Bengal Convective Outlooks
March-May 2006
Jonathan D. Finch

Sign Guestbook

Main Page
Elevated Mixed Layer
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 products on this website are provided on a voluntary basis. The forecaster(author of this website) is not
    responsible for the correctness or availability of products on this website.

    The purpose of this website is to forecast high-end severe storm episodes in Bangladesh and East India. High-end
    severe weather (1.75 inch or larger hail and destructive tornadoes) is usually produced by supercell thunderstorms,
    so supercell storms will be the main focus of these forecasts.

    Forecasts will normally be updated between 3 UTC and 6 UTC(0900-1200 BST) since surface based convective initiation
    is generally between 9 and 11 UTC. These are updated as needed through initiation time.  In particularly active periods
    I may update several times per day. If violent tornadoes are expected, I will try to issue tornado watch boxes before events
    unfold--time permitting.

    Surface maps, upper air charts and soundings may be posted whenever possible during active periods and time permitting.
    Please note that such maps can be easily generated with Digital Atmosphere. This is a fantastic graphical program that
    retrieves meteorological data from the internet and displays these data on customizable maps for any place on the globe
    where data is available.  This program ingests data straight from the internet from sites listed on my links page such as
    College of Dupage or  Albany.

                                                          2006 Convective Outlooks

updated 0950 UTC May 10  2006

The severe weather season for 2006 is over. Although some severe weather may occur
for the remainder of May, extremely large hail and strong tornadoes are very unlikely.
The mid and high level flow will be too weak.

 updated 0856 UTC May 05  2006

The mid to high level flow will be too weak for high-end severe storms through May 15. If this
indeed persists through mid May then the high-end severe weather season would be over.

updated 0122 UTC May 01  2006

The mid to high level flow will be too weak for high-end severe storms through May 08. If this
persists through mid May then the high-end severe weather season could be over.

updated 0326 UTC April 26  2006

The mid to high level flow will be too weak for high-end severe storms through May 1.

updated 1430 UTC April 21  2006

Storms developed on April 21 in western BD and far east India. But the surface dewpoint was only in
the upper 60s to near 70 with fairly high cloud bases and weak surface winds. This storm was probably
non-tornadic. The main problem the next few days will be the fairly light surface winds and exiting high
level jet. There is no sign of a strong jet core approaching Afghanistan and Pakistan in the models.
Nevertheless, with moderate mid to high level flow over the Bengal region and weak flow from the
bay of Bengal, a few severe storms will be possible just about any day. But the chance of major tornadoes
does not seem very high in this pattern.

updated 1446 UTC April 17  2006

Weak wsw mid to high level flow will persist ahead of an upper trough through April 18. Widespread
storms will occur across Bengal. These storms will not be tornadic and will not contain extremely large
hail WSW mid to high level flow is not conducive to high-end severe storms. However, by April 19
and 20, the flow will become more westerly in the base of the trough, albeit not very strong. There will
be a slight chance for high-end severe storms on these days.

I am on vacation through April 24 so updates will be few and far between.
updated 1340 UTC April 11  2006

The flow will be fairly weak at mid levels and the high level jet will be north of Bengal through April 14.
In fact, the mid to high level flow will become more west-southwesterly by April 15 and 16. This may
preclude high-end severe storms through April  19.

updated 1358 UTC April 08  2006

Storms exploded between 2 and 3 pm again today.    9 UTC sfc map is here.  12UTC surface map.

updated 0413 UTC April 08  2006

A tornado watch has been issued for much of Bangladesh until 9 pm local time.
Watch graphics will not be posted today as the author of this page needs a decent
nights sleep. :)

updated 0330 UTC April 08  2006

Supercell storms are likely again today with 40+ kt flow at 500mb and 50-60kt high level flow. 
This will be the last day of high-end storms for a while. The high level jet will retreat north of the
region with
weak flow over Bengal.

updated 1319 UTC April 07  2006

Storms were still raging as of 1438 UTC.

updated 1319 UTC April 07  2006

More storms have developed

updated 1000 UTC April 07  2006

Storms have erupted along the dryline in far east India.

updated 0656 UTC April 07  2006

Well, get ready for another day of severe storms. I suspect based on the surface observations
that the day will unfold similar to yesterday, with storms exploding 
near or north of Ishurdi
just east of the dryline/front intersection and possibly
north and northwest of Dhaka along
the surface boundary. 

