Design and assessment of legume-supported cropping systems

Legumes provide high quality protein for food and feed as well as other ecosystem services, but it is still challenging to use them to meet the growing global demand for protein, partly because European farmers consider their cultivation unprofitable and risky. This thesis aims to design legume-supported cropping systems and assess their environmental and economic impacts along with their production risks in European agriculture. The approaches used included (i) the development of a framework to design cropping systems and to assess impacts of management, (ii) modelling the impact of integrating legumes into cropping systems and assess trade-offs, (iii) the development of a statistical method to quantify crop yield stability independent of the mean yield, (iv) assessing grain legume yield stability statistically compared to other crops using data from long-term experiments, and (v) participatory methods to re-design legume-supported cropping systems. The framework consists of a rule-based rotation generator and algorithms to calculate impact indicators, following a three-step approach: (i) generate rotations, (ii) evaluate crop production, and (iii) assess cropping systems. It was used to design and assess legume-supported cropping systems in five case study regions in Europe and to identify trade-offs between economic and environmental impacts. On average, the generated cropping systems with legumes reduced N2O emissions by 18 % and 33 % and N fertilizer use by 24 % and 38 % in arable and forage systems, respectively, compared to systems without legumes. Grain legumes increased gross margins in two of five regions and forage legumes in all three study regions. A scale-adjusted coefficient of variation was developed as a stability measure that accounts for mean yield differences. Using data from five long-term experiments in northern Europe, this method showed that yield instability of grain legumes (30 %) was higher (P < 0.001) than that of autumn-sown cereals (19 %), but lower (P < 0.001) than that of other spring-sown broad-leaved crops (35 %), and only slightly greater (P = 0.042) than spring-sown cereals (27 %). The combination of on-station and on-farm trials with crop rotation modelling was useful when re-designing cropping systems. Nine agronomic practices were identified for improving grain legume production at the farm level. In this thesis, it is shown that legumes can provide both economic and environmental benefits, the instability of yields is similar to other spring crops and that cropping systems can be re-designed effectively in a co-learning process with farmers