The recent Kosi floods have proved once again that inadequate control measures have been responsible for the recurring disasters. Typically flood control and riverine studies focus on hydrological information, whereas a much more integrated approach that pays attention to specific morphological factors is required. Since Kosi is a dynamic river with a unique morphology and because it is a river which has always carried high sediment loads, flood management strategies must be attuned to such specific parameters of the river, besides being much more than mere "river control" through embankments.
BIHAR FLOODSnovember 15, 2008 EPW Economic & Political Weekly42caused by drainage congestion. It recog-nises the basic fact that the flood plains are essentially the domain of the river and as such all developmental activities there must be compatible with the flood risk in-volved. However, due to inability of the governments to prevent or check the en-croachments of the flood plains, no begin-ning could be made so far in flood plain zoning. The government of India circulated a model bill on the flood plain zoning in 1974 among all the states of the country with the request to enact the same, but no state except Manipur could do so and this state also could not implement it. Flood proofing is essentially a combination of structural change and emergency action. Raising of some flood-prone villages above predetermined flood level and connecting them to nearby road or high land has been done, but the enthusiasm was lost very soon as the method did not provide any protec-tion to surrounding agricultural areas. The flood forecasting and warning system is already working in Bihar satisfactorily.The most common adjustment to the flood hazard in flood plains is simply to bear the loss and live with flood in areas where there is some serious constraint (as in north Bihar). Under the situation, floods become a matter of concern for all sections of society living in the affected areas and flood management needs active involve-ment and participation of all of the stake-holders to fulfil its objectives. The National Water Policy (2002) also advocates a parti-cipatory approach with the involvement of various governmental agencies, users and other stakeholders in an effective and de-cisive manner, in various aspects of plan-ning, design, development and manage-ment of water resources schemes.The problem of flood in Bihar is very com-plex and so is the task of its management. A strong political force is required for seeking and ensuring cooperation of Nepal without which sustainable long-term flood manage-ment in Bihar is not possible. Competent and capable engineers, well acquainted with the latest technology and enlightened and efficient bureaucracy should beentrusted with the job of policymaking and its imple-mentation. Their work should not be inter-fered with. Last but not the least,the involve-ment of the stakeholders in all stages of the process is a key factor for the success of the effort and optimisation of benefits.Kosi: Rising Waters, Dynamic Channels and Human DisastersRajiv SinhaThe recent Kosi floods have proved once again that inadequate control measures have been responsible for the recurring disasters. Typically flood control and riverinestudies focus on hydrologicalinformation, whereas a much more integrated approach that pays attentionto specificmorphological factors is required. Since Kosi is a dynamic river with auniquemorphology andbecause it is a river which has always carried high sediment loads, flood management strategies mustbeattuned to such specific parameters of the river, besides being much more than mere “river control” through embankments.Rivers play a critical role in human society and history as they are the major source of fresh water, trans-portation, and resources. However, this relationship is often “troubled” because changes in river discharge (floods or droughts) or position can play havoc with permanent settlements. Such changes can be caused by both natural forcing as well as human interventions or a combination of both. Natural processes may include short-term changes in sediment load, water volume or, long-term changes in relative sea level or climate change. Human interventions could impact in changes in sediment load or run-off through water resource management schemes such as dams, barrages and embankments. Human alterations of river systems can have many important consequences primarily because river systems are dynamic and highly integrated systems and any change in any part of the river can easily propagate and affect the whole system. The recent flood in the Kosi is certainly one of the biggest “human” disasters in recent years and it has sent out a strong signal that our flood management strategies are questionable and our preparedness to face such events is far too inadequate. Further, there has been a paradigm shift in flood management globally from “river control” primarily involving an engineering approach addressing the “effect” at a local scale to “river management” which empha-sises an integrated approach at a crossover of scales and addresses the cause rather than the effect. Even though India is a country drained by several large rivers, our river management strategies are rather rudi-mentary and our planners are yet to embrace modern approaches such as satellite-based monitoring and multi-criteria decision support system. This situation needs to be corrected to save a large population from repeated miseries of floods year after year.The Kosi: A Dynamic SystemThe Kosi river in north Bihar plains, eastern India is a major tributary to the Ganga river system and has long been considered as a problematic river due to recurrent and extensive flooding and frequent changes in its course. The gently sloping alluvial surface of the Kosi has been described as “inland delta”, “cone” and “megafan” by various researchers owing primarily to build-ing up of a very large positive topography caused by deposition of enormous quantity of sediments carried by the river which it is unable to transport. During the last two centuries, for which records are available, Rajiv Sinha (rsinha@iitk.ac.in) is with the Department of Civil Engineering, Indian Institute of Technology, Kanpur.
