Use of near infrared spectroscopy to discriminate between and predict the nutrient composition of different species and parts of bamboo: Application for studying giant panda foraging ecology

Giant pandas (Ailuropoda melanoleuca) are specialist feeders, dependent upon bamboo as their main dietary resource. Due to the difficulty of many captive facilities to meet the natural qualitative diet changes in bamboo species and plant parts consumed seasonally by giant pandas, it is important to understand the nutritional quality of this forage and the differences among plant parts for improved husbandry. Near infrared (NIR) reflectance spectroscopy has been used as a tool to measure forage quality for both domestic and free-ranging species. The objective of this study was to determine the capability of NIR spectroscopy to: [1] discriminate between bamboo parts, (2) discriminate between bamboo species and [3] to predict the nutrient composition of bamboo. All bamboo samples were received from the Memphis Zoo Bamboo Farm (Memphis, TN, USA), dried at 60°C and ground to pass through a 1 mm screen before analysis. Discrimination between a total of 722 branch, culm and leaf samples resulted in an R² of 0.88 and SECV of 0.18. Spectra from a total of 756 samples of four different species were used to create a discriminant equation among bamboo species. This resulted in an R² of 0.47 and SECV of 0.29. Validation sets were correctly predicted at the following rates: (part) branch 94%, culm 100% and leaf 100%; [species] Phyllostachys aurea 10%, P. aureosulcata 98%, P. glauca 80% and Pseudosasa japonica 73%. Calibration equations for crude protein [CP], neutral detergent fibre [NDF], acid detergent fibre [ADF] and organic matter [OM] were created using all bamboo samples. For each nutritional constituent, the calibration R² values exceeded 0.96. The average SEP across alt constituents was 0.21% for CP, 2.35% for NDF, 3.62% for ADF, 0.84% for DM and 0.25% for OM. NIR spectroscopy was used to predict nutrient characteristics and discriminate between bamboo plant parts and species. The inability to discriminate among bamboo species is most likely due to a close physiological similarity between at least two of the species. Results suggest that NIR spectroscopy can be used to analyse bamboo forage quality which may have applications to captive giant panda husbandry.