The Use of FTIR Spectra for Classifying Plant Items in a Vertebrate Herbivore's Diet

Availability and quality of vegetation are critical factors influencing herbivore nutrition and population dynamics. Fourier-transform infrared spectroscopy (FTIR) offers a promising approach to analyze herbivore diets using spectral properties of phytochemicals to identify plant items. We evaluated the potential of FTIR to identify plant taxa and parts consumed by an herbivore species. Crop contents from 236 rock ptarmigan (Lagopus muta MONTIN) individuals from Iceland, collected over nine years, were separated into pure fractions of plant taxa and parts (e.g., berries, leaves) and analyzed using FTIR in the mid-IR region (4000 -400 cm⁻¹). We classified plant taxa and parts with PCA and Random Forests (RF) based on spectral signals. FTIR revealed distinct chemical fingerprints for plant taxa and parts, consistent with previously established variation in lipids, proteins, carbohydrates, and chemical defenses. RF yielded high classification accuracy for plant parts (96.7%) and moderate accuracy for taxa (85.5%), confirming the method's reliability. FTIR overcomes limitations of traditional genetic analyses by identifying plant parts with varying nutritional quality within species. FTIR provided insights into biochemical properties of plant items but could not distinguish chemically similar items. Future research should expand spectral reference libraries combining FTIR with quantification of phytochemicals and DNA metabarcoding.