|Canner, S - OWN BUSINESS|
Submitted to: Weed Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: May 7, 2002
Publication Date: November 20, 2002
Citation: Canner, S.R., Wiles, L., Mcmaster, G.S. 2002. Weed reproduction model parameters may be estimated from crop yield loss data. Weed Science. 50:763-772. Interpretive Summary: Studies measuring weed seed production as a function of the number of weeds present in a field are expensive and difficult to conduct. The resulting lack of these data is a severe restriction in modeling weed population dynamics that are critical components of decision support systems such as GPFARM. We used data from the literature to determine if we could use the more readily-available data on crop yield loss due to weeds to estimate weed seed production as a function of weed number using a hyperbolic model. We found that this was a feasible approach and can reasonably be used in modeling weed seed production when data are limiting.
Technical Abstract: Studies quantifying weed seed production as a function of weed density are expensive and difficult, and the lack of these data is a common bottleneck in the modeling of weed population dynamics. Data from literature sources was evaluated to determine if functions describing crop yield loss due to weeds could provide information useful for estimating weed reproduction as a function of weed density. For each of 161 data sets, a shape parameter (N50) and a scale parameter (U) were estimated for an increasing hyperbolic model for both crop yield loss as a function of weed density (N50YL, UYL) and for weed yield (either total biomass yield or seed yield) as a function of weed density (N50WY, UWY). N50 YL was strongly correlated with N50 WY across all data sets, with an apparent 1:1 relationship between the two, suggesting that they may substitute for each other in weed population models when data are limited. UYL was weakly correlated with UWY. This information can be useful for modeling of weed populations for systems where data are limited.