Associations among weed communities, management practices, and environmental factors in U.S. snap bean (Phaseolus vulgaris) production

Abstract

Weed species that escape control (hereafter called residual weeds) coupled with changing weather patterns are emerging challenges for snap bean processors and growers. Field surveys were conducted to identify associations among crop/weed management practices and environmental factors on snap bean yield and residual weed density. From 2019–2023, a total of 358 snap bean production fields throughout the major U.S. production regions (Northwest, Midwest and Northeast) were surveyed for residual weeds. Field-level information on crop/weed management, soils, and weather also were obtained. To determine associations among management and environmental variables on crop yield and residual weed density, the machine learning algorithm random forest was utilized. The models had 24 and 22 predictor variables for crop yield and residual weed density, respectively, and both were trained on 80% of the data with the remainder used as a test set to determine model accuracy. Both models had pseudo-R2 values of over 0.50 and accuracy over 80%. The models showed that crop yield was higher in the Northwest compared to the Midwest region, while higher average temperatures during early season growth and planting midseason (June-July) predicted greater crop yield compared to other time periods. The use of row cultivation was associated with lower snap bean yield and weed density, suggesting row cultivation had less-than-ideal selectivity between the crop and weed. Moreover, multiple spring tillage operations prior to planting were linked with an increase in weed density, implying that excessive tillage may favor the emergence of residual weeds in snap bean. Over the coming decades, climate change-driven weather variability is likely to influence snap bean production, both directly through crop growth and indirectly through weeds that escape control practices that also are influenced by the weather.