The right of the river   March 2011

Among the most frequent natural disasters in Southasia are floods in major rivers, many of which are trans-boundary and therefore require coordination between neighbouring countries that share the waters. Yet, river-water management in Southasia remains a highly contentious issue. Dipak Gyawali, research director of the non-profit Nepal Water Conservation Foundation and former water-resources minister, who has been a vocal critic of the handed-down paradigms of river-water management as they presently stand. He spoke to Himal Southasian about the need to re-orient our thinking if we are to manage our rivers in a way that reduces the risk and damage caused by river flooding in the region.

Tell us about your involvement with Duryog Nivaran, an early attempt at evolving a region-wide vision of disaster management.

The idea was to evolve an alternative perspective on disasters, and we came up with some principles. One: Disasters are the unfinished business of development. Two: Disasters are a recurring phenomenon, rather than one-time events, and will occur more frequently and intensely due to climate change. Therefore, what happens between disasters is more important than what you do in the immediate aftermath of the disaster. The focus should be on the interim period to re-think development from a disaster risk reduction perspective. Three: You have an extreme physical event, which interacts with social vulnerability – that is what makes for a disaster. You cannot avoid the former, so work on reducing the latter. Four: It is the structures within society and the relationships between different social groups that determine why certain sectors are more vulnerable to disasters than others. Five: Disasters are also a forensic moment, a window of opportunity, for changing the way things are to reduce overall social vulnerability.We chose the term Duryog Nivaran, rather than some English acronym, because it meant in most Southasian languages from Colombo to Kathmandu ‘avoiding misfortune’, i.e., mitigating disasters. It was set up in 1994, and was a mix of academics and practitioners. Apart from the Intermediate Technology Development Group (ITDG) in Sri Lanka, other members came from Bangladesh, India, Nepal; later on, Oxfam-Pakistan, also joined the network.

You were also part of a study on adaptive strategies for dealing with disasters?
We carried over some of the lessons that came out of Duryog Nivaran into our research on floods and droughts. To understand how to respond to disasters, it is vital to study strategies adopted by disaster-affected people themselves. The first strategy is livelihood diversity. We found that people who have diversified livelihoods are better able to adapt to disasters. If you are partly reliant on farming and partly on labour, for instance, you will be able to cope better. Non-farm components of livelihood are particularly helpful, as they are less affected by floods and droughts. If a member of a family has a government job or is working outside the region, then the family’s resilience in times of adversity is greater. Flowing from this is a second point, one that goes against the standard development gospel: it is not the poorest of the poor who are worst affected by disasters, but the lower middle classes. For example, a marginal peasant with just enough land to provide food for six months out of the year will have labour as an alternative strategy, and he can resort to labour work for twelve months instead of six if a flood or drought wipes out his crops. However, a farmer who has just enough land to provide food year-round will have no fallback strategy.

Access to communication – both physical and in terms of information – is also vital. For example, in the Kosi floods of 2008, the loss of human life was less than in the aftermath of previous floods due to the ubiquity of mobile phones, through which people could warn each other. As for physical communication, in the Gujarat drought in 2000, when cattle were dying due to lack of fodder, the existence of road networks howsoever kachha enabled fodder to come to the cattle when the cattle could not go to where the fodder was.

We also found that dedicated flood-warning systems are a disaster in themselves. What happens is that between disasters, the equipment lies unused and unattended; you find that a rat has eaten through the wires or the battery is dead when needed most. What works much better is communication systems that are integrated into normal life, such as community FM stations. Similarly, if you build dedicated shelters for use in times of disaster, they run the chance of crumbling and being unusable when needed most. Flood-proofing schools, on the other hand, works much better, because they are in regular use between disasters, and hence maintained. When a disaster strikes, the entire community can use it. Thus, the period between disasters is vital for re-tooling development; and although solutions are sought within their domain by water managers, it is usually outside the water box – for instance, communications or architecture – that the most effective ones are to be found.

Many of the major rivers prone to flooding in Southasia are crossborder. What kinds of problems crop up in relation to such rivers?
In a sense, trans-boundary rivers are ‘convenient’, because one can always blame the other side for problems. The upper-riparian state in particular is always easy to blame. For instance, every time there is a flood in Uttar Pradesh or Bihar, every politician blames Nepal. When I visited Bihar as Nepal’s water minister, the first question I was asked was ‘Nepal ne bandh ke patra khol kar baadh kiun bhej diya?’ [Why did Nepal open the dam gates and sent the floods?] I responded that Nepal did not have any storage dam to open any gate: the two barrages, on the Kosi and the Gandak are under India’s control and they store no water that can be released! But it is much easier for politicians to blame the upstream state, rather than look at their mismanagement of the riverbeds and the unscientific embankments that they have built. In the case of Kosi and Gandak treaties, Nepal can do nothing as all management powers have been retained by the Indian side. We should also note that the trans-boundary water issues are not only between countries, but also between states within a country. For instance, Tamil Nadu will blame Karnataka and vice-versa for floods or droughts on the Kaveri.

