Abstract
Field experiments were conducted in a moderately saline soil at regional station farm of Bangladesh Rice Research Institute (BRRI), Char Chandia, Sonagazi, Feni of Bangladesh. from March 2002 to October 2004 in order to study the feasibility of farm ponds for irrigation of Rabi crops. Three types of farm pond, such as unlined, soil-cement and polyethylene lined were constructed to conserve rainwater during wet season. Each pond received rainwater and runoff from predefined farm land. Stored water was used for irrigation of chilli in Rabi season in the specified land. For comparison, rainfed chilli was also cultivated in three separate plots. Water salinity of all ponds was less than 0.63 dS m during irrigation which was not harmful to chilli. Kelly's ratio and sodium absorption ratio of pond water were also found to be less than 1.0, which also indicated that stored rainwater was suitable for chilli irrigation. Peak salinity levels of rainfed and irrigated cropped soils were 34-37% and 42-44% lower respectively, than fallow soil salinity level which generally varied from 1.0 to 5.2 dS m'. Dry season cropping caused a lowering in topsoil salinity and plowing fallow land during the season was also found to lower the salinity buildup. The salinity of upper soil layer (0-15 cm) came down at the range of 1.20 to 0.69 dS m1 in wet season (July to October) due to leaching of salts by rainfall. Farmers can cultivate rice in Aman season without any salinity hazard. In Rabi season, three irrigations were applied to chilli at 20 days interval from the date of transplanting during the growing period. Yield of green and red chilli responded significantly to the application of irrigation. Irrigated chilli production was 30% more compared to that of rainfed chilli. In the subsequent studies, a farm pond storage simulation (FPSS) model was developed to simulate the water storage in the farm pond. The FPSS model optimized seepage and percolation losses for unlined and soil-cement lined ponds which were found to be 2.44 and 1.35 mm day' m', respectively. The model also showed that approximately 19% of rainfall becomes runoff from the command area which could be stored for supplemental irrigation of Rabi crops. The estimated crop evapotranspiration was found 74% of that of potential evapotranspiration as determined by pan evaporation method. Annual cost of unlined, soil-cement and polyethylene lined ponds were Tk. 3416.00, 19145.00 and 6960.00, respectively. Therefore, the unlined pond is economically profitable for clilli irrigation in Rabi season. The command area was optimized for a given size of pond using long term rainfall and evaporation data and fitted to the normal distribution and optimum command areas for desired probability of nonfailure were chosen from the distribution. With these data, farm pond length-command area relationship was developed. From these relationships, the lengths of square-sized unlined, soil-cement and polyethylene lined ponds having 1:1 side slope were estimated to be 32.30, 30.25 and 27.25 m, respectively for a nonfailure probability of 80% for 1 hectare farm land. The area occupied by unlined, soil-cement and polyethylene lined ponds were about 12, 10 and 8% of the command area. In the study area, Irrigated chilli-Rainfed T.Aman rice cultivation was found more profitable than that of Rainfed chilli-Rainfed T.Aman rice cultivation. Irrigated red chilli- Rainfed T.Aman rice pattern gave 3.61% higher net return than that of Irrigated green chilli-Rainfed T.Aman rice pattern. Therefore, rainwater harvesting in coastal saline areas using unlined farm ponds for irrigation is profitable for growing chilli in Rabi season.