Offered Papers Theme A: Efficient Production from Grassland
Description
Goats are used for brush control and ecological management of Mediterranean grazing lands. Farmers are willing to cooperate with communities but they need an easy method to evaluate the daily intake of nutrients. A calibration of the chemical attributes of goats' diets was set-up, based on faecal near infrared (NIR) spectra (Landau et al., 2004; Table 1). The accuracy of this methodology was estimated by using the standard error of cross-validation (SECV), which represents the variability in the difference between predicted and reference values when the equation is applied sequentially to subsets of data from the calibration data set. This procedure is justified in situations with calibration samples that are randomly selected from a natural population, but may give over-optimistic results, in particular if data are replicated. The standard error of prediction (SEP) represents the variability in the difference between predicted and reference values when the equation is applied to an external (i.e., not used in any step of the calibration) validation data set. (Naes et al., 2002). The aim of the present study was to test the robustness of predicting dietary CP, in vitro dry matter digestibility (IVDMD), and NDF percentages in goats' diets, using faecal samples totally external to calibrations.
Included in
Agricultural Science Commons, Agronomy and Crop Sciences Commons, Plant Biology Commons, Plant Pathology Commons, Soil Science Commons, Weed Science Commons
Validation of Faecal NIRS for Monitoring the Diet of Confined and Grazing Goats
Goats are used for brush control and ecological management of Mediterranean grazing lands. Farmers are willing to cooperate with communities but they need an easy method to evaluate the daily intake of nutrients. A calibration of the chemical attributes of goats' diets was set-up, based on faecal near infrared (NIR) spectra (Landau et al., 2004; Table 1). The accuracy of this methodology was estimated by using the standard error of cross-validation (SECV), which represents the variability in the difference between predicted and reference values when the equation is applied sequentially to subsets of data from the calibration data set. This procedure is justified in situations with calibration samples that are randomly selected from a natural population, but may give over-optimistic results, in particular if data are replicated. The standard error of prediction (SEP) represents the variability in the difference between predicted and reference values when the equation is applied to an external (i.e., not used in any step of the calibration) validation data set. (Naes et al., 2002). The aim of the present study was to test the robustness of predicting dietary CP, in vitro dry matter digestibility (IVDMD), and NDF percentages in goats' diets, using faecal samples totally external to calibrations.
