A hierarchical Bayesian approach to the classification of C-3 and C-4 grass pollen based on SPIRAL delta C-13 data

作者：Urban Michael A; Nelson David M^*; Kelly Ryan; Ibrahim Tahir; Dietze Michael; Pearson Ann; Hu Feng Sheng

来源：Geochimica et Cosmochimica Acta, 2013, 121: 168-176.

DOI：10.1016/j.gca.2013.07.019

Differentiating C-3 and C-4 grass pollen in the paleorecord is difficult because of their morphological similarity. Using a spooling wire microcombustion device interfaced with an isotope ratio mass spectrometer, Single Pollen Isotope Ratio AnaLysis (SPIRAL) enables classification of grass pollen as C-3 or C-4 based upon delta C-13 values. To address several limitations of this novel technique, we expanded an existing SPIRAL training dataset of pollen delta C-13 data from 8 to 31 grass species. For field validation, we analyzed delta C-13 of individual grains of grass pollen from the surface sediments of 15 lakes in Africa and Australia, added these results to a prior dataset of 10 lakes from North America, and compared C-4-pollen abundance in surface sediments with C-4-grass abundance on the surrounding landscape. We also developed and tested a hierarchical Bayesian model to estimate the relative abundance of C-3- and C-4-grass pollen in unknown samples, including an estimation of the likelihood that either pollen type is present in a sample. The mean (+/- SD) delta C-13 values for the C-3 and C-4 grasses in the training dataset were 29.6 +/- 9.5 parts per thousand and 13.8 +/- 9.5 parts per thousand, respectively. Across a range of % C-4 in samples of known composition, the average bias of the Bayesian model was <3% for C-4 in samples of at least 50 grains, indicating that the model accurately predicted the relative abundance of C-4 grass pollen. The hierarchical framework of the model resulted in less bias than a previous threshold-based C-3/C-4 classification method, especially near the high or low extremes of C-4 abundance. In addition, the percent of C-4 grass pollen in surface-sediment samples estimated using the model was strongly related to the abundance of C-4 grasses on the landscape (n = 24, p < 0.001, r(2) = 0.65). These results improve delta C-13-based quantitative reconstructions of grass community composition in the paleorecord and demonstrate the utility of the Bayesian framework to aid the interpretation of stable isotope data.

• 出版日期2013-11-15