Drainage & Irrigation in South Bengal, India - Full paper & Recommendations

Citation: Sen, H.S. and Ambast, S.K. (2009). Improving drainage and irrigation through OFR in rainfed rice lowlands of Sundarbans delta. Paper presented in Seminar on "Drainage System in South Bengal: Its Problems and Management", held at Geology Department, Calcutta University, 10-11 October, 2009.

Improving Drainage and Irrigation through OFR in Rainfed Rice Lowlands of
Sundarbans Delta

H.S. Sen1 and S.K. Ambast2
1Ex-Director, Central Research Institute for Jute and Allied Fibres (ICAR), Barrackpore, WB, PIN 700 120, and 2Central Agricultural Research Institute, Port Blair, A&N Islands, PIN
(Email: hssen.india@gmail.com,hssen2000@hotmail.com; Tele: 9874189762)

Abstract
In regions depending entirely on rainfed agricultural operations, rainwater harvesting and its recycling is the most effective means of increasing food production. About a third of the world's rice cultivated land is under rainfed conditions, which is both drought and submergence prone. The Sundarbans, one of the largest deltas in the world, shared by India and Bangladesh is one such region. The region is criss crossed by innumerable tidal rivers and creeks forming 54 islands. The region is predominantly mono cropped with more than 80% area under kharif (June-November) rice cultivation and about 62% of the area is lowlying. The region is generally flat with some undulations having an average ground level 1.2 m above mean sea level, whereas the average high tide level is 2.75 m above mean sea level. The area is surrounded by protective earthen embankments provided with one way sluice gates for drainage of excess inland water and to prevent ingress of tidal water. The field level drainage network is inadequate in the region. In the years of high rainfall, it gets severely waterlogged or occasionally flooded during monsoon period due to impeded natural drainage. On the other hand in some years, due to saline ground water and lack of assured means of water supply, kharif crop suffers from prolonged drought spells in the later period of crop growth. This compels the farmer's to resort to more certain local low yielding (1-1.4 tonne/ha) tall varieties. The practice of rainwater storage in an on-farm reservoir (OFR) is quite common but its design is not based on scientific principles. Also, there is no information available on the extent of surface drainage improvement due to rainwater storage in an OFR. The use of stored water for cultivation of second crop during rabi (December-April) is not judicious. This often results in crop failure or low net returns during rainfall deficit years.

Simulation studies using soil water balance approach for the purpose of land and water resource management have been attempted by several researchers. A soil water balance model for rainfed rice lowland is used for estimating excess rainwater to design an on farm reservoir (OFR) and to assess surface drainage improvement due to storage in OFR. Weekly rainfall at 2 and 5 years return periods are used to optimize the size of OFR and to simulate surface drainage improvement, respectively. It is recommended to convert 20% of the farm area into OFR to harvest excess rainwater. It is also estimated that rainwater storage in OFR (in 20% of the farm area) reduces surface waterlogging to the extent of 75% and thus, the cultivation of high yielding dwarf rice varieties in lowlying areas can be made. Further, a simple linear programming model is used to propose optimal land allocation for rabi (winter) crop cultivation to increase the agricultural profit under various limitations of land and water.

Key words: Rainfed rice lowland; On farm reservoir; Surface drainage; Optimal land allocation





