Luminescence age modeling of variably-bleached sediment: Model selection and input

Optically stimulated luminescence (OSL) dating of heterogeneously-bleached sediment by means of a minimum age model requires the input of a 'sigma_b' (sigma(b)) value describing the overdispersion of the single-aliquot D-e distribution expected for a well-bleached sample. We propose that sigma(b) and associated uncertainty can be accurately determined if a large dataset of D-e distributions is available and includes well-bleached samples. Our approach applies the bootstrapped Minimum Age Model (bootMAM) to a dataset of overdispersions in D-e distributions, to obtain quantitative estimates of sigma(b). Corrections are made for constant-diameter aliquots of different grain sizes, based on the published dependency of overdispersion on the number of grains per aliquot. These adapted sigma(b) values are then input to bootMAM to obtain robust paleodoses for the samples. We test the sensitivity of paleodose to sigma(b) and we demonstrate that with correct sigma(b), identical paleodoses are obtained using bootMAM and the Central Age Model on samples judged to be well-bleached. We conclude that for large datasets consisting of well- and heterogeneously-bleached samples, appropriate sigma(b) values can be obtained from the data, and that bootMAM can be applied to all samples within this dataset.