Vincent de Leijster

212 compost application resulted in higher production than conventional tillage and thus in a higher NPV. The difference in income between no tillage and green manure on the one hand, and conventional management on the other is referred to as opportunity costs. We evaluated which economic incentives may compensate for these opportunity costs, and in this way facilitate the adoption of vegetation cover practices. We found that public incentives (greening subsidies) and private incentives (price premiums) can be used to compensate for the opportunity costs, but this is only possible when compensation is 5-7 times higher than the amount currently paid. Furthermore, we found that the effectiveness of payments for ecosystem services (erosion control and carbon sequestration) is currently rather modest, as voluntary carbon prices are low and changes in carbon stock in Mediterranean agroecosystems are small and take time. Conventional tillage showed the highest shadow costs of erosion, but these costs did not counterbalance the opportunity costs. In sum, we found that compost application provides a win-win outcome, leading to both higher profitability and better ecological outcomes. Vegetation cover (no tillage and green manure) results in opportunity costs, which may be compensated by greening subsidies or price premiums. More research is needed on longer term economic impacts of agroecological practices and more research to understand whether a combination of compost and vegetation covers may optimize both ecological and economic outcomes while reducing dependency on external incentives. In Chapter 5 , we assessed the long-term developments of ecosystem services in response to agroecological interventions and did so using our second case study in Colombian coffee systems. To this end, my co-authors and I used a space-for-time substitution to reconstruct a chronosequence of the time since the adoption of agroforestry in coffee farms. We assessed the development of ecosystem services over time since adoption of agroforestry. We found that most canopy structure characteristics (such as canopy closure and tree height) increased over time, following an asymptotic shape. Indicators of carbon storage and habitat provisioning had similar trajectories to the development of canopy structure, whereas timber volume had a positive sigmoid trajectory. However, coffee yield did not significantly change over time. Overall, we found that ecosystem service supply changed most in the first ten years and largely stabilized after approximately twenty years. In the first ten years after adoption, coffee yield was negatively related to above-ground carbon, suggesting a trade-off. In the subsequent ten-year period, coffee yield was positively related to erosion control, which suggests a synergy. In addition, we found that above-ground carbon stock, erosion control and epiphyte abundance formed a bundle, which is defined as multiple ecosystem services that respond in a similar way to a driver of change. In the short