BRIDGING THE RICE YIELD GAP IN BANGLADESH - Sheikh A. Sattar^*

^* Agronomy Division, Bangladesh Rice Research Institute (BRRI), Gazipur, Bangladesh.

The almost uneven topography and humid tropical climate of the country with abundant monsoon rain offers a unique environment for the rice plant in Bangladesh. As such, rice is the staple food of the people of this country and is part of their culture. Once this land was capable of meeting the food demand but with passage of time the cultivated land started diminishing with the rapid growth of population. Feeding of these new mouths with rice became a heavy burden with the present level of food production and thus there was a necessity for food imports. To minimize food import emphasis was given to the research and development of rice since the birth of Bangladesh.

In spite of doubling rice production in the country since the introduction of modern varieties in the early seventies, Bangladesh has experienced a continued annual shortage of nearly 1.5 million tonnes of food grains (Karim, 1999). This shortage of food production will continue to increase even if the present level of population growth is maintained. In other words, rice production has to be increased by at least 60 percent to maintain the present level of rice requirements by the year 2020 (Bhuiyan & Karim, 1999). Increasing rice production further is a gigantic task since there is no scope for horizontal expansion of the rice area due to the gradual diminishing of cultivated land as a result of diverting its uses for houses, roads, industries and urbanization. Therefore, options available for increasing rice production are a) a breakthrough in the present yield potential of the varieties, b) full exploitation of the present yield potential of the existing modern varieties; and c) utilization of unfavourable but potential ecosystems for rice and or other systems of food production. This paper will elaborate on the possibility and means of exploiting yield potential of the existing modern varieties of rice and problems thereof.

2. STATUS OF RICE CULTIVATION IN DIFFERENT ECOSYSTEMS

In Bangladesh the rice-growing environment has been classified into three major ecosystems based on physiography and land types. These ecosystems are a) irrigated, b) rainfed, and c) floating or deepwater. The rainfed ecosystem has been further classified as rainfed lowland and rainfed upland. Thus, all rice varieties cultivated in the country are grouped into five distinct ecotypes such as a) Boro, b) Transplanted Aus (T. Aus), c) Transplanted Aman (T. Aman), d) Upland Aus (direct-seeded Aus), and e) Deepwater rice (Floating rice). Boro rice is grown completely under the irrigated ecosystem during the dry period (November to July) while T. Aman (during July to December), T. Aus (during April to August) and Upland rice (during March to July) are grown under the rainfed ecosystem. Of the total 13.8 million ha of cultivable land in the country (UNDP/FAO, 1988), 10.27 million ha (74.4 percent) are devoted to rice cultivation covering the above four ecosystems (BBS, 1993 & 1997; Hamid 1991). Besides these, special types of ecosystems like tidal wetland covering about 425 thousand ha and about 3.05 million ha of coastal saline soils are also included into the 10.27 million ha of rice land.

2.1 Area, Production and Yield Trend

The area, production and yield of rice in different ecosystems during 1998 can be seen in Table 1. Rice is cultivated on about 10.27 million ha including about 3.47 million ha of tidal wetlands (both saline and non saline). Modern varieties (MV) of rice cover about 31.4 percent of the total Aus area (lowland and upland) contributing 47 percent to the total Aus rice production, 51 percent of T. Aman area sharing about 59 percent of total Aman rice production and 92.4 percent of Boro area contributing 96 percent to the total Boro rice production.

Table 1. Area, Rough Rice Production and Yield under Various Ecosystems in 1998.

Ecosystem	Ecotype	Variety	Area (million ha)	Production (million tonnes)	Yield (t/ha)
Irrigated	Boro	Local Modern	0.22 2.67	0.52 12.03	1.54 4.51
Rainfed Lowland	T. Aman*	Local Modern	2.45 2.55	5.51 8.05	1.72 3.14
	T. Aus	Modern	0.49	1.35	2.76
Rainfed Upland	DSR Aus	Local	1.07	1.55	1.45
Floating	Deepwater rice	Local	0.81	1.46	1.80
Tidal Wetland Non-saline	Boro/T. Aman	HYV/Local	(0.425)
Saline	T. Aman	Local	(3.053)
Total			10.26	30.47

^* Including tidal wetlands

2.2 Production Constraints in Different Ecologies

2.2.1 Irrigated ecosystem

Irrigated rice is grown after the harvest of T. Aman rice or after harvesting a non-rice crop like potato, mustard or quick growing vegetables. Low temperature during the early vegetative stage of the crop prolongs growth duration and thus most of the existing modern varieties mature within 165 to 180 days. This requires use of a high level of inputs like irrigation, fertilizer and plant protection measures.

