Aristotelis C. Tagarakis, Animal Science, Cornell University, Ithaca, NY, Quirine M. Ketterings, 323 Morrison Hall, Cornell University, Ithaca, NY and Sarah E Lyons, New York, Cornell University, Ithaca, NY
Increasing home-grown forage production is important for the dairy industry in the northeastern United States. Double cropping of main forage crops such as corn (Zea mays L.) silage with winter cereals such as cereal rye (Secale cereale L.) or triticale (×Triticosecale spp.) can increase full-season yield and reduce the need for feed imports, thereby improving whole farm mass nutrient balances. However, inclusion of winter cereals can reduce the length of the growing season for corn. Brown midrib (BMR) brachytic dwarf forage sorghum (Sorghum bicolor L.) has great potential as an alternative to corn silage in double crop rotations, if sufficient nitrogen (N) is applied to the crop. Crop sensing is a promising approach in developing N application recommendation systems. In this study we evaluated the use of proximal sensing technology and timing of scanning to predict end-of-season forage sorghum yield, as a first step in developing algorithms for sensor-driven N recommendations. In addition, the impact of sensor orientation and distance from canopy on reflectance measurements were assessed. Four trials were implemented with N rates ranging from 0 to 224 or 280 kg N ha-1 at planting in four replications in 2014-2015. Sensor scans took place twice per week starting 19 days after planting (DAP) until 69 DAP. Yield was measured at soft dough stage (111 to 124 DAP). Sensor height and orientation impacted the Normalized Difference Vegetation Index (NDVI) prior to 45 DAP but not once the canopy was fully developed. Most accurate yield predictions were obtained 49 DAP when the sorghum was 0.76 m tall. In-season-estimated-yield (INSEY) expressed as plant growth per day (INSEYDAP) best predicted end-of-season yield (exponential model). We conclude that crop sensors can be used to accurately predict forage sorghum yields. Research is ongoing in 2016 to evaluate the algorithm for additional farm fields.