Cotton Yield Monitor Performance in Changing Varieties

Twenty-nine varieties of cotton replicated 3 times (4-rows per replicate) in a complete random block design were harvested by a 4-row spindle-type cotton picker. Two commercially available yield monitors, Agleader and Farmscan were mounted on 2 of the 4 cotton chutes to measure cotton yield. The primary objective was to determine if the cotton yield monitor could be used to assess varietal performance or would constant changing of varieties introduce significant error. Comparisons of the Agleader measured weights for each variety and cotton weighing wagon weights indicated that it is very important to use the same setup for weighing calibration loads and harvest loads to prevent introducing bias. The mean absolute error for the Agleader yield monitor was between 9 and 10 percent when compared to weigh wagon measurements. It was also determined that quality factors, such as micronaire, length uniformity and fiber strength in different varieties do not contribute to changes in the measurement of cotton yield with a cotton yield monitor. There was a slight linear correlation (r2=0.12) between % lint turn-out and yield monitor results. This may contribute to increased error in yield monitor results when changing varieties since % lint turn- out should be a function of lint seed mass, which is variety specific. However, other factors, such as using 2 chutes of 4 for sensing yield and weighing cotton loads with two different weigh wagons for comparison to the yield monitor results may affect yield monitor accuracy more significantly.