Flood Alleviation Schemes

ABM Faruquzzaman Bhuiyan
Ph D Student, School of Environmental Sciences
University of East Anglia, Norwich, UK

(A short essay submitted to the ENV school on request of Dr Richard Hey to state the relevance of flood alleviation schemes practiced in UK and Bangladesh, in January 1998. No conclusion is made with respect to Bnaglsdeh Flood alleviation)
 
Floods in Bangladesh: A Brief Review

Due to specific geographical features, Bangladesh experiences frequent invasion of flood in almost every year of different magnitudes. During the 38 years period from 1954 to 1991, 29 remarkable floods occurred in Bangladesh (Shilo and Siddique 1991). In relating time and size of flood events, there may be two types of floods in Bangladesh: annual floods which inundates some parts of Bangladesh almost every year (Barsha) and the low frequency catastrophic floods of high magnitude (Banna). The annual normal floods are very important for the overall growth of deltaic plain of Bangladesh and its agricultural economy but the low frequency floods of high magnitudes (e.g. 1954, 1968, 1970, 1974, 1980, 1984, 1987, 1988, 1991) cause serious damage to life and property of the people which have a long lasting effect on the national economy.

According to the causes and the sources, floods of Bangladesh may be classified as River floods, Rainfall floods, Flash floods and Drainage Congestion floods. River floods occur following snowmelt in the high Himalayas and sometimes combined with heavy rainfall in the upstream catchment. This causes severe floods especially when the peak flow through the Brahmaputra-Jamuna river system synchronizes with the Ganges River peak flow period or Meghna. The devastating flood of 1988 occurred mainly by this process. Rainfall floods occur due to the heavy rainfall over the hills and flood plains of Bangladesh. The maximum of the total annual rainfall of the country (approx. 2000 mm) occurs in the May-September period, which aggravate the flood situation. Flash floods occur in the hilly tracts in the northern and eastern part of Bangladesh. The rivers Punarbhaba, Karatoa, Dharalia and Teesta in northwestern Bangladesh, Someswari, Kangsa, Jadukata, Sarigoalu, Surma and Kushiara, Manu, Bhalal, Khowai, Gumti, Muhuri, Halda, Sangu, Matamuhuri and others of the northern and eastern part of the country are vulnerable to flash floods. They are caused mainly by rapid surface water runoff due to heavy monsoon and pre-monsoon rainfall in the lower Himalayan foothills and the hills of Meghalaya and Tripura. Storm surge floods are associated with cyclones and hurricanes which periodically move up from the Bay of Bengal (Cyclones of April 1991 caused this kind of flood). Drainage congestion floods occur due to human activities (e.g. construction of highway, railway, embankment, polders etc. without proper drainage facilities) which are very common localized floods in Bangladesh.

The major causes of the frequent floods of Bangladesh are excessive rainfall especially in the monsoon, synchronization of peak flows in the major rivers, deforestation and ecological imbalance, aggradation of river beds, unplanned and inadequate controlling of the river by structures like embankments, dams, sluice gates, and finally seismic (earthquake) and neotectonic activities etc.

The control or alleviation of floods in Bangladesh involves not only controlling the water flow or reducing the water level of the rivers. The country is a low lying delta situated in the confluence of the three great rivers of Asia- the Ganges, Brahmaputra and Meghna which constitute one of the largest active flood plains of the world. On the other hand only 7% of the river basin area falls inside Bangladesh. These rivers and their tributaries carry a huge load of water and sediment, most of it during the monsoon rainy months of May/June to September/October.

The Brahmaputra is a formidable river. So are the Ganges, Meghna and other rivers which all suffer bank erosion. Mean peak flows of the Brahmaputra and Ganges River are 65,500 and 51,500 cusec respectively; peakflows in 1987/88 were 50% higher than these levels. The rivers run through soft alluvial sediments, and can scour channels of 30-50m depth. There is no suitable rock for making structures. The cost of building groins and revetments is, therefore, very high.

So flood control in Bangladesh involves management of huge volume of sediments (1.5-2.5 billion tons per year, 10-15% of which is deposited on the flood plains and delta areas), control of erosion-deposition and shifting courses of rivers, removal or reduction of water logging in the flood plains and above all the management of the active delta which is nourished by the rivers while flowing through it.

