Study area and its ecological setting for rice and shrimp farming The village Damarpota is a medium sized village of 388 ha that had 1324 people living in 19 large holders (43 ha), 21 medium-holders (1–3 ha), 44 smallholders (0.2–1 ha), and 55 near-landless (0.2 ha) households. It is located at 4 km east of Satkhira town and 65 km north of the Sundarban forest in Southwestern Bangladesh. Situated at the tidal floodplain of the Betravati River, a distributary of the Ganges River, the village typifies a saline wet rice and shrimp farming ecosystem. Before its establishment in the early 19th century, the village land formed the northernmost flank of the Sundarban forest and was covered by mangrove vegetation grown on dark gray tidal alluvium sediments deposited by the river. Over time, both sedimentation and human modification gave rise to four land elevation and floodibility levels; each is suitable for specific use and cultivation practices. The highest land, bhita, is flood free and occupied by homesteads and gardens. Medium high land, danga, is flash flooded in high monsoon rain, and is suitable for rice, jute, legumes, oilseeds, and vegetable cultivation. Low and very low fields, hora and beel, are tidally inundated regularly and are suitable for rice and shrimp farming. The ox-bow lakes and man-made ditches serve as water reservoirs and natural habitats for fish and other aquatic species. Both non-calcareous gray floodplain and acid sulfate soils are identified in the village. According to the USDA Soil Taxonomy, the non-calcareous gray floodplain soils are Typic Haplaquept (18%) and Thapto Histic (17%); and the acid sulfate soil (65%) is Thapto Histic Haplaquept type (Brammer, 1996, p. 154; Soil Resource Development Institute (SRDI), 1993; USDA, 1975). The Typic Haplaquept soil is non-saline and non-acidic; and both the Thapto Histic and Thapto Histic Haplaquept soils are extremely saline (EC411 dS/m), rich in sulfur (4225 mg/l) and highly acidic (pHo4, SRDI, 1993, p. 45).1 The river water changes from extremely saline (EC415 dS/m) in dry months of April and May to moderately saline (EC ¼ 8–10 dS/m) in monsoon months of July and August when increased fresh water off-take from the Ganges River reduces the salinity (SRDI, 1993, pp. 53–54; BWDB, 2003, unpublished data). Ecological data In 1985, based upon the proportional distribution of three types of soils in the village, 15 sample plots from each of Typic Haplaquept and Thapto Histic, and 30 plots from Thapto Histic Haplaquept soils (a total of 60 plots) were randomly selected as experimental plots then used as rice fields, and soil samples were collected from each of them. Since the beginning of shrimp farming in 1986, major land use/land cover changes occurred in the village. In 2003, four land use types were identified on three soil types: type I, rice fields in Typic Haplaquept soil those were never used for shrimp farming; type II, rice fields in Thapto Histic soil used as shrimp pond for 5 years (1995–2000) and now under rice–shrimp rotation; type III, rice fields in Thapto Histic Haplaquept soil used as shrimp pond for 10 years (1990–2000) and now under rice–shrimp rotation; and type IV, rice fields in Thapto Histic Haplaquept soil used as shrimp ponds for 15 years (1986–2001) and now under rice–shrimp rotation. From each land use type, those sample plots selected in 1985 were selected again in 2003 to collect soil samples and ecological data to perform comparative analysis. (i) Soil properties: During 1985 and 2003 field works, from each of 60 sample plots, about 250 g of soil sample was collected and sealed in plastic bags to protect soils from moisture and salt loss. Soil samples were collected both in July and March to observe the effect of rainfall on soil salinity and acidity. Each bag was labeled with plot number and land use/land cover type to maintain distinct identification. Soil samples were taken to the SRDI, Dhaka, where they were air dried for 2 weeks, finely ground, sieved through 2 mm sieve to remove large stones and plant roots, and prepared for saturated extracts. Standard laboratory analytical procedures were followed to determine soil pH, salinity, Ca, Mg, K, S, N, P, and OC content. (ii) Rice cultivation data: During 1985 and 2003 fieldwork, field size, labor inputs, and total production of HYV aman and boro rice in each type of land use were directly and carefully observed and recorded from each sample plot. The total rice production was divided by the plot size to obtain per hectare yield. (iii) Shrimp farming data: During the 2003 fieldwork, each shrimp farm was visited twice and the owner was interviewed to obtain data on pond (gehere) size and uses of inputs including feed, chemical fertilizers, and antibiotics for shrimp disease. At the time of harvest, each shrimp pond was also visited to record the production of headless shrimp. (iv) Aquatic and non-aquatic species data: During both 1985 and 2003 field works, data on the number of aquatic, non-aquatic species present in rice fields, and the number of livestock animals owned by the farmers were collected. In this study, changes in the number of aquatic species and livestock population were computed to assess the impacts of shrimp farming on those species.