Item

Abstract

Climate change poses a substantial risk to agricultural production in Peru. Nationally Determined Contributions (NDCs) are currently developed and outline Peru's mitigation actions and adaptation plans to climate change in various sectors. To support the implementation of adaptation measures in the agricultural sector, information on weather-related risks for crop production and the effectiveness of adaptation options on the local scale are needed. We assess weather influences on starchy maize yields on different scales in Peru based on statistical crop models and a machine learning algorithm. The models explain 91% of yield variability (55% based on the cross-validation) on the regional scale. On the local scale, weather-related yield variation can be explained in some areas, but to a lower extent. Based on these models, we assess the effectiveness of adaptation measures which increase water availability to protect against negative impacts from dry weather conditions. The results show that a higher water availability of 77mm in the growing season would have regionally different effects, ranging from an increase of 20% to a decrease of 17% in maize yields. This large range underlines the importance of a local assessment of adaptation options. With this example, we illustrate how a statistical approach can support a risk-informed selection of adaptation measures at the local scale as suggested in Peru's NDC implementation plan.