Dan Li

Dan Li is a Postdoctoral Research Fellow at Commonwealth Scientific and Industrial Research Organisation (CSIRO). She specializes in applied econometrics and computational statistics, currently focusing on developing data-driven approaches to project climate change impacts on crop production.

Innovation Title:

Machine learning projection of climate and technology impacts on crops key to food security

Abstract:

Climate change significantly threatens all economic sectors, with agriculture being especially vulnerable due to extreme conditions, rising temperatures, and shifting precipitation patterns. The impact on crop production will have critical implications for food security policies. Projecting the prospective impacts of climate change on crop production necessitates a comprehensive modelling system outlining crop responses to future conditions. We present a multivariate autoregressive econometric model that incorporates a time-varying variable to reflect the diminishing impact of technology on crop yields. This model examines the interplay between technology, climate variables, and annual crop yield growth globally. By analyzing historical crop production and climate data from 1961 to 2018, our model outperforms traditional panel regression methods. Additionally, a novel machine learning climate model emulator enables efficient estimation of crop production under a multitude of carbon emission scenarios. Our findings reveal that technological advancements may increasingly fail to counterbalance the adverse effects of climate change on wheat and rice production. This suggests that simplified climate damage functions in economic models can lead to significant inaccuracies in production estimates and flawed policy decisions.