Bangladesh in Flood'

1
Bangladesh in Flood.
2
Ganges/Brahmaputra/Meghna compared with other
large rivers (Hofer Messerli, 2006)
3
Floods in Bangladesh Key Questions (Hofer
Messerli (2006)
4
Main physiographic units of Bangladesh (Hofer
Messerli, 2006)

• Floodplain
• Often unstable, 2-5m relief.
• Deltaic Plain Low gradients and complex river
  channels
• Tidal Floodplains
• Poldered and level. Often flooded by rainfall.
• Piedmont
• Gentle alluvial plains subject to flash floods

5
Causes of Floods in Bangladesh (Hofer Messerli,
2006)

• Floods are often produced through combinations of
  factors
• Hughes (1994) a key feature of flooding in
  Bangladesh is that each flood is different

6
Types of Flooding in Bangladesh

• Flash Floods Due to heavy rainfall over the
  hills. Damage may be due to rapid flow
  (especially if embankments are breached) or crop
  submergence.
• River Floods From Himalayan snow melt and
  monsoon rainfall on mountains and plains. This
  especially affects the active floodplain. Causes
  most crop damage when early (June) - especially
  on the Brahmaputra-Meghna - or when late (damage
  along all rivers).
• Rainwater Floods due to heavy rainfall in
  Bangladesh- effects meander floodplains and old
  estuarine land.
• Coastal floods cyclone related tidal surges.
  Decreased sediment influx to the delta, natural
  subsidence ans sea level rise exacerbate.
• Timing Early floods (April-June) destroy boro
  rice before harvesting and young aus rice.
• High Floods Late July/August - damage aus rice
  Jute. Good for aman rice unless flow rise gt
  15cm/day.
• Late Floods after mid-August - affect
  transplanted aman rice.

7
Types of flooding in Bangladesh (Hofer
Messerli, 2006)

• River floods most common.
• Ganga
• Brahmaputra floods long rise times.
• Meghna-
• shorter rise times

8
Normal Flood extent in Bangladesh (Hofer
Messerli, 2006)

• On average 20 country is flooded annually.
• 66 cultivable land is susceptible to flooding
• 25-33 Cultivated land flooded annually
• 1994 0.28 land reported as flooded 1998 68.
• Annual floods (Barsha) are generally regarded as
  beneficial.
• Deep floods (bonna) are not!

9
Cropping Cycles in Bangladesh

• Related to the monsoon
• Kharif1 - pre monsoon to monsoon
• Kharif2 Monsoon to post monsoon
• Rabi (dry season) farming.
• Crops
• Aus rice (sown April) and harvested in the full
  monsoon during peak flood risk.
• Aman Rice comes in two varieties
• Deep water Aman - planted on lowland
• Transplanted Aman (HYV) which is transplanted
  half way through the cycle as floods recede, then
  rain-fed. If transplantation is too late yields
  are low, too early and flood damage risks.
• Boro rice - dry season, flood free - uses
  residual moisture.
• Fisheries are also strongly dependent on the
  flood cycle

10
Rice Wheat crop calendar in relation to
flooding (Brammer)
11
Effects of flood timing on agriculture
12
Bangladesh Flood affected areas in 1987
13
1998 Flood
14
1998 Flood
15
1998 Flood
16
Bangladesh Flood Affected areas in 1998

• The map looks at the WHOLE monsoon season in
  1998. Humid monsoon associated with La Niña.
• The floods were the longest duration and most
  devastating in 100 years. 50 country was
  inundated by up to 3m for 67 days or more.
• Simultaneous high river peaks, 7-11 Sept. High
  tidal surges flash floods river floods
• Continuous/widespread rains in July-August.

17
Extent of flooded areas in Bangladesh, 1954-2004
(Hofer Messerli, 2006)
18
Maximum daily Q per year for rivers in the basin
(Hofer Messerli, 2006)

• Many of the trends are NOT statistically
  significant though.

19
Downstream response to Severe Bagmati River
floods in Nepal (Hofer Messerli, 2006)

• Reaction to floods in Nepal (600km upstream) was
  small. Flood wave had been attenuated.
• Extraordinary flood in Himalayan foothills had
  little effect on the floods in Bangladesh.
• Qp for Meghna coincided with peak rainfall in
  Bangladesh.
• Tista Brahmaputra remained at average levels
  associated with Bangladesh rainfall.

20
Relationship between Flow and suspended load in
the three rivers for different time periods.
(Hofer Messerli, 2006)

• Complex relation between Q and Qsusp. Best
  relation for Ganga, less so for Brahmaputra and
  non-existent for the Meghna.
• Ganga Four trend lines quite different. Rising
  curve is NOT chronological.
• Brahmaputra Position of the trend lines is
  similar- highest the earliest.
• Meghna Trend lines almost horizontal gt no
  trend/relationship.

