Discriminant Analysis as a Tool to Classify Grasslands Based on Near-Infrared Spectra

TL;DR

This study uses near-infrared spectra and discriminant analysis to classify different grassland types based on their botanical and chemical characteristics.

Contribution

The novel application of discriminant analysis with near-infrared spectra on fresh herbage samples for grassland classification is introduced.

Findings

Discriminant analysis achieved a 77% overall success rate in classifying grassland types.

Pure alfalfa crops and old meadows derived from them were classified with up to 80% success.

Grass–legume mixtures and their derived meadows showed lower classification success (~52%).

Abstract

Knowledge of grassland system characteristics is of primary importance in livestock feeding to apply proper management strategies and maintain specific territories and their biodiversity. This study aims to test the application of discriminant analysis based on principal components to near-infrared spectra derived from intact fresh herbage. Samples were collected from recently sown (pure and mixed) grasslands and old meadows that were naturalized from the previously sown north-central Apennine (Italy). Classification achieved an overall assignment success rate of 77%, and the discrimination seemed to be applicable with success (up to 80%) for pure alfalfa crops and old permanent meadows derived from the same. Grass–legume mixtures and permanent meadows originating from old grass–legume mixtures achieved lower assignment success and seemed more similar. The application of discriminant analysis combined with near-infrared spectra is promising for grassland types, which differ in their botanical and chemical characteristics, for quick assessment of pasture qualitative levels, considering that it can be performed without drying and milling.

This study aims to classify plant communities by applying discriminant analysis based on principal components (DAPC) on near-infrared spectra (FT-NIRS) starting from fresh herbage samples. Grassland samples (n~156) belonged to (i) recent alfalfa pure crops (CAA), (ii) recent grass–legume mixtures (GLM), (iii) permanent meadows derived from old alfalfa stands that were re-colonized (PMA), and iv) permanent meadows originated from old grass–legume mixtures (PLM). Samples were scanned using FT-NIRS, and a multivariate exploration of the original spectra was performed using DAPC. The following two scenarios were proposed: (i) cross-validation, where all data were used for model training, and (ii) semi-external validation, where the group assignment was performed without samples of the training set. The first two components explained 98% of the total variability. The DAPC model resulted in an overall assignment success rate of 77%, and, from cross-validation, it emerged that it was possible to assign the CAA and PMA to their group with more than of 80% of success, which were different in botanical and chemical composition. In comparison, GLM and PLM obtained lower success of assignment (~52%). External validation suggested similarity between PLM and GLM groups (93%) and between GLM and PLM (77%). However, a dataset increase could improve group differentiation.

Figures