Robust predictions of weed seedling emergence from the soil seedbank are needed to aid weed management. A common seed accession (Illinois) of giant ragweed was buried in replicate experimental gardens over 18 site years in Illinois, Michigan, Kansas, Nebraska, Ohio, and South Dakota to examine the importance of site and climate variability by year on seedling emergence. In a nonlinear mixed-effects modeling approach, we used a flexible sigmoidal function (Weibull) to model giant ragweed cumulative seedling emergence in relation to hydrothermal time accumulated in each site-year. An iterative search method across a range of base temperature (Tb) and base and ceiling soil matric potentials (ψb and ψc) for accumulation of hydrothermal time identified optima (Tb = 4.4 C, ψb = -2,500 kPa, ψc = 0 kPa) that resulted in a parsimonious regional model. Deviations between the fits for individual site-years and the fixed effects regional model were characterized by a negative relationship between random effects for the shape parameter lrc (natural log of the rate constant, indicating the speed at which emergence progressed) and thermal time (base 10 C) during the seed burial period October through March (r = -0.51, P = 0.03). One possible implication of this result is that cold winter temperatures are required to break dormancy in giant ragweed seeds. By taking advantage of advances in statistical computing approaches, development of robust regional models now is possible for explaining arable weed seedling emergence progress across wide regions.
Nomenclature: Giant ragweed, Ambrosia trifida L.