1. Introduction
In regions depending entirely on rainfed agricultural operations, rainwater harvesting and its recycling is the most effective means of increasing food production. About a third of the world's rice cultivated land is under rainfed conditions. In Asia, about 8% of the rice area is under rainfed humid condition which is both drought and submergence prone (Mackill, 1986). The Sundarbans, one of the largest deltas in the world, shared by India and Bangladesh is one such region. In India, it is located between 21o32' to 22o40' north latitude and 88o05' to 89o00' east longitude in the state of West Bengal. The region is criss crossed by innumerable tidal rivers and creeks forming 54 islands (Fig. 1).
The region is predominantly mono cropped with more than 80% area under kharif (June-November) rice cultivation and about 62% of the area is lowlying (Anonymous, 1988). The region is generally flat with some
undulations having an average ground level 1.2 m above mean sea level, whereas the average high tide level is 2.75 m above mean sea level. The area is surrounded by protective earthen embankments provided with one way sluice gates for drainage of excess inland water and to prevent ingress of tidal water. The field level drainage network is inadequate in the region. In the years of high rainfall, it gets severely waterlogged or occasionally flooded during monsoon period due to impeded natural drainage. On the other hand in some years, due to saline ground water and lack of assured means of water supply, kharif crop suffers from prolonged drought spells in the later period of crop growth. This compels the farmer's to resort to more certain local low yielding (1-1.4 tonne/ha) tall varieties. The practice of rainwater storage in an on-farm reservoir (OFR) is quite common but its design is not based on scientific principles. Also, there is no information available on the extent of surface drainage improvement due to rainwater storage in an OFR. The use of stored water for cultivation of second crop during rabi (December-April) is not judicious. This often results in crop failure or low net returns during rainfall deficit years.
Simulation studies using soil water balance approach for the purpose of land and water resource management have been attempted by several researchers (Harikrisna, 1981; Kemachandra and Murty, 1992). Jensen et al., (1993) reported the supplemental irrigation requirement using soil water balance technique and implications for small-scale water harvesting for rice lands in Bangladesh. In this study, a soil water balance model to optimize the size of an OFR on the basis of excess rainwater availability in rainfed rice lowlands is developed. Simulation study is made to evaluate reduction in weekly surface waterlogging by providing the optimum size of an OFR in a unit farm area. The need of supplemental irrigation during kharif season is also assessed. In order to improve the irrigation intensity through judicious use of stored water, an optimal land allocation for growing second crop under various constraints is proposed using linear programming model.
2. Materials and Methods
2.1 Analysis of rainfall and evaporation data
Historical daily rainfall (1963 92) and evaporation (1967-92) data for Canning, which is a representative block for Sundarbans delta, are collected and analyzed. The average monthly rainfall and evaporation are shown in Fig.2. The region receives an average annual rainfall of 1768 mm
ranging from 1030 mm to 2462 mm with a coefficient of variation of 21.2% (Fig.3). Of that, 82% rain occurs during monsoon season (June-October). Weekly rainfall and evaporation data are analyzed for different probability distributions, i.e., Normal, log Normal, Extreme value I and Log Pearson type III (Chow, 1964). The Extreme value I distribution is found suitable for the observed rainfall data, whereas Log Pearson type III is found closely fitting to the evaporation data, and therefore, used as input

parameter for further estimation. The weekly rainfall at 2 (50% probability) and 5 years (20% probability) return periods are given in Fig.4. It is estimated that on an average about five week

drought (no water stands on surface) for rice crop may occur during the season and about three week continuous

drought may be expected during the ripening stage compared to none during both vegetative and reproductive stages (Ambast et al., 1998).
2.2 Model formulation for optimizing OFR
A soil water balance model based on physical parameters is used to estimate the excess rainwater availability in lowland rice paddies. The schematic diagram for a typical rainfed rice lowland system is shown in Fig.5.
The soil water balance equation to estimate the depth of standing water in rainfed rice lowlands is expressed as follows:



SDWi = SDWi 1 + Ri – ETi DPLi + SIi (1)
where, i is the time index, week; SDW the depth of standing water at surface, mm; R the rainfall, mm; ET the crop evapotranspiration, mm; DPL the deep percolation loss, mm; SI the depth of supplemental irrigation, mm.
The following conditions are used to estimate the amount of excess rainwater in rainfed rice lowlands:
At the beginning of the computation, SDWi 1 is set to zero.
When SDWmin < SDWi < SDWmax, SDWi becomes actual water depth for the period.
When SDWi > SDWmax, the amount SDWi SDWmax is diverted to the OFR and SDWi is set to SDWmax.
When SDWi < SDWmin, the SIi (50 mm), if available in OFR is provided.
- When VOL > VOFR, the SDWi becomes actual water depth up to the height of field dyke (700 mm) and any excess rainwater beyond SDWi spills to drainage system.
SDWmin is the minimum depth of standing water for optimal growth of rice, mm; SDWmax the maximum depth of standing water for rice crop, mm; VOL the volume of rainwater storage in OFR, m3 and VOFR the volume of OFR, m3.
In Sundarbans region, the onset of monsoon occurs at 24th standard meteorological week (SMW) (Ambast and Sen, 1994a). It is assumed that the soil cracks may be filled during pre-monsoon rains. The weekly rainfall values at 2 and 5 years return periods are taken as R from 24th week onward to design OFR and to assess surface drainage improvement, respectively. As water availability is fairly well during kharif season weekly ET values are estimated using the pan evaporation method as proposed by Doorenbos and Pruitt (1977).
ETi = Kci Kp Evpi (2)



