Using n-alkanes and their carbon isotope enrichments (δ13C) to estimate the botanical composition of pasture mixes from the Mid Rift Valley grasslands of Ethiopia

The present study was conducted to assess the accuracy of n-alkanes and their carbon isotope enrichment to estimate the diet composition of herbivores in the Mid Rift Valley rangelands of Ethiopia. Five common grass species which are abundantly available in the area were selected, from which several composite grass mixtures were prepared containing all the five species in varying proportions (from 0.05 to 0.35). The n-alkane profile and n-alkane isotopic ratio (13C/12C) of each grass species and composite grass mixtures were determined by gas chromatography/combustion isotope ratio mass spectrometry. The botanical composition of the composite mixtures was estimated using the n-alkane profile and n-alkane 13C/12C-ratio of individual species using least squares optimization and linear programming procedures and compared to the actual botanical mixture composition. Three alternative scenarios (inclusion of an additional 0, 5, or 10 of species in addition to those that made up the mixes) and two options where additional botanical species were included were simulated. There was close alignment between estimated and measured botanical compositions with significant relationships (P < 0.001). The percentage contribution of the species in the simulated pasture mixtures was accurately estimated when the five species making up the grass mixes were used as the only inputs in the calculation. However, when additional botanical species were introduced, the accuracy declined with a significant increase (P = 0.003) in the mean square error of the prediction. The type of species in the extra inputs did not influence the results. In all scenarios, the combined use of n-alkanes and their δ13C values improved the linear relationship and reduced (P = 0.002) the mean square error between estimated and measured botanical compositions. The best fit equation (R2 = 0.996, P < 0.001) was obtained when n-alkanes and their δ13Cvalues were used together and no extra species was included as input in the calculation. The present study shows that a high degree of accuracy can be obtained in estimating the botanical composition of grass mixtures using n-alkanes composition and n-alkanes 13C/12C-ratio. It is important to increase the number of markers used or limit the number of potential diet components to improve the quality of predictions.