Usefulness of near-infrared spectroscopy to assess the composition and properties of soil, litter and growing media

Terhoeven-Urselmans, Thomas

kassel university press, ISBN: 978-3-89958-270-3, 2007, 123 Seiten

URN: urn:nbn:de:0002-2702

Zugl.: Kassel, Univ., Diss. 2006

| Preis und lieferbare Formen -->

Inhalt: Many challenges are present in agricultural and horticultural sciences, which call for the availability of fast, reliable and cheap measuring methods. Spectra of the VIS-NIR region (400 - 2500 nm) of (i) soil and litter samples from agricultural and forest sites (n = 56), (ii) soil samples from organically managed farms (n = 56) and (iii) a population of growing media with a great range of humification degrees and qualities (n = 320) were recorded from samples, which were fresh as received, or pre-treated. Modified-partial-least-squares regression and cross-validation were used to develop equations over the whole spectrum (1st to 3rd derivation). These equations were tested for predicting chemical and biological contents and properties and the yield characteristics.

(i) Near-infrared spectroscopy predicted for dried samples (60°C) the percentages of O-alkyl C and alkyl C well. The ratio of standard deviation of the laboratory results to standard error of cross-validation (RSC) was greater than 2, the correlation coefficients (r) of a linear regression (measured against predicted values) were greater than 0.9 and the regression coefficients (a) ranged from 0.9 to 1.1. The percentages of carbonyl C and aromatic C were assessed satisfactorily (0.8 ≤ a ≤ 1.2, r ≥ 0.8 and 1.4 ≤ RSC ≤ 2.0).

(ii) The soil chemical characteristics pH (CaCl2), contents of organic carbon (Corg), total nitrogen (Nt) and extractable P (Olsen), P (CAL) and Mg (CaCl2) were predicted well. The predictions were better for pre-treated (quick-freezing, freeze-drying and grinding) samples than for field-moist ones. The RSC ranged between 2.5 (pH) and 4.1 (Mg), (r) between 0.9 and 1.0, and (a) between 1.0 and 1.1. Predictions with NIRS of biological characteristics were slightly better for field-moist samples than for pre-treated. An exception was the content of ergosterol. Basal respiration and contents of microbial biomass C, N and P (Cmic, Nmic, Pmic) were predicted well. The RSC was between 2.5 (Pmic) and 4.6 (ergosterol), (r) between 0.9 (Pmic) and 1.0 (Cmic) and (a) between 0.9 and 1.1. Basal respiration was predicted satisfactorily. Nitrogen mineralisation rate was not predicted satisfactorily (RSC < 1.4). Predictions of grain yield of winter cereals and grain nitrogen uptake were good for pre-treated samples (RSC = 2.4 and r = 1.0 and RSC = 2.7 and r = 0.9, respectively).

(iii) Near-infrared spectroscopy predicted the chemical characteristics of growing media and the yields of fresh weight of Chinese white cabbage and rating at harvest (overall plant impression) better for samples, which were fresh as received, than for dried (60 °C) and ground ones. The pH and contents of total carbon (Ct) and Nt, salt, P, K, mineral nitrogen, NO3-, NH4+ and the NH4+ : NO3- ratio were predicted well: the RSC ranged between 2.0 (NO3-) and 4.4 (Ct), (r) was equal to or higher than 0.9 and (a) was between 0.9 and 1.1. Fresh weight yield of Chinese white cabbage was predicted well for the sub-population of the growing medium with a degree of humification of H-2 to H-3 on the von Post humification scale (RSC = 2.0, r = 0.9 and a = 0.9).

Near-infrared spectroscopy has shown its potential and proved its usefulness for the prediction of the contents and the properties of a wide range of constituents in agricultural and horticultural sciences. The sample state for the NIRS measurements played a crucial role for the prediction quality. Thus, the state of the sample has to be chosen carefully.

Die Publikation ist in folgenden Formen erhältlich:

Volltext (pdf-Datei, nicht ausdruckbar - 1.58 MB)
PDF ansehen (view)
Volltext (pdf-Datei, ausdruckbar, kostenpflichtig - 1.58 MB) 10.00 Euro
(kostenfrei im Netz der Universität Kassel - Im Netz der Uni Kassel befinden Sie sich, wenn Sie z.B. an einem Rechner im ITS, Ihrem Arbeitsplatz an der Uni oder auch in der Multimediathek der Bibliothek befinden.)
PDF erwerben (download) - Achtung kostenpflichtig, da Sie sich zur Zeit nicht im Netz der Uni-Kassel befinden