Where Kc is the dimensionless crop coefficient value for rice (Table 1); Kp the dimensionless pan coefficient value (0.8 valid for humid climate with moderate wind velocity) and Evp the open pan evaporation, mm.

The DPL beyond the rootzone occurs when soil moisture exceeds field capacity. The equivalent depth of soil moisture at field capacity is estimated by taking rootzone depth for rice and percentage moisture content at field capacity for silty clay loam soil (Michael, 1978). A maximum of 14 mm weekly DPL (subject to moisture availability) is taken on the basis of daily DPL (2 mm/day) measured in the farmer's field under puddled rice condition by ponded basin method (Rao and Dhruvanarayana, 1979). A simple book keeping approach is used to estimate the soil moisture storage in the rootzone. As rice requires an optimal submergence of 50±20 mm for potential production (Biswas et al., 1982), the SDWmax and SDWmin are taken as 100 mm and 0 mm (moisture at saturation), respectively. SI is provided only when water equivalent to a SI available in the OFR.
The low cost of construction has made OFR increasingly popular in the Sundarbans region. The design parameters are shape, size, depth, side slope, capacity and location. The prevailing dugout trapezoidal OFR with and without bund are considered for optimization. The length-width ratio of OFR is taken 1:1 as it is having minimum perimeter and therefore, attains maximum storage. As ground water is at shallow depth and saline in nature in this region, the depth of OFR is restricted to 3 m on the basis of observed water quality in the OFR at the end of rabi season. The standard side slope of 1:1 is taken for silty clay loam soil. The following equations are used to optimize the size of an OFR:
RA = WA (S / 100) (3)
W = (A / Y)½ (4)
L = W Y (5)
Ls = L - 2 ((2 Zb Hb) + TWb + Wb) (6)
Ws = W - 2 ((2 Zb Hb) + TWb + Wb) (7) ALs = [Ls + {Ls - (2 Zp D)}] / 2 (8)
AWs = [Ws + {Ws - (2 Zp D)}] / 2 (9)
VOFR = ALs AWs D (10)
Where RA is the surface area of OFR, m²; WA the farm area, m²; S the size of OFR, % WA; W the width of OFR at surface, m; Y the ratio of length to width, m/m; L the length of OFR, m; Ls the length of submergence at surface of the OFR, m; Zb the side slope of the bund, m/m; Hb the bund height, m; TWb the top width of bund, m; Wb the width of berm, m; Ws the width of submergence at surface of the OFR, m; ALs the average (at surface and bottom of OFR) length of submergence, m; AWs average width of submergence, m; Zp the side slope of OFR, m/m and VOFR the volume of OFR, m3.
The rainwater storage in OFR is computed using the equations expressed as:
Rexci = SDWi SDWmax (11)
n
VOLi =  {(Rexci / 1000) (WA - RA) + ((Ri / 1000) RA)} (12) i=1
Where Rexc is the excess rainwater diverted from field to OFR, mm.
The evaporation and seepage losses from OFR are not considered during kharif season because shallow watertable (0.5-1.5 m below surface) contributes water to OFR whereas, losses takes place on account of evaporation. However, these losses are considered to compute available water storage in OFR for irrigation during rabi season due to deeper watertable (2-3 m).
A flow chart for development of computer program is given in Fig.6. The model is verified by comparing simulated and observed depth of standing water in rice lowlands without OFR. The weekly rainfall and evaporation data for kharif 1994 are used for generation of simulated depth of standing water. Two fields; one at the tail end (lowland) and the other at the head end (upland) in the catchment


















of sluice no 2 of East Mograhat Drainage Outfall Division in Sundarbans delta, are selected. The maximum depth of standing water in lowland and upland are 600 mm and 100 mm, respectively
beyond which excess water spilled out to the drain. The depths of standing water are observed at different growth stages during kharif 1994. The regression analyses indicated a good agreement (r² =
0.97 for lowland and 0.95 for upland) between simulated and observed values (Fig.7) and therefore, used to simulate excess rainwater availability in rice paddies.
For computation purpose the initial size of OFR is assigned the maximum area (40% of farm area) which may be converted into OFR, whereas the location of OFR is varied to allow runoff from 10-100% of farm area. At the end of computation if storage in OFR is found less than the capacity, a decrement to the size of OFR is provided for further computation. When OFR gets filled the size is considered as an optimal and the process is terminated.

