Flood Risks and Reduction Approaches in Bangladesh

Abstract

Bangladesh is one of the most flood prone areas of the world because of its unique geographical setting and physiographic features together with a massive and unique hydraulic system. The ever increasing population, ill-planned infrastructural development and massive flood control interventions in a floodplain environment have resulted in flood disasters becoming larger and more frequent in recent times. Socio-economic impact of floods is profound; the flood prone zones represent areas with the highest incidence of the extreme poor, and the number of poor living in high flood risk areas is on the rise. The damage to infrastructure constitutes the major proportion when it comes down to economic damage resulting from floods. Since mid 1960s there has been a steady growth of flood control and drainage projects in Bangladesh through the construction of embankments, drainage channels, and sluices and regulators, with the total coverage area standing at 5.37 million ha. While the projects yielded a number of positive impacts such as increase in agricultural production, increase in economic activities, and reduction of damage to infrastructure inside protected area, the projects in general could not attain the desired objectives because of lack of consideration of interdependence of land, water, ecosystems and socio-economic development. Projects were mostly formulated with single objectives, aiming to solve the immediate problems without giving adequate attention to potential, undesirable long-term consequences. Ill-planned growth of projects have impacted the hydraulic stability of the system, with projects being able to provide protection against normal floods, while largely failing to provide protection during moderate to extreme floods. The cost of environmental degradation has been rather very high. Storage functions of floodplains have been lost, impacting important hydrologic functions of floodplains such as moderation of flood peaks, augmentation of dry season river flows and replenishment of groundwater storage. There has been substantial damage to capture fisheries because of disruption of hydraulic connection between river and floodplain. Disruption of links between river and floodplain by embankments has impacted country boat transports. Livelihood activities of different groups, especially the marginalized, have thus been compromised. In the light of the experiences with the flood control projects, the necessity to harness the beneficial impacts of floods especially in maintaining soil fertility and sustaining fisheries resources has become apparent. All the socio-economic and environmental concerns of earlier practices gave rise to exploration of alternative management strategies which are more resilient and environment friendly. This chapter provides a review of risks associated with different types of floods in Bangladesh. It also gives an account of historic floods and their impacts. Flood risk reduction approaches undertaken in Bangladesh and their impacts and consequences are discussed at length. The chapter also discusses three innovative flood risk reduction approaches that have evolved in Bangladesh, which tend to address the socio-economic and environmental concerns of typical flood control measures.

1 Overview of Flood Risk

Bangladesh is a deltaic country located at the lower part of the basins of the three mighty rivers—the Ganges, the Brahmaputra and the Meghna. The country is well known as one of the most flood prone areas of the world because of its unique geographical setting and physiographic features together with a massive and unique hydraulic system. About one-fifth to one-third of the country is annually flooded by overflowing rivers during monsoon (June to September) when the rainfall within the country is also very high. While normal floods are considered a blessing for Bangladesh-providing vital moisture and fertility to the soil through alluvial silt deposition in floodplains, moderate to extreme floods are of great concern, as they inundate large areas (more than 60 % of the country are inundated in large flood events), and cause widespread damage to crops and properties. Increased exposure due to growing population size and development in hazardous areas has made disasters in recent times larger and more frequent. Today’s flood hazards are already difficult for Bangladesh to cope up with considering its socio-economic conditions; the climate change impacts (the country ranks high in the list of vulnerable countries in South Asia, the most vulnerable region of the world to climate change impacts), would reinforce the baseline stresses that already pose a serious impediment to the country’s economic development. In the light of the experiences with flood control interventions implemented since mid 1960s, new insights have been gained and consequently new concepts of flood management have evolved. A key learning has been that flood risk reduction approaches cannot be sustainable without giving adequate attention to the interdependence of water regime, flood management interventions, other physical infrastructures, socio-economy and ecosystems in floodplain landscapes.

1.1 Geographical, Physiographic and Hydro-Meteorological Factors Responsible for Floods

1.1.1 Surrounded by Mountains

The country is surrounded by hills on its three sides, Rajmahal hills in the west, the Himalayas and the Meghalaya Plateau in the north, and Tripura–Chittagong hills in the east. The rainfall-runoff from this vast hilly area coupled with snowmelt in the Himalayas brings a huge inflow of water to Bangladesh during the wet monsoon season. A large area including Bangladesh and adjoining areas in India are under the influence of monsoons. From June to October large quantities of warm moist air travel from Indian Ocean north over Bangladesh and then to the Himalayan slopes as monsoon winds. Upper air turbulence and the long mountainous barriers stretching east–west make this moist air rise and as a result cool off and bring about enormous amounts of orographic rains which enter into Bangladesh territory in the form of runoff (Rashid 1991).

1.1.2 Lower Riparian Country

The country is located at the lower parts of the basins of the Ganges, the Brahmaputra and the Meghna (Fig. 4.1). The total area of the three basins stands at 1.75 million km² covering areas of India, Nepal, Bhutan, the Tibetan region of China and Bangladesh, of which only 7 % lies within Bangladesh. The excessive rainfall in these three river basins is the principal cause of riverine floods in Bangladesh. There are 57 rivers which originate outside the boundary of Bangladesh. About 1.18 trillion cubic meters of water flows annually to the sea, of which 1.07 trillion cubic meters or 91 % enters Bangladesh from upstream catchments and the rest are contributed by total internal rainfall (Rashid 1991). The annual volume of flow past Baruria just below the confluence of the Brahmaputra and the Ganges is 795,000 million m³ which is equivalent to 5.52 m of depth over 14.40 million ha of land area of Bangladesh. In contrast, the annual average rainfall for the country is 2.32 m (Harza et al. 1991).

