The role of climate on crop yield per unit area across the contiguous United States

The yield per unit area for the field crops is linked to several factors, such as technological changes, farming practices, and climatic variations. Empirical modeling of the historic yield per unit area allows us to quantify the role of these different factors. Previous research in this direction has been limited at smaller spatio-temporal scales (e.g., analyses over few fields) and for fewer crops. We analyze six major field crops across the contiguous United States (CONUS) to study the role of technology, climate variability and other factors. We use unpooled linear regression and linear mixed effects models to study the role of technology, climate, fertilizers, and harvested area respectively. The Palmer Hydrological Drought Index (PHDI) and Growing Degree Days (GDD) have been used to represent the climate. We find strong positive trend in yield per unit area for all the crops with varying magnitudes. The contribution of technological changes usually dominates the other variables such as climate and fertilizer application. However, regional variations exist, and climate plays important roles for rainfed crops (e.g. hay) in selected regions (e.g. southeast US). The limited role of climate is attributed to irrigation practices for a given crop, whereas rainfed crops show higher contribution of climate. The analysis at coarser resolution such as climate division yields better model performance than a finer resolution model at county level. The linear mixed model performs marginally poorer than the unpooled regression but offers greater flexibility in handling the data efficiently particularly with predictors available at different spatial resolutions.