Stream power Sediment supply
4 – 2 – 0 – Unit stream power (Mean annual flood) Unit stream power (Bankfull discharge)
BIHAR FLOODSnovember 15, 2008 EPW Economic & Political Weekly44Tectonically stable catchments would con-tribute significantly less quantities of sediment to the river systems. Similar to the rainfall distribution, spatial variation in the uplift rate along an east-west transect has also been observed. Research by several workers has shown that the eastern parts of the Himalaya are charac-terised by higher uplift rate (15-20 mm/yr) along the mountain front in Nepal, based on the analysis of deformation of terraces along Baghmati river system. The present models on Himalayan seismotectonics predict westward decrease inuplift rate along the Himalayanfrontin Dehradun(~7mm/yr). Therefore, not just that sediment production is low in their catchments, there is obviously no remobi-lisation of sediments due to incised nature of the river channels. It appears therefore that high sediment supply from upstream area coupled with low stream power inEGP rivers has resulted in aggradation of channels, frequent floods and rendering channels prone to avulsion. To the contrary, inWGP, high unit stream power and lower sediment supply to theWGP rivers may be responsi-ble for degradation and incision of river channel. Such hydrological differences triggered by climatic and tectonic variations may have existed for a fairly long time and this has produced marked geomorphic diversity across the plains. This means that every river has its own peculiarities and therefore needs to be dealt with care-fullyin terms of its management. Unfortu-nately, these factors are yet to become a part of river management in India.History of Flood ManagementKeeping in view the hydrological and geomorphological perspectives outlined above, let us now briefly examine the flood management strategies in Bihar plains with particular reference to the Kosi. Initial suggestions for flood control for the Kosi river during late 19th century included construction of marginal embankments, high dam at upstream section, river train-ing in lower reaches and a series of bar-rages and canals. The famous Calcutta Conference (1896-97) concluded against any major flood control measures except for short embankments. The Patna Flood Conference in 1937 also voted against the embankment strategy and G F Hall, chief engineer noted that “embankments merely transferred trouble from one area to another and that they give rise to false sense of security”. A later investigation by C C Inglis also highlighted that the cause of movement of the Kosi river is the build-ing up of the submontaneous delta (fan) due to very large influx of sand from the hinterland and suggested detailed investi-gations in terms of river discharge, silt load, ground elevations and subsoil water levels before finalising any scheme. No such investigations ever followed and no definitive action could be taken until 1953. The debate on “embankment or no embankment” continued and a certain section of engineers kept pushing the embankment strategy on the grounds that constraining the river between the embankments would increase the waterway of the river through increase in velocity and erosive power. The arguments against theembankment strategy that these would affect the natural flow of water and trap the sediments apart from creat-ing problems of waterlogging prevailed until this time.A very severe flood in 1953-54 and the subsequent social and political pressure led to the formulation of the “Kosi project” in 1954. The project consisted of (a) a barrage at Bhimnagar and afflux bund, (b) embankment downstream of barrage on both sides, (c) eastern and western canal system, (d) hydroelectric power station in eastern canal, and (e) a high dam at Barahkshetra. This project was primarily aimed at flood control and to provide irriga-tion for increasing agricultural productivity. The project started in 1959 and the river was diverted through the barrage in 1963. Barring the high dam at Barahkshetra, most of the components of the project have been completed or are still in progress. Apart from the huge irrigation potential of the project, the embankments on both sides of the river formed a very important component – designed to protect about 2,800 sq km of land in north Biharand Nepal from floods. While the success of the entire project is debatable, it has certainly not served the objective of flood control. Several large floods and frequent breaches in the embankments havecontinued to occur in the region. In addition,several adverse effects of the Kosi project have been noted, viz, drainage congestion and waterlogging, rise of river bed level, and reduction in crop produc-tivity due to reduced silt flux on the floodplains. These issues have been dis-cussed in great detail in a recent book titled Trapped! Between the Devil and Deep Waters (Mishra 2008) and have also been reported in the pages ofEPW (Mishra 2008). The bottom line however is that flooding problems have been further aggravated by the construction of embankments on both sides and the barrage in the upstream reach. This is not surprising if one recalls the statement of F C Hirst in 1908 that …it is more than probable that the heavy floods which in every recent year have de-vastated several of the North Bihar districts are mainly, if not entirely, due to prevalence of embankments…..An embankment, with little or no waterway through it