Using soil parameters to predict weed infestations in soybean

An understanding of environmental factors governing patchy weed distribution in fields could prove to be a valuable tool in weed management. The objectives of this research were to investigate the relationships between weed distribution patterns and environmental properties in two Mississippi soybean fields and to construct models based on those relationships to predict weed distribution. Two months before planting, fields were soil sampled on a 60- by 60-m coordinate grid, and samples were analyzed for calcium, magnesium, potassium, sodium, phosphorus, zinc, cation exchange capacity, percent organic matter, and soil pH. The relative elevation of each sample location was also recorded. Approximately 8 wk after planting, weed populations were estimated on a 30- by 30-m grid over the soil sample grid. Punctual kriging was used to estimate environmental values at each weed sample location. Discriminant analysis techniques were used to evaluate the associations between environmental characteristics on weed population densities of sample areas within each field. Generally, as sicklepod and pitted morningglory infestations increased, the prediction accuracy of the discriminant functions also increased; however, horsenettle infestations were not closely correlated to the environmental properties. Discriminant functions reasonably predicted presence or absence of sicklepod and pitted morningglory within the field. However, validation of the functions across years within the same field and with data collected from the other field resulted in poor classification of all species infestations. Prediction of weed infestations with environmental properties was specific for each field, year, and species. Nomenclature: Horsenettle, Solanum carolinense L. SOLCA; pitted morningglory, Ipomoea lacunosa L. IPOLA; sicklepod, Senna obtusifolia (L.) Irwin and Barnaby CASOB; soybean, Glycine max (L.) Merr. ‘DPL 3588’, ‘DPL 3519s’.