Soybean yield in relation to environmental and soil properties

Highlights

• Solar radiation and precipitation over the growing season were associated with planting date and were the main drivers of soybean yield.

• The soil respiration to soil organic carbon ratio correlated positively with soybean yield.

• In-situ measured saturated hydraulic conductivity was positively correlated to soybean yield and can be a promising indicator of soil condition.

Abstract

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.

Introduction

Average soybean [Glycine max (L.) Merr.] yield has increased steadily in the United States (McGrath and Lobell, 2011; Hoffman et al., 2020). Each year, the average yield for a region reflects a wide range of within field and between fields yield variation. These yields are obtained frequently in seemingly uniform fields that are managed similarly. In Pennsylvania, producers report yields ranging from 3 to 6 Mg ha^-1 (Voight, 2016). In some fields, exceptional yields above 7 Mg ha^-1 have been documented in yield contests (Frankenfield, 2017). Understanding the causes of the yield variation as well as the factors that contribute to high yields can help charting pathways for increased productivity, profitability and environmental care.

Identifying the drivers of yield variation can be challenging due to management x environment interactions. Frequently, soil water holding capacity explains year-to-year crop yield variation (Williams et al., 2016). Nonetheless, Rattalino Edreira et al. (2017) indicated that planting date was the most consistent soybean yield predictor when considering different regions and years: as the planting date is delayed soybean yield decreases. The reasons for such response are well understood (Mooers, 1908; Egli and Cornelius, 2009). In summary, soybean yield decreases due to the hastening in development caused by shortening daylengths and warmer temperatures in the summer, which leads to lower availability and capture of solar radiation for photosynthesis. Therefore, earlier planting can set a higher yield potential, and a key for realizing that potential is increasing our knowledge of local soil environments.

The relationship between soil nutrient levels and grain yield depends on multiple factors. For example, Sawchik and Mallarino (2008) found that soybean yield in Iowa was positively correlated to soluble soil P and K, while Cox et al. (2003) in Mississippi did not, suggesting that in the latter case factors other than P and K limited yield. In an environment known for low soil P availability such as in Brazil, Santi et al. (2012) used Principal Component Analysis (PCA) to evaluate the effect of 63 chemical and physical variables on soybean yield and suggested that K content and soil infiltration rates were, respectively, the greatest chemical and physical soybean yield limiting factors. Certainly, soil-plant-climate interactions differ regionally, resulting in yields being most responsive to different soil variables depending on the location.

The soil quality concept, originally conceived as a measure of the soil’s ability to yield crops (Mausel, 1971), has been extended to include the broader ecosystems services concept (Doran and Parkin, 1994) and referred to as soil health (Karlen et al., 1997). Laboratories are now offering soil testing packages that aim at evaluating comprehensively soil chemical, physical and biological properties. Among these soil testing packages is the Cornell Analysis of Soil Health (CASH, Moebius-Clune et al., 2016). Bünemann et al. (2018) showed in their review paper about soil quality that most studies have not tested or reported how soil biological and physical indicators relate to crop yield. For example, Karlen et al. (2017) summarized results from the on-farm Soil Health Partnership (65 farms), but without including crop yield. When researchers report the response to yield, it can be subdued or applicable to a narrow slice of the sampling space. Nunes et al. (2018) found some CASH indicators that were positively influenced by no-till management and plant diversification strategies which related to crop yield, but the yield response was noticeable in loamy sand and silt loam soils and not in clay loam soil.

Our goal was to explore the relationship between soybean yield, climate and soil properties in two regions of Pennsylvania. The southeastern region (deep soils, low elevation) is broadly representative of soil-climate clusters in Ohio, Indiana, Illinois and part of Iowa and Missouri, while the central region (variable soil depths, higher elevation) is representative of more northern latitudes in the Midwest (Rattalino Edreira et al., 2017; Kukal and Irmak, 2018). Our specific objective was to distill soil and climate attributes closely related to soybean yield while accounting for interactions with management. To fulfill this objective, we did an observational study of 22 site-years where, in addition to soybean yield, we collected descriptive attributes of soil profiles and measured a suite of environmental, physical, chemical and biological indicators in individual sampling units per field. Observational studies are an alternative to controlled experiments when the conditions of interest are not in place in experiments across regions (Seddaiu et al., 2013; Ernst et al., 2016).