Ground-based optical canopy sensing technologies for corn–nitrogen management in the Upper Midwest

Optical canopy sensing tools may improve corn (Zea mays L.) nitrogen (N) management, but their usefulness in far northern latitudes remains unclear. For this reason, the utility of SPAD, GreenSeeker normalized difference vegetation index (GS-NDVI), RapidSCAN normalized difference vegetation index (RS-NDVI), and RapidSCAN normalized difference red edge (RS-NDRE) were evaluated to predict corn grain yield, plant N accumulation, and plant N deficiency in 12 site-years throughout Minnesota. Six to seven N rates (35−45 kg urea-N ha⁻¹ increment) were pre-plant applied. Canopy sensing measurements and aboveground plant N accumulation were obtained at V4, V8, V12, and R1 stages. Regardless of the tool, low predictive power of grain yield, plant N accumulation, and N deficiency occurred at V4, likely because of low crop N demand and sufficient N supply. At V8, sensors provided good estimations of grain yield (R² =.75−.85) but underestimated the agronomic optimum nitrogen rate (AONR) by 33, 94, 102, and 46 kg N ha⁻¹ with the SPAD, GS-NDVI, RS-NDVI, and RS-NDRE, respectively. At V12 RS-NDRE measurements provided the most accurate estimations of grain yield (R²=.92) and AONR [R²=.84 and N rate differential from agronomic optimum nitrogen rate (dAONR) at −2 kg N ha⁻¹]. At R1 SPAD also provided good estimations of grain yield and N deficiency. The mismatch between the best timings for predicting N fertilizer requirements (V12 and R1) and the best timings for sidedressing (V4−V8) highlight that sensing tools may have limited utility to improve the standard maximum return to N approach in the Upper Midwest.