Therefore a TORNADO WATCH  has been issued for much of
central Bangladesh until 9 pm local time.
The mainthing that will distinguish another day
of supercell storms, from the major
tornadic storms is the strength of the low level inflow.
The 00 UTC sounding from Agartala, IN(60 east of Dhaka) can be found here.

updated 0532 UTC April 07  2006

The warm sector has been shunted into southern Bangladesh with light winds across the country.
60kt(300mb) to 80kt(200mb) high level flow will continue today with continued strong flow at 500mb
(40-50kts). The
main forecast problem is how much airmass recovery will occur and what boundaries
focus convection. Stay tuned to the 07z update once I look at the 06 UTC observations.

updated 1610 UTC April 06  2006

A supercell storm still rages west of Dhak at 1438 UTC.

updated 1154 UTC April 06  2006

A storm has exploded in western Bangladesh near Ishurdi. The 9 UTC surface map is here.

updated 0715 UTC April 06  2006

The 06 UTC surface map showed a wnw to ese oriented trough and a dryline in its familiar position. Storms will
likely develop near Ishurdi(west-central BD over the next 2 hours.  Therefore a TORNADO WATCH has been
issued for much of central Bangladesh until 9 pm local time.

updated 0225 UTC April 06  2006

Bangladesh will be in the entrance region of the jet today as shown by the 18hr UKMET 250mb fcst valid
12 UTC April 6  2006 as a weak shortwave trough undercuts the Tibetan Plateau.  Vertical wind shear will
be quite good with 40kts at 700, 50 kts at 500mb and 80 kts at 250mb, along with southerly winds at the surface.
There is an outflow boundary across central Bangladesh which will serve as a focus for storms. Isolated storms
may also form on the dryline and at the intersection of the dryline and the wnw-ese oriented trough.

The 00 UTC surface chart shows 3 distince boundaries. A dryline was in its climatological position. An outflow
boundary was further south than in previous days and the wnw-ese oriented trough was in its familiar position.

Wow, there was actually a sounding from Dhaka this morning. This shows a convective temperature of about 90F.
T/TD of 90F/75F would give a lifted index of -10 which of course is extremely unstable. The low levels have been
worked over by convection but this will quickly change.  I am not sure where the boundary will end up this
afternoon, but probably just south of Dhaka??

Major tornadoes usually do not occur in the entrance region of a jet in Bangladesh. The surface pattern tends to be
less robust with weaker surface winds???  This is just speculation.

A tornado watch will be required by 6 UTC (noon) for much of central and southern Bangladesh.

updated 2230 UTC April 05  2006

Storms were continuing across central Bangladesh even at 1438 UTC.

UK initialization for April 5 can be found here.


updated 1045 UTC April 05  2006

The 09 UTC surface chart is here. The aqua MODIS pass was way late but is now
available. 1 km


Towering cumulus clouds could be seen along the dryline. The outflow boundary is clearly visible
with stable stratocu to the north and cumuliform clouds to the south.

updated 0930 UTC April 05  2006

Storms are erupting about 20 to 30 miles north and northwest of Dhaka. These storms will quickly
become severe. The 20 million people that live in Dhaka could be impacted by these storms.
The Zmax showed a storm developing off the dryline to the southwest of Calcutta as well as the
storms north of Dhaka.

updated 0635 UTC April 05  2006

A tornado watch has been issued for much of central and northern Bangladesh until 9 pm local time.
An outflow boundary from morning thunderstorms will be the focus for new storms.  Initiation
will probably be near Jamalpur, Mymensingh or Siraganj by mid afternoon. The storms will then
move SE toward Dhaka and Agartala toward evening.

updated 0635 UTC April 05  2006

We are in luck today. There will be an aqua MODIS (polar orbiting) pass at 0730 UTC near 26N, 88E
with 250m resolution. This may show
a festering cu field and any boundaries that may lead to storms. I
will use this to help in decision making about a tornado watch.

The 00z UKMET shows the high level jet core near or just north of Meghalaya, India. The outflow boundary
was located from 24 to 25N. The wind field will look like the following:

700  29035kt
500 29050kt
300 29080kt
200  29085kt

Storms rolled across northeast Bangladesh this morning leaving an outflow boundary. These storms had exited
Bangladesh as of 0238 UTC leaving an outflow boundary behind. The 06 UTC surface map shows this boundary.

The surface flow will be southerly except easterly just north of the outflow boundary. Given the convergence
along the boundary, storms will probably develop despite the jet being just north. A tornado watch will be issued

updated 0548 UTC April 05  2006

The high level jet core is located across far northern Bengal today. The best chance of severe
would  be north of 25N. But will the airmass recover that far north?  The 6 UTC surface obs
and the hourly observations from Dhaka-Kurmitola will be key.