2.3 Model formulation for optimal land allocation during rabi
Rainwater harvesting through OFR can make available only limited quantity of water. To keep the economics of OFR in favorable range, it is imperative that water is used judiciously for crops having high water productivity in terms of economic returns. Linear programming formulations are normally used to arrive at the optimal land and water allocation amongst different crop activities (Palanisami, 1992). A simple linear programming model is used in the present case.


2.31 The objective function
The objective function maximizes returns from the irrigated area under different crops through OFR and cost of water stored in OFR (includes cost of OFR and a penalty cost for making OFR area out of production during kharif season). It is assumed that inputs other than land, water and labour are at fixed level.
n
Max NP =  (( Ci Yi ) Pi ) Ai – Cs S (13)
i=1
Where i is the index for crop; NP the net profit, Rs/ha; A the area under crop, ha; C the unit sale cost of crop, Rs/tonne; P the unit cost of crop production, Rs/ha; Y the yield per unit area, tonne/ha; Cs the unit cost of water in OFR, Rs/m3/year and S the total stored volume in OFR, m3 .

2.32 Irrigation requirement constraint: The sum of water demand of crops should not exceed available water in OFR (Initial storage Seepage Evaporation) during the period of growth.
n
 GIRi Ai (S WL) < 0 (14)
i=1
Where GIR is the gross irrigation requirement for crop, m3/ha; WL the water losses from OFR observed at the rate of 5 mm/day for 135 days, m3.

2.33 Land-use constraint: The percentage acreage under different crops is limited to 80% of the total available area as20 % area is used for OFR.
n
 Ai – L < 0 (15)
i=1
Where L is the total available land area, ha

The acreage under chilli (A3), cotton (A4) and cucumber (A5) are restricted to a given value due to the risk of crop damage by heavy rainfall at the time of harvest. Rice is the most demanding and safe crop against the weather uncertainty, and therefore, a minimum acreage under rice (A1) crop is placed. The crop activities and the limits placed on crop area are given in Table 2.
A (3) < Amax (3) (16)
A (4) < Amax (4) (17)
A (5) < Amax (5) (18)
A (1) > Amin (1) (19)
Where A denotes the area under crop, Amax and Amin denote the maximum and minimum area under the crop, respectively.
2.34 Labour constraint: Most of the non rice crops are labour intensive because irrigation is often applied through pitcher, and the inter-culture operations also require considerable labour. However, labour availability in the area is low. Therefore, a constraint is placed on labour requirements.
n
 Ai Lri – La < 0 (20)
i=1
Where Lr is the labour requirement for unit area for crops, man-days/ha; La the total labour availability, man-days.

Two OFR (with and without bund) at Nikarighata village in South 24 Parganas district of Sundarbans delta are selected for model application. The hydraulic and storage features of OFR are given in Table 3. The storage in OFR with and without bund is estimated 1100 and 5225 m3, respectively. Taking the surface area of OFR, crop duration and observed daily water losses into account, seepage and evaporation losses are estimated 270 and 1355 m3 for OFR with and without


bund, respectively. The water requirements of different crops are estimated using pan evaporation method (Doorenbos and Pruitt, 1977) and the prevailing water application methods. On the basis of 2 years (93 94 and 94 95) experiment in the farmer's field, the input on labour, cost of cultivation (includes pumping cost) and net profit per unit area for different crops are generated (Ambast and Sen, 1994b). The inputs on water use, labour requirement and net profit are given in Table 4.
3. Results and
Discussion
3.1 Design of an on-farm reservoir (OFR)