Fig. 4.1

(a) The Ganges–Brahmaputra–Meghna basins; (b) the intricate network of rivers in Bangladesh

1.1.3 Floodplain Country

The aerial extent of flooding within Bangladesh is principally related to the unique physiographic feature of the country, which is the low-lying and extremely flat floodplain of the major rivers and their tributaries and distributaries. Floodplains cover 80 % of the country, while hilly areas in the northern and eastern regions occupy about 12 % and terrace areas in the center and north-west occupy 8 %. Mean elevations range from less than 1 m on tidal floodplains, 1–3 m on the main river and estuarine floodplains, and up to 6 m in the Sylhet basin in the north-east. The capital city Dhaka which is about 225 km from the coast is within 8 m above the mean sea level (Rashid 1991). The rivers assume a minimum gradient owing to the flatness of the land surface and flood water spreads more evenly and accumulates on the plains. A relatively small increase in flood level results in inundation of a wide area in the floodplains.

1.1.4 Unique Hydro-Meteorological System

There is a wide spatial variation of rainfall in the basins of the Ganges, the Brahmaputra and the Meghna, as well as the whole of Bangladesh. There is also a highly skewed temporal variation of rainfall; about 80 % of the rainfall occurs during the months from May to September. As a result, unlike other deltas, the seasonal variation in river flow is highly skewed with abundant water during monsoon while very small flow during the dry season. The country, therefore, faces two major hazards: floods during the wet season and scarcity of water during the dry season. The hydrodynamic characteristics during flood flows in alluvial rivers in Bangladesh are quite different from that during low flows.

The country is crisscrossed with an intricate network of around 230 rivers. A remarkable aspect of the river system is that all the rivers, except those of the Chittagong sub-region, are hydraulically linked to each other, all rivers being either tributaries or distributaries of the three major river systems. Flood hydraulics in Bangladesh is dominated by the major rivers. High water level of the major rivers slows down the flow of their tributaries resulting in backing up of water in the tributaries. The Ganges, the Brahmaputra and the Meghna, discharge about 142,000 m³/s into the Bay of Bengal during high-flow periods (Rahman et al. 1990). The Brahmaputra and Ganges carry about 85 % of flood flow that enter Bangladesh. The Brahmaputra has the largest flood flow followed by the Ganges and the Meghna with a flow ratio of 4.4:2.5:1. Although the Meghna has the lowest flood flow, it is by no means less important for flood processes in Bangladesh (Hofer and Messerli 1997). However, as far as area flooded is concerned, flow in the Brahmaputra has the strongest correlation with the extent of flooding in Bangladesh (Rahman et al. 2006; Salehin et al. 2007).

1.2 Key Determinants of Extent of Flooding

The important elements that determine the extent of flooding are the magnitude, synchronization of peaks, and duration of floods in the major rivers. The first two elements are directly related to the amount of rainfall in the upstream catchments, while the last element is related, in addition, to the downstream control provided by the coast in the form of the spring tide and the monsoon wind set-up in the Bay of Bengal.

1.2.1 Peak Flow

The magnitude of peak flow has a direct bearing on the extent of flooding in Bangladesh. Smaller differences in peaks of major floods can make a big difference in terms of flood affected area, since it is the spreading of floodwater evenly over a wide and flat floodplain that slows down the rate of rise in water levels. Frequency analysis of 92 hydrological stations showed that the difference in annual maximum flood levels of different return periods is not large. In general, the difference between 20- and 2-year annual flood levels remains within 2 m while the difference between 100-year and 20-year annual maximum flood levels remains within 1 m (Chowdhury et al. 1997a).

1.2.2 Duration of Floods

The downstream control that exacerbates the flooding condition is offered by the Lower Meghna (Fig. 4.1), which acts as the single drainage outlet of the three major river systems. Spring tide and monsoon wind setup in the Bay of Bengal cause strong back water effect in the Lower Meghna, and slow down drainage building up high water level and causing increase in the severity and duration of flood. This was the case in major flood years including 1998 and 2004, when the combined flood flow reached the Lower Meghna during spring tide, resulting in an extremely high water level and consequently severe flooding (Rahman et al. 2006).

1.2.3 Time of Peak Flows in Major Rivers

Synchronization of peak flows in the Brahmaputra and the Ganges is a major determinant of the extent of flooding in the country. When the peaks of the two rivers coincide, severe flooding occurs as it was the case in 1988, 1998 and also 2004 (Rahman et al. 2006). It has long been recognized that occurrence of simultaneous flood peaks in the Ganges and the Brahmaputra is not a rare event (French Engineering Consortium and BWDB 1989). Flood in 1998 is an example of simultaneous occurrence of several extreme hydrological phenomena such as coincidence of flood peaks of major rivers, high magnitude, occurrence of high flood for long duration, and arrival of peak flood at the time of spring tide.

1.3 Types of Floods

Five main types of natural floods occur in Bangladesh: river flood; rainfall flood; flash flood; tidal flood; and storm surge flood. Areas prone to different types of floods are shown in Fig. 4.2. In addition, some floods result from human activities. The principal sources of floods, as discussed at length in previuos sections, are the river floods from the major river systems, the Brahmaputra, the Ganges, and the Meghna, in the monsoon months. A broad strip of land, amounting to about 30 % of the country, extending beyond the active river floodplains (flooded during average normal flood) to parts of the adjoining meander floodplains (normally inundated by rainwater), is subjected to this type of flood. The timing of the flood and sometimes the duration of flooding are as important determinants of crop damage as is the absolute height reached by a particular flood. River floods in June can damage aus and deepwater aman paddy as well as jute crops, and floods in late August and September can be particularly damaging to deepwater aman paddy in low lands as a result of submergence at the panicle-initiation or flowering stages and to transplanted aman paddy on higher land because of drowning of seedlings or prevention of them from being planted (or replanted after an earlier flood loss) (Brammer 1999).

