Cash or crop? On-going adoption of rice-fish fingerling production contributes towards food and nutrition security in Northwest Bangladesh

The potential for rice-fish farming as an approach to sustainable intensification of food production that supports food security has been debated within both academic and development communities. Based on producer household surveys and community group interviews, this work aimed to understand the long-term development and drivers of adoption, adaptation and retention of a decentralized rice fish fingerling seed production (RFFSP) in Northwest Bangladesh in order to guide nutrition-sensitive agriculture development policies. The study identified three key stages since its introduction as part of integrated pest management in irrigated rice systems in the 1990s (1) an initial adoption and NGO-assisted diffusion across the region, (2) a parallel process of disadoption for some and retention and consolidation for others, and (3) a period of intensification and expansion, with new adopters in areas that continued RFFSP. This fingerling production system shows evidence of geographic concentration of smallholders and product diversification and specialization liked to markets: polycultures of fingerlings of mixed carp species dominated practices but some farmers also specialized in monosex tilapia fingerlings from hatchery-produced fry; both were linked to a proliferation of hatcheries in the Northwest region over the past 20 years. Only a small minority now produce mixed-sex tilapia from their own broodfish. Farmers have innovated with interlinked pond and rice-field systems, largely driven by a need for more water to overwinter carp fingerlings. Compared with disadopters, rice farmers producing fish fingerlings had better food and nutrition security: they ate fish more frequently, sourced mostly from their farms, and had increased food purchasing power in markets. This system has evolved with farmers' ability to pivot, innovate, and diversify in the local dynamic and rapidly changing fish production value-chain. • RFFSP has commercialised immanently since its introduction in the late 1990's. • No evidence of consolidation was apparent, despite evidence of commercialisation. • Farmers have diversified species and specialized in large fingerling production. • RFFSP farmers eat more fish (mainly sourced from their own farm) than disadopters. • Inadequate land and incompatibility with off-farm activities drove disadoption.