A nomograph is developed to determine the hydrologic and hydraulic features of OFR in unit farm area for various combinations (Fig.8). It represents three type of curves i.e. the design dimension curve,
OFR capacity curves and rainwater availability curves. Since OFR is considered square in shape, the design dimension curve represents the equal length and width of the selected size of OFR. The OFR capacity curves give the information of the total volume that can be stored in the OFR of different sizes. The capacity curves are plotted for OFR with and without
bund for 2 and 3 m depths. The rainwater availability curves for different size of OFR i.e. 10, 20, 30 and 40% of farm area are generated for construction of OFR at different location i.e. 10, 50 and 100% below farm area which contributes runoff from 10, 50 and 100% farm area, respectively.

The basic steps to use the nomograph are i) note the volume of a particular size of OFR by projecting vertical line to OFR capacity curves and extending it towards volume, ii) note the location of OFR to meet the rainwater availability curve for that particular size of OFR and for known deep percolation rate and iii) note the dimension of OFR by projecting vertical line to design dimension curve. For example, the dotted line in Fig.8 represents an optimal 3 m deep OFR with bund in 20% farm area in soils with percolation loss rate 2 mm/day to be placed at 75% below the farm area. Similarly the solid line represents an optimal size of 3 m deep OFR without bund in 17% farm area for such soils and placed at 83% below the farm area to meet rainwater to its designed capacity of 4268 m3. The length and width of OFR in 20% area of a hectare should be 44.7 x 44.7 m. Therefore, it is suggested to convert 20 or 17% of the farm area into OFR with or without bund, respectively. However, for recommendation purposes, it is suggested to convert 20% of the farm area into OFR. The design features of OFR for a unit farm area in Sundarbans delta is shown in Fig.9.
3.2 Surface drainage improvement due to OFR
In lowlying areas of the region, the scope for cultivation of high yielding dwarf rice varieties (HYV) is almost negligible due to surface waterlogging during kharif season. The simulation study is made to assess the reduction in weekly surface waterlogging by diverting excess rainwater into the OFR (with and without bund) in 20% of farm area over lowlying area without OFR.

The surface water depth hydrographs are generated with probable weekly rainfall at 5 years return period (Fig.10). It is estimated that in a lowland catchment with no drainage, the water depth may reach as high as 0.63 m and therefore the cultivation of HYV rice crop is not possible under existing condition. The water depth in case of OFR with bund, remains at required water depth for rice up to 33rd SMW till OFR gets filled. However, the water depth increases to 0.35 m at panicle initiation and flowering stages, which are critical to excess water and affect crop production adversely. It is estimated that surface waterlogging during the season may reduce to about 45% in case of OFR with bund. In case of OFR without bund, the depth of water remains at required for a major growth period up to 37th
SMW. The maximum depth of water may reach to 0.18 m at the beginning of grain formation stage but for a short period. Though, some reduction in crop yield may occur due to excess water at this stage but that may not be significant. In this case the reduction in surface waterlogging
is to the extent of 75%. Therefore, construction of OFR not only provides water for irrigation but also improves surface drainage and thus, makes it possible to grow HYV of rice in lowlands of Sundarbans delta.
3.3 Supplemental irrigation demand
In order to assess the requirement of supplemental irrigation (SI) during kharif rice and to formulate water availability constraint for optimal land allocation in rabi season, simulation study is made. The preconditions for supplemental irrigation (50 mm) are defined, as 1) storage in the OFR should be more than one SI for cultivated land and 2) soil moisture in subsurface should reach below saturation level. On the basis of weekly values at 80% annual rainfall (no rainfall after 37th week), it is estimated that at least one supplemental irrigation is needed at the time of grain formation stage in two out of ten years (Fig.11). This will stabilize the crop production against the weather uncertainty.

3.4 Optimal land allocation
In the existing cropping pattern, rice crop is cultivated through OFR that causes shortage of water or crop failure in rainfall deficit years. The optimal area allocation to different crops is shown in Table 5.