Fig. 4.2

Map of flood prone areas (Brammer and Khan 1991)

Local rainfall floods often accompany river floods, which result from runoff of high intensity and long duration rainfalls over Bangladesh itself that cannot be drained because of high outfall water levels. Heavy pre-monsoon rainfall (April–May) causes local runoff to accumulate in floodplain depressions. Later in the monsoon (June–September), local rainfall is increasingly ponded on the land as the rivers flow at high stage due to huge inflow of water from catchments outside the country, impeding drainage of accumulated rainwater. Rainfall floods affect crops in the same way as do the river floods in monsoon. The impacts of this type of flood are on the rise because of a change in hydrological regime in the floodplains due to unplanned construction of different types of infrastructures, such as roads, bridges, culverts, etc. The importance of rainfall floods in terms of damage has been highlighted in recent studies such as the study by Messerli and Hofer (2007) who argue that they have been more damaging in Bangladesh compared to the overbank flooding from the rivers.

The northern and north-eastern trans-boundary hill streams are susceptible to flash floods from the adjacent hills in India in the pre-monsoon months of April and May. Flash floods cause extensive damages to dry-season boro rice crop in the haor areas in the northeast region just before or at the time of harvesting crops, as well as to properties and infrastructures. Damages to boro rice and breaching of embankments are very common in some part or other of the eastern foothill regions and damages to property, especially road and railway embankments and bridges, and buildings alongside river channels, occur during very high flash floods (Brammer 1999).

The areas adjacent to estuaries and tidal rivers in the southwest and southcentral parts of the country (where they are not empoldered) experience tidal floods twice a day due to astronomical tide from the Bay of Bengal. Tidal water is mainly fresh in the monsoon, when flooding within polders is by rainwater. During spring tide, which occurs fortnightly, large area is flooded by tidal water, which can be damaging, especially if the water is then saline. Tide is experienced upto 225 km inland in the wet season and 325 km inland during the dry season. A considerable area in the southwest region is below the high water level of spring tide.

Approximately 12,000 km² of coastal land is prone to occasional cyclonic storm-surge floods due to tropical cyclones in the Bay of Bengal during April to June and September to November. Cyclones have the most dramatic consequences among the different hazards in Bangladesh. In comparison, the number of deaths during monsoon floods, even during extraordinary events, is small (Hofer and Messerli 1997).

Anthropogenic activities in the form of construction of infrastructure (mainly road) without sufficient drainage capacity through them, road alignments transverse to the main drainage paths, blocked drainage channels due to siltation, cross-dams or fishing activities and inadequately sized drainage sluices are increasing flood hazards (WARPO 2001a). For example, ill-planned construction of roads was one of the main reasons for rapid increase in flood depth and excessive duration of flood in September to October of 1995 in the north-west region of Bangladesh (Chowdhury et al. 1997a). Reduction of flood storage area due to filling up of low-lying floodplains and natural depressions is also exacerbating the flooding condition during major floods. Another cause of concern is the damage caused by sudden floods due to failure of flood control embankments. Flooding due to breaching of embankment does much harm to the agricultural land by depositing coarse sand, while natural flooding of agricultural land by overflowing river contributes to fertility of the land by depositing silt.

2 Historic Floods and Impacts

That the country has experienced floods since ancient times is quite evident from historical evidences in published literature on ancient flood management practices in the Bengal. A translation of a Persian writing during the second half of the eighteenth century by Akbaruddin (1974) showed that people used to construct earthen embankment around cultivated land to protect from floodwater, the local rulers of the Bengal built their homesteads by raising earthen mounds, and people used to grow long-stem rice which rises with the rise of floodwater. Construction of dikes along the rivers and irrigation by planned excavation of canals were the practices during the Mughal rule in the sixteenth to eighteenth century, as reported by Wilcocks (1930). However, the first systematic study of floods in Bengal was done by Mahalanobis (1927), a member of the North Bengal Flood Committee, formed by the Government of Bengal after a devastating flood in the northern part of Bengal in September 1922. It was reported that flood occurred in 25 of the years between 1870 and 1922, and were severe in 8 of these years. In 1922, the flood was caused by rainfalls of unprecedented magnitude. The rainfall in 1 week was about 10 times the normal weekly precipitation. The report observed that railway embankments hampered quick draining away of the flood water, and thus served to prolong the duration of flood.

No dependable data are available from 1923 to 1953, except a report prepared by the Irrigation Department of the Government of Bengal on the flood of 1931. The area affected by the 1931 floods was reported to be around 2,000 km². The flood data are, however, well recorded since 1954. A historic overview of variability of annually flooded area since 1954 is presented in Fig. 4.3. The figure also includes the growth of flood control and drainage projects in Bangladesh over the last 50 years. There is a decreasing trend of approximately 80,000 ha per year in the flooded area during the period from 1954 to 2004. This is in keeping with the increased coverage of flood protected areas, a growth of about 120,000 ha of area brought under flood control and drainage projects since 1964. However, despite implementation of all these flood control projects, the flooded area increased during major floods. There is clearly an increasing trend in year-to-year variability in the annually flooded area from mid 1970s.

