Demonstrating the Benefits of Community Water Management and Improved Crop Technologies in the Coastal Zone of Bangladesh Isn’t Simple
This project, ‘Implementing community level water management in coastal Bangladesh’, was funded through CPWF’s Innovation Fund.
The densely populated coastal zone of Bangladesh is home to 38 million people (approximately 780 people per square kilometer). There, most farmers grow a single rice crop during the rainy (aman) season using tall, local varieties that can survive stagnant flooding (water depth of 0.3 to 0.5 meters for weeks to months at a time), but which have low yield (2 to 3.5 tons per hectare) and are slow to mature. Traditionally in the dry
season following the aman crop farmers will plant a late, low-input and low-yielding (approximately 0.5 tons per hectare) legume crop or allow the land to lie fallow.
There is huge potential for increasing cropping system productivity in the coastal zone. Trials conducted under the CPWF-Ganges Basin Development Challenge have shown that use of high yielding aman varieties (HYV) coupled with good management can double yield, provided that water is managed to avoid stagnant flooding because HYVs are less tolerant to such flooding than traditional varieties. HYVs mature earlier, allowing an earlier rice harvest and earlier sowing of a dry season crop. Our on-farm trials have also demonstrated that the earlier maturity of HYV enables double- and triple-cropping with rice or with a combination of rice and high yielding or high value rabi crops (dry season non-rice crops such as maize, sunflower and water melon).
The net result is that system productivity can be increased from 3-6 tons per hectare to 11-19 tons per hectare, depending on location. The key to making the change to HYV rice and cropping system intensification and diversification is improved water management.
The rivers of the coastal zone of Bangladesh are tidal, extending up to 200 kilometers inland from the coast. The tidal fluctuation is more extreme (up to several meters) in the rainy season, resulting in tidal flooding of adjacent lands. As the dry season progresses, the rivers become increasingly saline due to low river flows and seawater moving further up the rivers.
Most of the agricultural lands of the coastal zone of Bangladesh are protected by polders that were constructed by humanitarian projects in the 1960s and 1970s (Map 1). Polders prevent inundation during the rainy season, thus enabling production of a tall, low-yielding aman crop, and also prevent saline water intrusion during the dry season. The construction of polders involved building large embankments around the perimeters of the islands formed by the system of of rivers in the delta. There are currently 139 polders in coastal Bangladesh, varying in size from a few hundred hectares to over 30,000 hectares.
Despite enormous investment polders are still home to some of the world’s poorest and most vulnerable people. A recent CPWF survey showed that almost 80% of rural people in the polders of Bangladesh are living below the poverty line ($1.25/person/day), compared with the national average of about 40%.
The polders contain dense, natural drainage networks. The larger of these internal canals (khals) are connected to the surrounding rivers by sluice gates (photo 1) installed in the polder embankments. With judicious management of the sluice gates, water can be taken into or drained out of the polders as needed. Operation of the sluice gates generally follows the following pattern:
- Closed during the dry season (November to March/April) once the rivers become too saline for irrigation (except in areas that grow brackish water shrimp)
- Opened in June/July to bring in fresh water to start the aman crop
- Closed during the monsoon season (May to September/October) to prevent flooding due to high water level in the rivers at high tide
- Opened after the end of the monsoon rains (October) to bring in fresh water to finish of the aman rice, if needed, as well as for multiple other uses.
While the land surface in the delta is relatively flat, in fact, there is variation across the landscape. The lands are typically classified as low lands, medium lands and high lands based on maximum flooding depth during the rainy season, which typically ranges from never flooded on the high lands to 50 cm or more in the low lands. When it rains, water flows from the high lands to the lower lands where it accumulates to greater depths.
There is tremendous opportunity for improving productivity in the polders through improved water management and crop technologies. At low tide (twice a day!) during the rainy season the river level is generally lower than the land level within the polders, creating opportunities for drainage of excess water from heavy monsoon rainfall to a level that would allow HYVs to flourish. Furthermore, drainage of earlier maturing HYVs just prior to harvest allows the soil to dry sufficiently for timely establishment of rabi crops. However, these opportunities have not yet been recognized. Furthermore, there is fresh water in the large rivers for part to much of the dry season, currently untapped, in much of the coastal zone.
Achieving good water management within a polder requires an integrated approach. Its implementation requires investment in earthworks, such as excavation of khals (many of which are now silted up), separation of lands of different elevation by small embankments to prevent accumulation of water in lower lands, and the installation of small drainage/irrigation canal networks among the fields. In addition to facilitating drainage, excavation of khals would increase the capacity to store freshwater for irrigation during the dry season. Achieving good water management within polders requires community involvement, at times supplemented by infrastructure investments by local and national institutions.
