Lowest LI ever?

All that Lifted Index, and no tornadoes occurred within the map.

Tornadoes that occurred July 13, 1996 according to NCDC Storm Events.
Colorado had 4 tornadoes: 2:40 pm, 5:40 pm, 9:45 pm and 9:55 pm,
Maryland had 4 tornadoes:L 12:45 am, 1:05 am, 1:45 am and 2:35 am
Viriginia had a tornado at 12:30 am.

Two comments:

1. The map indicates 20Z, and there are no radiosonde observations at that time, so the results stray from the realm of hard data and into estimated territory. It's unknown what environment we're comparing the parcel against. There are also unknowns regarding how the starting parcel is computed.

2. The 00Z soundings for that date (actually 14/00Z) don't indicate any LI near -17, or even -7 for that matter. I would start looking at the source of where the info came from... maybe it was a test case that proved erroneous.

KILX 7/14/96 00Z
What's a lifted index?
Lifted index: Galway developed it, back in 1956.

Lifted Index. Temperature difference between the environmental and parcel temperatures at the 500 mb level.

LI= 500 mb envir. Temp - 500 mb parcel Temp
LI = T500 - TP500
Source: NWS A Comprehensive Glossary of Weather Terms for Storm Spotters

Lifted Index (or LI) - A common measure of atmospheric instability. Its value is obtained by computing the temperature that air near the ground would have if it were lifted to some higher level (around 18,000 feet, usually) and comparing that temperature to the actual temperature at that level. Negative values indicate instability - the more negative, the more unstable the air is, and the stronger the updrafts are likely to be with any developing thunderstorms. However there are no "magic numbers" or threshold LI values below which severe weather becomes imminent.
Source: NWS Springfield
The LI is a measure of potential instability from the surface to 500 mb. Lift a parcel with an average mixing ratio and dry adiabat in the lowest 100 mb of the sounding. It is very similar to the Showalter Index, but better considers available low level moisture below 850mb.
Source: NWS Louisville
Lifted Index (LI)
The LI is a commonly utilized measure of stability which measures the difference between a lifted parcel's temperature at 500 mb and the environmental temperature at 500 mb. It incorporates moisture and lapse rate (static stability) into one number, which is less vulnerable to observations at individual pressure levels. However, LI values do depend on the level from which a parcel is lifted, and rally cannot account for details in th environmental temperature curve above the LCL and below 500 mb. LI was originally intended to utilize average moisture and temperature properties within the planetary boundary layer.

LI = T(500 mb envir) - T(500 mb parcel)

in degrees C, where T (500 mb envir) represents the 500 mb environmental temperature and T (500 mb parcel) is the rising air parcel's 500 mb temperature.

LI over 0: Stable but weak convection possible for LI = 1-3 if strong lifting is present.
LI = 0 to -3: Marginally unstable.
LI = -3 to -6: Moderately unstable.
LI = -6 to -9: Very unstable.
LI below -9: Extremely unstable.

These LI values are based on lifted parcels using the average lowest 50 to 100 mb moisture and temperature values (i.e., the boundary layer). Variations exist on how LI values are calculated,

Surfaced-based LI: Surface-based LIs can be calculated hourly, and assume a parcel is lifted from the surface using surface-based moisture and temperature values, as well as assigned environmental temperatures at 500 mb. This method is valid for a well-mixed nearly dry adiabatic afternoon boundary layer where surface characteristics are similar to those in the lowest 50 to 100 mb layer. However, these values would not be representative of the ambient elevated instability if a nocturnal inversion or shallow cool air to the north of a frontal boundary is present. In these cases, more instability resides above the surface, and parcels may be lifted to form thunderstorms from the top of the inversion.

Best LI: The Best LI represents the lowest (most unstable) LI computed from a series of levels from the surface to about 850 mb. This index is most useful during cases when shallow cool air exists north of a frontal boundary resulting in surface conditions and boundary layer-based LI values that are relatively stable. However, the airmass at the top of the inversion, from which lifting may occur, is potentially unstable. An example of this would be elevated ("overrunning") convection (possibly a nocturnal MCS).

OK, if we're gonna have a CAPE/LI slugfest here's the two granddaddies that I'm aware of.

Peoria IL 8/29/90 - 00Z (Plainfield tornado)

Del Rio TX 5/28/97 - 00Z (no real "proximity" sounding to Jarrell due to giant RAOB hole in Texas)

For the newbies - the dates are a day off from the events since these are 00Z soundings and the UTC date rolls over in the evening.

The highest CAPE / lowest LI's I've seen are definitely not in the U.S. but in Bangladesh... just cruise on over to John Finch's site...
Not too many places can claim to have 9000-10000 CAPE ... Granted, the surface and upper-air network over there is poor in the least, but still...

Geez, look at the depth of the troposphere... that EL is off the chart! (That would put it about 55,000 ft). One would have to go to the TTCC/TTDD data to estimate storm tops but extrapolating this I'd say they're certainly in the 70-80,000 ft range.
Are such high values of CAPE/LI possible without a very strong cap. I'm not a Meteorologist, but you would think the imbalance between the surface and upper air would correct itself before LI's got to large(without a strong stable layer near the surface). Hence, the plotted charts do not necessarily reflect reality. :)
Originally posted by Steven Williams
Are such high values of CAPE/LI possible without a very strong cap. I'm not a Meteorologist, but you would think the imbalance between the surface and upper air would correct itself before LI's got to large(without a strong stable layer near the surface). Hence, the plotted charts do not necessarily reflect reality. :)

I don't see why it was not be possible... The atmosphere is conditionally unstable when lapse rates are between ~6-9.8C/Km (between moist and dry adiabatic lapse rates). There will be absolute instability and thus "automatic" convection should lapse rates become superadiabatic -- say 11C/km. This happens sometimes (oftentimes?) out west in the deserts, where surface heating can happen quickly enough or on an already-mixed parcel to create superadiabatic conditions. Whatever the case, the Bangladesh sounding is dry adiabatic from around 800-600mb, which isn't too uncommon, indicative of EMLs. The tropopause is extremely high, yielding a much "taller" sounding from which to contribute to CAPE.

The sounding above has 0 convective inhibition/cap, so only the slightest forcing would likely be necessary to initiate deep moist convection... This would be an issue of this sounding was representative of a large area, but I don't know if this is/was the case.