for carry-ing off the floodwaters, is a glove thrown in Nature’s face – an insult which she has yet not been known to leave unavenged. Lessons of 18 August The Kusaha breach on 18 August 2008 was not totally unexpected but the disasterthat followed was unprecedented. The breach occurred at a discharge of 144,000 cusecs in the river and more than 80-85% of the flow of the river passed through a new course east of the original course. On 24 August, the river was flowing as an approximately 22 km wide channel and swelled to about 35 km wide in later weeks. Till today the river flows through this new course and has inundated large areasinmorethan 1,000 villages affecting nearly 350,000 people in the region. This breach was dif-ferent from all previous breaches which have kept occurring repeatedly along the western or eastern embankment along the Kosi during the last 4-5 decades. During the Kusaha breach, two unusual events occurred: (a) the river moved east of the modern course – unlike the westward mi-gration trend over the last 200 years, and (b) the total movement wasoftheorderof about 120 km – an orderofmagnitude higher than any single movement record-ed in historical times.Reports available suggest that the eastern embankment around Kusaha has been under pressure for some time. The repetitive
BIHAR FLOODSEconomic & Political Weekly EPW november 15, 200845satellite images show that the river has been moving towards the eastern em-bankment at least since 1979. A breach in the embankment at Kusaha was detectedas early as 5 August 2008 and perhaps a timely action could have averted such a disaster. However, the fact remains that embankments have not produced the de-sirable results and that they have already outlived their effective lifespan. Thehumanintervention with a river like Kosi had perhaps reached a thresh-old and the lateral movehave occurred sooner or later. The basic principlesof earth surface processes governthatlarge changes happen after a threshold is ex-ceeded. The eastward shift of the river and that too by about 120 km suggests that the river may have reached the threshold of its westward movement and the natural fan building processes would demand an eastward sweep. Most analysts agree that the confinement of the Kosi within the embankment further worsened the situa-tion and has caused significant aggrada-tion within the channel belt. The river was possibly flowing at a higher elevation than the surrounding areas outside the em-bankment. There have been reports about the rising bed level of the Kosi from many parts in the plains in downstream reaches as evidenced from siltation of canals and sediment budgeting at downstream stations. In the upstream reaches, the Kosi barrage has also accentuated the aggradation. In a study carried out byIIT Delhi for a stretch of 167 km from Chatra to Koparia suggested that all reaches changed from a “degrading” to “aggrading” in the post-embankment period with rates as high as 150 mm/year creating a volume change of the order of 20-30MCM/year. Therefore, unlike the previous move-ments and flooding history, this disaster seems to have a strong human component in terms of our intervention and ill-planned, outdated flood management strategies. Time has come to realise that a long-term solution to floods does not lie in “controlling” the river through embankments but in “managing” the rivers through inte-grated planning and understanding of riverine processes. This event has once again raised doubts about the embank-ment strategy although we alreadyhadthe deplorable results from Chinese rivers and from Damodar embankments. Itistimeto develop a process-based understanding of rivers and encompass all physicalattributes of the earth’s surface involved in water cycle for flood management. It calls foranational policy for flood management to reduce risks to the people and to the developed and natural environment from flooding.What Next?If there is any lesson to be learnt from the Kosi disaster – unless this event passes off as yet another natural disaster – it is that we need to move towards a strategy which emphasises on “river management” rather than “river control”. The embankment strategy has been questioned at the inter-national level citing the failure in Missis-sippi and three major Chinese rivers and alternative methods such as small-scale irrigation strategies are now favoured flood control measures in many flood-prone countries such as Bangladesh. Our experi-ence shows that there has been no appre-ciable flood moderation in the Kosi and other rivers of north Bihar even after the construc-tion of embankments and dam and there will be very little effect on the river stage. It is time to adopt an integrated river basin management which requires a rigorousunderstanding of the physical processes by which river channels are formed and maintained. An understanding of the historical and site-specific conditions is critical for successful river management including floods. To achieve this, there is a strong need to include the geomorphological parameters of the river basin in flood analysis, as these parameters govern the hydrological response of river basin. In a traditional approach of rainfall-runoff analysis the effect of geomorphology is not considered. However, some recent works have focused on incorporating drainage network parameters and river morphology in flood analysis. More work on this aspect particularly on the Indian rivers is needed for better