Fig. 4.3

Growth of flood control projects and variability of annually flooded area (Source of data: Bangladesh Water Development Board)

2.1 Flood Damage

The main objective of flood control projects was to protect the crop land from flood damage. However, there occurred considerable crop damage even in years of moderate flood because the embankments were also damaged. The earthen embankment was unable to give protection against severe floods in 1987, 1988, 1998 and 2004 and even against some medium floods in 1991, 1993 and 1995 (Salehin et al. 2007). However, the crop loss due to flood damage does not seem to have substantial effect upon the total rice production of the country. This is because the crops are replanted and a part or even the total amount of loss may be caught up during the later part of the season. It also happens that the crop damage by flood is largely compensated by substantially higher yields due to higher residual moisture available to the following crops (Chowdhury et al. 1997a, b). For example, aman production in 1998 fell short by about 21 % below the normal trend production, while the production of dry season boro crop was 10 % above the normal trend in the same year (Islam 2006).

However, it is the damage to properties and infrastructure that is of major concern. When converted to monetary units, the damage to infrastructure outweighs the damage to crops (Chowdhury et al. 1997a, b). As can be seen in Fig. 4.4, there was substantial damage in 2004 even though the area flooded was much lower (38 %) compared to that in 1998 (67 %). The non-agricultural sector suffered loss for as high as 74 % of the total loss, with the remaining 26 % accrued to the agriculture (crop plus non-crop) (Islam 2006).

Fig. 4.4

Flood damage in moderate to major flood years (Source of data: Ministry of Diaster and Relief)

2.2 People at Risk in Flood Prone Area

With the increase of population, more and more people are settling in the flood-prone areas, making them more vulnerable to floods. An analysis conducted with 2001 population census data revealed that some 45.5 million people were exposed to severe and moderate floods (including river flood, flash flood and tidal flood), of which 22 million were male and 23.5 million were female (CCC 2009). Studies have also shown that flood-prone zones are the worst off among different disaster-prone areas in terms of food shortages, the incidence of extreme poor, insufficient income, illiteracy, and a high concentration of wage laborers (BIDS 2006). Vulnerability of the poor to floods has increased as a result of environmental deterioration and the increase in the number of poor living in high flood risk areas (Rahman 2004). Poor households suffer proportionately more from flood than richer households. Islam (1996) reports that an average poor household suffered (in terms of proportionate losses to values) four, five and three times as much compared to that suffered by a richer household, in a river flood, flash flood and tidal flood, respectively. Traditional flood damage estimates cannot capture numerous intangible effects of floods, which have serious implications on socio-economic pattern of living particularly in rural areas. Hossain et al. (1987) observed that the economy of flood affected region remains perpetually depressed. Very few have been able to improve their economic condition in the flood prone areas.

2.3 Growth of Flood Control Projects: Structural Measures

Following severe floods in 1954 and 1955, the Government adopted a policy of protecting agricultural land from the river flood in order to secure agricultural “production. A Master Plan for water resource development identified flood control and drainage measures as key requirements to increase agricultural production, by providing more stable conditions and reducing the depth and duration of normal and unusual flooding. Hence, since mid 1960s there has been a steady growth of flood control and drainage projects in Bangladesh through the construction of 12,850 km of embankments, 25,580 km of drainage channels and 4,190 sluices and regulators as per BWDB list of completed projects as of 1998 (WARPO 2001a). Currently the total coverage area stands at 5.37 million ha, which is about 37 % of the total area of the country and 56 % of the total cultivable lands.

Many conventional flood mitigation measures like flood control reservoirs, flood diversion or flood by-passes are infeasible inside Bangladesh because of its extreme flat topography. The main approaches that have been exercised are: (1) full protection of agricultural lands and urban areas against river flooding by constructing embankment along the rivers and providing appropriate drainage structures to minimize internal flooding; (2) partial protection against river flooding by constructing low height submersible embankments which are designed to delay pre-monsoon floods so as to ensure a safe winter crop (boro rice) harvest and to be overtopped during monsoon and remain submerged during the entire monsoon season; (3) evacuating unwanted rain water from behind the embankment or from within the poldered area by gravity through drainage regulators and sluices incorporated with the embankments (regulators and sluices prevent back flow from high river levels into the low-lying areas during monsoon and drain the water from the area to be protected when the river water level gets below than that inside); and (4) providing drainage by pumps in selected large projects to pump out accumulated water from the project area. The types of implemented flood control projects include flood control (FC), flood control and drainage (FCD), flood control, drainage and irrigation (FCDI), and drainage (D) projects. Other structural measures included dredging of rivers and canals at critical locations such as offtakes and confluences to remove sand bars, and hard and soft recurrent measures for bank protection and river training works.

Urban flood protection involved conventional method of constructing embankment and drainage regulator, with the exception of pumped drainage facilities in three drainage zones of the capital city Dhaka, with major part of the city dependent on detention storage-based storm drainage system. Approximately 123 FCD polders have been constructed since 1960s by Bangladesh Water Development Board under the Coastal Embankment Project (CEP) covering approximately 1.5 million ha of area in southwest, south central and Chittagong regions of the coastal zone. The objective was to create favorable condition for monsoon rice cultivation by preventing inundation of agricultural lands by saline water during high tides.

2.4 Non-structural Measures

Non-structural measures were considered as a means for mitigating flood damages. The Flood Forecasting and Warning Center (FFWC) of Bangladesh Water Development Board (BWDB) established in 1972, is responsible for making flood forecasts and flood warning during the flood seasons. The monitoring of floods and issue of flood forecasts are carried out in relation to Danger Levels (above which the flood is likely to cause damage to crops and homesteads) specified for each river gauging stations. The FFWC collects real-time data of water level (3 hourly) from 55 observation stations and rainfall from 56 observation stations. Currently a hydrodynamic mathematical model (MIKE-11) is used to forecast water levels. At present, the FFWC issues river stages forecast for 50 stations in the flood prone areas, formulated for lead time of 24 h and 48 h. The forecasts in the form of daily water level bulletins are transmitted to national radio, television, news agencies, newspapers, concerned ministries and government offices and field wireless stations.

