The Elevated Mixed Layer
Jonathan D. Finch

Sign Guestbook

Main Page
Convective Outlooks
Special Cases for the United States
Bengal Tornadoes--background information
Historical Tornado Tracks for East India and Bangladesh
Meteorological Charts for Historical Tornado Cases for Bengal
Latitudinal Comparison of the Geostrophic Wind Approximation
Assessing Instability on the Front Range Without Upper Air Data
Potential Temperature and Mixing Ratio--Contributions to CAPE on Elevated Terrain

        The elevated mixed layer (EML) was first discussed in pioneering work by Dr. Toby Carlson in the mid to late 1960s. Carlson
        did additional work on the EML in the 1980s. He noted that the EML results in a "lid" which prevents thunderstorm activity. His
        synoptic meteorology book entitled "Mid-latitude Weather Systems" has a chapter devoted to the "lid". This book is a must read for
        meteorology students and forecasters. Most meteorology textbooks are highly theoretical and are frankly no fun to read. Carlson's
        book actually reads like a book and is interesting throughout. I was fortunate enough to enjoy Carlson's advanced synoptic
        meteorology course while attending Penn State University in 1991-92. While many of the courses I took in graduate school were
        just about deriving equations, this course was very applied and I learned a lot. Yes, we learned the components of the important
        equations and applied the equations to practical problems,  but the whole course was not based on proving one's mathematical

        The EML is important for the following reasons:

        1. The EML prevents deep, moist convection until high instability is achieved. In the absence of deep, moist convection,
            warm, moist air can flow poleward in an unimpeded manner. Daily evapo-transpiration also adds moisture
            to the boundary layer further enhancing theta-e.

        2. The EML tends to keep storms isolated. When deep moist convection occurs in a capped environment, it tends to be in
            localized areas of enhanced convergence such as along out flow boundaries, dryline and terrain features, or of course along
            frontal boundaries. Isolated storms tend to be more severe than widespread storms since there is less competition for available
            warmth and moisture.

        3. The EML along the southern edge of the westerlies prevents deep vertical mixing. Deep vertical mixing is a CAPE
            destroyer. It is very difficult to maintain high mixing ratios when very deep mixing is occurring. The cap provided by the EML
            confines the moisture to a shallow layer, preventing the mix-out effect. This effect is most apparent in late-spring and summer
            when the southern edge of the westerlies retreats to 40-45N. The high dewpoints will usually be along the southern edge of the
            westerlies where the lid is the strongest and where cold fronts stall out. Moisture convergence is also greatest along the southern
            edge of westerlies, typically just poleward of stalled out fronts or outflow boundaries where evapo-transpiration is at a maximum
            from vegetation and previous rains. The mixout effect can also occur beneath the strong capping inversion in cases where moisture
            return is extremely shallow, particularly when strong synoptic-scale disturbances are involved.

        EMLs develop when arid regions heat up and deep, dry adiabatic lapse rates extend from the surface to between 450mb and 600mb.
        Read more about elevated heating and its contribution to instability on the high plains hereEMLs can occur any time of the
        year. Of course an EML can occur along with a very stable boundary layer too. But this page is
devoted to EMLs that result in
        capping inversions above a warm and moist boundary layer. I  have found EML soundings in all
 seasons and all areas east of the
        Rockies. EMLs are slow to modify after leaving their source region since lift and subsidence have
little effect on dry-adiabatic layers.
        The EML can be carried well  downstream
  without changing character much at all.

        In general, the surface dryline marks the southwestern edge of the EML in the Great Plains. To the east of the surface dryline, the
        moist layer is capped by the EML. To the west of the dryline, the mixed layer extends all the way to the ground and is not "elevated".
        In general,  the capping inversion increases as one progresses east, away from the surface dryline. Near the dryline at peak heating,
        convective inhibition tends toward zero, but convergence must still be present on the mesoscale to develop a storm.  If storms do
        not develop on the dryline, do not assume that the dryline is "capped". This explains the laminar look to the lower part of supercell
        storms that have progressed into the capped region east of the dryline. Storms that move rapidly into a capped region may die before
        becoming a  supercell. Keep in mind that just because there is convective inhibition out ahead of the dryline, does not mean that storms
        cannot survive in this region. Once storms develop into high based supercells in the strongly heated air near the dryline, the storms
        tend to maintain themselves after moving into the "capped" but more moist environment to the east.

        So forecasters should not be fooled by point model soundings 50 miles east of the surface dryline that show a capping inversion. Of
        course the moist layer is capped at locations east of the dryline. This should be no surprise. This does not mean that storms will not
        develop along the dryline with intense daytime heating on elevated terrain.

        If the EML did not exist, then the Plains severe storm environment would be entirely different. For one, there would be
        no dryline.
For the EML not to exist, one would have to remove the Rockies and desert southwest. Then of course you
        wouldn't get a lee trough.
So like the dryline, the EML is an integral  part of  the Plains' severe storm environment.
        The EML is a basic concept that must be
understood and appreciated by forecasters.

