Drivers and barriers to climate-smart aquaculture adoption: insights from Mwanza and Mara Regions, Tanzania

Climate-Smart Aquaculture (CSAq) optimizes resource use, minimizes environmental impacts, and enhances resilience to climate stresses. However, adoption remains limited in African countries, including Tanzania, due to socio-economic, institutional, and environmental barriers. This study investigated drivers and enabling conditions for CSAq adoption in the Lake Zone of Tanzania. A cross-sectional survey of 384 smallholder aquaculture farmers in Mwanza and Mara regions was analyzed using a multinomial logit regression model. The results show that socio-economic factors significantly influenced adoption, particularly education level (higher education increased integrated farming adoption; χ² = 132.48, p < 0.001), gender (male farmers were more likely to adopt integrated farming; χ² = 5.239, p = 0.003), and household size (larger households favored integrated systems; χ² = 10.838, p = 0.021). Institutional factors, including access to finance (χ² = 10.258, p = 0.013), extension services (χ² = 14.888, p = 0.008), and training (χ² = 19.564, p = 0.003), facilitated adoption. Environmental conditions, such as proximity to water (χ² = 6.351, p = 0.048) and lower climate variability (p = 0.011), also influenced uptake. Integrated farming, which combines aquaculture with agriculture and livestock, emerged as the dominant CSAq practice. Polyculture and monoculture were less commonly adopted. The regression results confirmed that integrated CSAq adoption is significantly driven by income diversity, household size, training, and climate risk perceptions. These findings directly inform the study’s recommendations, which emphasize targeted financial and technical support, including access to credit, extension services, and gender-responsive policies. Strengthening institutional capacity and embedding CSAq into national climate adaptation frameworks will be critical for scaling sustainable aquaculture in Tanzania and similar contexts.