Of all the constraints of Boro rice cultivation, the most pressing one is the availability of irrigation water followed by farmers incapability of using the required amount of fertilizer in a balanced dose. Farmers of some regions delay planting in order to shorten growth duration vis-à-vis the production cost, particularly of irrigation. This delayed planting, however, reduces yield significantly. Recently BRRI released relatively shorter duration Boro varieties. But some farmers without being fully aware of the appropriate technologies for such varieties often stick to their traditional practices of early transplanting, subjecting the crop to cold injury during the flowering stage and thus realize poor harvests.

2.2.2 Rainfed lowland ecosystem

The rainfed lowland rice - T. Aus, the wet season first crop, is grown when sufficient rainfall occurs during April to August. This is the period experiencing higher temperatures with minimum diurnal fluctuation, moderate humidity during the reproductive stage, but with occasional scanty rainfall during the early vegetative growth period. Such a climate is very much conductive to higher vegetative growth of the crop with the lowest partitioning coefficient and development of pests and diseases. Rice varieties grown are all insensitive to photoperiod and mature within 115 to 130 days. Therefore, climatic limitation is the most important constraint for this rice.

The wet season second crop grown in the rainfed lowland ecosystem is known as T. Aman, cultivated during July to December, the full monsoon period. The crop experiences high rainfall and temperature during the vegetative stage and low temperature often associated with drought during the reproductive stage. Though the occurrence of severe drought is found to be about once in five years, and annual drought of various intensities affects about 2.3 million ha of T. Aman rice (Karim, 1999). In the medium flooded area, harvesting of Aman rice, that is not or is less sensitive to photoperiod, often becomes difficult due to standing water at the harvesting time. To face this problem, varieties strongly sensitive to photoperiod were grown in the past which mature after the field dries up. Since the Aman crop experiences two extreme climates at two ends, planting time is very important for this rice but often farmers cannot follow the appropriate planting schedule due to various socio-economic factors and thus planting gets delayed. This late planting causes yield decline. To save the crop from low temperature stress at the reproductive stage and also to establish a wheat crop timely after the harvest of Aman, shorter duration varieties with less or no sensitivity to photoperiod have been evolved recently for cultivation in shallow flooded areas.

2.2.3 Tidal wetlands

This ecosystem includes both saline and non-saline ecologies. Mostly T. Aman rice is grown in the non-saline area, and MVs cover only 15 percent of the area (Nasiruddin, 1999). There is little scope for further expansion of MVs unless varieties with relatively higher growth rate in the nursery bed, sturdy culm and profuse root system are evolved.

2.2.4 Rainfed upland ecology

The yield potential of this crop is the lowest due mostly to the unfavourable weather. The second important constraint is the lack of high yielding varieties. Unpredictable distribution of rainfall hinders timeliness of some management practices, particularly fertilizer management. Thirdly, the climatic conditions are very much conducive for rapid growth of weeds and pest and disease infestation. Sometimes, incessant drizzling for days just after the emergence of both rice and weeds, makes weeding difficult resulting in complete failure of the crop.

2.2.5 Deepwater

This is a very long duration crop sown in March/April and harvested in November/December. This rice requires a special habitat of prolonged flooding. The varieties are strongly sensitive to photoperiod and low tillering, producing a very high amount of biomass but with the least harvest index. The most important constraints of this rice are lack of varieties with high yield potential, unpredictable flooding, and low response to fertilizers.

Besides the constraints discussed above, floods are a common natural phenomenon and affect almost all rice crops in an area of about 2.6 million ha in a normal year (Karim, 1999). In the flood-prone areas, matured Boro and the ripening Aus crops are affected causing either partial to complete decomposition of grains or viviparous germination of the seed. T. Aman crop is also affected by flash floods in the early vegetative stage and by late floods during the maximum tillering to panicle initiation stages. The extent of crop damage depends on the duration and level of flooding and the stage of the crops.

2.3 Yield Potential of Released Varieties

BRRI has so far evolved 38 modern varieties of rice for cultivation in different ecosystems except for the deepwater environment. Table 2 gives the names of the most popular modern varieties with their potential yields under actual field conditions. Yield potential of the irrigated rice is the highest due to the most favourable weather conditions. The recently developed variety, BRRI Dhan 29, often gives grain yield as high as 9-10 t/ha under the irrigated ecosystem. Some other popular Boro varieties have yield potential of about 8 t/ha. Yield potential of the most popular T. Aman varieties, on an average ranges from 5.0 to 6.5 t/ha, and in some cases (like yield potential of varieties such as BR11) was found to be 7.4 t/ha. However, physiological potential yield of most of these varieties has not been worked out as yet.

