Australian Severe Weather Forum

On this site, and in other places, I have seen terms such as CAPE, lifted index and wind shear etc.

Firstly, could anyone tell me whether there are any other measures of conditions which may effect the potential for a severe storm, and what do they mean.

Secondly, does anyone know how I might be able to find current figures, what charts I could use to use to work it out (and possibly how they should be read), or how I can observe it myself?

Really there are a whole swathe of values that can be used to help interpret the potential for severe storms...: the three you have mentioned are some of the more popular values but others include:

TOTL: Total Totals= (T(850hPa level)-T(500hPa)) +(Td850 - T500)
Where Td is the dewpoint temperature at 850hPa. Values >~50 indicate possiblity of deep convection.

SHOW: Showalter index:=T500-Tparcel
where T parcel is the temperature of a parcel of air lifted from 850 to 500.
Negative values indicate deep convection.

LIFT: Lifted index = T500- Tparcel. T of a parcel lifted from lowest 500m to 500hPa values <-6 indicate possible deep convection

CAPE: Convective available potential energy: the energy available to a parcel when it becomes buoyant. The positive area between the sounding curve and the parcel trajectory on a sounding diagram.

Windshear is simply the change in wind speed with height calculated from vector components.

Now CAPE, SHOW, LIFT, TOTL are all calculated from sounding diagrams(you can access these through either the University of Wyomings Upper Air Sounding division, or Brisbane storm chasers also have a good sounding program available), and basically are indicators of the possibility of convection: CAPE is notoriously difficult to calculate by hand, and hence unless available to you on a sounding is near useless, however when available gives a good idea as to the strength of a convective updraft maximum if it forms: Max vertical velocity = Sqrt(2*CAPE). Also in general CAPE over about 1000J/kg indicates the possibility of convection. If you need to calculate by hand then one of the others might be for you, although lifted indexs are usually pretty reliable.

Another important number is the CIN: The convective inhibition: or the energy require for a parcel to reach its level of free convection(become buoyant). Values less than 100J/kg are usually a good sign that convection is liable to occur if you have strong daily heating(EG a summers day), Large values (up to 300J/kg) mean that convection will struggle to break the cap unless strong forcing occurs(say a cold front). Small values/negative values can infer the atmosphere is freely convecting: in other words, the whole surface is positively buoyant.

Finally Windshear is important for further development of storms and defines whether convection is going to be a pulse storm, air mass, squall line or supercell.

So really there are plenty of values that can be used to help predict storms: A sounding diagram(a Skew-T is the most common form) will provide all but the wind shear, and if you desire you can calculate it too by looking up the side of the diagram where you see the wind barbs. Key things to look for are few barbs lower down and strong winds aloft, or a gradual increase in winds but an important change of direction. These not only allow you to work out the types of storms that should form but also their direction of movement.

Hope this helps, and have fun chasing

Rather than supplying links, would you like to compile explanations of the terms mentioned above:)

Nice post thanks John.

I've got a GFS forecast table on our weather site that attempts to show the convection around my home town, eg 'isolated showers' 'severe thunderstorms' etc, but at the moment it only looks at CAPE, so it is obviously somewhat deficient.

Does anyone have suggestions on combining some of these other parameters, such as shear and maybe LI, CIN, cloud cover, to come up with a useful forecast of the storm activity?

Best wishes,
  Dave

Hi David,

Hopefully there is something you are able to find.

I guess though that is the fun of forecasting - to combine your own experiences as well as current conditions and compare with forecast models. Even if there was something that combined all parameters, how accurate are they?

This year, more than any other recent years at least, the models have on the whole been disappointing in pinpointing where the best area of convection will occur. Success at forecasting and storm chasing has been attributed to past experiences in my case at least.

Regards,

Jimmy Deguara

Cheers Jimmy. I guess my question was more on the best way to interpret the models rather than the usefulness of the models themselves.

I was thinking of something grossly along the lines of:
potential=CAPE/500
if ( shear(850,600) > 20 ) potential=potential + 1
if ( CIN > 200 ) potential=potential - 1
etc.

While no algorithm will encapsulate what's really happen, at the moment I rely on CAPE alone and there has to be something better than that. Maybe I should just use SWEAT instead?

Cheers,
  Dave

Theres no real way to analyse potential as such, as the conditions resulting thunderstorms are so diverse. Potential, Conditional and Free instability are all possible forming conditions, and not all of these conditions are demonstrated in CAPE and other indices. Ill work on a post to explain these alternative formative methods.