understanding of flooding processes and their causative factors. A proper integration of fluvial geomor-phology and river engineering aided by high resolution aerial and satellite data is the need of the hour to understand the flood-ing behaviour of the rivers such as those draining thenorth Bihar plains. Complete understanding of related fluvial processes such as lateral shifting by avulsion andcut-offs would be extremelydesirabletoplan the flood controlstrategy in north Bihar plains. The detailedgeomorphological in-vestigations must be coupled withlong-term hydrological data to develop a better under-standing of the causative factors of floods in the area. The flood mitigation procedures must take into account the geomorpho-logical factors to derive long-term benefits. Anassociatedproblem in flooding is the ex-tensiveerosionin the hilly catchment area which contributes excessive sediment load to the Himalayan rivers thereby influencing the flow parameters. A better understanding of the sediment supply in theupper catchments of the riversisneces-sary particularly in relation to developing long-term solutions to flood mitigation as afforestation and landuse changes. This is where an “integrated management” of riverbasin becomes crucial.Another aspect is to develop basin-scale flood risk maps and to improve the decision support systems. Traditional methods of flood risk mapping are based on ground surveys and aerial observations, but when the phenomenon is widespread, such methods are time consuming and expensive. Furthermore, timely aerial observations may be impossible due to prohibitive weather conditions. We need a multi-parametric approach for delineating the flood risk areas in a geographic information systems (GIS) environment. In a recent research with students at IIT Kanpur, we attempted to produce flood risk map in parts of the Kosi river basin, north Bihar using one of the multi-criteria decision-makingtechniques, Analytical HierarchicalProcess (AHP). The basic aim of this research was to create easily-readable and rapidly-accessible flood risk mapsbasedon morphologic, topographic, and demographic data. A combination of different data sets such as remote sensing images (IRSLISS-III data), census data (1991), and topographic maps obtained from government agencies was used to compute a composite index of flood risk based on multi-parametric analysis. Finally, all data was integrated in aGIS environment to prepare a flood risk map which not only defined the suscepti-bility of each area to inundation but also provided means for assessment of flood risk in terms of loss of life and property.
BIHAR FLOODSnovember 15, 2008 EPW Economic & Political Weekly42caused by drainage congestion. It recog-nises the basic fact that the flood plains are essentially the domain of the river and as such all developmental activities there must be compatible with the flood risk in-volved. However, due to inability of the governments to prevent or check the en-croachments of the flood plains, no begin-ning could be made so far in flood plain zoning. The government of India circulated a model bill on the flood plain zoning in 1974 among all the states of the country with the request to enact the same, but no state except Manipur could do so and this state also could not implement it. Flood proofing is essentially a combination of structural change and emergency action. Raising of some flood-prone villages above predetermined flood level and connecting them to nearby road or high land has been done, but the enthusiasm was lost very soon as the method did not provide any protec-tion to surrounding agricultural areas. The flood forecasting and warning system is already working in Bihar satisfactorily.The most common adjustment to the flood hazard in flood plains is simply to bear the loss and live with flood in areas where there is some serious constraint (as in north Bihar). Under the situation, floods become a matter of concern for all sections of society living in the affected areas and flood management needs active involve-ment and participation of all of the stake-holders to fulfil its objectives. The National Water Policy (2002) also advocates a parti-cipatory approach with the involvement of various governmental agencies, users and other stakeholders in an effective and de-cisive manner, in various aspects of plan-ning, design, development and manage-ment of water resources schemes.The problem of flood in Bihar is very com-plex and so is the task of its management. A strong political force is required for seeking and ensuring cooperation of Nepal without which sustainable long-term flood manage-ment in Bihar is not possible. Competent and capable engineers, well acquainted with the latest technology and enlightened and efficient bureaucracy should beentrusted with the job of policymaking and its imple-mentation. Their work should not be inter-fered with. Last but not the least,the involve-ment of the stakeholders in all stages of the process is a key factor for the success of the effort and optimisation of benefits.Kosi: Rising Waters, Dynamic Channels and Human DisastersRajiv SinhaThe recent Kosi floods have proved once again that inadequate control measures have been responsible for the recurring disasters. Typically flood control and riverinestudies focus on hydrologicalinformation, whereas a much more integrated approach that pays attentionto specificmorphological factors is required. Since Kosi is a dynamic