Table 2. Yield Potential of the Popular Rice Varieties Grown in Various Ecosystems (BRRI, 1989, BRRI, 1999).

Ecosystem	Ecotype	Variety	Potential Yield (t/ha)
Irrigated	Boro	BR1 BR3 BR8 BR9 BR14 BR15 BR16 BRRI DHAN 28 BRRI DHAN 29 Purbachi	5.50 6.9 6.0 6.0 7.9 5.5 6.0 7.5 9.1 5.5
Rainfed Lowland	T. Aus	Purbachi BR1 BR3 BR14	3.5 4.0 4.0 5.0
	T. Aman	BR11 BR22 BR23 BRRI DHAN 32	6.5 5.0 5.5 5.0
Rainfed Upland	B. Aus	Local varieties	2.1-2.8
Deepwater	Deepwater rice	Local varieties	2.8-3.2

2.4 Yield Gap in Different Ecosystems

Although rice varieties with yield potential of about 8.6 t/ha of rough rice under most favourable environment are available, the researchers have been able to achieve, on an average, only 6.5 t/ha in the Boro, 6.0 t/ha in the Aman, and 4.0 t/ha in the Aus seasons (Figure 1). Such yield gaps are attributed mostly to the annual variations in weather conditions and ability of the variety to withstand a certain amount of pests and diseases pressure which varies from season to season and also from year to year. Surveys conducted in eight northern districts on farmer’s perception of rice yield and the associated management practices indicate that under real farming conditions this gap is much more wider. The yield gap between experiment stations and the farmers’ fields varies from 1.02 to 2.53 t/ha in Boro with an average of 1.63 t/ha, 1.56 to 3.39 t/ha with an average of 2.32 t/ha in Aman, and from 1.24 to 2.62 t/ha in Aus with an average of 1.69 t/ha (Table 3). Physical and socio-economic factors responsible for this large yield gap are a) low levels of management, b) lack of price stability, c) loss of farmers’ interest in investment due to unbalanced land tenure system, and d) other socio-economic factors.

Fig 1. Yield Gap in Different Ecotypes

Table 3. Rough Rice Yield (t/ha) achieved in experiment stations and national average yield by seasons

Season & Variety	Location	Experiment Station	National Average	Yield Gap
Boro BR3	Joydebpur Comilla Habiganj Barisal Rajshahi Rangpur	5.92 5.62 5.98 6.10 5.37 6.88	4.35	1.57 1.26 1.63 1.75 1.02 2.53
Aman BR11	Joydebpur Comilla Barisal Rajshahi Rangpur	6.60 5.43 5.38 4.77 5.47	3.21	3.39 2.22 2.17 1.56 2.26
Aus BR14	Joydebpur Comilla Barisal Rajshahi	4.2 3.9 4.0 5.28	2.66	1.54 1.24 1.34 2.62

Management factors that pull down farm level rice yields are wrong time of planting, use of poor quality seed, unbalanced use of fertilizers and other inputs, failure to control weeds during the critical competition period, and ineffective control of pests and diseases. Soil related factors associated with low farm level rice yield are highly reduced organic matter content, particularly in the north and north-western zone of the country (Table 4), and widespread occurrence of sulphur and zinc deficiencies. Among the other physical factors that often affect rice yield are unfavourable temperature, flood and drought.

Table 4. Soil Organic Matter Content (%) in some Rice Soils.

District	No. of Samples	Range	Mean
Barisal Chandpur Chittagong Comilla Faridpur Feni Gazipur Habiganj Jamalpur Jessore Jhenaidah Khulna Kishorganj Kushtia Mymensingh Naogaon Pabna Rajshahi Tangail	86 1 53 2 36 3 40 3 5 1 1 43 2 29 1 42 1 1 10	0.91-2.95 0.77-3.15 1.40-1.70 0.86-3.12 1.47-2.28 0.64-2.61 3.86-4.04 1.52-2.13 1.17-5.03 1.80-2.00 0.87-2.99 1.01-2.26 1.11-3.11	1.82 1.70 1.82 1.54 1.80 1.90 1.66 3.96 1.81 2.10 2.30 2.08 1.90 1.84 2.00 1.86 2.30 2.10 1.94