After the major floods of 1950s in Bangladesh, few important studies (e.g. studies by Krug Mission 1957, Hardin 1963, Thijsse 1964-65) were done to investigate the flood conditions in Bangladesh. Finally a twenty year (1965-1985) master plan was formulated. This resulted construction of a large number (58 originally) of Flood Control and Drainage (FCD) or Flood Control, Drainage and Irrigation (FCDI) schemes throughout the country. Up to the end of 1980s a lot of embankments and polders have been constructed under the FCD/I projects. By that time it was realized by all that the construction of these types of projects is not the solution of existing flood problems in Bangladesh. The negative sides of these projects like water logging, increase in the flood events, larger damages of crops by normal floods, effects on the river regimes and environment are quite obvious now. The poor operation and maintenance of these projects are thought to be the main reason for the poor performance of these projects. It is just after the severe floods of 1987 and 1988 that the flood problems of Bangladesh became a national and international discussion issue. After a number of preliminary studies by the international bodies, in December 1989, a five year (1990-1995) Action Plan for Flood Control was launched which later became known as Flood Action Plan (FAP). This is supported by 15 international donors and coordinated by World Bank. The FAP is consisted of 26 components, comprising of 11 main components and 15 supporting studies which were increased in number afterwards. The main components include five regional plans, studies concerned with options for strengthening the Brahmaputra Right Embankment, a cyclone protection study, flood control and river bank protection of Dhaka city and other urban areas, flood forecasting and warning, and disaster preparedness study. Up to now maximum number of the studies have been completed and some supporting studies and pilot projects are continuing on extension basis. A flood management model has been developed for flood forecasting and warning services with the help of DHI’s computer packages (MIKE11 and GIS). But the ultimate impact of the FAP on flood control of Bangladesh is not still now clear.

Flood Alleviation Schemes

To reduce the damages from flooding there may be strategies like to keep the flood away from the people or keep the people out of the way of the flood and, finally if there is really a flood, clean up afterwards. To keep the flood away from the people is tough but this is the method which people tried first by modifying or changing the flow pattern of the river or by confining flood in a narrow band. On the other hand to keep people away from the flood is also impossible. From the ancient time, people always liked flood plains for their valuable services even despite the danger of destruction. There are already many valuable developments including big cities in flood plains which can not be removed elsewhere. Although with the increasing world population people are forced to make the maximum utilization of their land resources. So there must be taken some measures to contain or reduce the floods. In most countries and societies people have used dykes or levees to keep the flood away from them. But those dykes and embankments can never contain all floods in long run. There are certain adverse impacts of these on the river regime itself. Recently people are more conscious about these impacts and trend is to think the whole land, water and aquatic system when a modification in one of these is done. Thus, now, the flood control or alleviation schemes are divided in two groups: the Traditional methods like Resectioning, Realignment and Embankments and the Environmental options like Relief Channel, Partial Dredging, Distant or Set Back Flood Banks and Two Stage Channels. More recently for flood alleviation other methods like Flood Control Reservoirs and Soil and Water Conservation at the upstream reaches or sources of the rivers are also thought as options. As for example for alleviation of floods in a flood prone country like Bangladesh and a part of India, an alternate thought is to construct large reservoirs in the upstream reaches (Nepal) of the large river, the Ganges. The ultimate aim is to holding back a portion of flood water in those reservoirs near the source of the river, through which the snow melting water from the Himalayas come down to cause devastating flood in the lower floodplains. The stored water is thought to be useful for other purposes like huge amount of hydropower generation, supply of water in lean period, recreation etc.

In the following sections the Traditional methods and Environmental options of flood alleviation schemes are discussed on the light of the study by Hey et.al. (1990). The usefulness of these methods in Bangladesh context are also cited in each cases.

Traditional Methods

Resectioning

In this method the flow section is modified in the required reach/reaches of the river to increase the conveyance of the section. The river reach is dredged to widen, straighten and reshape it. Theoretically, a circular section is the most efficient hydraulic section. But practically it is difficult to make and maintain such a circular section so that it is used only in special cases. In UK this circular type of culverts are used in the urban rivers such as Redhill Brook where it flows under the town of Redhill. Trapezoidal section with flat beds and with symmetric side slopes, selected according to the cohesiveness of the bank material type, are normally used for earth lined open channel. This can be easily excavated with mechanical dredgers.

As the flow section is modified the flow velocity is normally reduced in the engineered reach. So a compensation to the flow velocity is made by steepening the channel slope to increase it. This introduces instability of the river in both upstream and downstream sections. Specially in the upland areas, there may be considerable erosion in the upstream section and subsequent deposition in the modified reach as the river attempts to reestablish its original width, depth and planform. This instability and erosion may propagate further downstream. The river Ystwyth in West Wales is an example where the river is regularly straightened only for a meandering channel to reestablish itself.

