Predict bean production according to bean growth, root rots, fly and weed development under different planting dates and weed control treatments

These two-year experiments determined the best predictors of bean growth and fly, root rots, and weed development in different cultivars, planting dates and weed treatments across 256 plots. Root rot diseases were naturally caused by Fusarium solani and Rhizoctonia solani. Weed management treatments involved: herbicide (Imazethapyr or Trifluralin) application, hand-weeding and control. Parameters estimated by exponential, Gaussian and linear-by-linear models fitted to bean, disease and weed datasets were considered as progress curve elements. Factor analysis detected the most predictive variables to characterize bean growth and production, disease, fly and weed development over season. There were significant correlations between considered plant, disease, pest and weed descriptors. Based on principal component analysis, considered bean-disease-fly-productivity-weed predictors justified 70% of total variance in datasets. Finally, multivariate regression model involving eight selected predictors explained a noticeable part (63%) of yield variations. Such information may improve accuracy of future efforts to monitor bean, disease, fly and weed development, predict bean yield and develop integrative field management programs.