Contextual dependencies for digital knowledge sharing to support farmers and nitrogen use efficiency in South Asia

Abstract

Growing more grain with less nitrogen fertiliser mitigates climate change, improves food security, supports rural livelihoods, and reduces fertiliser subsidy costs in South Asia. Meanwhile, increased access to mobile networks offers novel opportunities for farmers to circulate knowledge of improved farming practices. But how might these two opportunities be beneficially combined? A dataset of more than 31,000 rice fields across South Asia was harmonised and analysed. This analysis found that disincentivising nitrogen fertiliser overuse was the most impactful mechanism for improving rice nitrogen use efficiency in South Asia, followed by addressing yield constraints unrelated to crop nutrition. A mixed methods randomised controlled trial was then implemented in Bihar, a state in eastern India with a particularly large opportunity to increase cereal nitrogen use efficiency. This trial explored the contextual dependencies for YouTube and WhatsApp to support the scalable use of cereal agronomy videos in Bihar and other contexts. It was found that the capacity for these digital extension tools to increase cereal NUE depended on the social inclusivity, goals, and reach of extension actors and networks that shared farming videos. From a theoretical perspective, these findings highlight the need for socio-technical theories, like affordance theory, to accurately conceptualise the use and impacts of digital tools in agricultural extension. From a practical perspective, the findings altogether highlight three contextual dependencies for digital tools to help farmers grow more grain with less nitrogen fertiliser at scale in South Asia: (1) economic incentives for cereal farmers to curb nitrogen overuse, (2) socially inclusive extension networks, and (3) equitable access to smartphones and mobile networks. The research methods only examined how human extension actors shared and discussed digital extension tools. This approach neglected algorithmic extension actors, like YouTube and chatbots powered by large language models. Further research should examine the goals and behaviours of these increasingly influential intermediaries in agricultural innovation systems.