PREDICTION OF WEED SEEDLING DENSITIES FROM BURIED SEED RESERVES

Bioeconomic models for weed management ultimately require knowledge of weed densities. Weed seedling populations may be predicted by multiplying emergence rates by seedbank densities. However, emergence rates vary according to species, year and management. Furthermore, seedbank estimates may vary with sampling technique, size, number and date. These variables must be quantified before bioeconomic models can be used profitably. Consequently, two experiments were initiated, both in conventionally managed maize. The first experiment documented proportional seedling emergence across years and sites for three taxa; foxtail (Setaria glauca [L.] Beauv. and S. viridis [L.] Beauv. combined), pigweed (Amaranthus retroflexus L.), and lambsquarters (Chenopodium album L.). The second experiment was devoted to methodology for estimation of seedbanks: sampling date (autumn or spring), technique (seed extraction or glasshouse germination), and soil sample size for the seed extraction technique. For emergence rates, the following order was observed; foxtail > lambsquarters > pigweed. Emergence rates for each species were related in a parabolic manner to growing degree-days in April. The glasshouse technique appeared to be more reliable than seed extraction for correlation with field seedling densities. For the seed extraction technique, a minimum soil sample size of 100 g was necessary. A spring sampling date appears to be more reliable than an autumn date, probably because many seemingly viable seeds die during winter.