In the urban areas the section may be required to make more rigid by harder engineering works like bank revetments, grade control structures, check weirs and bed load traps in order to control the transport regime of the river. Sometimes rectangular flume like channels is used in the urban areas due to space limitation. In this case the bank slope is maintained by vertical sheet piling, concrete or masonry structures. When the lining is used, the lining material must allow for groundwater seepage into the channel to prevent an increased hydraulic head developing during periods of rapid drawdown. Check weirs control the local gradients and prevent bed scour during flood and headword erosion is controlled by bedload traps which ultimately reduce the supply of sediment to the engineered reach. Although these structures control the regime of the engineered reach but erosion may occur further downstream as the sediment loads are substantially reduced. The results from the study of Hey et. al. (1990) reveal these processes in the UK rivers. For example, deposition occurs in the urban section of the Kent at Kendal due to not maintaining the bed load traps and erosion takes place further downstream. Similar case is the Usk at Brecon where serious aggradation occurs at Brecon due to absence of sediment trap at upstream with subsequent erosion at further downstream. On the other hand, in the case of Sid at Sidmouth deposition is prevented by limiting the sediment transport to the reach by regular dredging while weirs prevent the bed erosion.

In lowland areas instability problems are reduced due to lower sediment transport rate where sediments are mainly sand, silt and organic matter in suspension. The rivers would slowly attain its original condition through reed encroachment from the bank and by bed degradation. The main problem is the deposition of fine materials in the dredged reach which requires routine dredging or removal of the deposition in every 5-10 years.

This type of flood control measure considerably limits habitat diversity for both instream and bank species particularly if the banks are lined with concrete blocks, gabion baskets, block stones, rip-rap etc. So this type of measures are not recommended generally unless there is ultimate necessity for space constraint etc. Then stability, sediment transport capacity and the conservation or architectural merits of weirs and bank revetment should have to be checked.

In a low lying floodplain like Bangladesh this method is not suitable because the rivers are wide and carry huge load of sediments while their banks are also made of loose sediment materials. Thus full resectioning is not a stable solution for almost all the major rivers. Alternately the method has been used in some cases of flood control and irrigation schemes for smaller canals (e.g. in Teesta Barrage Project in the north-western part of Bangladesh) where the irrigation or drainage water with very low contents of sediments have to be transported via the existing natural small channels. A large scale resectioning program of the smaller canals of Bangladesh was launched in 1980s naming Khal Kata Karmachuchi (Canal Digging Program) by the then president of the country. His plan was to increase the conveyance of all the smaller canals which are ultimately connected to the rivers or other large water bodies. The objective was to transport water from the big water bodies at dry season for irrigation and drain out the water logged area and the localized flood water at the monsoon period. But ultimately this was not a successful project.
 
 

Realignment

Here the channel is aligned such that it follows a shorter course than previously. An example of this is cutting of meander loops to construct an entirely new channel and filling the old one. Thus channel sinuosity is decreased and the channel slope is increased. Flow resistance is reduced in the straightened reach as the secondary flow cells generated in the meandering river are removed. Finally the flow rate is increased. This method is normally practiced with resectioning.

The method when practiced in upland rivers, may cause serious instability like headcut erosion, deposition at downstream and re-establishment of a meandering channel in the engineered reach (Hey 1987a). To prevent the headcut erosion, a bed sill at the head, and a series of check weirs in the engineered reach may be required to maintain channel gradient and prevent bed scour. To stabilize the banks, bank revetments may also be required. On lowland rivers the problem is less but the bank protection is required depending on the depth and bank material. Like resectioning in this case also the instream bankside habitats and flora are totally destroyed.

In Bangladesh, this type of measures have been practices in few cases, specially in small tributaries of large rivers. A special case of this, which is used in Bangladesh, is the closure of a tributary of a big river to protect a certain area or certain structure. It is reported that there are above 1000 river closures of different sizes in Bangladesh (Hughes et. al. 1993). There are no much investigations about the environmental consequences of these closures.
 
 

Adjacent Flood Banks

This is widely used and the oldest method of flood control. Here embankments are used to contain the overbank flows in a limited area of the floodplain. In traditional method, they are constructed near the river to save space behind the embankment. This normally increases the in-bank flood capacity and the sediment transport capacity of the channel relative to the natural condition. Due to the increased bank height the bank habitat diversity is also increased. During construction of such embankments, the strength, permeability and cohesiveness of the soil should be checked to control seepage and subsequent breaching. The berms with proper side angle and sufficient freeboard increase the safety.