As per the allocation during rabi season in the years of normal rainfall, rice should be grown only in 20% of area and remaining area should be allocated to non rice crops in farms having OFR with bund. Under non rice crops 50% area is allocated to chilli crop, whereas 15% area allocated to each cucumber and watermelon. In case of OFR without bund, 40 and 60% area to rice and non-rice (chilli-50% and cucumber-10%) crops, respectively show an optimal allocation to maximize net profit. The increase in cost of rainwater storage in OFR does not change water allocation, and therefore, indicated the stability of the suggested crop allocation.
3.41 Effect of deficit rainfall on crop allocation: In rainfall deficit years it is stipulated that a SI of 50 mm is to be provided during kharif season and thus, water availability in OFR would reduce by same amount during rabi season. In case of OFR with bund, the optimal allocation indicates that the percentage rice area remains the same because of minimum area constraint but the irrigation intensity (percent area irrigated) reduced by 20%. In case of OFR without bund, the percentage acreage under rice crop reduces by 10% but the irrigation intensity remains same as 10% land now allocated to cucumber.

3.42 Effect of labour inadequacy on crop allocation: Labour is the major constraint in the region particularly when non rice crops are irrigated manually with water applied through pitcher. Therefore, the effect of labour availability on crop allocation is studied. Sufficiency of labour is estimated on the basis of average demand for chilli and wheat crop as the most and least labour intensive crop, respectively. In case of OFR with bund when labour sufficiency is reduced by 20%, the area allocated to rice remains the same due to minimum area constraint but the area allocated to chilli and cucumber crop reduced by 15 and 5%, respectively. However, the irrigation intensity remains the same as the area of watermelon is increased by 20%. In case of OFR without bund, the optimal land allocation to rice crop remains the same but the area under chilli crop reduces by 10%. However, now 10% area is allocated to wheat crop, as it requires minimum labour.
3.43 Effect of interaction of water and labour inadequacy: The effect of interaction between reduction in available water during rabi due to its use during kharif as SI and labour (by 20%) is studied. It is noticed that the rice area remains the same due to lower limit but the area under chilli and cucumber reduced equally by 10% in case of OFR with bund. Therefore, irrigation intensity reduced by 20%. In case of OFR without bund, the reduction in rice area is 5% and chilli and cucumber both are reduced equally by 10%. Since 20% area is allocated to watermelon, a reduction of 5% in irrigation intensity is observed.
3.5 Economic evaluation
To evaluate the economic feasibility of OFR, technical efficiency in terms of cost of water development (includes cost of OFR and a penalty cost for making OFR area out of production during kharif season) and productivity per unit water storage are estimated. The life of OFR is taken as 25 years. The annual cost of water harvesting and supply are estimated Rs 582/103 m3 and Rs 557/10 3 m3 for OFR with and without bund, respectively. The total agricultural profit per hectare area are estimated at Rs 1295/103 m 3 and Rs 1020/103 m3. Thus, the benefit cost ratio is around 2 in both the cases that justify investment in OFR.

4. Conclusions
The following conclusions can be drawn from this study:
1. It is suggested to convert 20% of the farm area into OFR to harvest excess rainwater during kharif season in Sundarbans delta.
2. Storage of excess rainwater in OFR may lead to reduction in surface waterlogging to the extent of 75% during kharif season. This provides scope for introduction of high yielding rice varieties in rainfed lowlands.
3. A supplemental irrigation at the time of grain formation in two out of ten years may be required during kharif season.
4. The judicious use of stored water in OFR through optimal land allocation indicated a cost benefit ratio around 2, which justifies investment in OFR.
5. The procedure suggested for optimal design of OFR may be used in similar agro-ecological conditions. It may also be used to assess the performance of a controlled surface drainage project through simulation study to draw policies for surface drainage system regulation in humid rice lowlands.

------------------------------------------------------------------------------------


“Improving Drainage and Irrigation through OFR in Rainfed Rice Lowlands of Sundarbans Delta” by H.S.Sen and S.K.Ambast

Recommendations

Background information

• The Sundarbans area comprises of large number of islands criss-crossed with rivers and tidal cannals/creeks. More than 50 islands of the Sundarbans is inhabited, is flat and low-lying with heavy textured soil. Average annual rainfall is 1768 mm, 82 % of which is received during 5-month long rainy season, and the rest during rest of the year. Average tidal level is 2.75 m above mean sea level (maximum recorded up to about 5 m above msl).
• The area is generally monocropped with tall indica low-yielding rainfed rice, while only about 20 % of the area under double cropping, mainly due to scarcity of good quality irrigation water needed for irrigation and presence of moderate to high soil salinity during dry period. The yield of rainfed rice during kharif is poor mainly due to high water congestion during majority of the crop growth period, while there is as much probability for drought (soil moisture below field capacity) during ripening stage of crop growth. High water depth on the soil surface is itself detrimental for good crop growth (tillering etc.), besides it does not permit scientific crop management including fertilizer, insecticide application, etc. Drought, on other hand, during flowering or ripening is likely to cause non-fulfillment of grains.

