Validation of a portable near‐infrared reflectance spectrometer to determine harvest maturity in grass seed crops

Abstract Seed moisture content (SMC) is the most reliable indicator of seed maturity and the optimal harvest timing in grass seed crops. Current SMC testing methodologies used in grass seed crops are slow or inaccurate, making it difficult to make timely harvest decisions. Harvesting too early can result in low seed weight and poor seed germination. Delaying harvest past the point of physiological maturity reduces seed yield by increasing losses due to shattering. Our objective was to validate the feasibility of using portable near‐infrared reflectance spectroscopy (NIRS) as a field‐based alternative to the oven method for determining SMC in cool‐season grass seed crops. Eight cool‐season grass species were used in field testing of the portable NIRS sensor over eight harvest seasons. Daily testing of SMC began when grass seed crops were at Biologische Bundesanstalt, Bundessortenamt and Chemische Industrie (BBCH) growth stage 69 and continued until windrowing. Seed samples were collected from each crop by cutting ∼40 inflorescences, then stripping the seeds into airtight containers until ready for estimation of SMC with a portable NIRS sensor, using SMC measurement by laboratory air‐oven (130°C) as the reference method. The SMC estimates made by the portable NIRS sensor were predictive of the actual SMC determined by the oven reference method across all eight grass species. These SMC predictions by the sensor closely followed the seasonal loss of SMC as the seed matured. Spring agronomic practices (mowing, plant growth regulators, foliar fungicides, and nitrogen fertilization) did not influence NIRS predictions of SMC compared with untreated controls. The portable NIRS sensor is a promising tool for determining harvest timing in grass seed crops by using predicted SMC values.