/model/index.php?model=eta

On this site, you may choose one of three different models, choose the forecast hour you want (get familiar with z times!) and then select the forecast parameter you want to look at.

Everyone has their favorite forecast site(s), they're all slightly different as far as navigation, graphics, etc. so you'll just want to experiment around.

 

[ChristofferB]

http://weather.rap.ucar.edu/model/index.php?model=eta

Oh, yeah! I've seen that before but had forgotten about it. Thanks!