river with auniquemorphology andbecause it is a river which has always carried high sediment loads, flood management strategies mustbeattuned to such specific parameters of the river, besides being much more than mere “river control” through embankments.Rivers play a critical role in human society and history as they are the major source of fresh water, trans-portation, and resources. However, this relationship is often “troubled” because changes in river discharge (floods or droughts) or position can play havoc with permanent settlements. Such changes can be caused by both natural forcing as well as human interventions or a combination of both. Natural processes may include short-term changes in sediment load, water volume or, long-term changes in relative sea level or climate change. Human interventions could impact in changes in sediment load or run-off through water resource management schemes such as dams, barrages and embankments. Human alterations of river systems can have many important consequences primarily because river systems are dynamic and highly integrated systems and any change in any part of the river can easily propagate and affect the whole system. The recent flood in the Kosi is certainly one of the biggest “human” disasters in recent years and it has sent out a strong signal that our flood management strategies are questionable and our preparedness to face such events is far too inadequate. Further, there has been a paradigm shift in flood management globally from “river control” primarily involving an engineering approach addressing the “effect” at a local scale to “river management” which empha-sises an integrated approach at a crossover of scales and addresses the cause rather than the effect. Even though India is a country drained by several large rivers, our river management strategies are rather rudi-mentary and our planners are yet to embrace modern approaches such as satellite-based monitoring and multi-criteria decision support system. This situation needs to be corrected to save a large population from repeated miseries of floods year after year.The Kosi: A Dynamic SystemThe Kosi river in north Bihar plains, eastern India is a major tributary to the Ganga river system and has long been considered as a problematic river due to recurrent and extensive flooding and frequent changes in its course. The gently sloping alluvial surface of the Kosi has been described as “inland delta”, “cone” and “megafan” by various researchers owing primarily to build-ing up of a very large positive topography caused by deposition of enormous quantity of sediments carried by the river which it is unable to transport. During the last two centuries, for which records are available, Rajiv Sinha (rsinha@iitk.ac.in) is with the Department of Civil Engineering, Indian Institute of Technology, Kanpur.
Stream power Sediment supply 
4 – 2 – 0 – Unit stream power (Mean annual flood) Unit stream power (Bankfull discharge) 
BIHAR FLOODSnovember 15, 2008 EPW Economic & Political Weekly44Tectonically stable catchments would con-tribute significantly less quantities of sediment to the river systems. Similar to the rainfall distribution, spatial variation in the uplift rate along an east-west transect has also been observed. Research by several workers has shown that the eastern parts of the Himalaya are charac-terised by higher uplift rate (15-20 mm/yr) along the mountain front in Nepal, based on the analysis of deformation of terraces along Baghmati river system. The present models on Himalayan seismotectonics predict westward decrease inuplift rate along the Himalayanfrontin Dehradun(~7mm/yr). Therefore, not just that sediment production is low in their catchments, there is obviously no remobi-lisation of sediments due to incised nature of the river channels. It appears therefore that high sediment supply from upstream area coupled with low stream power inEGP rivers has resulted in aggradation of channels, frequent floods and rendering channels prone to avulsion. To the contrary, inWGP, high unit stream power and lower sediment supply to theWGP rivers may be responsi-ble for degradation and incision of river channel. Such hydrological differences triggered by climatic and tectonic variations may have existed for a fairly long time and this has produced marked geomorphic diversity across the plains. This means that every river has its own peculiarities and therefore needs to be dealt with care-fullyin terms of its management. Unfortu-nately, these factors are yet to become a part of river management in India.History of Flood ManagementKeeping in view the hydrological and geomorphological perspectives outlined above, let us now briefly examine the flood management strategies in Bihar plains with particular reference to the Kosi. Initial suggestions for flood control for the Kosi river during late 19th century included construction of marginal embankments, high dam at upstream section, river train-ing in lower reaches and a series of bar-rages and canals. The famous Calcutta Conference (1896-97) concluded against any major flood control measures except for short embankments. The Patna Flood Conference in 1937 also voted against the embankment strategy and G F Hall, chief engineer noted that “embankments merely transferred trouble from one area to another and that they give rise to false sense of security”. A later investigation by C C Inglis also highlighted that the cause of movement of the Kosi river is