500mb chart 00 UTC  4 April 2006

updated 1240 UTC April 04  2006

storm near Dhaka

updated 0945 UTC April 04  2006

Storms were developing near Dhaka along the outflow boundary. An isolated storm was developing near
the triple point in India. The tornado watch was expanded westward.

updated 0645 UTC April 04  2006

Explosive thunderstorm development is immiment. A stationary frontal boundary(noon map) and dryline
will be the
focus for storms. Intitial development will be in fat east India or western Bangladesh where the
intersects the boundary.  Thus a tornado watch has been issued for central and southern Bangladesh
and extreme east India. The 0538Z radar image from Calcutta shows a few leftover storms around Dhaka,
which were serving to reinforce the outflow boundary.

updated 0520 UTC April 04  2006

An outflow boundary has been left over from previous thunderstorms. The nose of the high level jet has passed to the
northeast and the Bengal region will be south of the jet core today. But the mid level flow will remain strong with
50kts at 500mb. Despite the less favorable high level jet pattern, surface boundaries may provide enough lift for
thunderstorms. Given a favorable shear profile, these storms may be tornadic.

I am keeping the forecasts for April 3 on here for your information. When I receieve information on
damage or fatalities I will post.

For upper level charts for yesterdays severe storm outbreak, go here. I used the 12 hr
UKMET as the initialization.


updated 0520 UTC April 04  2006

Storms are still going at 1138 UTC.

updated 0930 UTC April 03  2006

Explosive thunderstorm development has occurred over the past 2 hours. Several supercell storms
        are now in progress. 

updated 0840 UTC April 03  2006

An isolated cell is developing west of Rajshahi. This cell is on the edge of the tornado watch but will
        move southeast into the watch area. The PDS has been shifted further west.

updated 0705 UTC April 03  2006

The latest Terra MODIS pass over Myanmar at 0425 UTC confirms my initial analysis of an
        outflow boundary over central Bangladesh. Explosive thunderstorm development should occur
        between 9 UTC and 11 UTC. Unfortunately, there is  no sounding information in the Bengal region
        today which makes forecasting even more of a challenge. It appears that the most likely place for
        storm initiation is western Bangladesh along the outflow boundary and just east of the dryline.
        In other words, storm will likely develop just ahead of the dryline/outflow boundary intersection.
        or along the west-east oriented trough.  Thus the tornado watch has been upgraded to a
        PDS(particularly dangerous situation) for much of central Bangladesh.

        Calcutta radar 0530 UTC.

updated 0550 UTC April 03  2006

tornado watch has been issued for much of central and northern Bangladesh and
        adjacent areas of extreme east India until 9 pm local time(15 UTC). Some cities
        in the watch include Dhaka, Mymensingh, Pabna, Bogra and Faridpur.

Discussion: updated 0150 UTC April 03  2006

        A supercell storm developed in west central Bangladesh about 4 pm(10 UTC on April 2).
        The 12z surface map showed an intersection of a dryline and trough in far east India.
        The approximate sounding for the area showed a lifted index of -11.  This is modified
        from the upstream Patna sounding.
        An outflow boundary has been generated by widespread nocturnal storms over
        northern and central Bangladesh.  Thunderstorms will explode near the intersection of
        this outflow boundary(if it remains intact) and the dryline. If the outflow boundary
        does not remain intact, then the trough/dryline intersection would be the initiation
        By 12 UTC April 3, the exit region of a jet streak will be nosing into Bangladesh. This is the prototypical
        example.  This will help increase the low level flow into Bangladesh, thus leading to a favorable shear
        profile for tornadoes.

        With increasing flow at all levels, any storms that develop today could produce extremely large
        hail and violent tornadoes. Surface dewpoints should be even higher today then yesterday
        which will result in lower cloud bases and even higher instability.

Discussion: updated 0500 UTC March 30  2006

    My composite reanalysis for 62 Bengal tornado cases can be found here. Note that tornadoes tend to
    occur when the exit region of the high level jet is over Bengal.