As for Pakistan and India, the former always blames the latter for taking away the ‘good water’ during the dry season and letting out the ‘bad water’ in the wet season, thus causing floods. The problem is that the Indus Treaty is not a water-sharing treaty at all, but a partition treaty – first, British Raj lands were partitioned, and then the rivers. Three tributaries of the Indus went to India, and the remaining three to Pakistan. When you partition a river, you don’t really share it – that would require sharing the good and bad times on a continuous basis. So when there is a drought, the two countries should share the impact and distribute water equitably. If there is a flood, they should work together to ensure that the damage on both sides is minimised. But when you partition a river instead of sharing it, the attitude is, ‘You do what you like with your river, and I will do what I like with my river.’ Blame is something that comes structurally with the way the treaty is drawn up. The only reason the Indus Treaty still stands is that once you try to renegotiate it, you raise a whole lot of issues relating to partition of lands. So, the treaty remains un-reviewed, and does not really address issues of integrated basin management.

What are the specific problems in the current system of river management that lead to floods?
In post-colonial Southasia, water management has been disastrous. Rather, construction empowerment has become a runaway technology demolishing the environment and the lives of poor villagers in its path. For instance, in Bihar, 300 km of embankments at the time of Independence has burgeoned to 3600 km or so; and the construction is not guided by hydrological principles, but by distribution of political patronage to contractors. Similarly, in Pakistan after Independence there was a mad rush to start building. All the while, river basin management in the true sense has not happened. What you have instead are attempts at controlling rivers through embankments and dams – the nature and needs of a river or the poor dependent upon it are never considered. The construction is not even controlling floods, but rather blocking all the drainage and causing water-logging outside the embankments. The most damning statement against this kind of paradigm is the Kosi River, where more land has gone out of production through water-logging in Bihar than has come under irrigation over the last half century.

While there can be some economic gains as is the case with the Tarbela Dam on the Indus which has helped dry season irrigation, Tarbela cannot control all floods. Last year, for example, a significant amount of rain fell outside the dam’s catchment area, the easy outflow of which was blocked by drainage congestion from poor planning. The problem is technology-driven ‘solutions’ that fail to see the larger picture. We have established numerous departments that only know how to build dams and embankments, but there is no similar investment in assessing whether these have actually brought in adequate benefits to those that most needed it.

If the same level of precipitation as fell in Pakistan last year had actually occurred 200 years ago, it would have been less of a problem. But now it became a catastrophe due specifically to drainage congestion through embankments, railways, roads, dams, canals, and drained marshes and wetlands. In previous eras, most villages would have been well protected, on higher land. But now, due to uncontrolled urbanisation and population growth, cities have expanded into the vulnerable low areas – a process that is happening with urban sprawl throughout the region. The intensity and frequency of the extreme events is also increasing due to climate change, which magnifies the problem, as do the political systems in Southasia that are myopic and genetically allergic to long-term thinking.

How can we forecast flooding better?
The Pakistan floods brought attention back to conjunctive land and water management. Although there was excessive precipitation, we don’t know whether this is unusual – only that it has not happened in recent memory. When planning, what you do is study the limited data for the past whatever number of years available, and extrapolate into the future the probability of different levels of flooding – whether it will happen in fifty, 500 or 1000 years. The trouble with this kind of analysis is that it relies solely on past data and, most of the time, developing countries don’t have that kind of
data available.

Furthermore, climate change is essentially telling us that the future will not be like the past. So it makes such hydrological regression analysis increasingly pointless, rendering useless much of the technical aspects of, for instance, dam building – after all, dams are designed for the worst case scenario from a certain level of floods. When there is a high dam you have to make sure it lasts for at least 100 years to get your investment returns back, so you design it for the kind of event that may happen only once in 500 years or more. But with climate change, what is a once-in-500-year flood now could become more frequent.

Experts have pointed to an increased number of disastrous cloudbursts, such as took place recently in Leh. Is this also a result of climate change?
Climate change is a given. Cloudbursts – meaning intense rainfall of 70 or more mm in an hour – occurred in the past, but are now occurring with greater frequency and increasing intensity. In July 1993 there was a cloudburst just south-west of Kathmandu Valley in the Rapti-Agra-Kulekhani catchments: according to the official records, 540 mm of rain came down in 24 hours, but locals said that most of this fell in just nine hours! In the example you brought up, that was a cloudburst on a massive scale, as opposed to a more localised phenomenon.

