Severe Weather Forecasts and Outlooks

Severe Storm Parameters

The "Cap"

    A lot of folks have been asking me how we know that there is a layer of warm air aloft. We do that through soundings of the upper atmosphere, accomplished twice a day by sending balloons up through the atmosphere with equipment to measure temperature and humidity. This information is then transmitted back to a ground station where meteorologists plot the data on a chart called a skewT-log p chart. Here's a skewT-log p chart from Fort Worth for the 7am CDT sounding on May 11, 2000:

On this chart, temperature is shown by the solid red line, the dewpoint is the solid pale green line, and the height above the ground is shown by the horizontal solid pale blue lines. The height is not really height, however, it is pressure (in millibars). Roughly speaking, 850mb is about 5,000 ft above mean sea level (msl), 700mb is about 10,000 ft above msl, and 500mb is about 18,500 ft above msl.
Temperature and dewpoint are in degrees Celsius. Where the temperature and dewpoint traces are close together, humidity is high. Where they are far apart, humidity is low. To read temperature, you have to look at the solid blue lines running from lower left to upper right (i.e. the temperature part of the chart is "skewed").

The "cap" on this chart is the part of the temperature trace (solid red line) which juts sharply to the right between 900 and 800 mb. A parcel of air might rise from the surface to that point, but would be constrained from rising above that level because the air would be much warmer than the parcel. That's the cap! Some indices important to forecasting convective storms are generated automatically and appear across the top of this chart. Some of these are discussed in the next section. We can compute the important indices either from actual soundings, like the one shown here, or from computer output from which model soundings are created.

One other point of interest: because of the "cap" shown on this diagram, the T(c) shown along the top of the chart means that the surface temperature would need to reach 110F to break the "cap" on this particular day.


    Let me introduce you to two new parameters that are important to the issue of whether tornadoes are likely to occur in association with severe thunderstorms. These are: lifted condensation level (LCL) and level of free convection (LFC). These parameters are computed using the skewT-log p chart discussed in the previous section. To understand these two parameters and their importance, we need to cover some fundamental principles of thermodynamic theory. Don't let that big word ("thermodynamic") discourage you from reading on. This stuff gets easier.


    Temperature in the atmosphere is subject to what is called a lapse rate. This refers to the change in temperature with height above the surface of the earth. Typically, temperature declines with altitude, although you know from the previous section that there are exceptions to this rule. The rate at which temperature declines with altitude or height above the earth is called the lapse rate. The lapse rate is not a constant, however. In dry air, the rate is ~5.5 degrees F per 1,000 feet; for moist air, the rate is ~3.5 degrees F per 1,000 feet. Another rule that applies to lapse rate is that the air temperature will decline at the dry lapse rate until the air is saturated (i.e. the relative humidity is ~100%), then the temperature declines at the moist lapse rate (3.5 degrees F per 1,000 feet). You're probably accustomed to hearing about temperature and dewpoint. "Dewpoint" is the temperature to which the air must be cooled in order to condense the moisture in the air from vapor to liquid form. If air parcel is lifted at the "dry" lapse rate (5.5 degrees F per one thousand feet) and hold the amount of moisture steady (neither decreasing nor increasing it), at some point above the surface, the two values (temperature and dewpoint) will converge, and the air will become saturated (or nearly so). That point is the lifted condensation level (LCL).

    After a parcel has reached the LCL, if we continue to lift it higher into the atmosphere, it will cool at the saturated rate (3.5 degrees F per 1,000 feet). Depending on the thermal structure of the atmosphere, if the saturated parcel reaches a level at which it is warmer than the surrounding air parcels (called the environmental temperature), it will begin to rise on its on (because warm air is less dense than cool air). The point at which an air parcel begins to rise on its own (because it is warmer than the surrounding air parcels) is the level of free convection (LFC). And the extent to which the saturated parcel is warmer than the environmental air determines its buoyancy. The higher the buoyancy (i.e. the greater the temperature difference between the parcel and the environment), the greater the upward acceleration due to buoyancy.