To meet the food requirement for the growing population, Bangladesh agriculture has been subjected to the highest cropping intensity of about 180 percent. Farmers often cannot follow the appropriate cropping scheme due to the shortage of draft power, unavailability of labour due to sudden high demand during the peak period when farm activities overlap across the seasons. Thus planting of rice becomes late resulting in reduction of yield at the rate of 60.0, 55.4 and 9.6 kg/ha for each day delay during Boro, Aus and Aman seasons, respectively. Contribution of other production factors depends on seasons and ecosystems. Fertilizer largely contributes to rice yield in all the ecosystems followed by weed control. Contribution of insect control varies with the level of infestations. In most of our studies it remained below the economic threshold level for which a negative contribution was found in some cases (Table 5). Moreover, rice yield decline due to over-mining of soil nutrients, organic matter depletion, floods and droughts, and the use of poor quality seed has not been critically analyzed so far. Therefore, an in-depth assessment of the effect of these factors on rice yield has yet to be understood clearly.

Table 5. Relative Contribution of Production Factors to Rice Yield by Ecosystem

Production Factor	Upland Aus	Transplant Aus	Transplant Aman	Boro
Fertilizer Weed Control Insect Control	0.42 0.27 0.05	0.52 0.53 -0.05	0.77 0.46 -0.23	0.52 0.44 0.04

Although the farm level rice yield remains below the attainable yield because of the factors discussed above, their influence is often overshadowed by the most important socio-economic factors that often remain latent. Systematic field monitoring is very much needed to assess the actual productivity that is affected by the socio-economic factors. Such studies were very rare in the past and thus more long-term studies of this nature are essential involving various ecosystems, farm sizes, and both availability and use of credit.

3. PROGRAMMES FOR NARROWING THE YIELD GAP

3.1 Historical Perspective

With the introduction of modern varieties of rice in the country during the early seventies, rice production increased to a great extent but not up to the expectation of the rice scientists. Causes for this low production were not known until 1975-76. Response of rice to fertilization, particularly to nitrogen fertilizers, was studied during the early days of BRRI. Results show that in most cases a very poor agronomic efficiency of nitrogen fertilizer was attained that ranged from as low or slightly over 6 to 28 kg grain per kg of applied nitrogen. BRRI scientists identified for the first time the factor responsible for such low response during the mid-seventies as being the deficiency of sulphur and zinc in many soils of Bangladesh. However, inherent sulphur and zinc deficiency was not widespread in many places, but an induced deficiency was found as a result of stagnation of water. Correction of these deficiencies narrowed down the poor agronomic efficiency further because of increased rice yield in the plots without nitrogen fertilizer.

3.2 Narrowing the Yield Gap

In order to meet the food shortage, there are no other options than bridging the rice yield gap in the country on the one hand, and breaking the yield ceiling of the existing modern varieties on the other hand. The existing modern varieties of rice have come to a stage of stagnation. Attempts are being made to break the present rice yield ceiling by developing and introducing a hybrid rice programme in the country and it will take some time to achieve a practical solution to this problem. Considering the complicated maneuvers and difficulty in launching a hybrid rice programme in a country like Bangladesh where most farmers lack formal education and financial resources, the immediate practical solution to the problem of food shortage would be to narrow the rice yield gap between the attainable and farmer’s rice yields. This gigantic task involves, apart from the present technological know-how, solving several physical and socio-economic constraints. The present activities for narrowing the rice yield gap in the national research institutions include:

- Development of shorter duration rice varieties with high yield potential so as to fit into the farmer’s cropping pattern. This will help farmers to free the land for the next crop in time and thus a balanced cropping scheme will be possible.

- To sustain soil productivity a programme of introducing integrated use of organic and inorganic fertilizers has been gaining momentum.

- Attempts are being made to strengthen the present linkage between research and extension in order to accelerate dissemination of available technologies among the farmers.

In the recent past Bangladesh had to import cereal food grain to meet the food shortage. Since the early 1980’s the volume of such imports tremendously reduced due to the use of high yielding modern varieties of rice, as a result of which the country achieved near self-sufficiency in cereal food grain production. In spite of this, an annual shortage of about 1.5 million tonnes of cereals (Hossain & Shahabuddin, 1999) is reported due to the increase of population at 1.8 percent per annum. With this rate of population growth, the country’s total rice requirement will be 35.5 million tonnes to feed 173 million people by 2020 (BARC, 1994). Since there is no scope for horizontal expansion of the rice area, it is the greatest challenge to the policy makers, administrators and rice scientists of the country to produce more food from the limited land available. In order to face this challenge the following issues have to be considered both on short and long-term basis.

