Soybean yield in relation to environmental and soil properties

Our goal was to identify soil, plant and climate attributes that are most closely related to soybean [Glycine max (L.) Merr.] yield variation in Pennsylvania. We studied 22 site-years over the 2016 and 2017 growing seasons in two regions. The average yields were 3.4 Mg ha^-1 in 2016 (range 1.4 to 5 Mg ha^-1) and 5.5 Mg ha^-1 in 2017 (range 3.5 to 7.4 Mg ha^-1). Solar radiation capture and water availability, both controlled by planting date, were the main predictors of soybean yield. Principal component analysis and Random Forest analysis revealed that the soil predictors of soybean yield were the content of zinc, copper, phosphorus, sulfur, potassium, as well as A horizon depth and total soil depth. The yield response to nutrients is likely a surrogate for a more complex response to animal manure additions. Soybean yield correlated positively with the ratio of soil respiration to soil organic matter, but did not correlate with the physical and biological soil metrics in the comprehensive Cornell Assessment of Soil Health (CASH). Saturated hydraulic conductivity (k_sat) and root depth correlated with both soybean yield and each other. Thus, while planting date sets the maximum achievable yield, only soils having the most water and nutrient availability (manured soils with high k_sat) expressed yields exceeding 7 Mg ha^-1. The k_sat appears to be a valuable indicator of soil condition that can be relevant well beyond its association with high soybean grain yield.