    Now that we have defined the two terms, let's look at how they are determined from the skewT-log p chart. We won't go into great detail because, although this can be done manually, there are computer programs and websites that will display this information on entry the proper commands. Take a look back at the skewT-logP chart shown above. Notice that along the left side of the diagram, there are tick marks identifying, among other things, the LCL and the LFC. Note that the LCL is just above the blue line which marks 900mb. Also note that the red line (air temperature) and the green line (dewpoint) converge at that level. That's where the air parcel "lifted" from the surface would become saturated on this particular day and at this particular station. Recall that on this diagram, we see a very pronounced cap, with very warm air just above the LCL. Because of that feature, the LFC is way, way up in the atmosphere on this particular day and at this particular station. For convective storms (thunderstorms) to occur, these two values must generally be much closer together. (Of course, if the cap erodes during the day, the LFC may gradually lower, but that's another story.)


    The relationship between these two parameters and the process by which tornadoes are formed (tornadogenesis) is the subject of intensive research at this time. It appears that tornadogenesis is favored when both the LCL and the LFC are relatively low (i.e. close to the ground). Informed speculation by tornado researchers (as well as analysis of detailed storm-scale data collection, as was done in the VORTEX program) suggests that buoyancy is maximized when the LCL is low, and that evaporative cooling of the rear-flank downdraft (RFD) is also reduced. Since the RFD is involved in "wrapping-up" the converging winds of a mesocyclone, it appears to be a positive factor (perhaps a necessary factor) that the RFD be relatively high in equivalent potential temperature (theta-e), and this condition is more likely if the RFD is not excessively cooled by evaporation as it approaches the surface.

    It also appears to be a positive for tornadogenesis if the LFC is not too far above the LCL. Although we don't completely understand why this is so, it is likely related to the fact that the environmental air above the LCL in severe storm situations is often quite dry. If the saturated parcel has a longer distance to travel between the LCL and the LFC, it may begin to become unsaturated (that is, the updraft containing the saturated air may begin to entrain drier environmental air, thereby reducing the relative humidity of the ascending air). In addition to the entrainment issue, recall that it is at the LFC that a parcel begins to accelerate upward as a result of its own buoyancy; thus, for buoyancy to contribute to upward vertical acceleration close to the surface, the LFC needs to be close to the surface.

Computer Output of Severe Storm Parameters

I am often asked how computer weather forecasts are used in making severe storms outlooks and forecasts. I don't have the time or space to go deeply into the various computer models and how they are structured. But a brief summary may be useful, so here it is.

Surface weather observations are taken around the world in virtually every nation as well as over international waters. These observations are standardized as to how they are taken and what parameters are measured, and the data collected is shared by countries with little or no restriction on the dissemination of the observations. In addition, twice per day, balloons are launched which carry instrument packages to directly measure weather parameters aloft. In addition, data is received almost continuously from weather satellites, some of which are capable of vertical soundings through the atmosphere as they "look" down to the earth. And many commercial aircraft are now equipped with sensors which constantly measure some weather parameters and periodically send this data back to a ground station.

All of this information flows into the National Center for Environmental Prediction (NCEP), as well as to other national or regional forecast agencies in other parts of the world. That center processes all of the data and periodically initiates computer analysis of the current state of the weather (both surface and aloft) and commences the running of computer models which take the current atmospheric conditions and step them forward in time and space, using physical equations which are applicable to all fluids, including air. These models are programmed to generate output on a three-dimensional grid which can be related to specific points on and above the earth's surface. NCEP runs several models: ETA, mesoETA, RUC, AVN and MRF, and variations of those main ones. The computers used are very fast, and the ETA (which extends currently to 48 hours but may soon extend to 60 hours) is usually available within three hours of the time the observations are taken!

The ETA model, which is run at NCEP, is one of the models of choice for forecasting in the 24 to 60 hour time frame, although there are other models which cover the same periods, including the meso-ETA  and the MM-5. From time to time one model may appear to have a better handle than the other on evolving weather systems. The output from these computer models is massive, so one has to decide what output is significant, and that depends, to some extent, on what kind of weather pattern you are dealing with. For convective storms (thunderstorms), I like to look at CAPE, Lifted Index (LI), Storm-Relative Helicity (SREH) and the Energy Helicity Index (EHI). But, I may also look at many other products in making a forecast.

