The following is an abstract for the selected article. A PDF download of the full text of this article is available here. Members may download full texts at no charge. Non-members may be charged a small fee for certain articles.
Prediction of Potential Mushroom Yield by Visible and Near-Infrared Spectroscopy Using Fresh Phase II Compost
Volume 59, Number 8 (Aug. 2005) Page 1054-1059
Sharma, H.S.S.; Kilpatrick, M.; Lyons, G.
Potential mushroom (Agaricus bisporus) yield of phase II compost is determined by interactions of key quality parameters including dry matter, nitrogen dry matter, ammonia, pH, conductivity, thermophilic microorganisms, C : N ratio, fiber fractions, ash, and certain minerals. This study was aimed at generating robust visible and near-infrared (Vis-NIR) calibrations for predicting potential yield, using spectra from fresh phase II compost. Four compost comparative trials were carried out during the winter and summer months of 2001-2003, under controlled experimental conditions employing six commercially prepared composts, with eight replicate (8 bag) plots per treatment (48 × 8 = 384). The substrates were prepared by windrow or bunker phase I, followed by phase II production. The fresh samples were scanned for Vis-NIR (400-2498 nm) spectra, averaged, transformed, and regressed against the recorded yield by employing a modified partial least squares algorithm. The best calibration model generated from the database explained 84% of yield variation within the data set with a standard error of calibration of 13.75 kg/tonne of fresh compost. The model was successfully tested for robustness with yield results obtained from a validation trial, carried out under similar experimental conditions in early 2004, and the standard error of prediction was 18.21 kg/tonne, which was slightly higher than the mean experimental error (17.94 kg/tonne) of the trial. The accuracy of the model is acceptable for estimating potential yield by classifying phase II substrate as poor (180-220 kg), medium (220-260 kg), and high (260-300 kg) yielding compost. The yield prediction model is being transferred to a new instrument based at Loughgall for routine evaluation of commercial phase II samples.