Real Time in Field Forage Quality and Quantity Characterization in Monocultures and Mixed Species.

An effective method for in-field estimation of forage dry matter mass (DM) and crude protein (CP) must be approximately as accurate as the accepted standard for destructive removal measurement and laboratory analysis in order for management or research utilization. Non-destructive methods for estimating DM and CP have been developed using physical plant or canopy measurements and field spectral analysis. Challenges though may be encountered in representation of variation due to the feasibility and magnitude of data collection, vegetation growth characteristics, and spatial variability. This can present difficulty in creating robust estimation models, which are appropriate for a comprehensive range of DM and CP that may be encountered. Alternatively, remote sensing strategies can overcome data collection limitations and a much larger area can be sampled. This increased magnitude in data collection provides opportunity for development of statistically robust estimation models as a more comprehensive representation of the area of interest (AOI) can be collected. Projects containing wheat, alfalfa, bermudagrass, and tall fescue were employed to develop and test a mobile sensor system which could collect data for rapid real-time characterization of forage DM and CP. Proximity sensors, active and passive spectral radiometers, and a custom software application were combined to rapidly acquire data parameters from which estimation models were derived via partial least squares regression with cross validation. These models were then used to estimate DM and CP on data not used for model construction. Estimates compared to laboratory analyzed CP and destructively measured DM were highly correlated (R² up to 0.79 and 0.85 respectively), and produced percent errors down to 10.5% and 30% respectively. Implementation of this technology reduced man hours by a factor of approximately 60 for data collection, and post processing of data by a factor of approximately 10. Using these types of systems for in field on-the-go data collection can increase efficiency for management decisions and research without incurring unacceptable inaccuracy.