Use of near-infrared reflectance spectroscopy to predict the percentage of dead versus living grass roots

We tested the potential of near-infrared reflectance spectroscopy (NIRS) to predict the percentage of dead versus living roots of five grass species grown in monocultures under field conditions. Root death was induced after total severance of aboveground vegetation. Root samples were collected immediately after this treatment to obtain predominantly live roots (L), and then one (D1) and two months (D2) to obtain dead roots. NIRS spectra of L samples were different from D1 and D2 samples for four of the five species. The percentage of live and dead roots and root C and N were significantly predicted by NIRS. Validation of live and dead root percentage calibration was achieved with an error of prediction of 15%. These results show the potential of NIRS to predict the percentage of dead and live roots under field conditions and open up new opportunities in estimating more accurately below-ground net primary production of grasslands.