Optically Stimulated Luminescence (OSL) age and supporting data for sample JQ200-OSL1.

aMean ± total (1σ) uncertainty, calculated as the quadratic sum of the random and systematic errors.
bSum of dose rate contributions from beta particles (0.113±0.006 Gy/ka), gamma rays (0.170±0.009 Gy/ka), cosmic radiation (estimated as 0.20±0.02 Gy/ka) and alpha emitters inside the quartz grains (assumed as 0.03±0.01 Gy/ka). Calculated for a water content of 5±2%, which is higher than the measured (field) water content (0.92%) to allow for the subsequent lake and swamp conditions. For each 1% increase in water content, the total dose rate decreases (and the OSL age increases) by ∼1%. The beta and gamma dose rates were measured by low-level beta counting and field gamma spectrometry, respectively, using the same equipment and procedures as described by Petraglia et al. [56], [58]. The cosmic-ray dose rate was calculated following Prescott and Hutton [125], taking into consideration the time and rate of deposition of the overlying lacustrine and aeolian units.
c3000 grains were measured and analysed as follows: preheat of 260°C for 10 s (natural and regenerative doses) or 220°C for 5 s (test doses); green laser stimulation at 90% power for 2 s at 125°C; OSL signal from initial 0.2 s and background from final 0.3 s of stimulation; use of saturating exponential function to fit dose-response curves and estimate the equivalent dose (De). The De measurement uncertainty includes variation in background-corrected OSL counts, laser repositioning error (2% per measurement) and curve-fitting errors estimated by Monte Carlo simulation. Forty-nine grains (1.6%) satisfied all of the criteria used to identify grains with suitable OSL properties for reliable De determination [68]. Of the 2951 grains rejected, 2640 were discarded because the OSL signals were too dim to discern above background, and 311 were rejected on the basis of feldspar contamination, high ‘recuperation’ or a poor ‘recycling ratio’. The weighted mean De was calculated for the remaining grains using the Central Age Model [70], after discarding three values (Figure 5) that differ from the weighted mean by >2σ and have normalised median absolute deviations of >1.5 [126]. Including these outliers increases the weighted mean by 0.38 Gy, which is within statistical error of the tabulated De value. The total uncertainty on the weighted mean includes the additional spread (‘overdispersion’) of 23±4% in De values beyond that associated with measurement uncertainties for individual grains [71] and a relative error of 2% for possible bias in the calibration of the laboratory beta source.