        Oftentimes you will see a saturated or nearly saturated layer at the top of the EML. In my opinion, this is where altocumulus castelanus
        is found.  This can be illustrated using the August 27, 1973 sounding and surface obs from Flint, MI. So where did the elevated mixed
        layer originate from? To answer this question, think of what the vertical potential temperature and mixing ratio profiles looked like in the
        EML source region. The sounding from Lander, WY 12 hours earlier was overlaid on the Flint sounding. Note that the thermal and
        moisture profiles are very similar. The moisture at 525mb in the Flint sounding originated in the boundary layer over the Rockies.
        Granted, it may have taken longer than 12 hours for the EML to advect from the northern Rockies to Michigan. With afternoon heating
        over the Rockpile, relative humidity will tend to increase with height. There is often a saturated or nearly saturated layer at the top of a
        deep, mixed boundary layer. So mixing can redistribute the moisture profile to such an extent that saturation occurs. During summer when
        mixing depth is greater, this moist layer can be as high as 450mb.
        On August 28, 1990(day of the F5 Plainfield tornado), the Flint sounding showed an elevated mixed layer. Although the observation
        forms from Flint were not available, the observations from Jackson, MI showed mid cloud that was likely accas.

        This vortex sounding taken by vortex near Friona  on June 2, 1995 at 21 UTC shows an EML with a nearly saturated layer from
        480 to 500mb. Overlaid is the El Paso sounding from nearly the same time. El Paso is roughly upstream from Friona.

        The 12 UTC June 8, 1995 sounding showed a well developed EML from possibly 2 different source regions. Accas was present
        over Oklahoma based on the surface observations that showed a scattered to broken cloud deck around 18000ft.

        The 00 UTC August 30  1995 sounding from Aberdeen, SD showed an EML with the top around 500mb. The surface obsertvations
        from Watertown show a scattered to broken cloud deck at mid levels which was probably accas.
        On May 22, 1981(day of the famous Cordell tornado) the 00z sounding from OKC showed a well formed elevated mixed layer
        and the surface observations showed ACCAS at about the same time.

         Some EML sources include:

        1. Dry, elevated terrain of the interior, western United States
        2. High plains of the United States
        3. Sierra Madre Occidental of Old Mexico
        3. Western desert areas of southern Africa
        4. Desert areas of northern India
        5. Parts of Spain
        6. Saharan north Africa

        In the cool season, the EML (with positive instability) can occur over the southern United States. Nothern old Mexico
        and the southern Rockies are the source region this time of year.

        Little Rock (Nov 27, 1994)  Tornado outbreak in eastern Arkansas and western Tennessee
        Longview, TX (Nov 27, 1994)
        Norman, OK  (Jan 3, 1998)  Numerous large hail reports across Oklahoma and Texas
        Little Rock (Jan 21, 1999)  Tornado outbreak in Arkansas and adjacent states
        Shreveport (Jan 21, 1999)

        The EML can be found from time to time along the east Coast. Here are a few notable examples from prominent tornado

        These EMLs originated from the high plains and Rockies.

        Albany( Aug 28, 1973)  F4 killer tornado in Columbia county, NY and Berkshire county, MA
        Washington Dulles (Aug 28, 1973)
        Washington Dulles (July 10, 1989) Tornado family moved SSE from Montgomery county, NY to New Haven,
                                               CT to eastern Long Island. Near baseball sized hail(2.5") fell on eastern Long Island
                                               Another tornado family moved SE across northern New Jersey.
        Hatteras (Mar 28, 1984)  At least 2 tornadic supercells, one of which produced many tornadoes over a 5 hour period.
                                               57 people killed

        Here are some additional EML soundings from the United States:

        North Platte (July 10, 1977)  F3 tornado in Cherry county, NE and Bennett county, SD
        Oklahoma City (May 22, 1981)   Tornado outbreak in Oklahoma and famous Cordell tornado
        Flint (Aug 28, 1990)   Plainfield F5 tornado
        Dayton (Aug 28, 1990)
        Green Bay (Sep 6, 1995)  F2 tornado Rice county, MN
        Dodge City (June 6, 1990)  Limon, CO tornadoes
        Friona, TX (June 2, 1995)  Large tornadoes near Dimmit and Friona, TX
        Amarillo, TX (Mar 19, 1982)  Long-track F4 tornado TX, OK panhandles
        North Platte (May 17, 2000)  Large tornado at Brady, NE
        Dodge City (May 16, 1995) Tornadoes in western Kansas
        Lake Charles  (Mar 2, 1999)  Tornado kills 1 person north of Lake Charles, LA
        Norman, OK  (Apr 7, 1995)
        Norman, OK  (Jul 26, 1995)  Large hail in Oklahoma. Large tornado the day before SE of Dodge City
        Aberdeen, SD  (Aug 29, 1995)
        Bismark, ND  (Aug 17, 1995) Several 4 inch hail reports in North Dakota
        Fort Worth  (Apr 30, 1995)  Large hail reports across Texas and Oklahoma
        Norman, OK  (June 8, 1995)  Large tornadoes in TX panhandle
        Norman, OK  (Oct 3, 1994)  Large hail in Colorado, TX and OK
        Sault Saint Marie, MI  (Aug 27, 1973)  Tornadoes the next day in NY and MA
        Flint, MI  (Aug 27, 1973)
        Huron, SD  (July 04, 1978) Killer tornadoes in in North Dakota and northern MN
        Oklahoma City (Mar 19, 1948)  Tornado outbreak from OK to IL with many fatalities
        Norman(April 22, 2004)  Tornadoes in northeast Oklahoma