There may be instability problems in the engineered reach in the upland rivers which transport high bed load. This is normally caused by the considerable erosion in the engineered reach and deposition in the further downstream reaches. In lowland rivers, however, there is less degradation and more stability.

In Bangladesh, flood control embankments have been used widely. There are about 7500 km of flood embankments throughout Bangladesh. Most of the Brahmaputra Right Bank and about half of the left bank are embanked. There are embankments on the both sides of Ganges and Teesta as well as a substantial part of Padma river (the combined Brahmaputra and Ganges) and Meghna river. Some of the eastern rivers, which experience flashfloods, are also controlled by embankments in many parts. Several thousand of kilometers of embankment exist around polders in the coastal zones. The concept of Compartmentalization which is being tested under Flood Action Plan Project in Bangladesh is also based on construction of embankments along the riversides and lowland sides with sufficient allowance for controlled flooding. One of such pilot project is going on near the left bank of Jamuna River namely Compartmentalization Pilot Project, in Tangail.

People are now much more serious about the negative impacts of the embankments. Because it is not possible to ensure that the embankments will not fail or overtopped in long run. There are numerous examples of failure of flood embankments in controlling flood. Most cited are the highly embanked Mississippi River floods of 1973 and 1993. From the frequent floods in Asam, India, where both sides of the Brahmaputra are embanked, the experts now opines that embankments are causing more problems than they mitigate.

In Bangladesh, there is a strong debate with the effectiveness and further construction of embankments. The negative effects are generally discussed on the specific geological and morphological conditions of Bangladesh. Due to the construction of embankments the delta building process of the lower floodplains by the large amount of carried sediments of the rivers are stopped. As a result the morphological balance between the river and the alluvial floodplain is disturbed. In the case of some medium and small rivers like Gumti and Bhadra, embankments have caused rise in the river beds. There are examples that the change in the local morphology has caused dying of perennial rivers (Choudhry, 1991). Another problem with the embankments in Bangladesh is water logging of the area behind the embankments and outside areas of the polders or compartments. During the flood in the river, there is also normally high precipitation in the floodplain. The rain water from the protected area can not be drained by gravity to the rivers due to proper drainage facilties or the high stage in the river at the same time. The embankments are mainly constructed of sandy/silty soil which are very sensitive to breaching. So maintenance cost is high enough for medium and long term success. The example of the Brahmaputra Right Bank Embankment, built in the 1960s is being corrected and maintained frequently in expense of huge amount of money. Still it was incapable of withstand a flood like 1987/88.

Despite of all the negative factors in the use of embankments in Bangladesh, it is not possible to exclude it from the effective options for flood alleviation. But the trend now is to the modification of the primitive ideas of flood controls so that it becomes more environmental friendly. The concept of controlled flooding through the embankments, set-back distances of the embankments from the rivers may be mentioned here. The later concept is important because the embankments proximity to the erosive rivers, in many cases, require heavy investment in river bank protection which causes high stage and greater impact on the morphology of the river. The setting-back of embankments further from the rivers might reduce the initial capital outlay and the recurrent financial costs of annual maintenance and may improve the long-term integrity of the embankments.
 
 

Environmental Options

Relief Channels

Relief channel, also called bypass, is used to divert the flow from the main river when the water stage of the main river exceeds a specified level. This is normally used at the upstream of a certain reach of a river where there is absolute necessity to limit the stage for certain reasons like some expensive installations or structures which are designed for certain level of water flow. The relief channel may be wet or dry in the low flow season. Where the river transports sufficient bed material, instability may occur at the intake of the relief channel. At the diversion point, the transport capacity of the river is reduced so that deposition of the coarse material occurs at the intake. This causes subsequent erosion at the downstream. Specially at the confluence of the main channel, erosion is significant if the confluence is obtuse. At the same time secondary flow generated by the flow from the relief channel can cause significant bed scour. In the lowland rivers these instability problems are less as long as the relief channel enter the river at an acute angle.

In this method main advantage is that excluding the deposition at the intake point, the main channel remain intact and its instream and bankside flora and habitats remain uneffacted. If the relief channel is wet, it may act like a closed water body which is a favorable environment for undesirable species like algae, pondweeds, water hyacinth etc.

If there is sufficient sediment load in the main river, then regular maintenance dredging will be required. In many cases relief channels become in operation during a high flood and other time the intake point remain closed with flexible construction. After certain pressure is developed in front of it then it is automatically opened and starts operating. The entrance of the relief channel is provided with stilling basin at required level to control the flow when it becomes in operation.

