Assessment of the potential of NIR spectroscopy for the estimation of nitrogen content in grasses from semiarid grasslands

Abstract

We tested the robustness of a near infrared reflectance spectroscopy (NIRS) calibration equation developed with grass samples from one sampling year to predict the nitrogen content of samples from the same area but subsequent years. The equation was developed with a population of heterogeneous and botanically complex grass samples from semiarid grasslands collected in 1993. This sample population included a great variability with respect to growth stage, place and the date of collection, soil-type, and botanical composition. The equation was validated with samples from the same population which were not included in the calibration, and also with samples from other populations collected in 1994-1996. The coefficients of determination and standard errors (g kg-1 of dry matter) of the Validation tests ranged from 0.90 to 0.98 and from 0.73 to 1.16, respectively. The results showed that our calibration equation is robust enough for an accurate prediction of the nitrogen concentration in samples of grasses collected in different years.