Premium
Prediction of model pools for a long‐term experiment using near‐infrared spectroscopy
Author(s) -
Michel Kerstin,
Ludwig Bernard
Publication year - 2010
Publication title -
journal of plant nutrition and soil science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.644
H-Index - 87
eISSN - 1522-2624
pISSN - 1436-8730
DOI - 10.1002/jpln.200800181
Subject(s) - standard deviation , chemistry , standard error , spectroscopy , organic matter , partial least squares regression , near infrared spectroscopy , analytical chemistry (journal) , derivative (finance) , biomass (ecology) , infrared , soil organic matter , environmental chemistry , soil science , environmental science , mathematics , soil water , statistics , agronomy , physics , optics , organic chemistry , quantum mechanics , financial economics , economics , biology
Fourty‐one soil samples from the “Eternal Rye” long‐term experiment in Halle, Germany, were used to test the usefulness of near‐infrared spectroscopy (NIRS) to differentiate between C derived from C 3 and C 4 plants by using the isotopic signature (δ 13 C) and to predict the pools considered in the Rothamsted Carbon (RothC) model, i.e. , decomposable plant material, resistant plant material, microbial biomass, humified organic matter, and inert organic matter. All samples were scanned in the visible‐light and near‐infrared region (400–2500 nm). Cross‐validation equations were developed using the whole spectrum (first to third derivative) and a modified partial least‐square regression method. δ 13 C values and all pools of the RothC model were successfully predicted by NIRS as reflected by RSC values (ratio between standard deviation of the laboratory results and standard error of cross‐validation) ranging from 3.2 to 3.4. Correlations analysis indicated that organic C can be excluded as basis for the successful predictions by NIRS in most cases, i.e. , 11 out of 16.