MODELING THE EFFECT OF DURATION OF WEED COMPETITION, WEED DENSITY ANDWEED COMPETITIVENESS ON SEEDED, IRRIGATED ONION

Weed removal experiments were conducted in growers' fields in northern Colorado to assess the effect of duration of competition, weed densit y, weed competitiveness and crop density on irrigated seeded onion (Al lium cepa L.). Duration of competition, expressed in thermal time unit s (TTUs) with a base of 7.2 degrees C, explained 65% of the variation in the reduction of onion relative yield. The first significant reduct ion in onion relative yield was at 90 TTUs, averaged over weed load (w eed density adjusted by competitiveness) and onion density. A polynomi al multiple regression model, accounting for duration of competition a nd weed load, explained 75% of the variation in onion relative yield. A non-linear multiple regression model, combining a gamma function res ponse of relative yield to duration of competition plus a hyperbolic r esponse of relative yield to weed load, was as good a predictor and a better description of the system. Onion relative yield was more sensit ive to the duration of weed competition than to weed load. Bulb size c lass distribution and the resulting average onion price were affected by weed competition. Polynomial models were used to describe changes i n bulb size class proportions as a function of duration of competition , weed load and onion density.