Future climatic impacts on soybean yields in South America and its consequences on agricultural markets

Moreno Dumont Goulart, H. (2021). Future climatic impacts on soybean yields in South America and its consequences on agricultural markets. IIASA YSSP Report. Laxenburg, Austria: IIASA

Preview

Text YSSP_Report_HenriqueMorenoDumont.pdf - Published Version Available under License Creative Commons Attribution Non-commercial. Download (3MB) | Preview

Abstract

Whilst used in all continents, the majority of soybean production takes place in specific regions. This causes the supply chain to be highly vulnerable to local climate shocks. With a changing climate, the impacts on crop productivity are expected to change as well. Crop models are commonly used to address the relation between climate and agriculture, simulating biophysical processes and estimating crop yields at a gridded location. However, they underestimate the impacts of extreme climatic conditions. In this study, a process-based crop model EPIC-IIASA is combined with a machine learning model to analyse the impact of future extreme weather events on soybean yields under different climate scenarios. The crop model outputs, such as simulated soybean yield, together with extreme climatic indices during the soybean growing season are used as input into a machine learning model that is trained on observed soybean yields. The coupling of crop model and machine learning model leads to a hybrid model that improves the crop model’s ability to represent extreme weather conditions and to translate these into yield anomalies and shocks. These events are subsequently provided as input to GLOBIOM to analyse the socio-economic and market impact of extreme climate events on soybean production. We show the hybrid model can significantly improve the overall performance of the crop model, with an increased interannual variability representation, especially for extreme climatic conditions. We find low precipitation values to be responsible for low yields in the region. The global warming scenarios suggest weather shocks will likely become more frequent and intense, especially for high-emission scenarios and especially in the medium to long term. The socio-economic impacts demonstrate a non-linear relationship to the bio-physical impacts, with metrics like price per ton of soybean being significantly amplified.

Actions (login required)

View Item