Suitability of Near Infrared Spectroscopy to study the effect of water management on rice characteristics and cooking behavior

Rice (Oryza sativa L.) is adapted to grow both in submerged and aerobic soils. Changes in the agro-environmental conditions, as well as the lower availability of water for irrigation in some rice-growing areas, have led to the introduction of new cultivation practices. The different irrigation conditions, however, may influence rice composition and its cooking quality. The aim of this research is the use of Near Infrared (NIR) spectroscopy to provide comprehensive information of the effects of water management practices on rice milling yield, biometric characteristics of grains, nitrogen and amylose content, pasting properties and cooking behavior. In two following years, four rice varieties (Baldo, Gladio, Loto and Selenio) were grown in three different agricultural systems, characterized by a progressively more intense use of water: dry seeding and delayed flooding (DRY); dry seeding and rotational irrigation (IRR); water seeded and continuous flooding (FLD). NIR spectral data were collected in reflectance mode using a FT-NIR spectrometer equipped with an integrating sphere on the raw samples and a resolution of 8 cm−1. Raw data were pre-processed using SNV and derivative transformations. A restricted spectral range was selected by statistical treatment (9083-4223 cm-1). The spectral data were correlated with biometric, chemical and rheological parameters by Principal Component Analysis (PCA) and PLS regression algorithm. PCA was performed to highlight the main changes of the different irrigation conditions. The PC loadings denoted that the main differences were related to the interaction of all parameters with a central role of amylose content and texture indices of cooked rice. PLS regression model developed for Nitrogen (%) was characterized by high R2 both in calibration (R2 = 0.98) and in cross-validation (R2 = 0.91) and by very low errors. Considering the overall dataset, it was not possible to highlight an univocal influence of the water-soil regime on rice quality, since the quality characteristics seem to be mainly influenced by variety.