Existing practices on flood protection & drainage and suggested improvements

• Owing to siltation in the river beds water flows above the cultivable land during high tide in most of the areas, but the latter is protected from inundation by earthen embankments, and for draining of excess water from cultivable land especially during low tides manually operated one-way sluice gates are installed at strategic points. Neither the embankments nor the sluices are scientifically designed to cater to the needs, and areas experience frequent flooding due to breaching of the embankments especially during severe cyclones or depressions.
• Since majority of the embankments are in the process of reconstruction at present after the disastrous damage caused by Aila a couple of months back, they may be reconstructed with the following designs based on the detailed analyses at Central Soil Salinity Research Institute (ICAR) located at Canning Town: (i) earthen embankments with 3:1 slope on the river end and 2:1 slope at the country end with at least 1 m free board above the high tide level, (ii) the embankments may be brick-pitched at the river end to provide more stability, (iii) provision of suitable wind breaks (plants identified) at the river end, and if possible, at the country end also, (iv) provision of two rows of embankments with minimum 100–500 m no-activity gap in between in areas likely to be breached more frequently than others, (v) the sluice gates should be much better designed as per recommendations made and operated much more frequently during low tides than what is normally done at present for effective drainage, (vi) primary and secondary drains (earthen) should be provided in the cultivable field as per design suggested to cater to the needs for field drainage (drainage coefficient 37.5 mm per day).
• It has been suggested that in order to prevent frequent breaching of the embankments, and keeping particularly in mind the increasing frequency of storms observed world-wide in the wake of global warming, the design of the embankments should be made with minimum 500 year return period based on probability analyses of the weather data for minimum 35 years.

Micro-watershed or OFR approach for integrated water management

• As probably a better option for on-field water management along with flood protection measures with the objective to harvest excess rainwater and utilization of the same during long dry period, together called as integrated water management approach, it has been suggested, based on 35 year weather data analyses and detailed test conducted at the farm as well as in the farmers fields, by CSSRI at Canning Town to (i) create on-farm reservoir (pond, etc.) on over 20 % area within the total farm area, either on individual farm area or preferably on community basis to store about 400 mm estimated to be in excess of the optimal requirement for high yielding rainfed rice crop during kharif, (ii) and utilize the same for a second and partially a third crop during dry season, and (iii) grow fishes in the OFR or pond and suitable plantation/ horticulture crops on the bund around the pond for additional benefit. A number of derelict channels, not in effective use at present in Sundarbans, may be properly utilized after reshaping for community OFR. For lowlying cultivated areas the soils so excavated for creating OFR may be used to raise the remaining 80 % farm area by a minimum 15 cm in height which will thus be conducive to high yielding rice and other crops. The OFR, with or without bund, may preferably be of trapezoidal in shape, with length-width ratio as 1:1, side slope as 1:1, and depth as 3 m. For other details of the design and location of the OFR nomograph may be consulted. It has been recorded, for which computer simulation model has also been prepared and tested with sufficient success, that 45 and 75 % of the water depth above soil surface during kharif season, can be reduced for OFR with and without bund, respectively, thereby creating a much better atmosphere for rainfed rice. A complete crop calendar with details of every important event for cultivation for different topo-sequences has been prepared with scientific cultivation programme for each. Linear programming approach has been followed to identify crops and related cultivation practices depending upon nature of constraints, viz. water, labour, etc. A user-friendly software has been prepared for detailed design recommendations and cropping practices depending upon location-specific conditions related to soil, weather, availability of water, choice of crops, etc. for application to a wide variety of situations using this approach for multiple cropping under rainfed conditions.