Interesting encounter between deep learning and corn yield

Understanding large-scale crop growth and its responses to climate change is critical to yield estimation and prediction. Remote sensing provides timely and cost-effective monitoring of crop progress on a large scale based on the interaction between electromagnetic radiation and plant material.  

Identifying and understanding the cumulative and nonlinear relationships between meteorological factors and corn yield is critical to learning how weather change affects agricultural production. However, taking full advantage of high dimensional and heterogeneous geospatial data for yield estimation requires an advanced and robust modeling approach.  

The research team led by LIN Tao from the College of Biosystems Engineering and Food Science at Zhejiang University developed a long short-term memory (LSTM) model that integrates heterogeneous crop phenology, meteorology, and remote sensing data to estimate county-level corn yields. Their research findings are published in a research article entitled “A deep learning approach to conflating heterogeneous geospatial data for corn yield estimation: A case study of the US Corn Belt at the county level” in the journal of Global Change Biology.

The study covers rain-fed corn production from 2006 to 2016 in nine Midwestern states within the US Corn Belt, including Illinois, Indiana, Iowa, Kentucky, Michigan, Minnesota, Missouri, Wisconsin, and Ohio. By conflating heterogeneous phenology-based remote sensing and meteorological indices, the LSTM model accounts for 76% of yield variations across the Corn Belt, improved from 39% of yield variations explained by phenology-based meteorological indices alone. It has the capacity to learn general patterns from high-dimensional (spectral, spatial, and temporal) input features to achieve a robust county-level crop yield estimate.  

The LSTM-based deep learning approach to learning spatiotemporal heterogeneity of crop growth holds great promise to gain an improved understanding of global climate change on agricultural production.