- The predicted effects of climate change call for methods to assess the performance of agricultural systems across economic, environmental, and social dimensions over different spatial and temporal scales. For research and development to have an impact on system efficiency, potential impacts of climate change and adaptation strategies need to be identified based on an understanding of these individual components of agricultural systems and their interactions in space and time. This dissertation contains four essays that develop and evaluate an Integrated Assessment framework for ex-ante assessment of climate impacts and related agro-technological innovations. The objective of this framework is to underpin assessment of agricultural systems at multiple scales (farm to regional), to provide systematic capabilities for economic, social, and environmental aspects of agricultural systems, and to develop a system that is sufficiently generic, so it can be applied to other places and problems.
The first essay provides a systematic comparison of crop yield predictions from different process-based crop models in Kenya and Ethiopia. Using three statistical indicators to measure model performance, we find that while some crop models perform better than others in estimating grain yields in specific environments, none is clearly more robust in terms of yield prediction accuracy across all regions. To this end, the use of multi-model to estimate crop yields can provide a valuable way of estimating the range of possible outcomes, and of unravelling the importance of different physical and biological processes written into crop models.
The second essay develops a generic bio-economic farm model for ex-ante assessment of technology adoption on agricultural systems. The applicability of this framework is demonstrated in an assessment of potential adoption and impacts of irrigation on Ethiopian smallholders. We combine data from a detailed household survey and a gridded crop model within a household-level impact assessment model to assess the potential adoption and economic impacts of irrigation. To uncover the distinct impacts of irrigation, we present evidence of heterogeneity in adoption rates and marginal treatment effects by agro-ecological zones and farm activity sets. Results indicate that predicted impacts of irrigation are significantly higher in the drought-prone zone compared to the moisture-reliable zone. Many farms and crops in the moisture-reliable zone see little benefits from irrigation alone, which suggests that although irrigation could be beneficial to farmers in this zone, complementary inputs are necessary to close water-stressed yield gaps. Moreover, we find negative selection among adopters, i.e., farms that would have lower returns in the old system are the ones who adopt irrigation, which suggests these technologies may potentially reduce income inequality. Overall, the heterogeneity in returns suggests that encouraging adoption of a technology based on large predicted average returns may be inefficient due to the lower returns for non-adopters. Furthermore, knowledge of the distribution of returns and adoption rates enables policy interventions that are cost effective.
The third essay extends the Integrated Assessment framework to study the potential impacts of climate change (in addition to impacts of adaptation measures) in Kenya. Our results predict average negative impacts of climate change on maize yields and net farm returns on current maize-based systems in Kenya. However, we find significant heterogeneity in these impacts – farms in the high potential maize zone are the most vulnerable to climate change because they are relatively more reliant on maize. In terms of potential adaptation, a large portion of farms in current maize-based systems may benefit from irrigation expansion in Kenya. The impacts of irrigation also show significant heterogeneity across zones; farmers in the low potential zone have lowest impacts on farm income and poverty despite having the highest adoption rates. Intervention aimed at improving livestock in addition to irrigation has similar positive impacts on farm net returns. The provision of multiple improved breed cows increases both milk production and milk productivity. As a result, maize and milk net returns tend to increase for farms across Kenya, leading to increases in per-capita income and decreases in poverty.
The fourth essay develops a two-way linkage between a farm-level model of agricultural systems and a partial equilibrium market model. The linkage is illustrated in a case study of potential adoption and distributional impacts of irrigation on smallholders in Ethiopia. In addition, we also simulate the impacts of changes in crop prices on overall household welfare based on a household’s position as net buyer or seller of crops. Our results show that potential irrigation adaptation could benefit Ethiopian farmers even if value of agricultural production declines from improved supply. Overall, the diﬀerences in predicted adoption rates and the diﬀerences in impacts of increased prices on these adoption rates support the need for combining the adoption behavior of heterogeneous farm populations with behavior modifying eﬀects of aggregate processes like prices.