The community water management pilot
With this in mind, using money from a CPWF Innovation Fund grant, we sought to implement a pilot water management unit with a small community (37 farmers) in an area of about six hectares. After discussing our ideas in several meetings, the local community agreed to implement the project and participated in selecting a suitable area, identifying lower and higher lands within that area, and constructing small levees and drainage canals. The area was bounded on most sides by built up roads so only a couple of small levees (Photo 2) needed to be constructed to separate this area from surrounding areas and prevent it from being flooded from outside. The farmers agreed to grow HYV rice during the rainy season and were provided with enough seed to plant all their land (Photo 3). They were also given information on recommended management practices for HYV aman rice.
Planting of the aman rice was postponed by about one week beyond the target date. Due to a delay in land preparation while waiting for the soil to dry sufficiently following the 2011 rice crop, the 2012 sesame crop harvest was planted and harvested late. This in turn resulted in a delay in bringing fresh water into the polder, which was necessary for preparing the seedling nurseries. Trapped in the cycle of an existing system, initiating a pilot community water management project in which timing of crop establishment and harvest are critical to successful implementation is particularly difficult. Nonetheless, the farmers did transplant the 2012 aman rice about one week earlier than their usual practice.
During the aman season, there were two heavy rainfall events (each in excess of 250 mm). The first occurred in August (shortly after transplanting), when the HYV plants were almost completely submerged, and the second in September. The farmers systematically operated the sluice gates to drain at each low tide and the excess water was removed within three to four days. As such, community water management within the mini-watershed prevented damage to the crops.
While the time of rice transplanting was early enough, only half of the farmers planted HYV aman rice. The rest grew local varieties that matured several weeks later than the HYV. Instead of draining fields in early November, shortly before the HYVs were ready for harvest, the community brought more river water into the fields to allow local varieties to mature, with the consequence that lands remained flooded until early December. By then the weather had become cold and foggy and the soil did not start to dry out until February, when the temperature started to rise. As a result, tillage and sowing of the rabi crops was delayed until the second half of February and early March, two to three months after our target sowing date of early December. Successful intensification depends on community members synchronizing planting and harvest dates and the timing of water supply and drainage.
Six farmers were selected to grow hybrid maize and sunflower, and improved varieties of mungbean and sesame during the rabi season on about 0.7 hectares of land in the watershed. The remaining farmers grew a local variety of sesame, the traditional local rabi crop. However, there was a huge amount of rainfall (total 420 mm) in May (May is a risky month for pre-monsoon rains and cyclones), and all the crops were virtually destroyed (Photo 4), apart from a couple of sunflower crops that were dibbled (seeds dropped into small holes made with a ‘planting stick’) into the moist soil on 31 December to demonstrate the benefits of early sowing. These crops were successfully harvested before the rains came. Unless all farmers in a mini-watershed grow HYV rice, allowing for early drainage of field, it is not possible to demonstrate or achieve the benefits of early rabi crop establishment and high yielding/high value alternative rabi crops.
Of course the farmers had many good reasons for not growing HYV rice, including the lower price of HYV in the local market; the need to invest in fertiliser to achieve the yield benefits of HYV; and the fact that tenant farmers have to give two-thirds of their crop to the land owners, while bearing all the production costs. (Twenty-five percent of the farmers in the pilot watershed area were leasing land.) Furthermore, it appears that the farmers were waiting for the project to provide full financial support (for fertilisers, labour etc.), a common practice amongst projects in the area. The combination of these factors prevented us from demonstrating the vastly superior production possibilities of improved community watershed and crop management.
How to successfully demonstrate the concept?
The bottom line is that agricultural productivity of the polders has to increase in order to feed the growing population of Bangladesh. The potential for increased yields is far higher in the polder areas than in other parts of the country, where they are already growing HYV and practising double- and triple-cropping (Photo 5).
The purpose of our research is to provide a “proof of concept”. By showing that it is indeed possible to dramatically raise productivity through better water control and farm system intensification, we can contribute to dialogues within farming communities and with policymakers. Widespread adoption will depend on communities agreeing to synchronize water control and crop management, and public sector institutions agreeing to help with investment in water control infrastructure. But the first step is to prove that the opportunities are truly there.
In 2013/14, we plan to continue with ‘proof of concept’ research. Working with only six farmers in a much smaller watershed of 1.2 hectares, we will attempt to demonstrate the potential benefits of improved water management combined with improved varieties of rice/rabi crops and improved agronomic management. We will distribute all the inputs. The information we provide will build a foundation for engaging with communities and policymakers on what it takes to improve water control infrastructure within polders and foster community water management, thereby paving the way for more productive and diverse cropping systems. We anticipate our research will provide information and insights essential for better-informed decision-making.