Modelling crop portfolios that minimize human macronutrient deficiency on subsistence farms in Burundi

In Burundi, many subsistence farmers employ mixed cropping systems in an attempt to provide food for their families in an environment with high risks, few safety nets and limited storage options. In this paper, we studied the crop portfolios that minimize energy and macronutrient deficiency. We accounted for yield variability, seasonality and storage availability. A mathematical programming approach was applied to four different farm types (based largely on farm size) to predict optimal mixes of crops for farmers to achieve the household year-round supply in calories, fats and proteins. The models predicted that farmers could best cover their household needs in terms of calories, fats and proteins by growing fewer crops in a more optimal combination. Bananas and cassava appear in the crop portfolios as sources of energy; growing beans add proteins to the diets and groundnuts and/or soybeans will supply the needed fats. Crop portfolios differ by farm type and change when yield variability and storage is accounted for. Some of the portfolios predicted by the models have a more diverse crop combination and include maize, rice and/or peas additional to the crops mentioned above. Results also highlighted the benefits of basic storage infrastructure in overcoming seasonal shortages. Nevertheless, even if they would choose to apply the proposed crop portfolios, farms with very small land size in Burundi will continue to struggle to supply sufficient food for the family. Further refinement for micronutrient supply, diet quality, food preferences and market purchases may change these findings.