the build-ing up of the submontaneous delta (fan) due to very large influx of sand from the hinterland and suggested detailed investi-gations in terms of river discharge, silt load, ground elevations and subsoil water levels before finalising any scheme. No such investigations ever followed and no definitive action could be taken until 1953. The debate on “embankment or no embankment” continued and a certain section of engineers kept pushing the embankment strategy on the grounds that constraining the river between the embankments would increase the waterway of the river through increase in velocity and erosive power. The arguments against theembankment strategy that these would affect the natural flow of water and trap the sediments apart from creat-ing problems of waterlogging prevailed until this time.A very severe flood in 1953-54 and the subsequent social and political pressure led to the formulation of the “Kosi project” in 1954. The project consisted of (a) a barrage at Bhimnagar and afflux bund, (b) embankment downstream of barrage on both sides, (c) eastern and western canal system, (d) hydroelectric power station in eastern canal, and (e) a high dam at Barahkshetra. This project was primarily aimed at flood control and to provide irriga-tion for increasing agricultural productivity. The project started in 1959 and the river was diverted through the barrage in 1963. Barring the high dam at Barahkshetra, most of the components of the project have been completed or are still in progress. Apart from the huge irrigation potential of the project, the embankments on both sides of the river formed a very important component – designed to protect about 2,800 sq km of land in north Biharand Nepal from floods. While the success of the entire project is debatable, it has certainly not served the objective of flood control. Several large floods and frequent breaches in the embankments havecontinued to occur in the region. In addition,several adverse effects of the Kosi project have been noted, viz, drainage congestion and waterlogging, rise of river bed level, and reduction in crop produc-tivity due to reduced silt flux on the floodplains. These issues have been dis-cussed in great detail in a recent book titled Trapped! Between the Devil and Deep Waters (Mishra 2008) and have also been reported in the pages ofEPW (Mishra 2008). The bottom line however is that flooding problems have been further aggravated by the construction of embankments on both sides and the barrage in the upstream reach. This is not surprising if one recalls the statement of F C Hirst in 1908 that …it is more than probable that the heavy floods which in every recent year have de-vastated several of the North Bihar districts are mainly, if not entirely, due to prevalence of embankments…..An embankment, with little or no waterway through it for carry-ing off the floodwaters, is a glove thrown in Nature’s face – an insult which she has yet not been known to leave unavenged. Lessons of 18 August The Kusaha breach on 18 August 2008 was not totally unexpected but the disasterthat followed was unprecedented. The breach occurred at a discharge of 144,000 cusecs in the river and more than 80-85% of the flow of the river passed through a new course east of the original course. On 24 August, the river was flowing as an approximately 22 km wide channel and swelled to about 35 km wide in later weeks. Till today the river flows through this new course and has inundated large areasinmorethan 1,000 villages affecting nearly 350,000 people in the region. This breach was dif-ferent from all previous breaches which have kept occurring repeatedly along the western or eastern embankment along the Kosi during the last 4-5 decades. During the Kusaha breach, two unusual events occurred: (a) the river moved east of the modern course – unlike the westward mi-gration trend over the last 200 years, and (b) the total movement wasoftheorderof about 120 km – an orderofmagnitude higher than any single movement record-ed in historical times.Reports available suggest that the eastern embankment around Kusaha has been under pressure for some time. The repetitive 
BIHAR FLOODSEconomic & Political Weekly EPW november 15, 200845satellite images show that the river has been moving towards the eastern em-bankment at least since 1979. A breach in the embankment at Kusaha was detectedas early as 5 August 2008 and perhaps a timely action could have averted such a disaster. However, the fact remains that embankments have not produced the de-sirable results and that they have already outlived their effective lifespan. Thehumanintervention with a river like Kosi had perhaps reached a thresh-old and the lateral movehave occurred sooner or later. The basic principlesof earth surface processes governthatlarge changes happen after a threshold is ex-ceeded. The eastward shift of the river and that too by about 120 km suggests that the river may have reached the threshold of its westward movement and the natural fan building processes would demand an eastward sweep. Most analysts agree that the confinement of the Kosi within the embankment further worsened the situa-tion and has caused significant aggrada-tion within the channel belt. The river was possibly flowing at a higher elevation than the surrounding areas outside the em-bankment. There have been reports about the rising bed level of the Kosi from many parts in the plains in downstream reaches as evidenced from siltation of