Flood proofing of homestead is a tradition of the rural settlements in Bangladesh. Homesteads are generally raised above unusual flood level. Ponds are dug in the homestead area and banks are raised to prevent intrusion of flood water. They are the source of domestic water supply. They are also used for fish culture and for conservation of water for irrigation during droughts. Schools or health centers built on high mounds serve as flood shelters where vulnerable people can take refuge. Cyclone shelters are constructed in the coastal zone where human lives are at high risk due to cyclonic storm surge floods. Currently there are a total of 2,583 cyclone shelters in the coastal districts (CEGIS 2009). Over the years, there has been a significant change of emphasis from designing the shelters for single purpose use as flood shelter to designing them for multipurpose use. Shelters are now designed as schools, health centers and other community service centers during normal life.

2.5 Experiences with Flood Control Projects

Although there have been a number of positive impacts of the flood control projects in terms of increase in agricultural production, increase in economic activities, reduction of damage to infrastructure inside protected area, increased opportunities for culture fisheries, and generation of employment, the projects in general could not attain the desired objectives because of lack of consideration of interdependence of land, water, ecosystems and socio-economic development. Many projects were implemented with a view to solving the immediate problems without giving adequate attention to potential, undesirable long-term consequences. Project design and implementation largely ignored people’s participation. There was lack of consultation across sectoral and institutional boundaries, and the projects did not fully take account of the potential impacts on fisheries, navigation, forests, domestic and industrial water supply, biodiversity, and salinity management.

2.5.1 Hydraulic Impacts

The increase in year-to-year variability in flooded area and the increase in the frequency of severe flood events in terms of flooded area despite the growth of flood control projects, as illustrated in Fig. 4.3, can be attributed to the hydraulic impact the projects have had on the river-floodplain systems. Ill-planned growth of flood control projects together with ill-designed transportation and drainage networks make the system unstable during extreme floods. The flood control projects provide protection against normal floods. However, during moderate to major floods the damages to infrastructure including embankments are very pronounced thus causing increased flooded area. Flood control embankments suffer substantial damage during large and moderate floods, while damage during smaller floods is also quite high (Chowdhury 2000). The resulting damage due to embankment failure is very high owing to the accelerated economic activities compared to that in “without embankment” situation.

Floodplains moderate the flood flow by acting as storage during monsoon. Floodplains also augment the post-monsoon river flow by gradually releasing water from storage during the recession phase of floods. The river flow during late October and November is thus increased. Floodplains are also a useful source of groundwater recharge. Rainfall and flood water over the floodplains infiltrate and percolate vertically through the pervious soil to reach the relatively shallow groundwater table and recharge the unconfined aquifer. Flood control embankments have disrupted these hydrologic functions of floodplains. The rising trend in the annual maximum water level series and a declining trend in annual minimum water level series for the Atrai river in the NW region is the consequence of preventing storage of flood water in the floodplain depressions (beels) by constructing polders (Chowdhury et al. 1997a, b). Reduced floodplain storage due to embankment leads to an increase in water levels and discharges in adjacent areas. Such transfer of flood risk generates social tension, often leading to upstream–downstream conflicts and consequently forced cutting of flood control embankments by the affected people.

The coastal polders have led to deterioration of river morphology and waterways in the southwest and south-central coastal region. Without the embankment, enormous volume of tidal water would have been stored in the floodplain during rising tide which would have allowed sediment in suspension to be deposited over the floodplain. During ebb tide, the stored water would have drained through the tidal rivers providing a flushing action. As a consequence of the decrease in tidal volume due to polders, the water level and tide velocity dropped during ebb tide, resulting in siltation in the river causing rise in the river bed (Halcrow et al. 1993). This has, in turn, caused severe water logging inside the polders, leading to serious damage to agriculture, forestry, fisheries, livestock, homestead and physical infrastructures. Many people had to leave their ancestral homestead by abandoning traditional livelihood activities.

2.5.2 Impact on Crop Production

While many projects have been successful in raising agricultural output, as found in project based evaluations such as the study of HTSL (1992) of 17 FCDI projects, the impact of these projects on the overall food grain production is less clear (Chowdhury et al. 1997a, b). Flood control projects with irrigation projects have been more successful.

2.5.3 Impact on Floodplain Ecosystem and Water Transport Function

The objective of increasing rice production by providing flood protection to agricultural land has got so much priority that the consequential stress on the floodplain ecosystem received little attention. Livelihood activities of different groups, especially the marginalized, have been compromised. There has been substantial damage to capture fisheries because of disruption of hydraulic connection and hence fish movement between river and floodplain. There was a decline of 81 % in catch per unit area and a reduction of 33 % in the total number of species recorded annually in the case of Brahmaputra Right Embankment project (ODA 1995), and there was a 91 % decline in capture fisheries inside full flood control projects in the northeast region (Shawinigan Lavalin Inc. et al. 1992). Closure of the outlet of floodplain khals by flood control embankments created obstruction to country boats. The study by HTSL (1992) showed that FCDI structures had severely impeded boat transport in half of the 17 projects investigated. Out of the 66 projects studied by Shawinigan Lavalin Inc. (1993), 19 had major and 14 had medium level negative impacts on boat transport. The depletion of fisheries and impediment to floodplain water transport have caused conflicts between different users of floodplain such as farmers, fishermen and boatmen.

