Use of near infrared reflectance spectroscopy (NIRS) to investigate selection and nutrient utilization of bamboo and to monitor the physiological status of giant pandas (Ailuropoda melanoleuca)
Wiedower, Erin Elizabeth
MetadataShow full item record
The objective of this study was to develop near infrared reflectance spectroscopy (NIRS) calibration equations from bamboo and fecal samples to predict diet composition and the physiological status of giant pandas. Discrimination between branch, culm, and leaf parts of bamboo resulted in an Rsquare (R2) of 0.88. The calibration equation for discriminating between 4 species of bamboo had an R2 of 0.47. Calibration equations were created for all bamboo species combined to determine the ability of NIRS to predict the nutrient constituents of CP, NDF, ADF, DM, and OM. No R2 was lower than 0.96, with the exception of DM at 0.63, which was consistently difficult to accurately predict due to variation in factors relating to difference in location of lab work (humidity, shipping, methods, etc.). Giant panda diets vary between seasons from eating primarily leaf to eating almost only culm. When bamboo part samples were compared between March and October, all resulting R2s were above 0.80. The sensitivity analyses for leaf and culm samples within diet season produced inconclusive results, but sensitivity analyses for fecal samples yielded an ability to more greatly discriminate between months that were further apart. For giant panda physiological status calibrations, fecal samples were collected from the Memphis Zoo, Smithsonian's National Zoo, Zoo Atlanta, and San Diego Zoo from 2006 to 2007. One-hundred fecal spectra were used to develop discriminant equations with which to predict between adults and juveniles. The resulting calibration was 100% correct for both age classes. Predictions between 252 male and female fecal spectra were 89% correct for females and 90% correct for males. A small number of samples (N= 60) were used to create a discriminant equation to differentiate between pregnant and non pregnant females. The exercise resulted in an R2 of 0.68 and a prediction of 100% for both pregnant and not-pregnant. It has been determined through these studies that NIRS has the potential to determine nutrient composition of bamboo and giant panda fecals, but increased sampling and equation development is needed before these calibrations are applicable in a captive or wild giant panda setting.