As for trying to look for target areas I like to look at a few different charts.
1) Ill look at a synoptic forecast/relevant chart and see if ive got a forcing mechanism for any areas(IE a cold front, triple point, trough or convergence zone)
2) I generally look for a map of LI, and identify locations of convective values(<-2) seeing if they correspond with the evident synoptic features if any
3) Ill then overlay(or compare) a Dewpoint temperature map, and look for rather high values and see if they correspond with likely convective locations(little moisture means high based storms more often than not, if convection occurs at all)
4) Ill look for the nearest available sounding, available on the BSC site and see how the moisture profile looks, the CAPE, and the shear profile as well as CIN.
5) Ill look at 500mb wind charts to see what the propagation of storm direction is going to be like over the target area.(steering flow is usually near the top of the troposphere)

That gives me a fair idea of the possibility of convection, and a likely target area for any day storm chasing. This is probably quite a technical way to do so, but it seems to work for me.

Hi David,

I appreciate what you are saying and had thought about this concept about 2 years ago. If only I was a good programmer, I would have come up with a solution years back.

When I came across the GFS (I guess then the AVN models) back in 1998, it was a gold mine. With these models alone we were able to confidently expand our storm chasing beyond the horizon.

CAPE is a pretty useful variable as well as is surface lifted index. The sensitivity of CAPE though with only slight changes in temperature and moisture should not be underestimated. But using it as a guide is still helpful.

So any programmers?

Regards,

Jimmy Deguara

Hi John,

If you are searching for surface lifted index values off the models of the order of -6 and lower, you won't be doing much chasing in Australia! Particularly this year.

I tend to look for values of less than -2 depending of course on available moisture and sunshine. This will often determine whether the models have underestimated or overestimated the specific variables. And of course, it depends on the distance travelled and the risk for failure.

One has to be extremely careful though. On the days leading up to the Muswellbrook hailstorm, LI values of -10C were forecast for the north east NSW region. The Bureau even had hail and squally winds in the forecast up there. In the region from Mudgee to Muswellbrook, Taree to just south of Sydney, conditions indicated less forecast CAPE and lower LI values. We opted for the region near Mudgee and Gulgong. The result was a long lived supercell that tracked for nine hours. The north coast struggled to break the cap with only late night lightning in the distance as a storm(s) tracked up the coast.

Regards,

Jimmy Deguara

I'm a programmer and I'm very keen to code up something like this.

I've already got some python code that rips down bits and pieces of the NOMADS GFS data, the formatted output for my site is here:
http://narrabriweather.net/forecast.shtml

It's just that I'm new to the storm game and need some advice about the right way to put the different parameters together. Some of the suggestions already made by John are useful.

Cheers for the encouragement,
  Dave

Thanks for the correction Jimmy: I put down the ideal supercellular LI reqs rather than standard convection reqs, my apologies.

John,

No apologies required. Perhaps there could be a an indication for supercells as well givent eh complex nature of the program.

This is an important topic.

Regards,

Jimmy Deguara

I would say you need to look lower than that for steering. I tend to use the average from about 850 to 500 depending of how strong the winds are at various levels in that range. No set rules of course.

regards, Michael

John why do find LI and DP more useful than CAPE? I see the input parameters are different but isn't CAPE already an attempt to combine the convective potential of a parcel and the DP? Do you see much difference between the two approaches?

I could use a combination of exisiting indicies, such as LI or CAPE, as a general convective indicator like I currently am, but then also look at the ratio of that and something that incorporates shear, such SWEAT, as the trigger for warning of severe storms. Is it also worth looking at CIN?

Although if I go back and look at the input parameters that make these different quotients there's probably an even simpler implementation of what I'm seeking. If anyone wants to suggest an interesting combination of the GFS forecast parameters I'll code it up and put it on my site with a few of our locations so we can assess it. My current code can't generate maps, just forecasts for specific locations.

Even though they're for specific locations, I do need to change my algorithm to look at the surrounding 0.5 degree forecast points as well. There's been a few times where the reported convection was fairly low, because a front has gone through faster than the 3 hour forecast interval of the GFS, such that you don't see the storms if you just look at a single forecast point.

Regards,
  Dave

This page describes how the existing stability indicies are calculated:
http://weather.cod.edu/sirvatka/si.html

Why do i prefer LI and DP? To calculate CAPE we require a sounding for a location, now unlike in the US where they have many many more sounding locations twice or more daily, in Australia we are limited pretty well to capital cities and major centres,and only twice daily. One is a 10am and the other a 10pm EST This generally does not reflect the part of the day at which the atmosphere is most unstable and hence is normally not on the mark very often. Also it is going to be a tough chasing season when you are relying on major centres to be the target areas: quite often the urban heat island effect disrupts storm organisation as well. LI and DP dont have this problem.