This thesis establishes the need for a novel enterprise-level crop yield model under climate change. A model for the approximation of enterprise-level crop yields under climate change is proposed, which integrates both economic and agronomic notions of crop production. A general computational strategy for the proposed model is provided. The strategy introduces a novel geometrical interpretation of the coefficient of risk-aversion, which is intrinsic to the farmer, as well as an approach to infer this coefficient of risk-aversion from farm-level accounting data. A new indicator to detect the nature states is also proposed as a methodological contribution based on the previous work by the author. On the empirical side, the thesis offers an estimation of a state-contingent production function for the crop-producing sector in the Federal State of Saxony-Anhalt, Germany. Based on simulated data the estimation rejects the hypothesis of an output cubical technology.