Forewarning can minimise the devastation of river erosion

Dr Maminul Haque Sarker, Deputy Executive Director, Centre for Environmental and Geographic Information Services (CEGIS), talks to Naznin Tithi of The Daily Star about how early prediction can help minimise the impacts of river erosion and the importance of developing a national strategy for river erosion.

We have been witnessing increasing incidents of river erosion this year, which has already devoured vast areas of croplands and homesteads of people across the country. Do you think river erosion has been causing more damage this year compared to previous years?

Recently, we came across a lot of news reports in both print and electronic media about river erosion—how it has already affected a large number of people and damaged vast agricultural land and properties of people. But it cannot be said based on these reports that river erosion has particularly increased this year.

When a char or agricultural land erodes, it does not get much attention from the media. But when big structures such as roads, bridges, hospitals, townships are devoured by the rivers, it becomes news. We often do not give much attention to small scale river erosion, but if we look at the erosion and damage done throughout the year, we would find that a lot of the damage remain unaccounted for.

Usually, we take account of the losses and damage at the beginning of each year using dry season satellite images. We publish a prediction report annually which also includes location-wise annual monitoring of erosion and damaged infrastructure. This report is for the use of the national level stakeholders, such as the Planning Commission, BWDB (Bangladesh Water Development Board), LGED (Local Government Engineering Department), Disaster Management Bureau, etc. Analysing the time-series satellite images, we found that the annual river bank erosion along different large rivers, such as the Ganges, Jamuna and Padma, increases with the upsurge of annual peak flood discharge.

Do we currently have any erosion prediction system or early warning mechanism? Also, what about developing a strategy for addressing the causes and consequences of river erosion?

We at the CEGIS have been making yearly predictions for the last 15/16 years with the financial support from different projects of BWDB, WARPO (Water Resources Planning Organisation), UNDP and CEGIS. We are even giving predictions two years in advance since last year. We have seen from experience that our predictions are 70/80 percent accurate. Brac had worked in some erosion-prone areas based on our predictions and found the predictions to be 90 percent accurate.

Such predictions should be a part of our river management strategy. Because these predictions and early warnings could minimise erosion or reduce the damage to properties. The government can take initiatives for river bank protection and also for the relocation of the local people who are supposed to be affected by river erosion. But it will only be helpful if the government makes a national strategy for river management and works based on the predictions.

Moreover, river erosion is a natural disaster and the ministry of disaster management should be handling the issue. It is no longer a problem for the developed countries of the world, because they had taken care of the issue long ago. They have managed their rivers in a way that the rivers do not change their course anymore. However, since it is still a major issue in Bangladesh, it needs urgent attention of the authorities concerned.

Could you elaborate on how the river erosion prediction system works?

Understanding the morphological behaviour of braided and dynamic rivers is very important. It was found in late 1990s that time-series satellite images are very useful to understand the main rivers of Bangladesh. At the beginning of this century, CEGIS research team developed a method to predict the river bank erosion of the Jamuna River. CEGIS also developed the prediction method for the Ganges (Godagari, Rajshahi) to Aricha and Padma (Aricha to Chandpur).

An interpretation of satellite images may provide huge information on the location of dry season channels, sandbar and chars (vegetated island). There are different parameters based on which the predictions are made, such as the curvature of meandering river bend, width of the river channel, approach of channel to the riverbank, location of the channel—upstream or downstream, right or left bank, starting and ending point of river bank erosion with respect to the meandering bend, etc. Several hundreds of data are extracted from the time-series satellite images and are analysed statistically.

The extent of river bank erosion largely depends on the characteristics of the bank materials. Before developing the predictive tools, we have to know the properties of bank materials. Erosion also depends on the curvatures, width of channel, terrain slope and shape of upstream sandbar. We relate the extent of erosion with different parameters which provide the rules for erosion prediction.

During the last one and a half decades, the accuracy of our prediction has increased significantly. By this time, we worked in India to develop tools for assessing vulnerability of different structures and river bank erosion along the Kosi River. Water Resources Department (WRD) of Bihar has been successfully using our tools for the last two years.

River erosion has been affecting thousands of people annually in Bangladesh and many of the landless people have no other option but to migrate to the cities in search of a livelihood. Do you think the government's rehabilitation programmes are sustainable?

We have a database of how many people have been affected and how much land and resources have been lost to river erosion every year. But sadly, we could not take advantage of this data for lack of a proper plan. Our estimates suggest that currently river erosion affects some 55,000 people annually. Although thousands of people fall victim to river erosion every year, the affected people have little access to institutional support, and the government's measures are restricted to mostly relief distribution to the victims. It would be more effective if the vulnerable people could be taken under the government's Safety Net Programmes. It is important that the government agencies as well as the non-government organisations plan their relief efforts as well as the rehabilitation programmes based on the predictions.

There are NGOs which have been doing a lot of work at the community level. They are building schools and healthcare facilities for the flood and erosion affected people. But often, such facilities are built in erosion-prone areas, meaning that these structures are often washed away by the rivers within a short time. In one instance, we predicted that a certain part of an area was at risk of erosion. But an NGO built a school in that area without taking into consideration the predictions. After a year or two, the school building was devoured by a river. Such incidents can easily be avoided if the early warnings are taken into consideration.

Moreover, protecting the river bank becomes much easier with an early warning mechanism in place. We can take small-scale river bank protection plans or, in some cases, go for big-budget plans. For example, in order to save the banks of the Jamuna near the Jamuna bridge, the government has undertaken some big-budget plans, because the bridge needs to be saved at all costs. However, we can undertake small-budget programmes every year to protect the river banks in the erosion prone areas.  The most important thing for the government to do now would be to formulate a national strategy. The government spends a large amount of money every year to rehabilitate the victims of river erosion across the country. By developing a national level prediction system and proper guidelines for addressing the causes and impacts of river erosion, the process could be made much easier and a lot of money could be saved, and most importantly, the sufferings of thousands of victims of river erosion could be reduced.