21
Changing courses of the Brahmaputra. (Messerli
Hofer, 1995)

• The long term variability of the Brahmaputra
  course is related
• High rates of sedimentation
• Unstable braided channel form
• The rivers migrate by avulsion and siltation of
  braided channels.
• In 1988 the Meghna created new channel 45m deep
  and 0.5 miles from its previous main channel.
• Earthquakes also cause instability of channels.
• Uplift and subsidence increase the likelihood of
  channel migration.
• The course changes of the Brahmaputra are most
  likely related to seismicity in the long term.
• The 1951-56 Assam earthquakes have been
  associated with increased sediment loads.
• Dynamically unstable river system is likely to
  show periods of increased flood instability.
  These natural factors are likely to outweigh the
  effects of upstream catchment degradation.

22
The Geomorphology of the Brahmaputra (Thorne et
al 1993)

• The Brahmaputra has a mean annual peak flow in
  early August of 65,500 cumecs.
• The Ganges peaks in late august at c
  51,625cumecs.
• The Brahmaputra channel is braided
• Largest islands called Chars. These divide the
  flow into anabranches .
• Three orders of channel exist in the braid
  system
• The whole channel first order
• Anabranches (2nd order) which change on an annual
  basis
• 3rd order channels branch around braid bars and
  change form within seasons. The channel
  migrates by right bank erosion.
• The dominant discharge (that which carries most
  sediment) is c 38000 cumecs.
• Bankfull discharge is 65000 cumecs.
• The 1987/88 floods did not give peaks in the
  sediment yields as they are relatively rare.
• Flows over 60000 cumecs carry about 8 total
  sediment sediment.

23
Dominant Discharge Plot Dominant Discharge is
that which, over the period of flow data, carries
the most sediment. (Thorne et al 1993)
24
(Thorne et al 1993)

• Upper Bar Level Chars - above river level and
  vegetated, used for agriculture.
• Braid Bars Loose sediment and annual immature
  vegetation cover plus seasonal cropping. No
  permanent habitation.
• Dunes c 15m high
• and 1000m long - submerged all but lowest flows,
  and migrate c 100m/day downstream.
• The dominant discharge covers the level of the
  braid bars but not Chars.
• The braid bars are thus adjusted to the
  dominant discharge.
• Flows between this level and bankfull carry
  most sediment.
• The Char and braid bar tops form a sinuous
  profile which includes areas of channel widening
  and convergence.

Dominant Q
25
The hourglass planform consists of bar
reaches where sediment is deposited and stored
and Nodal reaches which are narrow and deep and
where sediment is transported. Length of the
island reaches is scaled to the channel width.
NODE
BAR REACH
NODE
26
Hour-Glass morphology on the Brahmaputra
BAR REACHES
NODES
27
Brahmaputra river geomorphology-3 (Thorne et al
1993)

• Location of the spikes in retirement
  frequencies closely match the Char/Node pattern
  and scale at c 6-10km.
• This suggests that growth of braid bars in the
  Island reaches drives bank erosion and the
  location of embankment breaches. As mid-channel
  braid bars grow, flow is deflected to the
  Brahmaputra banks causing erosion.
• Thus the geomorphology of the river gives us
  insights into the distribution of erosion and
  flood hazard by embankment erosion.

28
Braid instability, erosion of banks displaces
population.
29
The Flood Action PlanProposals (Brammer, 1990)
30
The Flood Action Plan

• The main arguments for Flood Mitigation revolve
  around the issue of embankments.
• The 1989 UNDP study proposed preparation of a
  Flood Master Plan to embank the rivers and create
  Compartments along the river over a 15 year
  period. The World Bank produced the Action Plan
  for Flood Control.
• The Flood Action Plan aimed to improve quality of
  life and environment for agriculture.
• The plan proposes embankments to proceed
  downstream to allow lower channels to adjust to
  increased flows.
• Dykes would be installed in the Brahmaputra to
  cause siltation of braids and reduce the number
  and width of the total channel.
• Compartments would be built behind the channel
  embankments with sluices in the main embankments
  to allow artificial flooding to natural flood
  levels and to help fisheries.
• Deep flooding of compartments would occur with
  exceptional rains or floods greater than the
  design recurrence intervals.
• These are Dhaka - 500-1000 years main river
  embankments - 100 years Agricultural land -
  10-100years.