CAPE is (<c>onvective <a>vailable <p>otential <e>nergy), a measure of the potential buoyancy of atmospheric parcels if they are given a nudge upward in the atmosphere by some force. The higher the CAPE, the more buoyant the parcels, and the faster they will ascend after being disturbed. CAPE actually has units, joules per kilogram (j/Kg), but is often referred to without stating the units. Since the most intense storms occur when the air near the earth's surface is warm and moist, we tend to look at CAPE which is either surface-based, or averaged over some portion of the lower atmosphere.

CINH is (<c>onvective <inh>hibition) is a measure of the resistance of the atmosphere to upward vertical motion (you can think of it as anti-CAPE).  As long as there is substantial CINH, parcels will resist upward motion even if they are given a nudge upward in the atmosphere by some force. However, if the force applied is sufficient to lift the parcels above the level of CINH into a region which is not "capped", then the parcels can rise in response to positive CAPE. By implication, then, CINH has the same units (j/Kg) as does CAPE, but CINH will always be a negative number. As a very general proposition, various factors can overcome CINH <-75 j/Kg but as the numbers increase beyond that level (e.g. to >-100 or more) it becomes harder to overcome the negative bouyancy of CINH.

Lifted Index is a measure of instability in the atmosphere based upon the temperature of a parcel which is "lifted" from the surface to an arbitrary height above the surface. The "lifting" of the parcel can result from buoyancy along, or with the aid of a frontal surface, or as a result of topographical features. If the lifted parcel is warmer than its environment, it will continue to rise until the temperature of the parcel and the environmental temperature are equal. So, the atmosphere is positively buoyant and unstable (parcel will continue to rise) if whenever the environmental at a particular altitude is colder than the rising parcel, and the lifted index (LI) will be a negative number in those cases (i.e. the index is the difference between the parcel temperature and the environmental temperature). The units involved in <LI> are actually degrees celsius (C) but, as with CAPE, we often dispense with the units when mentioning this index. And, it is measured normally at 500 millibars (mb), which is a pressure level whose actual altitude varies from day to day.

Helicity is a measure of the change of wind direction and/or wind speed with ascent through the atmosphere. The computation of this  parameter is quite complicated, so I won't try to describe it. Most of the time, we speak of storm-relative environmental helicity (SREH), which is computed by taking the forecast wind profile up through the atmosphere and then assuming a storm motion based on this forecast flow, and then computing the helicity relative to the forecast storm motion. Generally, helicity needs to be over 150 for storm updrafts to rotate and at values over 250, the threat of tornadoes increases. However, recent studies show that there are so many variables resulting from local variations of conditions and deviant storm motion that helicity may not be as useful as it was once thought to be. And recent studies have shown that both CAPE and helicity are important and a deficit in one may be compensated by a surplus of the other, at least in some cases.

To try to account for the importance of both CAPE and helicity, we also look at a unitless parameter which combines both CAPE and helicity into one parameter, called the Energy Helicity Index (EHI). EHI normally ranges from 0 to as much as 5 or 6. As a general rule, values over 3.0 are associated with the more violent storms and values toward the higher end of the scale may be associated with tornadoes.

It is important to keep in mind that you can have elevated values of some or all of these parameters and still not have severe storms. If the atmosphere is "capped" by a layer of warm air, or if there is no mechanism for nudging that parcel of air into initial ascent, then storms may not occur. You may see me use the term "insolation" from time to time, and wonder what that means. It is not a misspelling of "insulation". "Insolation" is a contracted form of "<in>coming <sol>ar radi<ation>" and is defined in the Glossary of Meteorology (AMS, 1959, 1980) as meaning "solar radiation received at the earth's surface".

Insolation is an important issue in many severe storm cases because it is the heating of the earth's surface by incoming solar radiation which in turn heats the air closest to the earth's surface, causing that air to rise and thereby creating vertical currents which carry moist, unstable air aloft. In other words, the convective process often requires insolation in order to initiate the development of convective storms.

As time permits, I will try to add additional information on the subject of computer models and their application to severe storms forecasting.

This page was last updated on 4-6-2002.

f you have comments or suggestions, email me at

This page created with Netscape Navigator Gold