In Bangladesh, this type of relief channel or bypass is used sometimes to increase the safety of large structures like barrages, sluice gates against high flood of higher magnitude than the design flood of the structures. But the problem is with the space for the relief channel and protection of its surrounding areas during high floods. When the relief channels come in operation in that condition, the inhabitants of the surrounding areas may experience a new kind of unexpected flood which causes social and environmental problems. Similar happening occurred in the Teesta Barrage Project in Bangladesh in 1991. After a long time of construction of the main barrage, it was the first time when the bypass channel became in operation suddenly when a high flood arrived overnight from the upstream. Then a large area surrounding the bypass channel was flooded. A large cultivable area became covered with transported materials, which was a great concern of the local farmers.

Partial Dredging

In this method to increase the flood carrying capacity only a part of the main channel is dredged. Normally the shallow riffle sections are dredged. This creates a deeper section of channel at low flow but leaves a section of the bed undisturbed which act as a biological store for recolonizing species.

The alignment of the dredged areas should be along the main flow direction because the low flow channel can act as a gravel trap when the bed becomes mobilized at high flows. This has occurred in the river Wharfe at Conistone.

This partial dredging is frequently practiced in Bangladesh. As the most rivers carry huge amount of sediments which are deposited in a considerable amount in the riverbeds forming riffles, bedforms of different sizes, sometimes, as submerged chars. So within the limitation of the finance and equipment (dredger) availability, a regular maintenance dredging in some of the rivers are performed every year by Bangladesh Water Development Board, which has a separate division for dredging. Two main purpose of this dredging is to increase flood carrying capacity and navigability of some important rivers. But the rapid silting up of the dredged sections and the huge cost involving in maintenance dredging discourage to take large scale dredging in the main channels throughout the country which have been silting up for a long time.

Distant Flood Banks

In this method the flood embankments are set back from the rivers compared to the traditional embankments of the rivers. Flood is allowed to flow through a narrow strip of the flood plain. The main advantage of this method is the embankments can be built in relatively lower sizes and the natural regime of the main channel is kept relatively undisturbed. The instream and bankside habitats and flora remain uneffacted. Thus the meandering or braiding pattern of the river is allowed to occur as in the natural way within their bound. So there is a very little interference between the natural stability of the river and the embankment which are very important for the alluvial sand bed rivers.

The main constraint with this method is the space availability along the banks of the rivers to spare for flooding. In urban area this is a problem. But in rural areas it can work well with a modification in the landuse pattern in the flood plain areas which are inside the embankments and will be allowed for flooding.

This kind of flood alleviation scheme is preferred. In the low lying flood plain areas of Bangladesh this method seems to be more suitable than the traditional method of embanking which has already discussed earlier.

Two Stage Channels

This is the method where flood berms are constructed besides the main channel by excavating a part of the flood plain. This keeps the main channel intact and the flood flow is accommodated in the high stage channel. This method keeps intact or increases the instream species diversity but bank habitat is changed considerably.

In the upland rivers with high sediment loads, there may be appreciable aggradation and bank erosion or berm destruction during flood flows. On the other hand, if the berms are stable and constructed near water table then high vegetation growth may be a problem which reduce the flood carrying capacity of the rivers.

This type of two stage channel is not suitable for Bangladesh and not practiced still now. The rivers are wide with high sediment load. So if berms are constructed on the sandy or silty flood plain, these will be rapidly destroyed and the main channel will be silted up during the flood flows.
 
 

References

Asaduzzaman, M. (1994). The Flood Action Plan in Bangladesh: Some Lessons of Past Investments. CDR Working Paper 94.3.

Brammer, H. (--). Protecting Bangladesh. Internet Source.

Choudhry, J. H. (1991). Flood Action Plan: Some Points to Consider. GRASSROOTS, July-Sept., 1991,

Hey, R. D., Heritage, G. L. and Patteson, M. (1990). Design of Flood Alleviation Scheme: Engineering and the Environment. Research Report, School of Environmental Sciences, University of East Anglia, UK, 176pp.

Hughes, R. and Adnan, S. (-). Floodplains or Flood Plans..

Khalequzzaman, M. (1991). Recent Floods in Bangladesh: Possible Causes and Solutions. Presented Hazard 91 conference held in August 4-9, 1991 in Perugia, Italy.

Miller, J. (1997). Floods: People at Risk, Strategies for Prevention. Internet source of Department of Humanitarian affairs, United Nations.

Shilo, I. and Siddique, R. (1991). Legacy of Floods 1954-1991.GRASSROOTS, July-Sept., 1991, 65-67.

1