2.6 Lessons Learned and Paradigm Shift

The experiences of flood control interventions in a floodplain country have proved that it is difficult to attain stated objectives of interventions without giving due consideration to the hydro-morphologic features of the floodplain and the socio-economic conditions and cultural heritage of its inhabitants. It is seen that prevention of flood in floodplain has many negative environmental consequences. On the other hand, population, in general, do demand protection against flood damage. So an important issue is how to mitigate flood damage without causing degradation of floodplain environment. What it requires is a shifting of policy from flood control to flood management, with preservation of floodplain resources like soils, wetlands, capture fisheries, wild life, flora and fauna, and with indigenous production systems receiving their due priority. Floodplain zoning can be developed with the aim of preserving floodplain resources.

The nature of a water resources project is such that while it brings benefits to certain section of population, it may bring disbenefits to another section of the population. While resolution of conflict arising out of inequitable distribution of project benefits and disbenefits is difficult in any country, it takes on special poignancy in Bangladesh where the majority of the population living in the floodplains is overwhelming poor, holds small landholding and depends on many free resources of the floodplain. Hence, calculation of economic return should not be the only guide in selecting water resources projects. Decisions should be based on a multi-criteria framework of economic costs and benefits, and social and environmental impacts. Vulnerability of the society to large floods can be reduced by making communications lines, especially utility services and other infrastructures flood proofed. A floodplain land use regulation needs to be formulated so that planning, design, construction and maintenance of infrastructures account for flood factor and preservation of floodplain resources and environment.

The shifting of priority of the Government from “flood control” to “flood management” started towards the later stages of Flood Action Plan (FAP), initiated by the government after the severe floods of 1987 and 1988. At the initial stage of FAP, the focus was on flood mitigation. After a series of public debates followed by documentation of the concerns of flood mitigation in a number of published books, gradually it was recognized that FAP should pay attention to the complete hydrologic cycle and develop an integrated water management plan covering issues relevant to not only flood but also drainage, irrigation, navigation, environment and socio-economy. A real start in the paradigm shift in water resources management policy and practice towards “integrated” management took place with the preparation of the National Water Policy (NWP) in 1999 (MoWR 1999) and subsequently the National Water Management Plan (NWMP) in 2001 (WARPO 2001b). The NWP envisages full structural protection against floods in regions of economic importance (e.g. metropolitan areas) and reasonable degree of protection in other critical areas (e.g. district town), motivating the people to develop different flood proofing measures such as raising platform for homesteads and community facilities in the rural areas, with the exception of those already covered by existing FC infrastructure, and construction of all highways, railway tracks, and public buildings and facilities above the highest ever recorded flood level in future. The policy also stressed the importance of designating flood risk zones and taking appropriate measures to provide desired levels of protection for life, property, vital infrastructure, agriculture and wetlands. The NWMP includes a total of 84 programs grouped into eight clusters. One of the programs in the “agriculture and water management” cluster is to rationalize flood control and drainage projects considering the adverse impacts on the society and the environment and the institutional requirements for participatory management. The “disaster management” cluster includes programs related to improved warning and preparedness systems, social measures based on improved or more appropriate dissemination and response procedures, physical and social mitigation measures such as elevated platforms, cyclone shelters, raised highways and the like, and multiple use of infrastructure.

3 Innovative Flood Risk Reduction Approaches

In the light of the experiences with the flood control projects, the necessity to harness the beneficial impacts of floods especially in maintaining soil fertility and sustaining fisheries resources became apparent. All the socio-economic and environmental concerns of earlier practices gave rise to exploration of alternative management strategies which are more resilient and environment friendly. Three such strategies are tidal river management in the coast, compartmentalization in the central Brahmaputra floodplain and partial flood control in the north-eastern part of Bangladesh.

3.1 Tidal River Management

Polders in the southwest and south-central region implemented under the CEP had initially created good scope for growing agricultural crops by preventing the intrusion of saline water. However, as discussed previously in Sect. 4.2.4, obstruction of flood tide by CEP embankments and sluice gates caused reduction of tidal volume, leading to gradual silting up of rivers. As the land development of the beel stopped because of lack of sediment deposition, this land became lower than river beds.

After more than a decade of good productivity, drainage congestion began to increasingly affect the northernmost polders from the 1980s when the rivers and creeks silted up to such an extent that most of them became inoperative. This resulted in vast tracts of land remaining waterlogged round the year. In 1995, BWDB undertook the Khulna–Jessore Drainage Rehabilitation Project (KJDRP) (Fig. 4.5) to relieve drainage congestion in the Khulna and Jessore districts. The project embarked on the design of large regulators at the mouth of rivers in order to control tidal inflow into the project area. Such controlled inflow would reduce the possibility of sedimentation of the rivers as the sediment load is tide borne. Such regulator option was opposed by the local people as they apprehended sedimentation in front of the regulators and therefore feared that regulator based option will not be a permanent and sustainable solution. Instead, the local people came up with the concept of tidal river management.

Fig. 4.5

Location of innovative flood management projects in Bangladesh

According to the Tidal River Management (TRM) concept, beels are to act as tidal storage basins which allow natural tidal flows up and down in the river system. During high tides, a large volume of water along with sediments flows into the beels. These sediments settle down into the basin due to reduction of velocity and flocculation due to high salinity. This sedimentation would occur into the riverbed if the beels are not utilized for storage. This sedimentation in the basin lowers the river sediment capacity below the equilibrium condition and as a result, during ebb tide the river erodes the banks and scours the river bed. This continuous process gradually widens and deepens the river increasing the tidal flow volume of the river and thereby keeping the river alive. Tidal river management is in fact a natural water management process with very little human interventions but it needs strong participation and consensus with a great deal of sacrifice by the stakeholders for a specific period (3–5 years or even more depending on the tidal volume and the area of the beel ).