Water has tremendous scouring potential as it flows, and a cloudburst increases the scouring ability of a river or a stream manifold. As the flow in a stream increases from, say, two litres per second to three, the scouring potential of the water shoots up five times. So, when the intensity and frequency of cloudbursts increases, it will carve up the land completely differently, making water management much more difficult.

In Kathmandu – a bowl-shaped valley – the effects of a cloudburst would be disastrous. There have been constructions right on the Bagmati River’s natural flood plains, massively constricting the flow. If there were to be a cloudburst like in 1993 or September 1981, the city would face a major disaster. Indeed, if such a cloudburst were to happen north of Kathmandu and sweep into the valley floor, the results could be just as bad as a major earthquake.
 
How do we turn around today’s ineffective water management?
There has to be an integrated ‘basin management’ approach, but unfortunately the term has been hijacked by the construction-oriented hydrocracy and debased. It should mean that we start with preventing drainage constriction at the lower end and go upstream to study precipitation and how groundwater is feeding the springs that keep the rivers flowing in the dry months. We need to accept other uses of a river basin (navigation, fisheries, rafting, religious rituals, etc) that would actually keep a river healthy, rather than myopically thinking as we do currently of only extracting river water. How can we make our rivers, especially in capital cities and major metropolis, into matters of national pride? The hydrocracies of all Southasian countries need to work together, keeping the broader water management perspective of the full basin in mind.

What specific measures would you suggest to ameliorate flooding?
More than anything else, avoid drainage constriction. Even if a once-in-1000-year flood comes, we have to ensure that this water can flow to the sea as easily as possible, instead of being waterlogged in productive agricultural land. The river has a right to flow – that’s its natural property. Respect that. When you kill a river by blocking its flow, you kill its self-cleansing properties. Of course we still need highways, bridges, canals and dams, but we have to build these in such a way that they do not constrict drainage. There has to be a right of navigation on rivers – you will then at least have to maintain a flow, even if it is only a polluted flow. There is little water in the Yamuna downstream of Delhi. Without flow, in the dry season channels get mucked up – and so, when a flood comes the impact is far worse, as the water is not able to flush out and backs up.

Another important principle is planning from the perspective of the marginalised. The needs of the fisher folk are never considered in water management. Many of these people have no land titles; they fish for a living. We need to recognise these benefits of a river. We must look at different human needs, and examine the ways in which they can be met. For example, when a town is growing and its water needs increase, in accordance with the current approach we simply build a dam to divert river water. Instead, you can use simpler and more cost-effective technologies to reduce water demand – smaller flush cisterns, rainwater harvesting, recycling, re-using, changing building codes to have roof gutters recharge groundwater and ponds etc.

The idea is to use ‘many 10 percent solutions’, rather than one big, perfectly optimised solution that tends not to work anyway. These simpler solutions can be picked up by different social groups, market entrepreneurs, NGOs, communities and not just big hydrocracies as in the case of dams and embankments. They involve some expense, of course, but are far less than the alternatives. How expensive has been the once-in-a-century clean-up of the Ganga or the Yamuna, after all? Wouldn’t it have been cheaper in the long run to use such smaller approaches instead to build resilience and good development at the local communities as well? Indeed, that is how the Rhine, the ‘sewer of Europe’ was cleaned up. Are the ‘many 10 percent solutions’ not more practical, cheaper, faster and less prone to social conflicts than the never-never land of Melamchi water supply diversion with a 30-km tunnel for Kathmandu Valley that is ever receding into the future? We must also look at appropriate technology as well as traditional technologies of ponds, wells and stone water spouts. Currently, Southasian water management, unlike say Singapore’s, is not broad-viewed but based on mere civil engineering – and it’s all about cement and earthmoving contracts!

Finally, to manage water effectively, one needs to think ‘outside the water box’. The response to the Kosi floods was to build more and more embankments. But, as my colleague water expert Ajaya Dixit once wrote, ‘Embankments can be classified into those that have already breached, and those that will breach.’ If you go for out-of-the-water-box solutions, you have a better chance, especially in light of the tremendous uncertainties that climate change is about to visit upon us. To tackle the Kosi flooding, for instance, we can plan to build houses on stilts, or villages with raised plinth levels, equipped with rubber rafts, mobile phones and flood-proofed schools and village council buildings that people can take refuge in for the few days of inundation. In such cases, the flooding can take place and then recede leaving behind valuable silt without causing loss of life or property, or at least reduce such vulnerabilities considerably. But such a solution is not available within conventional river-management thinking – it comes from ‘outside the water box’.