Agrobiodiversity and climate adaptation: insights for risk management in small-scale farming

Agriculture is a dynamic activity that sustains food and other goods for global human population. Aiming to ensure global food security, the sector has evolved dramatically, especially over the last century with the introduction of high-yielding crops, improved technology and pathogen resistant varieties, to name a few. According to the Food and Agriculture Organisation of the United Nations (FAO), the food security issue is still present. In 2019, around 26% of the world population experienced either a moderate or severe level of food insecurity. Climate change makes the challenge of food security even more pressing. It is argued that increased agrobiodiversity through farm diversification and varietal selection can help farmers to cope with the negative effects of climate change while ensuring food security. However, such approaches have been difficult to scale up. One could argue that we often lack information to understand the contexts that drive farmers’ adaptation decisions and how to develop recommendations for adaptation. In this thesis, I developed methods and tools to support farmers and stakeholders in adapting to a changing climate. I present results from three continents to improve the understanding of the food systems at the farm level, and specifically in smallholder farming. I provide insights for the different biological levels: species level, focusing on trees as slow grower organisms for interspecific diversification; varieties level, looking for locally adapted phenotypes; and genotype level, focusing on G × E interactions to support crop breeding for intraspecific diversification. From the first part of the study, conducted in Central America, the results showed that farmers have a clear preference to a set of adaptation strategies, with reforestation (agroforestry) as the first choice (Paper 1). Crop variety management is the least preferred choice of the top-5. In the second part of the study, I assessed the impacts of climate change on the habitats of the 100 most common tree species used in coffee (Coffea arabica L.) and cocoa (Theobroma cacao L.) agroforestry in Central America (Paper 2). The results showed that the most preferred trees are the most vulnerable, but farmers could re-think the agroforestry designs using a portfolio of underutilised species already present in low densities at the current systems. In the third part of the study, I employed a citizen science approach that can scale variety testing and help farmers to select the right crop variety for their farms (Paper 3). I tested this approach with common bean (Phaseolus vulgaris L.) in Nicaragua, bread wheat (Triticum aestivum L.) in India and durum wheat (Triticum durum Desf.) in Ethiopia. The results showed that the approach reduces geographic sampling bias and could be scaled to provide tailored recommendations for crop variety management. I also show, with durum wheat genotypes in Ethiopia, that linking the farmer-generated data to scientist-generated data can support breeding programs targeting challenging crop production environments using a data-driven decentralised approach (Paper 4). The approach is fully replicable, and part of its workflow is presented in this thesis (Paper 5 and Paper 6). Overall, the results of this thesis should be seen as starting point to develop lines of research that support recommendations to adapt agricultural systems to a changing climate