31
The hourglass planform consists of bar
reaches where sediment is deposited and stored
and Nodal reaches which are narrow and deep and
where sediment is transported. Length of the
island reaches is scaled to the channel width.
NODE
BAR REACH
32
Flood Action Plan Proposals-2

• Two options for embankments were suggested
• 1) Close to the river, which would protect more
  of the population and land. Cost was estimated at
  10 billion in 1990.
• 2) set back 5km which would require less
  management and would cost less at 5 billion, but
  would sacrifice more land and people.
• The huge scale of the rivers - at 20km wide,
  unstable braid plains subject to major avulsions
  would give huge maintenance costs.
• The existing Right Embankment of the Brahmaputra
  is having to be repeatedly retired. If the whole
  river length of embanked this problem would be
  magnified.
• Confinement in a narrower embanked channel would
  increase velocity and depth of flow of floods and
  could thus magnify pressure on banks - so huge
  embankment protection works would be required.
• Failure would be even more catastrophic than is
  currently the case.
• Annual maintenance costs were estimated at 200m.
  Set back embankments would leave 6.3m people
  exposed to higher flooding.
• In 1987 the UNDP review ascribed embankment
  failures to inadequate design and maintenance.
  Problems of stone supply for embankment support
  are huge as Bangladesh has very little rock
  available.
• The Flood Action Plan came under attack for its
  cost, likely impact on river dynamics and
  projected effects on agriculture and fisheries.
  The 1993 floods of the Mississippi, the worlds
  most controlled river also undermined the
  technical fix- hard engineering approach.

33
The Flood Action Plan Problems

• Custers (1992) discusses problems associated with
  the Flood Action Plan.
• The International Rivers Network believe that
  embankments will increase the risk of devastating
  floods if they fail and that the high flood
  levels are pushed downstream (the replacement
  effect).
• They argue that in the long term embankments are
  self defeating as siltation may cause the need to
  keep raising the embankment height.
• Channel migration could also disrupt the flood
  action plan.
• The evidence of the success of compartmentalisatio
  n is mixed. Many have become waterlogged and
  the numbers of fish species has declined.
• Custers argues that the elite stand to gain most
  from the Flood Action Plan through construction
  firms, the richest farmers and the HYV rice
  users.
• He argues that the Flood Action Plan is part of
  the aid and consultancy game where autocratic
  imposition occurs.

34
Impacts of the Flood Action Plan Custers (1992)
35
Hydrological and fisheries impacts of flood
control projects Asaduzzaman (1994)
36
Alternatives to the Flood Action Plan

• 1. Upstream storage dams
• 40 billion would need to be spent on storage and
  7 dams the size of the Bhakra Dam in India would
  be needed to reduce peak flows by 10 entering
  Bangladesh. Potential conflicts between flood
  storage, HEP and irrigation functions would be
  likely.
• Dams would be located in seismically active zones
  and there would be a rapid siltation problem.
• Political co-operation would be needed as the
  catchments are largely outside Bangladesh.
• Also much of the damaging rainfall is below the
  potential dam sites within Bangladesh.
• 2. Floodplain storage
• 12-15 large compartments would be needed to
  absorb 10 excess flow and these would cover
  24,000-30,000 km2. (This is 15-20 the area of
  Bangladesh) and thus politically not feasible.
• Natural basins already absorb some floodwater.
• Diversion of floodwater would result in channel
  siltation.

37
Alternatives to the Flood Action Plan

• 3. Draw down of groundwater
• In the dry season to increase capacity for
  groundwater storage in the Monsoon.
• Full use would reduce Ganges monsoon flow by 50
  and increase dry season flow. Drawdown would
  allow water storage in alluvial fan gravels of
  the Terai.
• Jones estimates that 2500 wells flowing at 1500
  cumec/hr for 240 days/year would reduce 21
  billion cubic meters for groundwater storage and
  this would absorb 10 Ganges flow.
• This would be less effective for the Brahmaputra
  due to its higher rainfall and less irrigation
  potential in the basin.
• There is little evidence that this strategy would
  decrease peak flood flows.
• Catchment infltration rates would still be low
  compared to heavy rainfalls.
• Deep drawdown might also dry many wells and
  reduce flow to Boro rice crops.
• 4. Small Scale Irrigation
• This strategy is favoured by the World Bank.
  Large scale water control projects have low
  rates of return .

38
Effect of embankments on soil fertility
(Alexander et al 1998)

• Soil on top of the embankment is known as fuldi
  series and that on the lower elevations know as
  Matlab.
• Soil fertility elements were sampled for both
  soil types within and outside the embankments.
• Over the six years since the embankment scheme
  began significant differences in soil fertility
  had developed in many parts of the scheme.
• Significantly lower available phosphate,
  potassium and sulphur in both topsoil and subsoil
  outside the embankment.
• This result was despite addition of potash
  fertiliser within the embankments.
• The embanked areas were receiving less flood
  borne silts, which renew nutrient inputs.
• The study suggests that there could be long term
  soil fertility issues arising from flood
  protection schemes using embankments.