canals and sediment budgeting at downstream stations. In the upstream reaches, the Kosi barrage has also accentuated the aggradation. In a study carried out byIIT Delhi for a stretch of 167 km from Chatra to Koparia suggested that all reaches changed from a “degrading” to “aggrading” in the post-embankment period with rates as high as 150 mm/year creating a volume change of the order of 20-30MCM/year. Therefore, unlike the previous move-ments and flooding history, this disaster seems to have a strong human component in terms of our intervention and ill-planned, outdated flood management strategies. Time has come to realise that a long-term solution to floods does not lie in “controlling” the river through embankments but in “managing” the rivers through inte-grated planning and understanding of riverine processes. This event has once again raised doubts about the embank-ment strategy although we alreadyhadthe deplorable results from Chinese rivers and from Damodar embankments. Itistimeto develop a process-based understanding of rivers and encompass all physicalattributes of the earth’s surface involved in water cycle for flood management. It calls foranational policy for flood management to reduce risks to the people and to the developed and natural environment from flooding.What Next?If there is any lesson to be learnt from the Kosi disaster – unless this event passes off as yet another natural disaster – it is that we need to move towards a strategy which emphasises on “river management” rather than “river control”. The embankment strategy has been questioned at the inter-national level citing the failure in Missis-sippi and three major Chinese rivers and alternative methods such as small-scale irrigation strategies are now favoured flood control measures in many flood-prone countries such as Bangladesh. Our experi-ence shows that there has been no appre-ciable flood moderation in the Kosi and other rivers of north Bihar even after the construc-tion of embankments and dam and there will be very little effect on the river stage. It is time to adopt an integrated river basin management which requires a rigorousunderstanding of the physical processes by which river channels are formed and maintained. An understanding of the historical and site-specific conditions is critical for successful river management including floods. To achieve this, there is a strong need to include the geomorphological parameters of the river basin in flood analysis, as these parameters govern the hydrological response of river basin. In a traditional approach of rainfall-runoff analysis the effect of geomorphology is not considered. However, some recent works have focused on incorporating drainage network parameters and river morphology in flood analysis. More work on this aspect particularly on the Indian rivers is needed for better understanding of flooding processes and their causative factors. A proper integration of fluvial geomor-phology and river engineering aided by high resolution aerial and satellite data is the need of the hour to understand the flood-ing behaviour of the rivers such as those draining thenorth Bihar plains. Complete understanding of related fluvial processes such as lateral shifting by avulsion andcut-offs would be extremelydesirabletoplan the flood controlstrategy in north Bihar plains. The detailedgeomorphological in-vestigations must be coupled withlong-term hydrological data to develop a better under-standing of the causative factors of floods in the area. The flood mitigation procedures must take into account the geomorpho-logical factors to derive long-term benefits. Anassociatedproblem in flooding is the ex-tensiveerosionin the hilly catchment area which contributes excessive sediment load to the Himalayan rivers thereby influencing the flow parameters. A better understanding of the sediment supply in theupper catchments of the riversisneces-sary particularly in relation to developing long-term solutions to flood mitigation as afforestation and landuse changes. This is where an “integrated management” of riverbasin becomes crucial.Another aspect is to develop basin-scale flood risk maps and to improve the decision support systems. Traditional methods of flood risk mapping are based on ground surveys and aerial observations, but when the phenomenon is widespread, such methods are time consuming and expensive. Furthermore, timely aerial observations may be impossible due to prohibitive weather conditions. We need a multi-parametric approach for delineating the flood risk areas in a geographic information systems (GIS) environment. In a recent research with students at IIT Kanpur, we attempted to produce flood risk map in parts of the Kosi river basin, north Bihar using one of the multi-criteria decision-makingtechniques, Analytical HierarchicalProcess (AHP). The basic aim of this research was to create easily-readable and rapidly-accessible flood risk mapsbasedon morphologic, topographic, and demographic data. A combination of different data sets such as remote sensing images (IRSLISS-III data), census data (1991), and topographic maps obtained from government agencies was used to compute a composite index of flood risk based on multi-parametric analysis. Finally, all data was integrated in aGIS environment to prepare a flood risk map which not only defined the suscepti-bility of each area to inundation but also provided means for assessment of flood risk in terms of loss of life and property. 
