Evaluating Phenological Indicators for Predicting Giant Foxtail (Setaria faberi) Emergence

Abstract

We evaluated the use of ornamental plants as phenological indicators for predicting giant foxtail emergence and compared their performance with predictions based upon Julian day, cumulative growing degree–days (GDD), and the WeedCast program. From 1997 to 2001, we monitored giant foxtail emergence in a field experiment with and without fall and spring tillage to estimate the dates of 25, 50, and 80% emergence; we also recorded dates of first and full bloom of 23 ornamental plant species. Dates of weed emergence and ornamental blooming for 1997 to 2000 were compiled in a phenological calendar consisting of 54 phenological events for each year, and events were ordered by average (1997 to 2000) cumulative GDD (January 1 start date, 10 C base temperature). Bloom events occurring just before the giant foxtail emergence events were chosen as the phenological indicators for 2001. The Julian day method used the average (1997 to 2000) dates of foxtail emergence to predict 2001 emergence. The GDD model (October 1 start date, 0 C base temperature) was chosen by determining the combination of start date and base temperature that provided the lowest coefficient of variation for the 1997 to 2000 data. The WeedCast prediction was generated using local soil and environmental data from 2001. The rank order of the 54 phenological events in 2001 showed little deviation from the 4-yr (1997 to 2000) average rank order (R2= 0.96). The phenological calendar indicated that, on average, 25% of giant foxtail seedlings had emerged when red chokeberry was in first bloom, and 80% of seedlings had emerged around the time multiflora rose was in full bloom. We compared the phenological calendar predictions for 25, 50, and 80% emergence with those based on Julian day, cumulative GDD, and WeedCast. The average deviation in predictions ranged from 4.4 d for the phenological calendar to 11.4 d for GDD. In addition to being generally more accurate, the phenological calendar approach also offers the advantage of providing information on the order of phenological events, thus helping to anticipate the progress of emergence and to plan and implement management strategies.