ECMWF day 5 200mb and 500mb from 12 UTC March 29 and valid at 12 UTC April 03  2006
ECMWF day 7 200mb and 500mb from 12 UTC March 29 and valid at 12 UTC April 05  2006
GFS 114hr 500mb hgt and sfc wind from 18 UTC March 29 and valid at 12 UTC April 03  2006
GFS 138hr 500mb hgt and sfc wind from 18 UTC March 29 and valid at 12 UTC April 04  2006
GFS 162hr 500mb hgt and sfc wind from 18 UTC March 29 and valid at 12 UTC April 05  2006

ECMWF day 5 SLP and 500hgt from 12 UTC March 29 and valid at 12 UTC April 03  2006
ECMWF day 6 SLP and 500hgt from 12 UTC March 29 and valid at 12 UTC April 04  2006
ECMWF day 7 SLP and 500hgt from 12 UTC March 29 and valid at 12 UTC April 05  2006


    The climatological tornado peak is April 10-14. There is a sharp decline in  violent tornadoes after April 19. There have
    been a few notable tornadoes in May. However, 2 of the most violent tornadoes on record occurred outside of the peak
    period on April 26 and May 13, with 1300 and 700 fatalities respectively. The April 26th tornado path was 8 miles long
    but up to a mile wide. Interestingly, the violent tornado on March 20, 2005 that killed 65 people was very early. Very
    few violent tornadoes have occurred so early in the spring. The exceptions were March 19, 1961 when over 200
    people were killed by a single tornado and March 13, 1953 when over 20 were killed. Only 2 violent  tornadoes have
    occurred after May 13.

Several factors lead to a very short but active severe weather season across Bengal.
 North and central India heats up and dries out in late March or early April. A deep, dry mixed
        layer develops. Low level  flow from the Bay of Bengal increases markedly during this time.

        Westerly mid-level flow around the Tibetan Plateau advects the Indian mixed layer over
        the Bengal moist tongue. This leads to the elevated mixed layer. Note that parts of the Indian
        desert are  "elevated"(1-3000ft) compared to Bangladesh which is near sea level.

        The mid level flow is still fairly strong in April with 30-50kt 700mb flow and 35 to 50 kt 500mb
        flow fairly common.

        The high level jet is usually over or just north of the Bengal in April.

        The southern branch of the polar jet often retreats north of the Tibetan Plateau by May, leaving
        light mid to high level flow across the Bengal region.  By June the high level flow is light.

           All these factors result in a tornado maximum in early to mid April. In short, vertical wind shear
        and instability are maximized and the jet is in a favorable position during this time.

                       Severe storms or tornado hits Bangladesh killing over 20

    Over 20 people have been killed in a tornado or severe thunderstorm on May 17, 2005.

                                            Severe storms hit central Bangladesh

    Bangladeshi picking up hail stones on April 25, 2005. This picture appeared in the "Daily Sangram".


                                         Tornado hits northern Bangladesh killing 65

    A tornado killed at least 65 people in northern Bangladesh on Sunday March 20, 2005. This tornado hit Sadullahpur and Sudarganj
    upazilas of Gaibandha and Mithapukur upazila of Rangpur.  I am currently preparing a case study with surface and upper air charts.

    On March 18, before leaving town on a 2 day trip, I issued a slight risk for March 19 and the first moderate risk of the year
    for March 20. Unfortunately, I apparently overwrote this file by mistake since I cannot find the March 20 forecast showing the
    moderate risk.

                                           Storms kill 7

    High winds and lightning killed 7 people in scattered locations across Bangladesh on March 30, 2005.

                     Thunderstorm winds and lightning kill 20

    Storms developed on March 23, 2005 in a tornado watch. At least 16 people were killed across Bangladesh. The deaths were
    scattered in nature and were the result of strong straight line winds and lightning. Lightning killed 4 people working in a field.
    An outflow boundary can be seen on the 06 UTC  Mar 23 surface map. Surface dewpoints south of this boundary were in the
    mid to upper 70s(24 to 25C) and surface based lifted indices ranged from -9 to -12 along and south of this boundary. Some of
    these deaths may have been from storms on March 22.

    A jet streak was approaching Bengal as seen from the UKMET 250mb initialization. The 500mb chart at 12 UTC March 23
    showed a shortwave trough over western India. I analyzed this map despite the very poor data quality over India, and the paucity
    of upstream data. The large view and small view  12UTC March 23 UKMET initialization 500mb maps are also available.

                               Lightning kills 8 in India

    8 people were killed by lightning in 4 separate cases in India over the past few days(March 19-23, 2005).

                                                  Storm kills 2 in Meghalaya

    A late evening thunderstorm on Marh 19, 2005 killed a couple on the western end of the Khasi Hills in the Meghalaya state of India.

    Morning(6am BST) soundings from Dhaka--good  job BMD(Bangladesh Meteorological Department). These are
    high quality soundings(especially the T/Td data)--a dramatic improvement from years past. These are from April 2004.


                                            Killer tornado on April 14  2004

    A tornado hit in northern Bangladesh on April 14, 2004, killing 75 people.

                                            Killer tornado on May 4  2003

    Click here for details about the tornado in far eastern Bangladesh on May 4, 2003.