        Even in the hot, dog days of summer, the EML can be found. Here is an example from Topeka, KS from August 6, 1962. An
        F4 tornado hit not far from Topeka at 530 pm. The source region was likely the high plains since the top was 620mb. This is
        one of the most unstable soundings I have ever seen.

        The EML is present over East India and Bangladesh from late-March through May. The soundings below were taken at  Calcutta,
        IN unless otherwise noted. As stated above, the source region is interior north India--not the Himalayas. Any elevated airmass
        coming off the Himalayas would be detrimental to thunderstorm development. Instead of providing a low-level capping inversion,
        the Himalayan EML would provide a strong inversion around 500mb with 500mb temps from 0 to +5C.

        Apr 02, 1972 12Z
        May 21, 1972 00Z
        Apr 22, 1990 00Z
        May 03, 1992 12Z
        May 06, 1993 12Z
        May 17, 1994 00Z
        Apr 07, 1998 00Z
        Apr 21, 1998 00Z
        Apr 21, 1998 12Z
        May 28, 1998 12Z
        Apr 20, 2000 00Z
        Apr 20, 2001 12Z
        Apr 29, 2000 12Z
        Apr 30, 2000 00Z
        May 10, 2001 12Z
        May 08, 2003 12Z
        Mar 27, 2004 00Z  Dhaka, BD
        Mar 27, 2004 00Z  Agartala, IN
        Mar 27, 2004 12Z  Agartala, IN
        Mar 29, 2004 12Z  Agartala, IN
        Apr 03, 2004 12Z
        Apr 10, 2004 12Z
        Apr 14, 2004 00Z    Dhaka--75 people were killed by a tornado in north central Bangladesh from 1230-1400Z
        Apr 15, 2004 00Z    Dhaka
        Apr 16, 2004 00Z    Dhaka
        Apr 17, 2004 00Z    Dhaka

        The EML is present across the eastern part of southern Africa, especially in the warm season. Much of southern Africa is high in
        elevation, with a vast area above 4000ft. Within this elevated area, there is a large area centered around 30S, 28E which is very
        high(6-10,000ft). There is a large desert area in the western part of the country including much of Namibia, Botswana and the
        northern part of Northern Cape. Durban, South Africa, located near 29.5S, 31E, is just east of the very high terrain and EMLs
        are a common occurrence. Elevation really increases just inland from Durban. The marine layer really keeps Durban fairly stable
        when the EML is in place. Oftentimes Durban will be very moist at low-levels but quite cool. But just inland, elevated heating on
        the higher terrain is often enough to erode the cap. The area southwest of Durban(just east of the Drakensburg mountains) has a
        high incidence of  tornadoes. Storms in this area develop as moist upslope flow, elevated heating and local terrain effects help to
        break the cap. The last major tornado to hit this area occurred on January 18, 1999, killing 21 people.

        On December 15, 1998, a tornado hit Umtata in South Africa, killing 15 people. This event made world news since Nelson Mandela
        nearly became a victim. The 12Z(early afternoon) sounding from Durban(near the time of the tornado) showed an EML(although
        not exactly dry adiabatic) with a strong capping inversion. The T/Td at Durban were about 27C/24C. Inland at Underberg, only
        about 90 miles from Durban, the T/TD were 27C/16C. Despite identical surface temperatures, the surface potential temperature
        was 315K(42C) at Underberg and 299K(26C) at Durban. This is because Underberg is 5290ft ASL while Durban is near sea
        level. Despite the much higher surface dewpoint at Durban(24C versus 16C at Underberg), the theta-e was actually higher at
        Underberg(358K vs 355K). Please note that a 16C(61F) dewpoint at Underberg  has the same amount of moisture as a 66F
        dewpoint at Durban. These facts explain why the cap is virtually non-existent at Underberg and very strong at Durban. This example
        also highlights the importance of elevated heating to instability and convective initiation, demonstrates how point soundings should
        not be used to the exclusion of other data, and shows the importance of  potential temperature instead of temperature. In addition,
        as hinted at in red above, dewpoint temperature is not an absolute measure of moisture. A dewpoint at one elevation cannot be
        compared with a dewpoint at another elevation. Mixing ratio, on the other hand, is an absolute measure of moisture. For more on
        this, see my discussion on elevated heating.