The principle of a rotational tidal basin was proposed to share both the inconveniences and the benefits with adjacent beels. In this regard nine beels (Fig. 4.6) with average size in the range of 1,000 ha were selected to operate as tidal basin one after another. It was planned that a new tidal basin would be opened after every 2–3 years.

Fig. 4.6

Location of proposed tidal basins in KJDRP

The Environmental Impact Assessment (EGIS 1998) carried out for the KJDRP showed the viability of TRM for the project. The approach was adopted for the Hari river system, with a benefitted area of about 27,000 ha. The KJDRP started its activities in 1998 by first dredging the silted up rivers to restore their natural conveyance capacities. In the mean time, the first tidal basin was implemented and operated in an unplanned way by forced-cutting of the embankment by the local people at Beel Bhaina. This tidal basin operation continued for 4 years starting from 1998 to 2001. The high tidal volume generated by the Beel Bhaina tidal basin increased the flow area of the Hari River downstream of the basin. The riverbed deepened by about 10 m more than the design level, and the river widened as well by eroding both of its banks (CEGIS 2002). A huge amount of sediments deposited within the beel bringing the entire beel area under cultivation and navigability of the Hari river was restored.

The second tidal basin was implemented during the period from 2002 to 2005 in Beel Kedaria in a planned way without any obstruction from the local people. While deposition of silt in Beel Kedaria was not adequate as per expectation of the landowners, the drainage was perfectly well and the project was free from any water logging. The landowners gave their land for 3 years without any compensation. However, they did not want to run TRM for the fourth year in their beel without any compensation despite the efforts of BWDB to convince them. As BWDB could not finalize a new beel and complete the construction of peripheral embankment for operation of TRM in the new beel, it ultimately resulted in the closing of TRM operation in Beel Kedaria with resultant deposition of sediments in the river system and recurrence of water logging in vast areas in the KJDRP.

The third tidal basin could not be started in Beel Khukshia in the year 2005 due to opposition and non co-operation from the landowners as they were not willing to give their land without any compensation. On the other hand, attempt to continue the on-going tidal basin operation in Beel Kedaria for another 1 year was not possible due to non-cooperation and opposition from a section of people who cultivated crops in the high land of Beel Kedaria. As such, about 17.00 km stretch of the Hari river rapidly silted up within September, 2005; drainage route of the area became blocked and heavy drainage congestion resurfaced. In response, BWDB engaged dredgers, excavators and manual labours to dredge the river again. In the mean time, BWDB was able to prepare another tidal basin in Beel Khukshia, which started operating since 2006. Drainage congestion situation gradually started to improve and in year 2008, water logging in the project area virtually disappeared again.

In the year 2012, Beel Kapalia was due to be the next tidal basin in operation. But again the land owners of Beel Kapalia refused and resisted giving their land for TRM without compensation. Violent clash between local people and law enforcing agency occurred and TRM operation in the KJDRP again started facing uncertainty.

The biggest responsibility in proper operation of tidal basin lies with the beneficiaries. This is for two reasons: firstly, TRM was suggested by the project inhabitants themselves; therefore they are expected to own it and work accordingly, and secondly, if TRM does not work they have to face severe consequences. When a tidal basin operates, the inhabitants of other beels enjoy the benefits of the project and nobody wants to hand over their beel for next round of TRM. Attitude like TRM is necessary but “not in my beel” prevails everywhere. The spirit of TRM is to share long-term benefits and short-term dis-benefits among the beneficiaries. In this regard the vast number of beneficiaries should be able to arrange compensation for the small number of inhabitants of the running tidal basin. On the other hand, BWDB can arrange compensation from the Government fund which will be easier to implement and disburse. This is now being tried in the case of Beel Kapalia. If such mechanism works then recurrent problem of preparing next round of tidal basin can be solved and TRM can become a model of innovative approach to solving tidal flood problem in the coast.

3.2 Compartmentalization

From a historical point of view, floods have devastated large parts of Bangladesh, especially during the 1987 and 1988 floods. In the aftermath, various studies were conducted in the assessment of confining these floods or, at the other extreme, apply a strategy of “living with the floods”, which is the more traditional way of dealing with these floods in these circumstances. A compromise was made out of these two extreme points of view, and hence the concept of compartmentalization was developed.

Compartmentalization is an established technique in watershed management. The technique is applied when it is necessary to slow down the runoff of river systems or watersheds or to create independent systems to which different drainage criteria can be applied. By applying the technique, the risk of flooding and/or the size and cost of interventions can be reduced. In the compartmentalization approach it is foreseen that floodwater could be regulated in such a way that it gives maximum protection, minimum flood damages and maximum benefits to agriculture, fisheries and navigation. When the system is in full operation (entire floodplain covered with multiple compartments), decisions are needed to determine how the flood (and rain) water has to be distributed among the compartments, given the hydrological conditions (flood levels, flood duration, flood predictions and actual rainfall) and given the degree of liberty as determined by the technical features of the system.

The entire scenario of possible choices regarding the management of these compartments is heavily determined by the prevailing hydrological conditions in a particular year. In hydrologically “wet” years, the choice of flood protection and minimizing flood damages will prevail, while in “dry” years the choice of making use of the beneficial effect of controlled flooding will dominate. Risk management is an essential part of the concept and therefore an analysis of possible scenarios need to be worked out which would provide guidelines for management at (floodplain) regional scale.

