The Importance of Early-Season Weather Factors and Vegetation Indices on Prediction of Corn Yield at County Level

Corn is one of the most important agricultural commodities in the United States. For enabling informed agriculture decision-making, it is crucial to accurately predict corn yield early in a growing season. In many previous studies, weather conditions and vegetation indices are used for predicting crop yield. This study attempted to understand the importance of each parameter on the early-season crop yield prediction. The study extracted the monthly data from April to July over two decades (2003 to 2023) to analyze the most influential parameters affecting corn yields using the random forest model across all 102 counties in Illinois. The analysis identified top parameters, such as, monthly average dew temperature, soil temperature, solar radiation, evaporation, Leaf Area Index (LAI), and Enhanced Vegetation Index (EVI). Furthermore, standard deviation for each of these parameters was calculated for sensitivity analysis. The increasing or decreasing of each parameter from the averages by one standard deviation were input into the yield prediction model to evaluate the change of the predicted corn yield at the county level. The results indicated that an increase in solar radiation in June, along with improvements in Leaf Area Index (LAI) and Enhanced Vegetation Index (EVI) in July, each by one standard deviation, could accieve with an increase in corn yield. Specifically, fertilizing can lead to an increase in yield by up to 32.06 bushels per acre. It is important to note that LAI and EVI are measures of crop condition and reflect the overall health, which can be enhanced through proper fertilization and other management practices. This study provides a general framework for forecasting corn yield even when certain input features are missing or difficult to acquire, enabling fast and efficient yield predictions based on available data.