Development of a New Tool for Estimating Sorghum Yields at the Farm-Scale.

The estimation of crop yields before harvest can be erratic, but producers often like to know about the potential yield of their crops. The goal of this research project was to get simple but fairly good estimates of sorghum yield before harvest. From a physiological perspective, the main yield driving forces for sorghum are: 1- number of plants; 2- number of tillers per plant; 3- total number of seeds per head, and 4- seeds per pound. From all abovementioned components, the final number of seeds per head presented the greatest variation (3500%, Variation = [(Max. –Min./Min) x 100], with the smallest variation for the seeds per pound (220%). As related to the association with yields, the number of seeds per head presented a high and strong degree of correlation with final sorghum yields; while the seeds per pound component did not show any consistent trend (“flat trend”). Therefore, the development of a method for rapid estimation of seeds per head at the field-scale could significantly contribute to the final goal of estimating sorghum yields at the farm-scale before the harvesting time. The estimation of the final seed number per head was pursued via implementation of allometric models. A head volume calculation was utilized to predict the final size of the sorghum head organ. For evaluation purposes, >1,000 sorghum heads from diverse germplasm sources were measured. From each individual head, final head volume, grain number and size were quantified. A strong association was obtained between final grain number per head and head volume estimated via allometric equations. In summary, on-farm yield estimations based on predicting the final number of seed per head would make the estimates much more accurate. Thus, implementation of this technique at larger-scale would allow to get rapid and accurate estimations of sorghum yields at the farm-scale.