In order to test the compartmentalization concept in Bangladesh, the Compartmentalization Pilot Project (CPP) at Tangail was constructed during 1995–1998. The project is located (Fig. 4.5) in the Brahmaputra floodplain and is surrounded by embankments that can withstand a flood with a return period of 20 years. The CPP has established a single pilot compartment covering an area of 13,200 ha, divided by a seasonal river, the Lohajang, and bordered by two other rivers, the Dhaleshwari to the west and the Pungli to the East (Fig. 4.7). The Lohajang River is allowed into the project through a gated regulator at the northern side, not for the supply of water but for drainage, although the river also has a flushing function for the drains of Tangail Town. By lowering the water level of the Lohajang, the river will act as a drain for the numerous outlets that discharge into it. In the peripheral embankment on the northern, western and eastern sides of the project area, gated inlets are built to allow water into the compartment. The southern “embankment” is not an embankment: the road embankment is open and the Lohajang River and a number of other khals exit the Tangail Compartment without any control. In the case of long lasting high floods, flood water will enter the project from the southern side as back flow. Water control structures control water levels between sub-compartments and systems.

Fig. 4.7

Principal features of Compartmentalization Pilot Project

Early in the monsoon season water levels in the rivers are still low and only the occasional rain shower has to be drained off. Local drainage is thus the objective during this period. All gates are to be kept open, but can be closed if local considerations require. Once the river water levels start to increase, inlets and some outlets are opened or remain open and water enters the compartment. Controlled flooding is thus the objective during this period. Siltation is allowed to take place on the field, and the entry of fish into the compartment is facilitated. The water control structures between sub-compartments and systems are operated as required, to facilitate an even distribution of the silt laden river floods over the compartment, and to deal with rain showers. The key objective is thus to maintain a level of flooding which will give maximum benefits to both agriculture and fisheries.

At a later stage, the emphasis shifts from controlled flooding to water level control. The main regulator will be (partly) closed, lowering the water level in Lohajang River, inlets will be closed, and outlets will generally be open when drainage is possible. The water level will be lowered to such an extent that land preparation and planting can take place. The objective is to keep the water level close to an optimum level. These levels are dictated by crop requirements in compromise with the physical possibilities of the water management infrastructure and fisheries and environment aspects.

When flood levels in the main river system exceed certain levels, then flood control is called for, meaning that the main regulator is completely closed, the inlets are closed, the water control structures are operated as required to limit flooding in sub-compartments and systems, and the outlets are open, downstream water levels permitting. Some local inundation, however, may occur, particularly as a result of rainfall, which cannot be drained off.

Towards the end of the monsoon, once the water levels start to drop, early drainage of the compartment becomes possible. The improved drainage in the compartment will allow early land preparation for the dry season crops, and early harvesting of aman rice crops.

Due to high density of required structures in the CPP, the cost was relatively high. On the other hand, the incremental agricultural benefit was relatively low as the cropping intensity at 191 % was already very high in pre-project condition and the CPP attained a cropping intensity of 214 % in 1999. As a result, an EIRR of only 7.3 % and a negative net present value of $4 million were achieved as per the CPP Final Report (Lahmeyer International et al. 2000).

In a later study Rahman (2008) found that the cropping intensity improved further, and that the controlled flood component of the CPP has been highly beneficial for agricultural purposes and is in much demand. The Dhaleswari river is, however, almost dry now and carries very little water. As a result, the project is suffering from water shortage. The innovativeness of controlled flooding lies in the main purpose of the CPP, which was to retain beneficial effects of flood. Among many beneficial effects, two major purposes were under consideration in the CPP; one was to reinvigorate soil fertility each year by allowing flood water in agricultural land to deposit silt and the other was to retain the project area as a suitable ground for capture fisheries by allowing fingerlings migration through fish passes in the main inlet during rainy season. Although the project allows flood water to enter into the project area to a certain extent to take advantage of floods’ beneficial effect of increased fertility, but in reality water passing through the narrow inlets reduces sediment load to a great extent. Indiscriminate catching of fish near the main regulator is leading to a decrease in capture fisheries in the project area. Therefore, these two purposes are being served only to a limited extent in CPP area. However, if properly designed and adequately monitored and maintained, controlled flooding through compartmentalization remains an innovative flood management approach.

4 Partial Flood Control

The Haor Basin in the North east region of Bangladesh has an area of 6,000 km² which is deeply flooded region and is an important fishing ground of the country. During monsoon (June to October) the entire basin is flooded to a depth of 6–8 m. During dry season (November to May), the basin dries up with few pockets of water remaining. The basin is cultivated during dry season and the standing crop is harvested during late April–early May period. During this period the haor basin frequently experiences flash flood. In order to safeguard the only standing crop of the region, about 47 submersible embankment projects (SEP) were constructed during 1975–1990. The projects are small in size usually in the range of 5,000–6,000 ha. The concept of submersible embankments is illustrated in Fig. 4.8. While the typical FCD projects provide full protection throughout the year, the submersible embankments provide partial flood control as it provides protection with low height embankment only during pre-monsoon. During monsoon, the embankments remain submerged, and therefore do not disrupt the lateral migration of fish between river and floodplain.

Fig. 4.8

Illustration of protection from river floods by embankments. (a) Protection from monsoon floods; (b) protection from pre-monsoon floods

Saleh and Mondal (2007) found that the SEPs have achieved their desired objective of protecting the boro crop from pre-monsoon flood, especially after 1990s. The impact on fisheries production due to these types of projects has been minimal as found out by FAP 17 (1995). This is in contrast to traditional flood control projects where embankments obstruct monsoon flood and consequently the impact on fisheries has been very high. The submersible embankments result in a higher economic return; for example, HTSL (1992) found that in Halir Haor the average yield of boro was 19 % higher with the project than pre-project, and the estimated economic rate of return was on the higher side (about 65 %). On the other hand, the Brahmaputra Right Embankment (BRE) project (a full flood control project in nature), despite achieving a moderate increase in incremental value of monsoon paddy production, yielded a zero internal economic rate of return because of huge loss of capture fishery.