I have seen the two articles on the Kosi Floods by Kanak Mani Dixit and Dinesh Kumar Mishra. There is little to add to those analyses, with which I am in substantial agreement. May I, however, supplement them with a few remarks on (a) India-Nepal relations, (b) the idea of flood-control, and (c) the future course when the waters subside?
The old Kosi and Gandak agreements (1954 and 59), whatever the thinking behind them, have left a legacy of resentment and mistrust in Nepal which has persisted despite the amendments of 1966 and 69 to take care of Nepal's concerns. That sense of grievance was aggravated by the Tanakpur episode. The Mahakali Treaty of 1996, signed after extensive consultations, has remained a dead letter, contributing to a worsening of India-Nepal relations rather than opening a new chapter as had been hoped. The old debate has now been revived by the Kosi floods. The India–Nepal relationship has been badly mismanaged on both sides. Ham-handedness and insensitivity on India's part, and excessive touchiness and readiness to misunderstand on Nepal's part, have combined to create a convoluted and volatile relationship between the two countries, which resists repair. Perhaps the best course would be to wipe the slate clean and start afresh. There is now a new government in Nepal, and a comprehensive review of the old treaties and agreements (including the Trade and Transit Treaty) is in any case inevitable. Why not scrap the lot and explore a new relationship? In doing so, it might be sensible to avoid excessive closeness and aim for no more than friendliness, correctness and a reasonable distance.
Secondly, leaving aside the urgent humanitarian problem of rescue, relief, assurance of essential supplies, shelter, medical help and so on, may I venture a few tentative remarks on dams and embankments as instruments of 'flood control'? Given the friability and proneness to mass-wasting of the Himalayan system and the waywardness of the Kosi and the heavy load of sediment that it carries, it was probably a mistake to build a barrage and embankments on that river. Interventions in such a river are definitely fraught with risk. Even if the embankment had been properly maintained, it might have given way in an exceptionally heavy flood. That is the nature of embankments. Even if they do not break down, they might cause various problems, such as a rise in the level of the river-bed and the consequent elevation of the river above the level of the ground on either side, possible attacks by the river further downstream, and of course the emergence of waterlogging and even flooding in the areas 'protected' by the embankments because water cannot drain from those areas into the river. While it might not be possible to rule out embankments altogether, and while in some specific instances they might have done some good without doing much harm, they are in general remedies worse than the disease, particularly in the case of a river like the Kosi.
Floods are natural phenomena. They will occur from time to time, in varying magnitudes and intensities. When the flood waters come, the river needs space to spread and accommodate them until they recede. The natural flood-plain of a river is an integral part of the river. If we build on it, or if we try to contain the river within embankments, we are asking for trouble. It might be asked: what is wrong with dams? A dam will create a reservoir which will surely provide space for the temporary storage and gradual release of floods, thus moderating them. That seems very plausible, but a dam-and-reservoir project is rarely built exclusively for flood-control. It is generally built for multiple purposes (irrigation, power-generation, flood-control, etc), and there is a conflict in-built into such projects. Flood-control would require the intended space in the reservoir to be kept vacant for accommodating flood-waters, whereas irrigation or power-generation would require the reservoir to be as full as possible; and as the latter are gainful activities in an economic sense, they are apt to prevail over flood-control. If the space meant for accommodating floods is not available when the flood comes, the gates will have to be opened in the interest of the safety of the dam, and the downstream area might experience a greater flood than it would have done if the dam had never been built. This has actually happened more than once.
Let us assume that a flood cushion is built into a dam project and is operated as such; or alternatively, that a dam is built exclusively for flood-control and strictly operated for that purpose. Under those circumstances, the dam might moderate a flood up to a point. Even in such a case, a flood larger than the 'design flood' would raise safety concerns and necessitate the opening of the gates; and this can happen at any time. This is an inherent danger in all dams.