PREDICTING THE STAGE OF DECAY OF DECOMPOSING LEAVES BY NEAR-INFRARED REFLECTANCE SPECTROSCOPY

To study mineral cycling in forest ecosystems, it is essential to know the decomposition rate of the litter. This study attempted to predict directly, by near infrared reflectance spectroscopy, the stage of dec omposition of leaf litter expressed as the percentage of ash-free litt er mass remaining (LMR). Leaf litter of 10 different species, with var ied initial compositions and at different stages of decomposition prod uced by incubation in the laboratory under controlled conditions, were used in this study. The LMR calibrations were carried out on half of the samples of the various populations (all species, woody species, br oad-leaved species, trees, broad-leaved trees, oaks, deciduous trees, and evergreen trees). The standard error of cross validation varied be tween 1.69 and 3.01. Predictions were carried out on the other half of the samples of each population; the standard error of prediction vari ed between 2.35 and 3.77, with a r(2) (coefficient of determination) o f 0.97 to 0.99. The calibration equations obtained from the laboratory samples were applied to samples that had decomposed in the field in l itter bags. The standard error of prediction varied between 4.36 and 5 .97, with a r(2) of 0.90 to 0.93. Near infrared reflectance spectrosco py thus provides a direct prediction of the LMR in leaf litter of diff erent species, during the decomposition stage studied (i.e., between 1 00 and 20% of litter mass remaining). The possibilities of using near infrared reflectance spectroscopy in decomposition studies are discuss ed.