The use of near infrared spectroscopy to predict vegetation quality and faecal indices of diet quality of ruminants
Estimation of the quality of the diet of wild and free-ranging ungulates is of paramount importance for conservation and management purposes. The composition of faeces is a commonly used method to assess herbivore diet quality. The majority of research has been carried out using pen-fed domesticated species. We tested the utility of near infrared spectroscopy to assess diet quality of herbivores, by using it to predict quality of forage and faeces of both extensively grazing and wild ungulates. Specifically, we estimated quality of forage available to free-ranging sheep (Ovis aries) and red deer (Cervus elaphus) across different months of the year. We developed predictions of diet quality using faecal quality as a proxy via NIR calibrations of faecal chemical constituents for two data sets a) extensively grazing sheep and red deer and b) wild red and roe deer (Capreolus capreolus). We compared the seasonal changes in available forage and faecal indices of quality.
NIRS calibrations for the prediction of faecal proxies of diet quality [faecal nitrogen (FN), acid detergent fibre (ADF), neutral detergent fibre (NDF) and lignin] were developed and compared using two datasets: a) the local scale data set of 164 faeces of grazing red deer hinds and Scottish Blackface sheep, collected between 2007-08 in a 2x2 km fenced plot in Glensaugh Research Station, Aberdeenshire, and b) a national scale data set of 1878 faeces of wild red deer and roe deer, collected between 2006-08 at different locations across Scotland. A set of 101 hand-plucked vegetation samples collected in association with data set a in 2007 at Glensaugh Research Station was analyzed for N, ADF, NDF and lignin and used to test for correlation between the chemical composition of faecal samples and the available vegetation. NIR spectra of the faeces were recorded in reflectance mode at 2nm intervals (1100 to 2500 nm) using a FOSS NIRSystems 5000 spectrophotometer (FOSS NIRSystems, Silver Springs, MD, USA) with a transport module sampling attachment and a quarter cup sample holder. Infrasoft International WinISI III software, version 1.63 (FOSS Tecator AB, Hoganas, Sweden) was used to develop calibration equations.
NIRS calibration and validation results for the prediction of N, ADF, NDF and lignin achieved a successful performance (R2cal ≥ 0.85, RPD= 2.3-6.09) for faecal samples of the two datasets. We found a significant linear relationship between the monthly values of vegetation quality and all faecal indices (p<0.05), suggesting faeces are a valid proxy of diet quality. The calibration equation which originated from wild red and roe deer samples was applied to predict samples of unknown composition and the examination of the predicted diet quality proxies (N= 1629) revealed annual, geographical, inter and intraspecific variations in diet quality. These results confirm that NIRS allows accurate prediction of diet quality indices for ungulates at the herd level for extensively grazing herbivores, population and national scale; and represents a valuable tool for wildlife ecology and conservation research.