- The present available rice production technologies are enough to mitigate the present 1.5 million tonnes cereal food grain shortage. This necessitates a strong agricultural extension programme for dissemination of all these production technologies among the farmers. For example, if we are to mitigate this food shortage from a single rice crop like T. Aman, then we have to increase rice yield by only 0.3 t/ha (1.5 million tonnes/5 million ha = 0.3 t/ha clean rice or 0.46 t/ha rough rice). This can easily be done by increasing the present level of modern variety (MV) adoption from 51 to 55 percent, providing supplemental irrigation, using balanced fertilizers and adopting better plant production measures.

- Farm level irrigation infrastructures are developed during the Boro season which are usually dismantled after the season. These infrastructures should be maintained so as to provide supplemental irrigation in the following T. Aman rice crop, which covers about 47 percent of the total rice area. Although T. Aman is grown rainfed, often it suffers from drought during flowering stage, thereby lowering yield to a great extent. Therefore, the Government should have a policy to support a permanent irrigation infrastructure instead of a temporary one developed mostly by the farmers.

- A strong extension programme should be launched to popularize “Mini watershed” technology developed by BRRI for providing supplemental irrigation to the T. Aman crop during its flowering stage. These mini watersheds can also be used for shelter for fish in a system of rice-fish culture.

- Policy issues relating to production, import and distribution of fertilizer should be reviewed and positive steps be taken to encourage use of balanced fertilizer by the farmers.

Of the 13.7 million ha of arable land, rice is grown on 10.27 million ha (75 percent) producing 94 percent of total food grain requirement. This is not enough to feed the nation and 1.5 million tonnes annual shortage of food grain exists. To attain full self-sufficiency in food grain production, the rice yield has to be increased without delay by making use of the available technologies. Further work is needed to sustain this productivity by generating new technologies. One way of increasing rice production is to narrow the gap between research stations and actual farm yields. In order to fulfill this challenging job, an integrated effort of researchers, extension workers, administrators and planners is needed to solve issues like: a) government’s continued support for rice research to develop further appropriate improved technologies; b) faster dissemination of available technologies to farmers; c) government’s support for the development of permanent irrigation infrastructures for providing supplemental irrigation, particularly for the T. Aman crop.

REFERENCES

BARC (Bangladesh Agricultural Research Council), 1994. Strategic plan for national agricultural research system to the year 2010. Dhaka, BARC.

BBS (Bangladesh Bureau of Statistics). 1997 & 1993. Statistical Yearbook of Bangladesh, BBS, Dhaka.

BRRI (Bangladesh Rice Research Institute). 1989. About BRRI, 39 p.

BRRI (Bangladesh Rice Research Institute). 1999. Adhunik Dhaner Chash (in Bengali). 54 p.

Bhuiyan, S.I. and A.N.M.R. Karim. 1999. Rice production in Bangladesh: an overview. pages 1-11 in Bhuiyan, S.I. and A.N.M.R. Karim (ed.). Increasing Rice Production in Bangladesh: Challenges and Strategies. Dhaka (Bangladesh): Bangladesh Rice Research Institute and Manila (Philippines): International Rice Research Institute.

Hossain, M. and Q. Shahabuddin. 1999. Sustainable agricultural development in Bangladesh. Challenges and issues. Pages 65-78 in Bhuiyan, S.I. and A. M.R. Karim (ed). Increasing Rice Production in Bangladesh: Challenges and Strategies. Dhaka (Bangladesh): Bangladesh Rice Research Institute and Manila (Philippines): International Rice Research Institute.

Hamid, M.A. 1991. A database on Agriculture and Foodgrains in Bangladesh. Ayesha Akhter. 606 North Shahjahanpur, Dhaka. 425 p.

Karim, Z. 1999. Accelerating the growth of agriculture in Bangladesh. Pages 53-64 in Bhuiyan, S.I. and A.N.M.R. Karim (ed). Increasing Rice Production in Bangladesh: Challenges and Strategies. Dhaka (Bangladesh): Bangladesh Rice Research Institute and Manila (Philippines): International Rice Research Institute.

UNDP and FAO. 1988. Land Resources Appraisal of Bangladesh for Agricultural Development: Land Resources. Vol. 11 (1-13).